american gas

2003 PROCEEDINGSAMERICAN SCHOOL OF GAS MEASUREMENT TECHNOLOGY

Thirty-Eighth Annual School & ExhibitSeptember 15-18, 2003

Houston, Texas

SPONSORED BYGulf Coast Gas Measurement SocietyTexas A & M University – Kingsville

IN AFFILIATION WITHCorpus Christi Area Measurement Society

North Texas Measurement AssociationPermian Basin Measurement Society

Texas Gas Association


ADAM’S MARK HOTEL

Thirty-Eighth Annual School & ExhibitSeptember 15-18, 2003

Houston, Texas


AMERICAN SCHOOL OF GASMEASUREMENT TECHNOLOGY

P.O. Box 3991

Houston, Texas 77253

World Wide Web — http://www.asgmt.com

FUTURE SCHOOL DATES:

September 20 - 23, 2004September 19 - 22, 2005September 18 - 21, 2006September 17 - 20, 2007

American School of Gas Measurement Technology


TABLE OF CONTENTS

SECTION A

A1-A2 “Fundamentals of Gas Laws” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 1John Chisholm, Texas A&M University – Kingsville

A3 “Fundamentals of Orifice Metering” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 6Bill Buckley, Daniel Measurement and Control

A4 “Fundamentals of Orifice Recorders” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 9Robert Bennett, American Meter Company

A5 “Fundamental Principles of Diaphragm Meters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 11Don Jones, Schlumberger Gas Division

A6 “Fundamental Principles of Rotary Meters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 16Wayland Sligh, Instromet

A7 “Fundamental Principles of Gas Turbine Meters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 20Wayland Sligh, Instromet

A8 “Fundamentals of Natural Gas Safety” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 23Linton T. Lipscomb, PGE-GTT

A9 “Fundamentals of Pressure Regulation” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 26Kevin Shaw, Actaris

A10 “Fundamentals of EGM Electrical Installations” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 33Michael D. Price, CenterPoint Energy Field Services

A11 “Fundamentals of V-Cone Meters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 39Dr. RJW Peters and Dr. R. Steven, McCrometer

A12 “Basic Electronics for Field Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 46Rick Heuer, Emerson Process Management

A13 “Fundamentals of Ultrasonic Flow Meters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 50John Lansing, Daniel Industries

A14 “Multiphase Flow Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .no paperGerald L. Morrison, Texas A&M

SECTION B

B1 “Fundamentals of Natural Gas Chemistry” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 61Steve Whitman, Coastal Flow Measurement, Inc.

B2 “Techniques of Gas Spot Sampling” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 65Kris Kimmel, YZ Systems

B3 “Techniques of Composite Sampling” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 70David J. Fish, Welker Engineering Company

B4 “Operations of On-Line Chromatographs” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 75Charlie Cook, Daniel Measurement and Control

B5 “Calibration Standard Gases” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 79Ronald C. Geib, Matheson Tri-Gas

B6 “Methods for the Determination of Specific Gravity” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 82Myles J. McDonough, Jr., Chandler Engineering Company LLC


B7 “Devices for Field Determination of H2O in Natural Gas” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 86Borys J. Mychajiliw, MEECO, Inc.

B8 “Fundamentals of Energy Determination” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 90David Hailey, Instromet, Inc.

B9 “Effects of Entrained Liquid On Orifice Measurements” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 96Authors: William Johansen, Colorado Engineering Experiment Station, Inc.Speaker: Josh Kinney, Colorado Engineering Experiment Station, Inc.

B10 “Gas Flow Conditioning” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 100B.D. Sawchuk, P. Eng and D.P. Sawchuk, C.E.T., Canada Pipeline Accessories Company, Ltd.

B11 “Verifying Gas Chromatograph at Custody Transfer Locations” . . . . . . . . . . . . . . . . . . . . . . . . . page 111Mark E. Maxwell, Instromet

B12 “Advances in Natural Gas Sampling Technology” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 114 Paula Lanoux, A+ Corporation

B13 “H2S Detection and Determination” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 118Patrick J. Moore and Rodney W. Spitzer, Thermo Electron Corporation

B14 “Field Inspection and Calibration of Measurement Instruments” . . . . . . . . . . . . . . . . . . . . . . . . . page 124George E. Brown III, Reliant Energy-Entex

SECTION C

C1 “On-Site Proving of Gas Turbine Meters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 127Daniel J. Rudroff, Invensys Metering Systems

C2 “PANEL: Ultrasonic Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . no paperKevin Conrad, PanamtericsBurt Reed, InstrometJohn Lansing, Emerson DanielPeter Espina, ControlotronLars Farestvedt, FMC

C3 “Local and Wide Area Networking of Gas Flow Computers” . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 133King Poon, Thermo Flow Automation

C4 “Field Testing by Transfer Proving” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 135 Larry Wunderlick, Centerpoint Energy

C5 “Periodic Inspections of District Regulator and Relief Valves” . . . . . . . . . . . . . . . . . . . . . . . . . . . page 138John Johnson, CenterPoint Energy

C6 “Remote Meter Reading” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 140Arun Sehgal, Itron, Inc.

C7 “Principles of Odorization” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 143Tom Tucker, Odor-Tech

C8 & C10 “How to Perform a Lost and Unaccounted-For Gas Program” . . . . . . . . . . . . . . . . . . . . . . . . . . . page 147Rick Feldman, Quorum Business Solutions, Inc.

C9 & C11 “Applications of Telemetering in Natural Gas Distribution” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 156Doug Osburn, Automation Solutions

C12 “Meter Selection for Various Load Requirements” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 160Mike Haydell, Centerpoint Energy

C13-C14 “Design and Installation of a Complete Measurement and Control Facility” . . . . . . . . . . . . . . . page 163Thomas G. Quine, NorthStar Industries, Inc.


SECTION D

D1 “GRI Metering Research Facility Update” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 168Edgar B. Bowles and Marybeth Nored, Jr., Southwest Research Institute

D2 “AGA Calculations of Orifice Measurement – Old vs. New” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 173Brent E. Berry, ABB Totalflow Automation

D3 “A Review of the Revisions to API 14.3/AGA REPORT NO. 3 - PART 2” . . . . . . . . . . . . . . . . . . . . page 191Tom Cathey, JW Measurement Company

D4 “A New Perspective on Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 194Dave Wofford, Shell Gas Transmission, LLC

D5 “Problems Unique to Offshore Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 197Wayne T. Lake, Independent Measurement Consultant

D6 “Overall Measurement Accuracy” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 202Paul J. LaNasa, CPL & Associates

D7 “Pulsation Effects on Orifice Metering (Considering Primary and Secondary Elements)” . . . . . page 206Robert J. McKee, Southwest Research Institute

D8 “Pulsation Reduction by Acoustic Filters for Metering Applications” . . . . . . . . . . . . . . . . . . . . . page 215Robert J. McKee, Southwest Research Institute

D9 “Lessons Learned from the API 14.1 Gas Sampling Research Project” . . . . . . . . . . . . . . . . . . . page 219Eric Kelner and Darin George, Southwest Research Institute

D10 “Ultrasonic Flow Meter Calibrations” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 226Joel Clancy, Colorado Engineering Experiment Station, Inc.

D11 “Understanding the Different Standards that Govern Measurement” . . . . . . . . . . . . . . . . . . . . . page 233Ron Beaty, Premier Measurement Services, Inc.

D12 “Report on API 21.1 EGM Standard” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 236Brent E. Berry, ABB Totalflow Automation

D13 “Transient Lightning Protection for Electronic Measurement Devices” . . . . . . . . . . . . . . . . . . . . page 244Author: Patrick S. McCurdy, Phoenix Contact Inc.Speaker: Dick McAdams, Phoenix Contact Inc.

D14 “An Overview and Update of A.G.A. 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 249Author: Charles W. Derr, Daniel Division of Emerson Process ManagementSpeaker: John Lansing, Daniel Division of Emerson Process Management

SECTION E

E1 “Proper Testing of Odorant Concentration Levels” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 251Paul D. Wehnert, Heath Consultants Inc.

E2 “Pipeline Safety Improvement Act and Operator Qualification” . . . . . . . . . . . . . . . . . . . . . . . . . . page 255Author: Jesus Ramos, Texas Gas ServiceSpeaker: Tom R. Cheney, Anadarko Petroleum Co.

E3-E5 “PANEL: New Products in Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . no paperModerator – Becky Chambers, Reynolds Equipment Co.

E6 “Advance Communication Designs” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 264Bob Halford, PathTech, Ltd.

E7 “D.O.T. Requirements for the Transportation of Sample Containers” . . . . . . . . . . . . . . . . . . . . . page 269Tom Welker, Hydrocarbon Quality Association

E8 “SCADA and Telemetry in Gas Transmission Systems” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 272Chris J. Smith, Invensys Process Automation – The Foxboro Company

E9 “Terminology Used in Instrument Accuracy” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 279Rick Williams, Rawson & Company


E10 “Economics of Electronic Gas Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 285Tom R. Cheney, Anadarko Petroleum

E11 “Electronic Calibrators” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 288Betsy Murphy, M’nM Enterprises

E12 “Low-Power Flow Computers” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 291Greg Phillips, Bristol Babcock

E13 “Electronic Transducers and Transmitters” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 296Charlie Harris, Honeywell Industrial Automation & Control

E14 “Automating Gas Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 298Richard L. Cline, Integrated Information Technologies

SECTION F

F1 “Training Field Measurement Personnel” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 307Russell Treat, Gas in Certification Institute

F2 “Basic Communication Design” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 311Steven R. Cree, ABB Inc., Totalflow Division

F3 “Communication Between Office and Field Personnel” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 315Duane A. Harris, CMS Energy Corp.

F4 “Overall Measurement Accuracy” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . See D6Tom Cleveland, Enogex, Inc.

F5 “From Pen Tip to Volume Statement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 318David Pulley, Metron Gas Measurement

F6 “Chart Auditing” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 320Tom Tauer, Southern Flow Companies

F7 “North American Energy Standards Board” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 322Cynthia Corcoran, BTU Watch, Inc.

F8 “Requirements of an EGM Editor” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 326R. Michael Squyres, Flow-Cal, Inc.

F9 “Gas Contracts: Then and Now” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 329Mark B. Fillman and Gary P. Menzel, Coastal Flow Measurement, Inc.

F10 “Training Office Measurement Personnel” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 332Tom Cleveland, Hanover Measurement Services

F11 “Conversion from Volume to Energy Measurement” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 335Radhey S. Thakral, Texas Eastern Transmission Corporation

F12 “Electronic Gas Measurement Auditing” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 338Author: Kenneth W. Blackburn, Houston Flow Measurement, Inc.Speaker: Joe Landes, Southern Petroleum Labs

F13 “Internet Based Measurement Monitoring and Control” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 342Jim Griffeth, Emerson Process Management

F14 “Methods of Gathering EGM DATA” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 346Dennis Kline, TXU Business Services


HISTORYOF THE

AMERICAN SCHOOL OFGAS MEASUREMENT TECHNOLOGY

The First Annual Gulf Coast Measurement ShortCourse (GCMSC) was held in Houston, Texas inOctober, 1966, under the sponsorship of the GulfCoast Measurement Society and the University ofHouston. Following the 1973 Short Course, therelationship with the University of Houston wasterminated and the Short Course became affiliatedwith Texas A&I University. In 1974, the ExecutiveCommittee voted to accept participation by otheraffiliated organizations to enhance the growth ofthe Short Course. The Corpus Christi AreaMeasurement Society became affiliated in 1975, theSouthern Gas Association (SGA) and the VictoriaArea Measurement Society followed in 1976, theTexas Gas Association in 1978, and the NortheastTexas Measurement Society in 1981. In 1989,several changes were made. The Victoria AreaMeasurement Society was discontinued as anaffiliate and the Permian Basin MeasurementSociety was added as a new affiliate to the ShortCourse. One affiliate had a name change. TheNortheast Texas Measurement Society is nowknown as the North Texas Measurement Associ-ation. The Southern Gas Association has sincediscontinued their affiliation with the School.

In 1980, the Gulf Coast Measurement Short Coursewas incorporated and reorganized to further thegrowth of the School. The Gulf Coast MeasurementShort Course was operated by a Board of Directorscomprised of the then present and past GeneralChairmen of the Short Course, the President andVice President of the Gulf Coast Gas MeasurementSociety, and a representative from Texas A&IUniversity.

In 1991, the Board of Directors voted to changethe name of the School from the Gulf CoastMeasurement Short Course to the American Schoolof Gas Measurement Technology (ASGMT). With avision for international participation, continuedgrowth, and a broader industry recognition, theBoard of Directors set a course for future expansionby adopting a more comprehensive school formatwhile expanding upon the tradition and quality ofthe “Short Course.” The School is planned andoperated by the General Committee which iscomprised of approximately 55 members fromvarious operating and manufacturing companies inthe natural gas industry.

In 1994, while maintaining its co-sponsorship withthe Gulf Coast Measurement Society, Texas A&IUniversity became Texas A&M University-Kingsville.

PAGE i


GENERAL INFORMATIONOF THE

AMERICAN SCHOOL OFGAS MEASUREMENT TECHNOLOGY

The School is believed to be the largest gasmeasurement school in the United States devotedto natural gas measurement, pressure regulation,flow control, and other measurement relatedsubjects. The School is structured to focus on sevenmain subject groups: Fundamental Measurement,Gas Quality, Distribution, General and AdvancedMeasurement, Transmission, Office Procedures andAccounting, and Hands-on Training. Since 1966,the School has been held annually during the monthof September in Houston, Texas, the energy capitalof the world.

The philosophy of the School and those conductingit is intended to be an educational offering toindustrial and technical people. The primaryundertaking of the School is the responsibility ofthe General Committee and its members, who areselected by and serve under a Board of Directors.

The School is comprised of a two and one-half dayseries of 80 lecture classes, 55 hands-on periods,three plant tours, and exhibits by more than 97supplier companies. Suppliers are requested toteach the application and performance capabilitiesof their equipment/services, and refrain from salesactivities during the School.

At the time of registration, the registrants receive acopy of the “Proceedings” (hardbound/CDRom) ofthe American School of Gas MeasurementTechnology containing the technical papers to bepresented in the various lectures that comprise thecurrent program. The registrants also receive a“Pocket Program” to assist them with theirschedules.

Each registrant should plan their schedule carefullyand adhere to it in order to derive maximum benefitfrom the School.

PURPOSE

The purpose of the School, the sponsoringassociations, and the operating companies withinthe petroleum and natural gas industry, is to provideinstruction on technical subjects for people in theindustry. In this way, proper facility design,installation, operation, and maintenance ofmeasurement and regulation equipment, and thehandling of natural gas is presented and studied.Accurate and useful information is also developedand published for the benefit of the industry andthe general public.

ORGANIZATION

(1) Board of Directors

The School is governed by a Board of Directorscomprised of the present and past GeneralChairmen, the President and Vice President of theGulf Coast Gas Measurement Society, and oneappointed representative from the sponsoringeducational institution.

The Board of Directors elects from its membersa President, Vice President(s), Secretary, andTreasurer.

The Board of Directors also elects the GeneralChairman of the School and approves the Chairmanand Vice Chairman for each of the Subcommittees.

(2) General Committee

The General Committee consists of approximately55 members selected from the industry’s operatingand supplier companies.

PAGE ii


To accomplish the large amount of work involved,the School is supported by the General Chairman,Secretary, Treasurer, and six Subcommittees:Arrangements, Exhibits, Program, Publications,Publicity, and Registration. The Subcommittees areresponsible for planning and conducting the Schooleach year.

OPERATION

(1) Registration

Each person attending the School is required toregister. Payment of a modest enrollment fee isrequired of all registrants. Visitor passes areavailable for those who desire to view the exhibitsonly.

(2) Program and Instruction

The General Chairman selects the Principal Speakerfor the General Assembly. The instructors for thespecialized classes of the School are a diversifiedgroup. The group is comprised of members of thefaculty of major universities who offer engineeringinstructions pertinent to the School’s theme,representatives from the petroleum industry and ofsupplier companies, and others directly interestedin the science of gas measurement. The ProgramSubcommittee, from the available candidatesrecommended by the General Committee, selectsthese speakers and instructors.

(3) Educational Exhibits

The supplier companies assist the School by pre-senting educational exhibits and class instruction.They recognize the impropriety of competitiveselling during the School and at the request of theGeneral Committee have observed such. Exhibitpreparation and supplier cooperation is appreciatedby the School and is necessary for our continuedsuccess.

(4) Finance

The American School of Gas Measurement Tech-nology is a non-profit organization, which hasalways been self-supporting. The expenditures ofthe School are paid entirely from the proceeds ofthe School.

(5) Publications

The formal “Proceedings” are comprised of thetechnical papers presented at the School and arepublished by the General Committee. The“Proceedings” also contain opening remarkspresented by the General Chairman during theGeneral Assembly and a biography of the PrincipalSpeaker. Other useful information is included in the“Proceedings”. A copy (CDRom) is furnishedwithout additional cost to each paid registrant.

The “Pocket Program” contains a multitude ofgeneral information, schedules, classes, maps,exhibitor information, and other School relatedactivities. A copy is furnished without additional costto each paid registrant.

(6) Class Rooms

Lecture classrooms are Sections A through F in thesalons of the Adam’s Mark Hotel, Houston, Texas.Hands-on classrooms are sections G through K andare also in the salons. Please check the “PocketProgram” that you received when you registered,for the correct class date, time, and location.

Admission to hands-on sessions requires couponsavailable only at the Program Booth. Hands-onclasses are limited to 20 students. You will receivea coupon for each hands-on class that you desireto attend. You must present that coupon at theclassroom door. Coupons will only be distributedon the day of the class. To ensure availability, obtaincoupons early in the day.

(7) Information

There will be several staffed locations near theExhibit Hall where information can be obtained asnecessary: Registration Booths, ArrangementsBooth, Program Booth, Exhibits Booth, and theGeneral Chairman’s Lounge. Members of the Boardof Directors and the General Committee will be thereto assist you in every way possible with anyinformation that you need while attending theSchool.

PAGE iii


Henry DedekPuffer-Sweiven L.P.

2003 General ChairmanAmerican School of Gas Measurement Technology

Opening Remarks

Good Morning!

PAGE iv

On behalf of our Board of Directors and General Committee, it is my pleasureto welcome you to the 38th Annual American School of Gas Measurement Tech-nology. My name is Henry Dedek. I am employed by Puffer-Sweiven as Vice-President of Sales for Central & SW Texas and I am based out of Corpus Christi.I am honored to serve as General Chairman of the ASGMT for 2003.

The ASGMT is proud to be sponsored by the Gulf Coast Gas MeasurementSociety and Texas A&M University – Kingsville; as well as its affiliation with theCorpus Christi Area Measurement Society, North Texas Measurement Asso-ciation, Permian Basin Measurement Society and Texas Gas Association. TheASGMT is all about the natural gas industry and its people. The School is anon-profit organization comprised of 85 volunteers, some retired yet still givingback to the industry. In addition, we enlist numerous industry professionalsthat volunteer their time and expertise to help deliver a quality educational ex-perience to our students.

Although the natural gas industry has contributed to my livelihood for almost30 years, I was not exposed to the ASGMT until the early 1980’s. As I recall, theopening remarks of many past General Chairmen included comments on“change” and our need to “accept change.” In keeping with tradition, I will alsomention change. My message, however, is not about accepting change. Rather,it is about embracing change. Change is an integral part of life, but for mostdoes not come easily. It seems we have the natural tendency to resist beforewe accept, and then finally settle in. However, in the midst of change, you will


notice that some few rise to the top and new frontrunners are born. This isbecause those “early adopters” embraced change, while others were still busycomplaining about it or trying to decide how they would be inconvenienced byit. As an employee, embracing change provides both you and your companywith a competitive edge.

Our industry is embroiled in change. As such, the ASGMT is in a constant stateof change focused on the goal of helping you meet the challenges and technicaldemands that are placed on you by technology advances and industry trends.This year the School implemented a number of changes. First we expandedfrom 76 to 80 lecture classes and from 50 to 53 hands-on classes. We increasedour exhibitors from 97 to 108 and our booth spaces from 129 to 134. We changedour School format to include Vendor Bus Tours during normal class periods,enabling us to extend our School hours on Wednesday afternoon. Not only didthis allow us to offer additional classes, but also to repeat certain high demandclasses that were previously unavailable to some because they historicallyexceeded capacity. Finally, one of the most emotional and difficult changes weimplemented this year was the discontinuation of our printed Proceedings infavor of furnishing only a CD. Although this may not be a popular decision byall, please trust that it was necessary and helps the ASGMT offer our attendeesthe very best educational value in our industry.

Our individual Subcommittee Chairmen will discuss our School changes in moredetail, will provide information on School logistics, and will update you on anylast minute changes that are not included in the Pocket Program.

In closing, I would like to thank you for your attention, and particularly for yourattendance at the School. In these challenging economic times, your company’sdecision to support our School is a reflection of their confidence in the long-term health of the natural gas industry and of their commitment to the profes-sional growth of their employees. I congratulate you on being chosen to attendour School. Your willingness to enhance your value as an employee througheducation is one step in embracing changes that will pay dividends for you andyour company. I challenge you to maximize the opportunities that are availableto you this week and sincerely hope that your ASGMT experience is rewardingand memorable.

PAGE v


Jeryl MohnPrincipal Speaker

Panhandle Energy

Jeryl Mohn is senior vice president for Panhandle Energy, a Southern Union company. His responsibilities include operations andengineering, gas control, gas measurement, and LNG operations.

Mohn was named to the position in December 2000 after serving as vice president of operations and engineering since March1999. A career employee, Mohn joined Panhandle Eastern in 1974 as a transmission engineer at Newton, Kansas, and laterserved as a division engineer at Overland Park, Kansas. He then became an environmental engineer in the company’s energyconversion development department in Houston before being named superintendent of safety and security for the Trunkline LNGCompany liquefied natural gas receiving, storage and regasification terminal at Lake Charles, Louisiana, in 1978. He was pro-moted to operations superintendent at Lake Charles in 1982 and became region superintendent of transmission for Trunkline GasCompany at Lake Charles the following year. Mohn was made division manager for Trunkline in Houston in 1988 before movingto Liberal, Kansas, as division manager for Panhandle Eastern in 1989. He became general manager of operations for AlgonquinGas Transmission Company, based in Boston, in 1992 and in 1997 was named general manager of transmission for the company’sNortheast natural gas pipelines.

Mohn is a 1974 magna cum laude graduate of Wichita State University, where he received a bachelor’s degree in mechanicalengineering. He completed the general management program at the Harvard Business School in 1998.

Mohn serves on the board of directors of Boys and Girls Country of Houston, a residential care facility for children from familiesin crisis.

He was born November 25, 1951.

Panhandle Energy, a Southern Union Company, is comprised of Panhandle Eastern Pipe Line Company, Trunkline Gas Company,Trunkline LNG Company, Sea Robin Pipeline Company and Southwest Gas Storage Company. Panhandle Energy operates morethan 10,000 miles of mainline natural gas pipeline extending from the Gulf of Mexico to the Midwest and Canada. These pipelinesaccess the major natural gas supply regions of the Louisiana and Texas Gulf Coasts as well as the Midcontinent and RockyMountains. The pipelines have a combined peak day delivery capacity of 5.3 billion cubic feet per day, 88 billion cubic feet ofunderground storage facilities and 6.3 billion cubic feet of above ground liquid storage facilities.

Southern Union Company (NYSE:SUG) is engaged primarily in the transportation and distribution of natural gas. Through its localdistribution companies, Southern Union serves approximately 1 million natural gas end users in Missouri, Pennsylvania, Massa-chusetts and Rhode Island.

For more information on Panhandle Energy, please visit our website at www.panhandleenergy.com

PAGE vi


PAGE vii

C. Van MooneyASGMT Memorial Scholarship

Texas A&M University – Kingsville

The American School of Gas Measurement Technology (ASGMT) was founded on the belief that education in gasmeasurement and control was of prime importance to the gas industry. Education is the basis on which the schoolwas founded and is recognized as one of the leading schools in providing information to the students and gasindustry personnel.

As a co-sponsor of the school, Texas A&M University – Kingsville has played an active role in guiding and supportingour school since its inception. Faculty and staff have participated as lecturers, General Committee members, andserved on the ASGMT Board of Directors. One long-time representative was Dr. C. Van Mooney. As a faculty memberand department head of the Chemical and Natural Gas Engineering Department, and Professor Emeritus at TexasA&M University – Kingsville (then Texas A&I University), Dr. Mooney made many contributions to the gas industryand to ASGMT.

In 1999, to further the educational focus and to honor the long commitment with Texas A&M University – Kingsville,ASGMT established a memorial scholarship fund in honor of Dr. C. Van Mooney. The scholarship is in recognition tohis long time meritorious service to ASGMT and with appreciation to Texas A&M University – Kingsville for its long-term support of ASGMT.

Two scholarships are funded each year by ASGMT. The scholarship is regulated by Texas A&M University – Kingsvilleand granted to worthy undergraduates within the Chemical and Natural Gas Engineering Department.

Recipient Name

Audrey Hooks

A graduate of Harlingen High School in 2001, Audrey entered Texas A&M Uni-versity - Kingsville having already earned 23 hours of college credit at TexasState Technical College while in High School. She is majoring in ChemicalEngineering and currently has a 3.654 GPA. In 2002, as an intern at NASAAmes Research Center, she participated in a project that involves using non-intrusive infrared imaging techniques to examine exhaust plumes.

Extracurricular Activities* Treasurer - Resident’s House Council - Lynch Hall* Historian - American Chemical Society - participating in the Black Heritage Festival and Chemistry Olympics* Secretary - Society of Women Engineers* Society of Hispanic Professional Engineers and Mexican American Engineers and Scientists* American Red Cross - assisted in teaching a Lifeguard Training course

Recipient Name

Aimee Almarez

Aimee Almarez is the most recent recipient of the Van Mooney MemorialScholarship funded by the American School of Gas Measurement Technology.Born in Laredo, Aimee is a graduate of Hebbronville High School and is enteringinto her Junior Year in the Chemical Engineering Program at Texas A&MUniversity – Kingsville. She is on the Dean’s List because of her fine academicperformance while remaining active in the Student Professional Organizations.Her goal is to work in Texas in the Chemical or Petroleum Industries.


BOARDOF

DIRECTORS2003



E.D. (RUSTY) WOOMER, JR.Secretary

El Paso Pipeline Group

KEN BLACKBURNTreasurer

Hanover Measurement ServicesCompany, L.P.

BOARD OF DIRECTORS

WAYNE BURRELLPresident

Mercury Instruments

DON BACHLEEnron

(Retired)

JERRY BLANKENSHIPEmerson – Daniel Division

PAGE x

MAURICE CALAWAYNatural Gas Odorizing, Inc.

(Retired)

JOHN CHISHOLMTexas A&M University – Kingsville

WALLY PECKHAMVice PresidentYZ Systems, Inc.

CRAIG CHESTERWilliams Gas PPL Transco


GARY HANSONConsultant

BOARD OF DIRECTORS

PAGE xi

WINSTON C. MEYERCenterPoint Energy

DON SALYARDS2003 GCGMS President

Totalflow Measurement and Controls

JOHN STARCKDaniel Valve Company

PATRICK CUMMINGS 2003 GCGMS Vice-President

Amerada Hess

BILL LIGONThe Measurement Company

HENRY DEDEK2003 General Chairman

Puffer-Sweiven

LONNIE R. GRADYEl Paso Field Services

ROBERT A. (BOB) FOGLE Publicity Liaison

Union Texas Petroleum(Retired)

Not Pictured

Not Pictured


DAN TENNANTTennessee Gas Pipeline

(Retired)

PAGE xii

SENIOR ADVISORS

BOARD OF DIRECTORS

JOHN B. TALLEYCoastal Oil & Gas Corporation

(Retired)

GLENN TOLLERENEColumbia Gulf Transmission

(Retired)

GEORGE F. WHITE, JR.Tenneco Gas (Retired)

BRUCE SHRAKEDaniel Valve Company

(Retired)

DALE DONEGANColumbia Gas Transmission

(Retired)

ROGERS THOMPSONTexas Eastern Gas PPL

(Retired)


AFFILIATE REPRESENTATIVES

CORPUS CHRISTI AREAMEASUREMENT SOCIETYJim Lee, 2003 CCAMS President

Koch Pipelines

NORTH TEXAS MEASUREMENTASSOCIATION

David Pulley, 2003 NTMA LiaisonJ.W. Measurement Company

PAGE xiii

PERMIAN BASIN MEASUREMENTASSOCIATION

Robert QuattlebaumEmerson Process Management – Daniel Div.

GULF COAST MEASUREMENTSOCIETY (SPONSOR)

Patrick Cummings, 2003 GCMS Vice-PresidentAmerada Hess

TEXAS GAS ASSOCIATIONShoshanna Solomon, 2003 Liaison

TEXAS GAS ASSOCIATIONMike McCorstin, President

NORTH TEXAS MEASUREMENTASSOCIATION

Kenneth Smith, 2003 NTMA PresidentONCOR

GULF COAST MEASUREMENTSOCIETY (SPONSOR)

Don Salyards, 2003 GCMS PresidentTotalflow Measurement and Controls

Not Pictured Not PicturedNot Pictured

Not Pictured


GENERALCOMMITTEE

2003


PAGE xiii


MATT LIGONAssistant Treasurer

The Measurement Company

BILL BRANTLEY, JR.Treasurer

University of Texas – Petex

DUANE HARRISSecretary

Panhandle Energy

HENRY DEDEKGeneral Chairman

Puffer-Sweiven L.P.

GENERAL COMMITTEEOFFICERS

PAGE xvi

WILLIAM (BILL) WEBERAdvisor to Treasurer

(Retired)


JIM MICKLOSInstromet, Inc.

BURT REEDChairman

Instromet, Inc.

MIKE MCKAYVice Chairman

Odor Eyes Technologies

JERRY BLANKENSHIPLiaison

Emerson – Daniel Division

GENERAL COMMITTEEARRANGEMENTS SUBCOMMITTEE

PAGE xvii

LARRY QUICKThermo Electron Corporation

CRAIG CHESTERLiaison

Williams Gas Pipeline–Transco


PHIL LOCKWOODEl Paso Field Services

CARVEL JASMINThermo Electron

Not PicturedDARRELL CHERRY

Gajeske, Inc.

MAURICE CALAWAYLiaison

Natural Gas Odorizing, Inc.(Retired)

DANNIE MERCERVice Chairman

ONCOR Measurement Services

GEORGE F. WHITE, JR.Liaison

Tenneco Gas (Retired)

GENERAL COMMITTEEEXHIBITS SUBCOMMITTEE

PAGE xviii

TOM TRAMELChairman

Del Mar Scientific, Ltd.

JEFF GOETZMANCenterPoint Energy


ED BOWLESGTI MRF

GENERAL COMMITTEEPROGRAM SUBCOMMITTEE

PAGE xix

TOM CHENEYChairman

Anadarko Petroleum Corporation

TOM CATHEYJ.W. Measurement Company

Not Pictured

CHARLES LAMEVice Chairman

El Paso Production Co.

WAYNE HANERTransCanada Calibrations

CLIFF BURTONGulf South Pipeline Co.

To Come

LEE BURKETTJ.W. Measurement Company

To Come

BECKY CHAMBERSReynolds Equipment Company

JOEL CLANCYCEESI

Not Pictured Not Pictured

Not Pictured


GREG PHILLIPSBristol Babcock, Inc.

JOHN McDANIELHanover Measurement Services

WES PARTAINPuffer-Sweiven


PAGE xx

PAULA LANOUXA+ Corporation

DEE HUMMELOneok Field Services

SHARON MAYESEnergy Transfer Co.

ROYCE MILLERPrecision General, Inc.

Not Pictured

BRIAN McCLUREEmerson Process Management

Rosemount Division

STEVE HARLESSDuke Energy Field Services

Not Pictured



PAGE xxi

SHOSHANNA SOLOMONTexas Gas Association

TOM TAUERSouthern Flow Companies

ERIC THOMPSONChandler Engineering

RUSSELL TERRYEl Paso Production Company

Not PicturedMADELINE S. VECELLIO

Dresser Measurement

CHARLES STARTZDuke Energy

JAMES TABOREl Paso Production

MARSHALL SCHREVEDel Mar Scientific, Ltd.

JOHN SCHWEITZERRed Man Measurement

Not Pictured

Not Pictured


JOHN CHISHOLMLiaison

Texas A&M University – Kingsville

LONNIE R. GRADYLiaison

El Paso Field Services

PAGE xxii

DON BACHLELiaison

Enron(Retired)


LANCE WITTWelker Engineering Company

Not Pictured


JIM GRIFFETHBristol Babcock

GENERAL COMMITTEEPUBLICATIONS SUBCOMMITTEE

GLENN TOLLERENELiaison

Columbia Gulf Transmission(Retired)

GARY HANSONLiaison

Consultant

PAGE xxiii

PETER GRIMESC. H. Fenstermaker & Assoc.

DAVID PULLEYJ.W. Measurement Company

DEAN GRAVESChairman

Devon Energy

ED ROBERSONNatural Gas Odorizing, Inc.

KENNETH REEDOncor Gas Metering Services


PAGE xxiv

GENERAL COMMITTEEPUBLICITY COMMITTEE

STEVE CREEABB Totalflow

BILL LIGONLiaison

The Measurement Company

GREG METOYERAmerican Electric Power Company

KRIS KIMMELChairmanYZ Systems

GEORGE BROWNVice Chairman

CenterPoint Energy Houston

ROBERT A. (BOB) FOGLE Liaison

Union Texas Petroleum(Retired)

WINSTON MEYER Liaison

CenterPoint Energy

ACE ASTALAREPTEC

DON MORLEYEl Paso Production Co.


GENERAL COMMITTEEREGISTRATION SUBCOMMITTEE

JERRY CADRINTexas Eastern Gas Pipeline

(Retired)

ROGERS THOMPSONLiaison

Texas Eastern Gas PPL(Retired)

WALLY PECKHAMLiaison

YZ Systems

PAGE xxv

JOHN STARCKLiaison

Daniel Valve Company

JUDY PAWLIKChairpersonDevon Energy

STEVE RHODENThermo Electron

JIM KEATINGDuke Energy

MIKE SQUYRESVice Chairman

Flow-Cal, Inc.

DAVID BEASLEYCenterPoint Energy

Not Pictured


PAST GENERAL CHAIRMENof the

AMERICAN SCHOOL OF GAS MEASUREMENT TECHNOLOGY

2002 KEN BLACKBURN Hanover Measurement Services2001 E. D. (RUSTY) WOOMER El Paso Pipeline Group2000 WALLY PECKHAM YZ Systems1999 WAYNE BURRELL Mercury Instruments, Inc.1998 CRAIG E. CHESTER Williams Pipeline — Transco1997 WINSTON C. MEYER Entex1996 LONNIE R. GRADY Valero Energy Corporation1995 JOHN E. STARCK Daniel Valve Company1994 ROBERT A. FOGLE Union Texas Petroleum (Retired)1993 CHARLES W. McBRYDE El Paso Natural Gas Company (Retired)1992 JOHN B. TALLEY Coastal Oil & Gas Corporation1991 JERRY BLANKENSHIP Daniel Flow Products, Inc.1990 W. LEON CLAXTON Lone Star Gas Company (Retired)1989 GLENN L. TOLLERENE Columbia Gulf Transmission Company (Retired)1988 GEORGE F. WHITE, JR. Tenneco Gas (Retired)1987 WAYNE C. MELLOR Houston Flow Measurement, Inc.1986 DON BACHLE Houston Pipeline Company (Retired)1985 FRED PITTS Houston Pipeline Company (Retired)1984 DALE DONEGAN Columbia Gulf Transmission Company (Retired)1983 GARY HANSON Dresser Measurement Division1982 ROGERS THOMPSON Texas Eastern Gas Pipeline Company (Retired)1981 BILL LIGON Daniel Industries, Inc.1980 GEORGE NORMAN Tennessee Gas Pipeline (Retired)1979 MAURICE CALAWAY Natural Gas Odorizing (Retired)1978 M. L. WILLIAMS Amoco Production Company1977 LLOYD PETTICREW Southern Flow Measurement1976 HENRY J. (HANK) HENDRIX Transwestern Pipeline Company1975 BRUCE SHRAKE Daniel Industries, Inc.1974 R. VAN KERREBROOK Natural Gas Pipeline1973 DAN TENNANT Tennessee Gas Pipeline (Retired)1972 ROBERT WATSON Transcontinental Gas Pipeline1971 THOMAS JACOBS Tennessee Gas Pipeline1970 LEE BOYTER Trunkline Gas Pipeline (Retired)1969 HOWARD GRAY Tennessee Gas Pipeline (Retired)1968 ROBERT UTTER Union Texas Petroleum1967 EDWIN L. JONES Houston Pipeline Company

1966 HUBERT PRINGLE Amoco Production Company

IN MEMORIAM

George Armstrong Gordon Cook Henry J. (Hank) Hendrix Thomas JacobsEdwin L. Jones R. Van Kerrebrook Allan Owens Lloyd Petticrew

Hubert Pringle Robert J. Utter Robert M. Watson Howard Gray

PAGE xxvi


PAGE xxvii

AMERICAN SCHOOL OF GAS MEASUREMENT TECHNOLOGYEXHIBITORS INFORMATION

EXHIBITOR EXHIBITOR

A+ CORPORATIONABB, INC.ACTARIS METERING SYSTEMSALLIGATOR COMMUNICATIONSAMERICAN INNOVATIONSAMERICAN METER COMPANYAMETEK PROCESS INSTRUMENTSAMETEK TCIANALYTICAL SYSTEMS INTL.ARCCO INSTRUMENT COMPANY, INC.ATOFINA CHEMICALSBALON CORPORATIONBARTON INSTRUMENTS SYSTEMSBRISTOL BABCOCKBRUEST CATALYTIC HEATERSCANADA PIPELINE ACCESSORIESCATALYTIC HEATER COMPANYCEESI IOWACHANDLER ENGINEERINGCHECKPOINTCHEVRON PHILLIPS CHEMICAL COMPANY LPCONTROL MICROSYSTEMSDANIEL MEASUREMENT & CONTROLDCG PARTNERSHIP 1, LTD.DEL MAR SCIENTIFIC, LTD.DETCON, INC.DIA NIELSEN USADIXIE PRODUCTS, INC.DRESSER - BECKER PRECISIONDRESSER - GROVEDRESSER INSTRUMENTSDRESSER MOONEY CONTROLSDRESSER ROOTS METERS & INSTRUMENTSDUKE CONTROLS, INC.DYNAMIC FLUID MEASUREMENTEAGLE RESEARCHELYNX TECHNOLOGIES, L.L.C.EMC SERVICES, INC.ENERGY METER SYSTEMS, INC.ENERSYS CORPORATIONFISHER CONTROLSFLOW-CAL, INC.FMC MEASUREMENT SOLUTIONSFREEWAVE TECHNOLOGIES, INC.GALVANIC APPLIED SCIENCE, INC.HANOVER MEASUREMENTHEATH CONSULTANTS, INC.HONEYWELLHOUSTON CENTER VALVE & FITTINGINSTROMET, INC.INTEGRATED INFORMATION TECHNOLOGIES, INC.INTEGRATOR SERVICE, INC.INVENSYS METERING SYSTEMSJM TEST SYSTEMS, INC.

JW MEASUREMENTLOCUS, INC.LOMIC, INC.M&M INSTRUMENTS, LLCMARSHALL J. BROWN COMPANY, INC.MEECO, INC.MERCURY INSTRUMENTSMETCO SYSTEMS, INC.MICROWAVE DATA SYSTEMS, INC.MNM ENTERPRISESMTL, INC.NATIONAL METER, INC.NATURAL GAS ODORIZINGNORTH STAR FLOW PRODUCTSNORTHSTAR INDUSTRIES, INC.NUFLO TECHNOLOGIES, INC.ODOREYES TECHNOLOGIESOMNI FLOW COMPUTERS, INC.PGAS SYSTEMSPGI INTERNATIONALPMC GLOBAL INDUSTRIESPSA, INC.PUFFER SWEIVENQUORUM BUSINESS SOLUTIONS, INC.RAWSON, L.P.RED MAN MEASUREMENTRED RIVER INSTRUMENTS, INC.REFINERY SUPPLY CO., INC.REYNOLDS EQUIPMENT COMPANYROSEMOUNTSAVANT MEASUREMENT CORP.SCOTT SPECIALTY GASES, INC.SOFTWARE PERFORMANCE, INC.SOLARCRAFT, INC.SOUTHERN FLOW COMPANYSOUTHWEST PV SYSTEMS, INC.SPL, INC.STEP-KO PRODUCTS, LLCTEXAS ANALYTICAL CONTROLSTEXAS ENERGY CONTROL PRODUCTS, INC.THERMO ELECTRON CORPORATIONTRACE TECHNOLOGY, INC.TWIN CITIES TECHNOLOGIES, INC. (TCT)TYCO VALVES & CONTROLSTYCO/FIRE&SECURITY/GRAPHIC CONTROLSUNIVERSAL VORTEX, INC.VAL-TEXVERIS, INC.VORTABW. L. WALKERWELKER ENGINEERING COMPANYYOKOGAWA SOUTHWESTYZ SYSTEMS, INC.ZSYSTEMS / MRR


PROCEEDINGS2003


PAGE xxv

2003 PROCEEDINGS PAGE 1AMERICAN SCHOOL OF GAS MEASUREMENT TECHNOLOGY

FUNDAMENTALS OF GAS LAWSJohn Chisholm

Texas A&M University — KingsvilleChemical & Natural Gas Engineering

Kingsville, TX 78363

INTRODUCTION

In the gas industry a standard unit of measure is required.In the English system it is the standard cubic foot. In themetric, it is the standard cubic meter. This standard unitis the basis of all exchange in the gas industry. When theunit of purchase is the energy content (BTU) we achieveit by multiplying the BTU content of a standard cubic foottimes the number of cubic feet delivered to the customer.So we must obtain standard cubic feet or meters.

A standard cubic foot is defined as one cubic foot of gasat a pressure and temperature agreed upon by the buyerand seller. Common standard conditions are 14.73 psiaand 60° Fahrenheit. The gas passing through a meter israrely at standard conditions. It is necessary to convertthe gas in the meter from the metered conditions tostandard cubic feet. The tools we have for relating volumeto pressure and temperature are Equations of State or,simply, the Gas Laws.

The Gas Laws serve two purposes. They allow theconversion of a gas stream from metered conditions tostandard conditions. They also provide an understandingof what the gas is doing and why. This paper will brieflypresent the Gas Laws and the physical properties of gaswhich the Gas Laws describe.

ABSOLUTE UNITS

The very first thing you need to know is that all calculationsusing the Gas Laws are in Absolute Units. Absolutepressure starts at zero psia, an absolute vacuum. Absolutetemperature starts at zero degrees Rankine, approxi-mately -460°F, where theoretically no molecular motionexists. It is an error to perform calculations using psigor °F.

COMPOSITION OF NATURAL GAS

Matter can exist in three phases, solid, liquid, and vapor.The phase of a pure substance such as water is controlledby the energy of the molecules. For example, water atlow energy forms a crystalline solid, ice. At higher energy,it is liquid water. The molecules have too much energy toremain bound in a crystalline structure, but the attractiveforces between the molecules cause them to assume aminimum volume. Liquids will conform to the shape of thelowest parts of a vessel, but maintain the same volume.

As the energy increases the vibration of the moleculesovercomes the attractive forces and the water boils. The

steam molecules completely fill the confining vessel. Innormal measurement conditions natural gas is always inthe vapor phase.

The phase of a pure substance is determined by thetemperature and pressure. Figure 1 is a phase diagramof a pure substance such as ethane. At low temperaturesthe material exists as a liquid and will occupy a specificvolume at the base of its container. As a pure substancechanges from liquid to vapor there is a sudden change ofvolume as the liquid changes from its minimum volumeto the volume of the confining vessel. As the temperatureincreased across the line on Figure 1 the material changesfrom liquid to vapor and this sudden change of volumecan be observed. However, there is a pressure abovewhich no sudden change of volume is observed. The“liquid” completely filled the vessel and then the “vapor”completely filled the vessel. So, an observer cannot tellwhether the material is a liquid or a vapor. However, asthe material completely fills the vessel, we can treat it asa vapor either way. This critical pressure, Pc, is specific toa given material. There also exists a critical temperature,Tc, above which only one phase can exist. The criticaltemperature and pressure become important in thedetermination of the compressibility factor, z, which willbe discussed later.

Natural gas is not a pure substance. Figures 2 and 3 arephase diagrams for hydrocarbon mixtures. Whenever amixture of gases exists at intermediate energy levels, twophases, liquid and vapor, can exist at the sametemperature and pressure. The shape of the two phaseregion depends on the composition of the mixture. If theconfining vessel is transparent, an obvious liquid phasewill occupy the bottom of the vessel and the remainder ofthe vessel will be filled with the vapor phase. As the energyis increased, the mixture eventually goes to pure vaporcompletely filling the vessel as a single phase. There existsa temperature and pressure at which an observerwatching as the mixture reaches that point would reportthat the mixture went to a single phase occupying theentire volume of the containing vessel. The unique aspectof that temperature and pressure is that the momentbefore, any combination of liquid and vapor may haveexisted in the chamber. This is the point where all thelines in the two phase region come together. For example,if the mixture were 75% liquid and 25% vapor, the observerwould announce that the mixture suddenly became singlephase occupying the entire volume. Whether the materialis now 100% liquid or 100% vapor, the observer wouldbe unable to tell. (Given which side of the two phase regionthis point is on, one might hazard a guess.) However, as


The real gas law simply includes the compressibilityfactors.

Solving this equation for V2 yields

This equation is widely used to convert between twoconditions as in the calculation of line pack, for example,and to convert flowing conditions to standard conditions.

PRESSURE AND THE KINETIC THEORY OF GASES

Pressure is the force per unit area. The force on the wallof a container containing gas is the number of impacts ofgas molecules on the wall. So, anything that increasesthe number of impacts or the velocity of the impacts willincrease the force, and, hence, the pressure.

The Kinetic Theory of Gases states that the kinetic energy(KE) of a molecule is directly proportional to itstemperature.

KE(T) = 1 mv2 = cT2

Where m is the mass of the molecule, v is the molecule’svelocity, and c is a constant.

The Ideal Gas Law can be derived using the KineticTheory. And the derivation explains a great deal aboutwhat pressure is and how gases react.

Consider a cubic container with sides of length L,

In a container of fixed volume, the velocity controls howlong it will take a molecule to travel from one side of thecontainer to the other and back again.

Length of a round trip = 2L

The number of impacts on a given wall is then:

Number of impacts per unit time = v

2L

where v is the average velocity of the molecule. Eachmolecule of gas hits the wall at v and rebounds at –v. The

the single phase fills the entire volume of the vessel itcan be treated as a gas. These values of temperatureand pressure are called the pseudo-critical temperatureand pressure, pTc and pPc. The word pseudo is from theGreek and means false, counterfeit, or lie. However it hasthe property that it resembles the genuine. The pseudo-critical point is the point where all combinations of phasescan exist simultaneously. The pseudo-critical propertiescan be determined for any mixture of gases and are usedto determine the compressibility factor.

Natural gas consists of molecules of hydrocarbons.Usually this is predominantly methane. Other gases maybe present as contaminants such as water, oxygen,nitrogen, carbon dioxide, and hydrogen sulfide.

THE IDEAL AND REAL GAS LAWS

The ideal gas law is given asPV = nRT

and the real gas law asPV = znRT

where P = pressure, psiaV = volume, cubic feetz = compressibility factor, dimensionlessn = number of moles of gas, lb molT = temperature °RR = 10.732 psia ft3/(lb mol °R) for this system of

units. (The value of R will change dependingon the system of units.)

The difference between the ideal and the real gas law isthe compressibility factor, z. The ideal gas law assumesthat the molecules of gas have no volume and there areno attractive or repulsive forces acting between themolecules. At very low pressures z =~ 1.0 and gasesbehave as if they are ideal gases. At higher pressuresthe attractive and repulsive forces (often called thedynamic pressure) are significant and the behavior ofgases deviates from the ideal. The compressibility factoris often called the gas deviation factor or, simply, thez-factor.

Early experiments were conducted at low pressures andthe gases tested acted as ideal gases. Boyle’s Law statesthat at a constant temperature the product of the pressureand the volume is a constant, so any two conditions, 1and 2, of a gas were related by

P1V1 - P2V2

Charles’ Law states the ratio of volume to temperatureis constant at constant pressure,

These can be combined into one of the most commonexpressions of the ideal gas law.

V1 = V2

T1 T2

P1V1 = P2V2

T1 T2

P1V1 = P2V2

z1T1 z2T2

V2 = V1 P1 T2 z2

P2 T1 z1

LL

L


The Kinetic Theory shows that each impact contributesto the pressure. If the temperature increases, the velocityincreases, the round trip time goes down, and the numberimpacts goes up, hence, more pressure. If the number ofmolecules in the container increases, the number ofimpacts increase and the pressure increases. Finally, ifthe volume of the container is reduced, it takes less timefor the molecules to travel back and forth, so eachmolecule makes more impacts in a unit of time. Moreimpacts produced more pressure. The reverse of each ofthese events produces less pressure.

Both Charles’ and Boyle’s Laws have been derived fromthe Kinetic Theory. Some other very useful laws can bederived from the Kinetic Theory. (For the record, the Lawswere determined experimentally before the Kinetic Theorywas developed.) Avogadro’s Law states that, at the sametemperature and pressure, equal volumes of an ideal gascontains the same number of molecules. Also the volumecontaining one molecular weight of a given gas will beequivalent to the volume containing one molecular weightof another gas at the same temperature and pressure.There are 2.73 x 1026 molecules per pound mole of anideal gas.

When the Kinetic Theory is extended to mixtures of gases,Dalton’s Law of Partial Pressures can be found. Dalton’sLaw states that the pressure on a surface by a gas is thesum of the pressures that would be exerted by eachcomponent in the gas if it were alone,

Ptotal = PA + PB + PC + ...

and that the partial pressures are proportional to the molefractions of each component. The partial pressuresbecome especially important if liquid water is in contactwith the gas. The partial pressures can be used todetermine what fraction of each gas is in solution in theliquid. If significant quantities of CO2 or H2S are present,they will produce corrosive acids in solution with water.

THE DYNAMIC PRESSURE AND THE Z-FACTOR

The Kinetic Theory assumes no attractive or repulsiveforces are acting in the gas. This allows the Kinetic Theoryto derive the ideal gas law. However, these forces do existand at higher pressures cause gases to deviate from idealbehavior significantly. These forces are called the dynamicpressure. The gas deviation factor is just the ratio of theactual volume of a gas to that which would be predictedby the ideal gas law.

As the pressure increases from near atmospheric, themolecules are pushed closer together. Both gravitationaland electrical attractions cause the molecules to pulltowards one another with the result that the volumeoccupied by the gas is less than that predicated by thegas law. The z-factor becomes less than one.

momentum of a molecule is its mass times its velocity,so the change of momentum at the wall is given by:

∆ momentum = m’v - m’(-v) = 2m’v

where m’ is the mass of each molecule and the ∆ standsfor “change”. Consider a ball thrown at a wall. The wallmust first stop the bal’s momentum (m’v) and then sendit away again with momentum m’v. So from the point ofview of the wall it had to impart to the ball 2m’v, half justto stop it and half to send it away.

The change in momentum of a molecule per unit time isgiven by:

2m’v * v = m’ v 2

2L L

Only a third of the molecules, in the box are hittingeach wall.

The force acting on a wall is the number of moleculestimes the change in momentum per unit time.

F = n’ m’v 2

3 LAnd pressure is the force per unit area.

P = n’ m’ v2 1 = n’ m’ v 2

3 L L2 3L3

Now, L3 = V, the volume of the cube, so:

P = n’ m’ v 2 or PV = n’ m’ v 2

3V 3

The beginning assumption was that the kinetic energywas a function of temperature alone.

Then,m’ v 2 = 2cT and PV = n’ 2cT

3where c is a constant.

Before completing the derivation, consider someconsequences of the equation above. If the temperatureis held constant, then PV = a constant, which is Boyle’s

Law. If pressure is held constant, then V = a constant,which is Charles’ Law.

If we set A = Avogadro’s number, the number of moleculesin one mole, then

PV = n’ 2 cA T A 3

Now

n’ = n and we set 2 cA = R A 3

Then,PV = nRT

which is the ideal gas law.

n’3

T

( ) z =VACTUAL

VIDEAL


and molecular weights of the individual gases areavailable from any gas handbook.

EXAMPLE 1 – A turbine meter indicates that 4,000 macfof gas passed through it in one day. The line pressureand temperature were 500 psia and 122°F. The gasstream was 85% methane, 9% ethane, 4% propane, and2% butane+. What volume of gas in mscf was deliveredthat day?

The calculation of the critical properties of this gas is givenin Table 1.

122°F + 460 = 582°RpTr = 582°R / 384.5°R = 1.51

pPr = 500 psia / 664.2 psia = 0.75

z = 0.925

Vsc = Vline Pline Tsc zsc

Psc Tline zline

Vsc = 4,000 macf 500psia 520°R 1.0

14.7psia 582°R 0.925

Vsc = 131,417 mscf

EXAMPLE 2 – LINE PACK

Each day a pipeline must be balanced as to reciepts anddeliveries. A pipeline also represents a large storage unitfor gas. At a constant temperature, if the pressure of thepipeline is different at the end of the day than at thebeginning, then the reciepts and deliveries will not balancebecause the amount of gas stored in the pipeline haschanged. Determining the amount of that change is oftencalled a line pack calculation.

Consider the gas of Example 1. The pipeline runs 39 milesof 8 inch pipe with a total enclosed volume of 70,000cubic feet. If we start the day with an average pressure of500 psia and 122EF and end it at 600 psia and 122EF,how much additional gas is in the pipeline?

Initial gas in the pipeline:

Vsc = Vpipe

Ppipe Tsc zsc

Psc Tpipe zpipe

As the pressure continues to increase, the molecules,which have a real volume and velocity, begin to interferewith the other molecules in the gas. Personifying the gas,one can think of each molecule fighting for its own space.This results in repulsive forces. So the z-factor begins toincrease and returns to and then exceeds 1.0, so that athigh pressures the gas will occupy more volume thanthat predicted by the ideal gas law.

As many pure gases and mixtures of gases were testedit was observed that their behavior matched thatdescribed above. Although the z-factors for methanefollowed the same trends as that of ethane, the z-factorswere by no means identical. To unify these relations allthe tests were related to the reduced temperature andpressure.

For mixtures, the pseudo-critical properties can be usedin these calculations. When plotted for each and everyhydrocarbon gas the z-factor lines now tracked together.This discovery is commonly called the Law ofCorresponding States. This allowed a generalizedcompressibility factor chart to be developed.

Many mathematical correlations have been developedto calculate z-factors. The current methodology in AGA 8should be used unless another method is specified in agas contract.

The pseudo-critical properties can be calculated using

PTc = Σ yj Tcj and pPc = Σ yj Pcj

where yj is the mass fraction of each component of themixture. Values of yj are obtained from gaschromatograph readings.

The density and specific gravity of a gas can becalculated using

ρg = ρ MWg and SGg = ρg = MWg

zRT ρair MWair

where MWg is the molecular weight of the gas and ρg isthe density of the gas. Values of the critical properties

Component Mole Fraction Tc yjTcj Pc yjPcj

yj °Rankine °Rankine psia psiaC1 0.85 343.3 291.8 666.4 566.4C2 0.09 549.9 49.5 706.5 63.6C3 0.04 666.1 26.6 616 24.6

C4+ 0.02 830.0 16.6 482 9.6384.5 664.2

Table 1. Pseudo-Property Calculation for Example 1.

Tr = T and Pr = P Tc Pc


Vsc = 70,000 acf 500psia 520°R 1.0

14.7psia 582°R 0.925

Vsc = 2,127,311 cubic ft = 2,127mscf

At 600 psia, pPr = 0.903 and z = 0.905. So,

Vsc = 70,000 acf 600psia 520°R 1.0

14.7psia 582°R 0.905Vsc = 2,820,744 cubic ft = 2,821 mscf

So, 2,821 – 2127 = 694 mscf more gas is stored in thepipeline at days end than at the beginning.

REFERENCES

McCain, William D., Jr., The Properties of PetroleumFluids, Penn Well Books, Tulsa, Oklahoma, 1990.

Thompson, Roger G., “Fundamental Gas Laws,”Proceedings of the 29th Annual American School of GasMeasurement Technology, Houston, Texas, September19-22, 1994.

Gas Processors Suppliers Association, Engineering DataHandbook, Tulsa Oklahoma, 1972.

Standing, M.B. “Volumetric and Phase Behavior of OilField Hydrocarbon Systems,” SPE of AIME, Dallas, Texas,1977.

Standing, M.B. and Katz, D.L., “Density of Natural Gases,”Transactions AIME, 146, 1942.

John Chisholm


FUNDAMENTALS OF ORIFICE METERINGBill Buckley

Daniel Measurement and ControlP.O. Box 19097, Houston, Texas 77224

INTRODUCTION

The purpose to this paper is to discuss the fundamentalcomponents used in orifice measurement

BACKGROUND

The general concepts of head meters, which include theorifice, have been known for centuries. The orifice hasbeen in commercial use since the early 1900’s. Thedevice is used to create a differential pressure that relatesto the velocity of the gas from which a flow rate can becalculated. As the following gas passes through therestriction in the line caused by the orifice plate, thedifference in the upstream and downstream pressure canbe measured at set points, called taps, and a flow rateat the point can be determined

FIGURE 1. Head Meter

STANDARDS AND IMPORTANCE

Orifice measurement is guided by the standards ofseveral organizations. Primary among these is theAmerican Gas Association and the American PetroleumInstitute. The AGA #3 report is the standard that providesguidelines for the construction and installation of orificemeters. All orifice plates, holding devices and meter tubesshould be manufactured adhering to this standard inorder to help insure that the end product is an accuratemeasurement device.

ORIFICE PLATES

The most fundamental component of orificemeasurement is the orifice plate. This is typically a

circular, flat device, which is held in the flowing streamby a holding device. Typically, it is made of a durablemetal such as stainless steel. Orifice plates come inbasically two types, the paddle plate and the universalplate. The paddle plate is held in place by flanges, whilethe universal plates fit into the various types of holdingdevices. AGA #3 standards spell out specificrequirements for the orifice plates, including theconcentricity of the orifice bore, the surface finish,flatness of the plate, and edge thickness. While the orificeplate is the least expensive of the components in orificemeasurement, its importance should not be overlooked.

FIGURE 2. Orifice Plates

ORIFICE DEVICES

There are primarily three different types of devices usedto help center an orifice plate in the flowing medium.The first and least expensive is the orifice flange union.This is a pair of flanges, which has been tapped to providea differential reading. While it is the least expensive topurchase, it requires a higher maintenance level sincethe line must be bled down and the flanges spread apartin order to remove the plate.

The next device type is the single chamber orifice fitting.The single chamber device has an advantage overflanges in that it makes removal of the plate easier andsafer due to the prevention of spillage that occurs whenflanges are spread apart. Like flanges, however, thesimplex device requires that the line pressure be bledoff before the plate may be removed. The simplex deviceutilizes universal type orifice plates.


FIGURE 3. FIGURE 4.Orifice Flange Union Single Chamber Fitting

The third device is the dual chamber orifice fitting. Thisfitting allows for the removal of the universal orifice platewithout first bleeding down line pressure. This isaccomplished through the use of internal valves, whichisolate the upper (non-pressure) chamber from thebottom (pressured) chamber. The senior type is the mostexpensive of the plate holding devices to purchase, butcould be the most economical when compared to theoverall cost of the installation since isolation valving isnot required to allow plate removal.

FIGURE 5. Dual Chamber Fitting

METER TUBES

A meter tube basically consists of upstream tubing, theorifice fitting or flanges and downstream tubing. Thepurpose of tubing is to insure as smooth a flow profile,going into the orifice plate, as possible. The AGA #3standard has very specific requirements for meter tubepipe, including the smoothness of the inside surface ofthe tubing and minimum lengths required under particularinstallations. If these standards are not met in themanufacture of the meter tube, then degradation inmeasurement could result.

FIGURE 6. Three-Section Meter Tube

STRAIGHTENING VANES AND FLOW CONDITIONERS

Straightening vanes are bundles of small diameter tubing,which are placed inside the upstream section of a metertube. They are commonly of two types, flanged and in-line. The flanged types are held in the line between apair of flanges in the upstream. The in-line vane is heldin place inside the tubing by setscrews. Their purpose isto facilitate the smoothing of flow going into the orificeplate while allowing for shorter upstream tubing lengths.

The flow conditioners eliminate swirl like a straighteningvane and also generates a near fully developed flow profile.The conditioner also reduces the amount of requiredupstream tubing needed to meet AGA #3 requirements.

SECONDARY DEVICES

The orifice fitting with its orifice plate is known as theprimary devices in the orifice measurement package.There are other devices known as secondary devices,which translate the raw information from primary devicesinto more useable information. The most common ofthese are pneumatic chart recorders and flow computers.

FIGURE 7. Straightening Vanes

FIGURE 8. Flow Conditioner

FIGURE 9. Secondary Devices


The pneumatic chart recorder presents the informationfrom the differential pressure, static pressure andtemperature transmitters in a graphical form, usuallycircular charts. The chart usually represents a 24-houror 8 day time period, which can be integrated later toprovide volume figures.

Flow computers have increased in use in recent yearsdue to the requirements for measurements informationon a more “real time” basis. Flow computers, like thepneumatic chart, take the flow information from thedifferential pressure, static pressure and the temperaturetransmitter and calculates flow volumes. Unlike the chart,flow computers do not have to go through an integrationstep to come up with these figures. There are severallevels of sophistication available in flow computers. Thebattery-powered, solar charged devices have the bestutility as field devices, which can store the flowinformation on site, do the volume calculations and thensend that information on to a higher device such as amainframe computer. The higher-level flow computers

Bill Buckley

are usually AC or DC powered and provide not only thesame calculations capability as the solar –powered units,but also have advanced flow control and alarmcapabilities.

CONCLUSION

The new AGA #3/API 14.3 measurement standard hasgreatly tightened the tolerances for the manufacture oforifice devices and meter tubes. It is very much in thebest interest of the users of these devices to have soundmaintenance programs in place to insure that the like-new quality of the tubes be maintained for as long aspossible. The primary device, whether a fitting or flange,cannot be expected to provide accurate, reliable flowinformation if the orifice plate is bowed or otherwisedegraded in some way. The vast body of data supportingorifice measurement over the years becomesmeaningless if the guidelines for the design, manufacture,installation and maintenance of these devices are notfollowed.


FUNDAMENTALS OF ORIFICE RECORDERSDavid E. Pulley

J-W Measurement CompanyP.O. Box 2029, Tyler, TX

What is an orifice recorder? The answer usually dependsupon whom you are talking to. The term orifice meter isused to mean every thing from the orifice meter chartrecorder to the entire meter station. American GasAssociation defines the “orifice meter” as the completemeasuring unit comprised of primary and secondaryelements.

The “primary element” consists of an orifice meter tubeconstructed to meet the minimum recommendedspecifications of the measurement authority contractuallyagreed upon by two or more parties. The “secondaryelement” consists of equipment that will receive valuesproduced at the primary element. The values may bemeasured and recorded onto circular charts or receivedby electronic flow computers that calculate a volumeonsite, to be retrieved as desired.

In this paper I will address the “Orifice Meter ChartRecorder” and endeavor to explain its fundamentalworkings.

American Gas Association does not specify amanufacture or type of recording instrument to beemployed. The instruments installed at the meteringstation should answer the following questions:

1. Does it have adequate initial accuracy?2. Will it sustain adequate accuracy under

expected operating conditions?A. Range of ambient temperatures.B. Corrosion.C. Accumulation of dirt or foreign material.D. Vibration.

3. Will it require excessive maintenance?4. Are the available facilities adequate for proper

maintenance?5. Is it adequately flexible as to range, damping,

etc.?6. Does it adequately perform all of the required

functions?7. Is it adequately portable?

INSTALLATION

When locating the recorder at metering station attentionis required for placement and distance from the pressuretaps. Pressure values at the recorder must be the exactsame pressure values as pressure values at the metertube pressure taps. By this I am saying the distancebetween pressure taps and recorder is not limited ifpressures are the same values for the same moment intime. The location of recorder relating to elevation

comparison to meter tube pressure taps is a not a definedrequirement but a requirement of the gas condition.

When the gas is pipeline quality (separated anddehydrated for removal of free liquids and water vapor)location of the recorder varies, the orifice recorder maybe installed in a walk in house. The piping of manifoldlead lines may be of a configuration not consideringdrainage of free liquids into meter tube. But it must benoted gas of pipeline quality should not have free liquids,which accumulate in the lead lines.

The orifice meter chart recorder is an instrument thatmeasures the difference in pressure between a point oneinch upstream from the face of the orifice plate to a pointone inch downstream from the face of the orifice plate.This measurement is converted to inches of water columnand recorded onto a chart.

DIFFERENTIAL PRESSURE MEASUREMENT

The industry today uses two types of orifice meter chartrecorders. One is referred to as a bellows and the other isa diaphragm. Both are liquid filled and meet all theconditions required to measure and record the differentialpressure across an orifice plate. The differential shaftrotates in the bellows unit as an increase or decrease indifferential pressure causes the liquid to travel from oneside of the bellows to the other while the differential shaftof a diaphragm travels from side to side. A damping valvecan control the speed at which liquid travels from side toside, closing of the valve to restrict liquid movement whenthe flow characteristics are of a varying or fluxing natureand will result in a false recording of differential pressuresonto a chart. Generally this damping of liquid movementwill result in a false recording as the differential pressure


variations will record at the highest value and not reflectthe low values. This will result in a computed volumegreater or less than the chart would normally record.

Today the industry generally uses the bellows type unit.Located on the downstream or low side of the bellows isa calibration spring that can be of various ranges. Themost used range is 0 to 100 inches of water column. Asthe liquid is pushed from the upstream side of bellowsinto the low side a differential range spring, located onthe bellows, can be calibrated to restrict the liquidmovement. This calibration span is equal to the chart spanin inches of water column. If the bellows unit is damagedand loses part of its liquid fill movement of the bellows isreduced resulting in a false differential recording onto thechart. Also if the range spring is damaged the liquid willnot have a restriction in its movement and can result in agreater or false differential recording.

Other conditions that effect or causes false differentialrecording is free liquid or freezing of liquids in the manifoldlead lines. A pressure leak in the manifold lead lines doesnot allow a static pressure condition and will cause liquidto travel up into the lines. Plus a leaking tubing connectionferrule will cause a refrigerating condition at the fitting withan ice accumulation inside of the lines resulting in a lossof pressure downstream of the block. This results in therecording differential pressure to travel to one extreme orthe other on the chart depending on the freeze location.Free liquid and freezing can be minimized by installingthe manifold lead lines to allow drainage of liquid backinto the meter tube and assuring that no pressure leaksare present in the lines at operating condition.

FLOWING PRESSURE MEASUREMENT

The measurement of flowing pressure is normally by apressure measuring device (static element) installed inthe orifice recorder so that the flowing pressure andorifice differential pressure is recorded onto the samechart. The flowing pressure is normally measured inpounds per square inch. Pressure can be measured ateither side of the orifice plate but normally it is measuredon the downstream side. The pressure measuring device(static element) consists of a hollow tube closed at oneend and twisted in the form of a spiral. A pen or penmechanism is connected to the closed end. An increasein pressure will cause the spiral to unwind and positiona pen recording onto a chart equal to a value within itscalibration range. The range of an element can vary withflowing pressures. Element range should allow recordingof flow pressures that allows a separation in the recordingof differential and flowing pressures.

The degree of chart travel between the two recording pensat zero pressure and degree of travel from zero to 100% ofchart range is set to match the same degree of travel asthe chart integrator pens. When they are not the sameintegration value will not be a product of differential pressureand flowing pressure at the same moment in time.

Located within the orifice chart recorder the chart isplaced on a clock that can be adjusted to make a 360-degree rotation in the time desired. The orifice recorderis a real time instrument that measures the orificedifferential pressure continually. Rotation of clock andrecording chart reduces ratio of real time to recordingtime by the rotation speed. By this I am saying that avarying pressure recorded onto a one day rotation chartmay allow the accurate tracing of a value by the integratoroperator where a multiple day chart rotation results in aintegrator operator subjective tracing to produce aintegrator count for volume calculation.

The accountability of volume computations is greatlyeffected by damping of the fluid movement in a bellowsunit and by chart rotation where the flow rates vary. Irecommend a valuation of flowing conditions at themetering station and adjustments made to the recorderfor accountability in measurement.

FLOWING TEMPERATURE MEASUREMENT

When desired a recorder can have a third measurementdevice that is measuring the flowing temperature onto thechart. Again it is desirable that the recording of flowingtemperature not distort the value of other recordings andbe of a range required by the flowing temperature. Thecalibration of recording accuracy is important to recordingthe flowing temperature onto the chart for volumecomputation. The degree of travel from 0 to 100 of chartrange and degree of travel at zero chart range betweendifferential pressure pen is set by the manufacture to agreewith the integrator pen movements for correct averagingof temperature to flow values at the same moment in time.If the temperature recording is not correctly calibrated forchart recording the volume computation may be in error.

The continued accuracy of he orifice meter chart recordingdepends on keeping the recorder in proper operatingcondition and establishing and maintaining a schedule ofinspections. The inspection schedule may depend oncompany policy, importance of the station, size in termsof daily flow volume, type of equipment, etc.

CONCLUSION

In closing it is important to remember that the recorderrecords historic information from which volumes arecomputed and money isexchanged between companies.The accountability of volumes areonly as good as the accuracy ofrecordings.

David Pulley


FUNDAMENTAL PRINCIPLES OF DIAPHRAGM METERSRobert Bennett

American Meter Company1350 Bayshore Hwy., Suite 200, Burlingame, CA 94010

INTRODUCTION

The first gas company in the U.S., The Gas LightCompany of Baltimore, Maryland, founded in 1816,struggled for years with financial and technical problemswhile operating on a “flat rate” basis. Its growth was slowwith the charge for gas service beyond the pocketbookof the majority.

By comparison, the New York Gas Light Company,founded in 1823, prospered and expanded. They hadbuilt their system on “the use of gas meters to measurethe supply of gas to customers, and a large one to registerthe quantity made at the station before it is conveyed tothe gasometers.

The pattern of operation used by this New York companywas quickly copied by other companies throughout theEast Coast, including the Baltimore company. Seeingthe success, New York businessmen formed new gascompanies in Albany, Boston, Philadelphia, New York,etc. and the new U.S. gas distribution industry began toflourish.

Since this early beginning, meters have been animportant, integral element in every phase of gas industryoperations. Various types of meters are used; diaphragm,rotary, turbine, and orifice; each serving a definitepurpose and meeting specific requirements.

These four common types of meters can be broken downinto two distinct categories: positive displacement andinferential. Diaphragm and rotary meters fall into thepositive displacement category because they have well-defined measurement compartments that alternately filland empty as the meter rotates. By knowing the volumedisplaced in each meter revolution and by applying theproper gear ratio, the meter will read directly in cubicfeet or cubic meters.

Turbine and orifice meters, on the other hand, have nomeasurement compartments to trap and then releasethe gas. These meters are inferential meters in that thevolume passing through them is “inferred” by observingor measuring some physical characteristic.

COMPONENTS

A diaphragm meter is physically composed of: 1.) A bodyto contain the gas pressure and form part of thecompartments that measure the gas, 2.) Diaphragms thatmove as gas pressure fluctuates on either side, 3.) Valve

covers and seats that control the flow of gas into eachside of the diaphragm, 4.) Linkage to connect thediaphragm with the valves and index, and finally 5.) Theindex which registers the number of revolutions of theentire mechanism. (See Figure 1.)

FIGURE 1. Diaphragm Meter

A diaphragm meter can be compared to a two-pistondouble-action engine in which the diaphragms correspondto pistons and the meter body to the cylinders. Each strokeof the diaphragm displaces a fixed volume of gas and thediaphragms operate 900 out of phase so that when one isfully stroked, the other is at mid-stroke.

This provides a smooth flow of gas to the meter outlet andinsures the meter will always start regardless of its staticposition. When a demand for gas is made on thedownstream side of the meter, a pressure drop is createdacross the meter and its diaphragms. This differential, whichamounts to 0.1" W.C., provides the force to drive the meter.

Above each diaphragm is a “D” shaped valve (See Figure 2).Under the valve are three port openings that direct theflow of gas in and out of the case and diaphragmcompartments. As the diaphragm expands, it forces thegas in the case compartment up through the case port.The valve directs the flow of gas into the center port thatleads to the meter outlet. A similar process occurs whenthe diaphragm contracts. The stroke of the diaphragmis controlled by linkage in the upper port of the meterand a rod (flag rod) that extends down into the diaphragmcompartment. The tangent link, as it is called, is attachedto the top of the meter crank and is adjustable in length.


Increasing the tangent length increases the diaphragms’stroke which increases the meter proof and vice versa(See Figure 3.)

The crank makes a certain number of turns per cubiccubic foot and transmits this motion to the front counter(index) by means of an axle shaft driven by a worn andwheel. The crank also drives the sliding valves, whichare timed to the motion of the diaphragm

STANDARDS

To provide some common base for rating the capacityof various manufacturers’ gas meters, the ANSI B109.1gas meter standard establishes a set of guidelines formeters with a capacity of less than 500 Cubic Feet PerHour. ANSI B109.2 is the standard for diaphragm metershandling more than 500 Cubic Feet Per Hour and ANSIB109.3 covers the “Rotary Type Gas Displacement Meter.

Both B109.1 and B109.2 define diaphragm type gasmeter capacity as that volume of 0.60 specific gravitygas at an absolute pressure of 14.73 PSIA that will resultin an average pressure drop through the meter of 0.5inch of water column, using specified inlet and outletconnections. This capacity rating is not to be construedas a maximum capacity but as a common capacity ratingbase. Within the B109.1, meters are divided up intoseveral classes:

Class Capacity (ft3/hr)Minimum Maximum

50 50 174174 175 249250 250 399400 400 499

This standard calls for new meters to be +1.0% accurateand +2.0% after accelerated life tests. (See Figure 4. forsmall meter test apparatus.)

FIGURE 2. Movement of a Diaphragm Meter

FIGURE 3. Tangent Adjustments FIGURE 4. Test Apparatus


The B109.2 (500 Cubic Feet or More) also divides largevolume meters into several classes.

Class Capacity (ft3/hr)Minimum Maximum

500 500 899 900 900 13991400 1400 22992300 2300 34993500 3500 5599

(Figure 5. shows a typical test stand used to determinemeter capacities for large volume meters.)

In addition to the capacity of the meter, its body must bephysically constructed to withstand the internal pressureof the gas. Modern diaphragm meters have sufficientstructural integrity to withstand a minimum shell test of10 PSI. This is necessary to meet the federal rulescontained in Part 192 of Title 49, Department ofTransportation, Office of Pipeline Safety regulations.

The end connections must be of sufficient size to alloweasy passage of gas through the meter and can be sometype of insert, NPT screw connections, or flanges. (SeeTable 1 for capacities and connections.)

FIGURE 5. Large Meter Test Apparatus

Diaphragm meters meeting the requirements of bothB109.1 and B109.2 will measure the gas flow to a pilotlight starting at .25 ft3/hour for a Class 250 meter andsmaller and up to 7 ft3/hour for a class 3500. The allowableaccuracy at this flow is +10%. A typical accuracy curvefor a diaphragm meter is shown in Figure 6.

FIGURE 6. Diaphragm Meter Accuracy

Table 1. Diaphragm Meter Capacities

INSTALLATION

FIGURE 7. Typical Diaphragm Meter Installation


Diaphragm meters should always be shipped, stored,and installed in an upright position. Dust caps on theinlet and outlet connections should be left in place untilthe meter is installed. Caution should be used with metersthat have been removed from service since they maycontain gas within the diaphragm chambers.

The meter set should be in a location that is ventilatedand readily accessible for examination, reading,replacement, or maintenance. The set should beprotected from outside damage and at least three (3)feet from known sources of ignition or air intakes.Electrical isolation for cathodic protection purposesshould be maintained.

A diaphragm meter should be installed as close to levelas possible. Tests have indicated that tilting a meter willaffect its accuracy since, at high angles, the valves tendto come off the valve seat and let the gas bypass thediaphragm chambers. Therefore, the inlet and outletconnections should be within + 1/4" of each other.

The meter should be installed in a manner to avoid unduestress on the connecting piping or meter. The use of ameter bar may be a consideration. If there is foreignmaterial such as sand, rust scale, or welding beads inthe gas supply; then a filter or strainer should be providedon the inlet side. The meter should not be installed atthe low point since it may act as a trap for liquids. By-pass piping or some other type of testing components,pressure taps, and over-speed protection should beconsidered.

Avoid having the meter body come in direct contact withsoil or concrete walls since alkali in concrete can causepremature corrosion. Under no circumstances should themeter be buried.

Hand-tighten the swivels first and then tighten with awrench approximately three flats (approximately 20 ft-lb). Do not over tighten since damage to the rubbergasket inside the swivel cap may occur.

Check all connections for leaks.

Slowly pressurize the system.

MAINTENANCE

The diaphragm meter really does not require muchmaintenance other than a periodic proof test. Factorsthat affect meter accuracy include:

1. Internal Friction that increases meter differential.Excessively dirty or tacky valves or binds in the meterwill cause higher differential pressures.

2. Maintaining Constant Diaphragm Displacement. Aprecise and stable diaphragm displacement is requiredfor each stroke of the meter. Therefore, the effectivecross-sectional area of the diaphragm and the diaphragm

stroke must remain constant.

3. External Leaks. Any opening, such as cover gaskets,index seal box, or meter connections that lets gas escapewill affect its accuracy.

4. Internal leaks will cause the meter to run slow and areusually found in areas such as the diaphragm assembly,valves, or flag rod seals.

5. Wear can affect accuracy in several ways. Wear ateither end of the short flag arm or in the tangent bearingwill cause the check rate proof to decrease while notappreciably affecting the open rate proof. Wear of thecrank or crank arms will affect the timing of the valvesthat will increase the open rate proof.

To determine the accuracy of a diaphragm meter, it mustbe tested or proved using conventional testing equipmentsuch as a transfer prover, bell prover, or sonic nozzle.(See Figure 8. for examples.)

a. Transfer Prover

b. Bell Prover


CONCLUSION

Diaphragm Positive Displacement meters are and willcontinue to play a major role in natural gas measurementin the U. S. for the foreseeable future. Understandinghow they work, their specifications, and how to installand maintain them are critical to having them perform inan acceptable manner.

c. Sonic Nozzle

FIGURE 8. Proving Methods

REFERENCES

-Gas Measurement Manual, Part Two, DisplacementMeasurement, American Gas Association, 1985.

-ANSI B109.1, 2000, Diaphragm Type Gas DisplacementMeters, American gas Association

-ANSI B109.2, 2000, Diaphragm Type Gas DisplacementMeters, American Gas Association

-ANSI B109.3, 2000, Rotary Type Gas DisplacementMeters, American Gas Association

-Product Specification, Aluminumcase Meter, DomesticSize, American Meter Company

Robert Bennett


FUNDAMENTAL PRINCIPLES OF ROTARY DISPLACEMENTMETERS

Wayland SlighInstromet Inc.

Stafford, Texas 77477

Natural gas measurement today is accomplished throughthe use of two different classes of gas meters. Theseare inferential type meters, which include orifice andturbine meters, and positive displacement meters, whichinclude diaphragm and rotary displacement meters. Theinferential type meters are so-called because rather thanmeasuring the actual volume of gas passing throughthem, they “infer” the volume by measuring some otheraspect of the gas flow and calculating the volume basedon the measurements. The positive displacement typemeters are so-called because they measure the actualvolume of gas displaced through them.

The rotary positive displacement meter has been inexistence for over 75 years. Its reliability, rangeability,long-term accuracy, and ease of installation,maintenance and testing have made this meter a favoriteamong gas utilities for billing purposes in industrial andcommercial applications. Rotary Meters have alsogained popularity in the production and transmissionmarkets

This paper will present basic operating principles of rotarygas meters, accuracy and rangeability, installation ofmeters, maintenance and testing, meter instrumentation,and finally, a brief glimpse at the industry trends in rotarygas metering.

PRINCIPLES OF OPERATION

The lobed impeller type rotary meter consists of twofigure-eight shaped impellers, positioned at 90∞ fromeach other, which rotate in opposite directions inside acylinder of fixed volume (Fig. 1). Gas flowing through themeter causes the impellers to turn, creating ameasurement chamber bounded by the impeller, cylinderand the headplates. This known volume is thendischarged and another identical volume of gas istrapped by the other impeller, cylinder and headplates.Gas is alternately trapped and discharged four times foreach impeller revolution.

The displaced gas per revolution is multiplied by thenumber of impeller revolutions to determine the volumeof gas passed by the meter. A gear reduction system isused to totalize the displaced volume for instrumentdrives and counter readouts in engineering units (i.e.,cubic feet).

Position 1 Position 2 Position 3 Position 4

FIGURE 1.Rotary Positive Displacement Meter Operating Principle

POSITION 1. As the bottom impeller rotates in acounterclockwise direction towards a horizontal position,gas enters the space between the impeller and cylinder.

POSITION 2. At the horizontal position, a definite volumeof gas is contained in the bottom compartment.

POSITION 3. As the impeller continues to turn, thevolume of gas is discharged out the other side.

POSITION 4. The top impeller, rotating in oppositedirection, has closed to its horizontal position confininganother known and equal volume of gas.

SIZING OF METERS

The sizing of a rotary meter is simply the selection of theappropriate capacity meter for the given flow conditions.The actual flow rate should not exceed the rated capacityof the meter. Line pressure and line temperature shouldbe considered.

Applying the Basic Gas Laws, the following formula maybe used to size a rotary meter:

Qs = Qd X Fp X F t

Where:

Qs = Standard or corrected volume

Qd = Displaced or uncorrected volume

Fp = Pressure correction factor =

Gauge Pressure + Atmospheric Pressure4.73 psia

Ft = Temperature Correction Factor =


520γ R .460γ R + Gas Temperature

ACCURACY AND RANGEABILITY

The accuracy of a meter is defined as the degree to whicha meter correctly measures the volume of gas passingthrough it. Accuracy is determined by comparing thevolume registered by the meter with a known volumeregistered by a connected proving device.

The accuracy of a rotary meter is built-in through thecareful machining of its components and cannot beadjusted. Since the volume of the measurement chamberdoes not change, the only factor that can affect accuracyis an increase in internal friction within the meter whichallows gas to slip through the clearances. A typicalaccuracy curve is depicted in Fig. 2.

The rangeability of a meter is defined as the meter’smaximum rated capacity divided by a selected “minimumcapacity.” Assuming that a meter “runs” at 100%accuracy, minimum capacity is determined to be thepoint where the meter’s accuracy moves above or belowa specified tolerance (usually ± 1% or ± 2%) from the100% accuracy point.

Newer designs of rotary meters, using lighter materials,have greatly improved rangeabilities by reducing startrates and thus pushing the lower end (knee) of theaccuracy curve to the left.

INSTALLATION

Proper installation in accordance with the manufacturer’srecommendations is necessary to ensure optimalperformance for any rotary gas meter. For the mostsatisfactory operation, rotary meters should be placedwhere they have the best chance of remaininguncontaminated.

Prior to meter installation, the line should be cleaned ofpipe dope, weld slag, liquids, sand, valve grease or otherdebris. The meter should not be installed at the lowpoint in the piping system where liquids or particulatematter tend to collect, or behind a lubricated valve whosegrease could block the impellers and cause the meter tostop. If particulate matter is in the gas stream, a suitablestrainer, screen or filter should be used.

Particular care should be exercised in installing the meterwith proper support. Piping connected to the metershould be supported to prevent strains on the metercylinder. Placing one or two flexible couplings in thesystem can eliminate possible piping strain.

Proper leveling of the meter is important duringinstallation and will maximize meter life and efficiency.Leveling not only reduces the possibility of meter bodystress due to flange misalignment, but also may preventoverfilling of oil chambers that may be reading low onone end due to improper leveling.

Where a meter may be subject to possible overspeedingbecause of sudden pressure drops or flow surges, arestricting orifice should be installed downstream of themeter in accordance with the manufacturer’srecommendation. A properly sized orifice plate or nozzlewill protect a meter from damage due to overspeed byrestricting the gas flow to 120% (typically) of the meter’srated capacity.

Most manufacturers recommended installing the meterusing a bypass loop. This allows for easy maintenanceor testing without the interruption of service, as well as asmooth, controlled start-up of the meter. Typical pipingarrangements are shown in Fig. 3.

MAINTENANCE AND TESTING

Very little field maintenance is required for rotary metersif care and proper installation procedures are followed.A periodic check and maintenance of the proper oil levels

FIGURE 2. Typical Accuracy Curve


is all that is needed under normal operating conditions.The interval between checks can vary substantially andmay be dependent on utility requirements and/or thecondition (cleanliness) of the gas. Oil which may havebecome discolored because of dirt or emulsified due towater, should be changed. A normal or typical timeinterval is once every 3 to 5 years.

In addition to the visual check of the oil level andcondition, a differential rate test may be performed onthe rotary meter. The differential rate test is an accurateand acceptable method of comparing the performanceof a rotary meter to its original, manufactured or installed,performance.

This test is based on the premise that the differentialpressure between the meter inlet and outlet will not changesubstantially — over 50% of its original reading — unlessthe meter parts wear or become dirty, given that thepressure, temperature, specific gravity and flow rate asinitially tested remain relatively unchanged. For anaccurate baseline curve, the differential pressure shouldbe checked and plotted for several gas flow rates (at leastthree between 25% and 100% of capacity).

As with the oil level check, the recommended frequencyof a differential test is subjective. However, experiencehas shown that a five-year interval is usually more thanadequate. If the test indicates an increase in differentialpressure over 50% from the initial test, the meter shouldbe checked for the causes of increased resistance.

Possible causes may be worn bearings, build-up ofdeposits on impellers, casings, or mechanical parts,incorrect oil or volume of oil, or out-of-time impellers.Usually, the meter may be brought back into specificationwith a simple flushing of the cylinder, which removesdust and materials collected on the impeller and cylindersurfaces.

METER INSTRUMENTATION

The volume of gas which has passed through a meterand has been measured must be totalized and registered.This may be done with an index. The simplest type ofindex is a mechanical counter, which providesuncorrected volumes at line conditions. A counter is usedwhen the line pressure is low and temperature is assumedto be at base conditions. Gas volumes at constantpressures and temperatures other than base conditionsmay be corrected by applying fixed-factor multipliers tothe index reading.

Where gas temperatures vary significantly, affecting thevolume of the gas, a temperature compensated indexmay be used. The mechanical temperature compensator,or TC, corrects the volume counter to a standard volumeat a base temperature of 60˚F. In addition to the correctedreadout, the index also contains an uncorrected readout.As with the counter, a fixed-factor multiplier may beapplied to correct the volume of gas at a constantpressure other than base pressure.

On an Instrument Drive (ID) unit, a spur gear reductionwith the proper gear ratio rotates a drive dog. Onerevolution of the drive dog represents a certain displacedvolume — either the uncorrected volume registered witha counter or the temperature corrected volume registeredwith a TC.

Rotary meters may also be equipped for automatedmeter reading (AMR). Pulsers generate either high orlow frequency pulses which represent volumetricinformation for remote data collection units.

Because gas volumes are subject to the effects ofpressure as well as temperature according to Charles’and Boyle’s laws, instrumentation may be used tocompensate for these effects. With fluctuating pressuresand temperatures, it is desirable to use instrumentationwhich corrects the volume measured at line conditionsto base conditions. This instrumentation may bemechanical or electronic and may be integrally mounted,remotely mounted or mounted atop an instrumentmounting plate which is part of the instrument drive unit.

Today, rotary meters are available with integral electronictemperature and/or pressure correction. This designinnovation combines the high accuracy of the rotarymeter with the increased measurement accuracy, greatflexibility, and economy of electronic volume correctors.

FIGURE 3. Typical Piping Arrangements


With integral instrumentation, the elimination of gearing,bearings, seals and shafts has improved the combinedaccuracy of meter and instrument.

RECENT DEVELOPMENTS AND INDUSTRY TRENDS

The last two decades have brought about many changesin the rotary gas measurement business. Increasedcustomer demands, as well as changes in governmentregulations have driven these changes. The mostsignificant of these is the replacement of the long-usedmechanical corrector by electronic correction devices.Advances in electronic technology have made it possibleto provide cost effective electronic gas volumecorrection, resulting in increased measurement accuracy,better long-term stability, quicker calibration andelimination of the mechanical gear reduction system.

Significant improvements in rotary meters have beenrealized in the last few years because of improvementsin manufacturing techniques and the manufacturingequipment’s ability to hold closer tolerances. Bettermachinability has lead to improved rangeabilities.Additionally, new designs for high capacity meters havereduced the sizes and weights of once necessarily

foot-mounted, cast iron meters, and made them line-mounted aluminum meters.

Many utilities are using rotary meters with prefabricatedmeter sets to reduce field labor installation costs and times.Quick change rotary meter conversion kits, designed toeliminate the need for field welding, pipe cutting, and othercostly field exercises, are based upon existing meter pipesizes and flange dimensions to provide faster, lessexpensive change-outs of other types of meters.

SUMMARY

Rotary positive displacement gas meters have been inuse for more than 75 years. In that time, they havebecome the preferred method of measurement forindustrial and commercial gas loads. The rotary meteris designed to measure gas with a very high degree ofaccuracy and reliability over time. They also offer theunique-to-rotary-meters differential testing capability,which reduces the need to shut down for testing andlowers the whole-life cost of ownership. Advancementsin technology continue to improve meter performancemaking rotary gas meters a vital element in gasmeasurement today.


FUNDAMENTAL PRINCIPLES OF GAS TURBINE METERSWayland SlighInstromet Inc.Stafford, Texas

INTRODUCTION

Gas Turbine Meters have developed greatly since theirintroduction to the US 1963. From the mechanically geardriven version, meters have developed into fullyelectronic designs and self-correcting models. Althoughthese technological developments have greatly improvedthe application of the meter, the meters basic designand principles have remained very similar.

As an inferential meter, the gas turbine meter competeswith the rotary and diaphragm positive displacementmeters. Like these meters, the turbine meter is versatileand accurate over a wide range of flowing conditions.Unlike these meters the turbine meter provides lesspressure drop for equivalent flow rates. It also providesa digital pulse output for input with flow computers andlocal readout devices.

Accuracy, the meters ability to measure as close to actualflow as possible, are typically in the ±1% range or better.Repeatability, the meters’ ability to give the same readoutunder similar flowing conditions, is listed as ±0.1%.

The meters ability to hook up with local pressure andtemperature correctors as well as flow computers allowsthe full versatility of the meter to be utilized; to providecorrected flow output and higher accuracy’s across itsflow range.

OPERATION/METER CONSTRUCTION

As previously stated the Gas Turbine meter is aninferential meter that infers volume by sensing flowvelocity through a known area by measuring the rotationalspeed of a turbine rotor.

As gas enters the meter inlet, it is guided through a flowarea created by the meters’ inside diameter and theoutside diameter of a flow deflector (Figure 1). The flowdeflector, or nose cone, increases gas velocity to prepareit for impingement on the rotor blades.

The accelerated gas flows through the meter rotor,designed with a specific number of blades positioned ata precise angle, and impinges on the rotor blades,causing the rotor to rotate. Rotor angular velocity isdirectly proportional to gas velocity and consequentlyto volumetric flow rate.

By accurately measuring the rotor rotational velocity,volumetric flow rate can be determined:

FIGURE 1. Turbine Meter Construction

Q = VA

Where Q = Volumetric flow rateV = VelocityA = Flow area

ROTATIONAL MEASUREMENT

In the U.S., rotor rotation measurement is accomplishedin one of two primary ways:

• Mechanical• Electromagnetic Proximity

Mechanical meters use a gear train to determine the rotorrotation. Gear ratios are such that the rotor revolution ismechanically connected to a rotating mechanism on themeter top. This directly indicates total flow in actualvolume flowrate.

The electronic meter is equipped with two magneticpickup coils that sense rotor rotation through accuratelypositioned magnets in the rotor hub. As a magnet passesthe coil, a voltage pulse is produced that represents adiscrete unit of volume. The turbine meter output is ratedin pulses per unit volume.

Independent pickup coils can be used with a localtotalizer and/or flow instrumentation capable ofaccepting inputs from pressure, temperature and densitytransducers for corrected flow or mass calculations.

Bearings typically are lubricated through an external fittingor a pressurized connection to allow lubrication duringoperation. Permanently lubricated bearings are availablefor vertical installations or where access is restricted.

INSTALLATION

AGA Report No. 7 provides standard practices andprocedures for installing gas turbine meters and details


correction factors and operating requirements. Severalcommon installations are described in the standard.

In-Line meters. To maintain fluid velocity profile andmeter accuracy, it is recommended that an upstreammeter tube be installed. Minimum distances are 10 pipediameters upstream and five diameters downstream ofthe meter (Figure 2).

To reduce swirl and maintain velocity profile, astraightening vane, or tube bundle, should be installed aminimum five pipe diameters upstream of the meter.

Offset, short-coupled meters. This installation has twoElbows, or Tee sections, connected to the meter inletand exit. A shortened inlet pipe section, of four pipediameters, is equipped with straightening vanes.

FIGURE 2.

VERTICAL INSTALLATIONS

Turbine meters can operate vertically up or down.Manufacturers offer permanently lubricated bearings ora pressure connection for lubrication during operation.Accuracy relies on an installation without offsets at theflanges and without protruding gaskets.

To protect internal components from foreign materialsin the gas stream, it is important that a strainer of filterbe installed upstream of the straightening section. A 1/8-in. perforated cone or basket strainer may berecommended. Alternatively, a filter capable of trappingdust particles up to 10 microns can be used.

A pressure tap usually is provided on the meter body toallow static-pressure measurement at the rotor. Otherinstruments taps should be installed a minimum of twoto five pipe diameters downstream.

To control over-ranging the meter and damaging the rotormechanism, a downstream control orifice or venturinozzle is recommended. Although over-speeding therotor to 120 % of maximum capacity is permissible forshort duration’s, it is recommended that the meter berun within its normal working range.

Blowdown valves should be installed downstream of themeter and, ideally, downstream of the control orifice toprevent damage.

FIELD MAINTENANCE

Normal maintenance includes lubricating and spin testingthe meter. While a spin test does not produce absoluteproof of accuracy over the entire flow range, it does givean indication of the meter’s accuracy at low flowrates. Aspin test is performed at the factory before calibrationand shipment.

The spin-down time is the time the rotor takes to stopafter being spun by hand or air blast. A reduction fromthe original spin-down does not necessarily mean themeter no longer is accurate. It does indicate reducedrangeability or an accuracy loss at the minimum flowrate

A notable reduction in spin-down time means eitherdeteriorated bearings, or debris buildup in the bearings

Meter proving. Each meter is calibrated to a recognizedstandard at the factory. Calibration methods include lowpressure, sonic nozzle, transfer proving and bell prover.

Standard primary calibration methods are sonic nozzleand bell prover. Transfer proving is a common calibrationmethod that avoids resorting to the time-consumingprimary methods. In transfer proving, a master meterfirst is calibration curve with at least 15 points. Thismaster meter then is used to prove other meters. Typicalcalibration curves have five to ten calibration points atatmospheric pressure or the customer’s pressure choice.

The typical calibration curve shows the percentage flowrate versus the percentage error. Calibration curves alsoprovide the flow rate versus the K factor. A turbine meterhas an accuracy of ± to 1.0%. Its repeatability, the abilityof the meter to provide the same reading at similar flowconditions, is about ±0.1%.

CONCLUSION

Turbine meters are capable of handling wide flow rangesover a wide range of pressures and offer accurate, reliable

FIGURE 3.


measurement over many years. With the addition of newtechnology and the versatility of the meter it shall remaina major standard in the field of gas measurement.


FUNDAMENTALS OF NATURAL GAS SAFETYLinton T. Lipscomb

PGE-GTT7330 San Pedro, P.O. Box 400, San Antonio, TX 78292

INTRODUCTION

Natural Gas: A combustible mixture of methane andhigher hydrocarbons used chiefly as fuel and rawmaterial.

To safely produce natural gas and natural gas products,a basic understanding of the hazards of the material itselfand the processes required to bring it to market isessential. Let’s start out with the hazards of natural gasas it is in its raw field gas state:

1. Flammable/Explosive2. Hazardous Impurities

a. Hydrogen Sulfideb. Benzenec. CO2d. Liquid, Petroleum Gases (LPG’s)

Flammable/Explosive

Raw field gas can have a lower explosive limit less than0.5% in air and an upper explosive limit in excess of 40%in air. This is different from pure methane, which is themajor component of commercial natural gas.

Differences in raw field gas and methane are:

Raw Field Gas MethaneVapor Density 1.0 to 2.1 0.6Vapor Pressure 30 psi @ 60°F NA

To 765 psi @ 100°F

Solubility in Water Negligible Slight

Appearance Colorless Gas/ ColorlessCan Produce VaporCloud Resembling Fog

Odor None to Light NoneHydrocarbon Odor

Boiling Point Varies with Mix -258°F

Lower Explosive 0.4% 3.6-5%Limit

Upper Explosive 40% 15%Limit

Because of the densities of raw field gas components, itcan settle and form explosive mixture at ground level.

The colder the air temperature, the greater the probabilityof this occurring. Should leaks occur or line depressuringbecome necessary, extreme caution should be exercisedbefore entering the area. Use of a L.E.L. meter isrecommended to determine if gas has settled in the area.Gas concentrations in excess of 10% L.E.L. should stopentry and all activity in the area.

Methane is lighter than air and rarely settles to the surfaceunless strong winds influence its ability to rise. It is stillan advisable step to check all work areas with a L.E.L.meter prior to starting work.

Hazardous Impurities

A. Hydrogen Sulfide (H2S): Gas containing this chemicalis often called “sour gas” because of its characteristicrotten egg odor. This gas is an extremely fast-actingtoxic substance. The NIOSH 8 hour exposure levelis 10 ppm. A maximum allowable short-term exposurelevel, (STEL), of 50 ppm for 15 minutes is allowable,if no other exposure occurs. The reason for no otherexposure if you are exposed at the STEL level is thatthe effects are cumulative. The toxic effect will destroyyour ability to detect the odor at concentrationsabove 100 ppm. At concentrations above 800 ppmto 1,000 ppm, a one-breath exposure can causeunconsciousness and death unless rescue isimmediate, followed by correct medical procedures.It is estimated that 50% of all hydrogen sulfide deathsare rescuers trying to make a rescue without properrespiratory protection. Never enter an atmospherethat is immediately dangerous to life or health withoutpositive pressure air-supplied respiratory protection.

B. Benzene (C2H2): Benzene is a liquid that may bepresent in small amounts in raw and pipeline qualitynatural gas. It rarely is found in quantities that wouldadd to the flammability or explosive characteristicsof the total mix. Its primary hazard is that of a long-term health hazard. OSHA has determined thatexposures in excess of the threshold limit valve(T.L.V.) of 1 ppm in the work place are hazardous.Extensive safety precautions described in 29cfr1910.1028, (Benzene), are all required by OSHA ifthe “Action Level” air-borne concentration of Benzenein the workplace exceeds 0.5 ppm based on an 8hour time weighted average. Long-term exposuremay cause liver, kidney, and brain damage, as wellas leukemia.


C. Carbon Dioxide (CO2): Inhalation exposures over5,000 ppm (0.5% in air) can produce overexposuresymptoms such as rapid breathing, rapid heart rate,headache, sweating, shortness of breath, mentaldepression, visual disturbances, shaking, uncon-sciousness, and death. Concentrations of 50,000 ppm(5% in air) are considered immediately dangerous tolife and death. The hazard of the CO2 at 5% in air iscompounded by its displacing the oxygen content inair. This gas is heavier than air and can accumulatein closed spaces and low places. This gas is colorlessand odorless. A relatively small leak of almost pureCO2 at an extraction plant can cause an extremelyhazardous situation at indoor equipment andcompressor locations.

D. Liquefied Petroleum Gases (LPG): Other chemicalsentrained in raw natural gas are referred to as naturalgas liquids. Butane, N-Butane, Propane, ISO-Butane, ISO-Pentane, and Ethane, as well as naturalgasoline can be present in raw field gas. Thesechemicals all have vapor densities heavier than air.Leaks or intentional blow-down can create explosivemixtures at ground level. Fires at or near LPGfacilities, storage tanks, and transport containers areextremely hazardous. Should the equipment ortankage be exposed to enough heat to weaken themetal of the equipment or tankage to the point offailure, the resulting release can be catastrophic.The energy released from the expanding LPG as itvaporizes and ignites is monumental. These specificfailures are referred to as a B.L.E.V.E. (Boiling LiquidExpanding Vapor Explosion). Equipment pieces, tankends, etc., can be propelled up to several thousandfeet and trailed by an extremely large fireball. Unlessspecific equipment training and large amounts ofwater are available and can be delivered correctlyon the involved equipment, evacuation should beginimmediately.

To produce pipeline-grade natural gas from raw field gas,certain processing must take place.

I. DehydrationII. H2S removalIII. CO2 RemovalIV. Liquid extraction

If you work with, in, or in close proximity to any of theseprocessing facilities, you should become familiar withthe following:

I. HazardsII. Operations’ ProceduresIII. Emergency ProceduresIV. What Personal Protective Equipment is

Required for Working In or Near TheseFacilities

At this point in time, technology is expanding our enve-lope of knowledge faster and faster. New equipment,

techniques, procedures, and applications are beingdeveloped with ever increasing speed.

To keep up with new technology and maintain a safe workenvironment, four areas must be addressed:

I. Knowledge

A. Basic knowledge of natural gas industryassociation standards API, AGA, NPRA.

B. In-depth knowledge of operations policies,procedures, and equipment.

C. Remain technically up to date.

1. Operational changes.2. Equipment changes and how they influ-

ence operations.

D. Governmental regulations

1. Federala. OHSAb. EPAc. DOT

2. Statea. TNRCCb. RRC

3. Local city and county regulations

II. Ability

A. Mental ability and skills to learn.B. Physically able to perform tasks requiring

strength, agility, etc., to match tasks.C. Ability to think through what the results of

your actions will be.

III. Accountability

A. Management Commitment1. Time2. Money3. Resources

To build, man, and operate safely.

B. Supervision1. Knowledgeable and well trained in all

aspects of his or her job.2. Has management backing.3. Is aware of the responsibility for his or her,

as well as employees’ actions.4. Can communicate well with his superiors

and employees.5. Sets good example for employees.6. Give employees time to complete job

safely.


C. Employee1. Well trained2. Understands his or her job and accepts

responsibility for getting it done.

IV. Attitude

The individual’s attitude about working safely isthe key to working safely in any industry.

A. Do it right every time.

B. If you don’t know, ask. Special emphasis tonew employees.

C. Use the correct tools and equipment to dothe job.

D. Work as a team—communicate.

E. You are your brother’s, or sister’s keeper. Yes,even contractors.

1. Help those who need assistance.2. Help correct improper actions of others.3. Create an atmosphere of looking out for

each other.

F. Housekeeping: Keeping your area clean,organized, and safe. It’s your home away fromhome — keep it clean, neat, and orderly.

G. Last but not least, common sense and cour-tesy to others.

1. Fellow employees2. Customers3. Contractors4. Public officials5. Land owners


FUNDAMENTAL PRINCIPLES OF PRESSURE REGULATORSPresented by Kevin ShawActaris Metering Systems

970 Hwy. 127 North, Owenton, Kentucky

INTRODUCTION

The following paper will concentrate on the fundamentalsand principles of natural gas pressure regulators. In thegas regulator’s conception it was mainly a device usedto reduce high pressure to a more usable lower pressure.Today, more is expected from the performance of thepressure regulator. Pressure reduction is no longer theonly function needed. The regulator is considered anintegral measurement instrument that must adhere tothe stringent codes put forth by the U.S. FederalDepartment of Transportation and many state PublicUtility Commissions.

In order to understand the principles of pressureregulation this paper will focus on:

I. OPERTIONII. DROOP VS. BOOSTIII. CHANGING IN INLET PRESSURE VS.

MECHANICAL ADVANTAGEIV. SAFETY MECHANISMSV. KEYS TO SELECTIONVI. CONCLUSIONS

REGULATOR CLASSIFICATIONS

With few exceptions, gas pressure regulators can beclassified into either of the following two categories:

1. Self-Operated Regulators (also known as Spring-Loaded)

2. Pilot Regulators• Constant-Loaded (a.k.a. Pilot-Loaded)• Pilot Operated (a.k.a. Two-Path Regulation)• Pilot Unloading (a.k.a. Pressure Unloading)

The Pilot Regulator category can further be classified intothe three sub-categories of Constant-Loaded, PilotOperated, and Pilot Unloaded. Make no mistake; thesethree design are vastly different and, thus, will exhibitsignificantly different performance characteristics. Eachof these designs is covered in more detail later in the paper.

REGULATOR DEFINITION

A pressure regulator is a feedback control mechanismdesigned to maintain a constant downstream pressurethrough the manipulation of gas flow. By definition, aregulator is composed of three essential components(see Figure 1 illustration):

1. Restricting Element — A restriction which allowsgas to flow through the regulator at a reducedpressure to meet downstream demand. In mostcases this consists of a resilient valve seat (plug)and a sharp edged orifice.

2. Measuring Element — A device that continuouslysenses changes in downstream pressure caused bychanges in downstream demand and transmits asignal to open or close the restricting elementaccordingly. This is typically an elastomericdiaphragm.

3. Loading Element — Adjustable force which iscontinuously compared to the downstream pressureby the measuring element to determine what signal(open/close) to transmit to the restricting element.In pilot regulators the loading element is gaspressure.

SPRING-LOADED REGULATOR OPERATION

The main function of any regulator is to provide a flow ofgas through the regulator to match the downstreamdemand while holding pressure constant. In a spring-loaded regulator three devices are employed to achievethis — a restricting device (usually an orifice); a sensingdevice (diaphragm); and a loading device (spring orpressure). Tying these three things together is a

Restricting Element

Sensing Element

Loading Element

FIGURE 1.


mechanical linkage (usually called the valve stem and/or lever) that acts like a “see-saw.” The linkage works tobalance the forces associated with the three abovedevices.

A pseudo free-body diagram gives a representation ofthis balancing act is shown in Figure 2.

FIGURE 2. Regulator Force Balance

In all regulators two types of forces exist: 1) openingand 2) closing forces. These two forces act on themechanical linkage with one trying to close the valve(shutting off gas flow) while the other works to open thevalve (increasing gas flow). Under steady operation thesums of the opening and closing forces are always equalbut opposite in direction giving a static equilibriumcondition. As in any static equilibrium condition, the valvewill remain in a fixed position until one of the forceschange, upsetting the equilibrium. The valve will thenreposition again until the forces are again in equilibrium.In a spring-loaded regulator, the sum of forces on themechanical linkage can be expressed as follows:

Opening Forces = Closing Forces

or

Fi + Fs = Fo

Where:Fi = Inlet Pressure ForceFs = Spring ForceFo = Outlet Pressure Force

This equation assumes there are no frictional effectsinside the regulator.

The opening forces consist of one mechanical and onepneumatic load. The high-pressure inlet gas creates apneumatic force (Fi) pushing on the face of the valveseat forcing open the valve. An adjustable spring force(Fs) assists the high inlet pressure by pushing on thesensing device (diaphragm), opening the valve tomaintain the set pressure. These forces can be calculatedas follows:

(Hooke’s Law)

Where: K = Spring Constant (lb./in.)x = Spring Compression (in.)

Where: P1 = Inlet PressureAo = Orifice Area

The downstream pressure under the diaphragm createsa closing force (Fo) pushing against the diaphragm tryingto close off the flow of gas. This force is calculated as:

In normal operation, the diaphragm will sense the outletpressure force change and provide a force to the linkage.The linkage moves to control the flow through the valveto maintain the set outlet pressure. For instance, if theoutlet pressure drops from the set pressure, the forceunder the diaphragm Fo decreases allowing the springforce Fs to reposition the diaphragm. This downwardtravel acts to move the valve seat away from the orifice,bringing the outlet back to approximately the desiredset pressure. If the outlet pressure increases, the reversehappens. The linkage responds to the increased outletpressure force and tends to restrict the orifice. The flowis reduced and the outlet pressure once again returnsapproximately to the set pressure.

In a perfect world our loading element (the spring) wouldsupply a constant force, there would be no friction withinthe regulator or material hysteresis. If this were the casethe regulator would supply a constant outlet pressureover an infinite range of gas flows. However, thisperformance is unattainable. Hindrances like 1) springforce linearity (Hooke’s Law) and 2) the change in effectivediaphragm area as the valve travels, effect theperformance of the regulator. In Figure 3 typical spring-loaded regulator performance is shown.

FIGURE 3. Typical Spring-Loaded Regulator Performance

The performance loss due to these effects is calledpressure droop. Droop, and outlet pressure loss belowset with increasing flow, is an ever-present problem in

Fs = [ ][ ]K x

F Ai o= [ ][ ]P1

Fo =

outlet

pressure

Effective

diaphragm

area

X


the performance of the regulator. To combat the effectsof droop, a method called “boost” is employed. A betterunderstanding of these principles will follow in the nextsection.

FIGURE 4. Inlet Pressure Effect on Performance

DROOP VS. BOOST

One contributor to droop is spring effect. Spring effectis a result of the reduction in spring force when thediaphragm and valve are in the wide-open position (fullydown). The reduction in the spring force requires a loweroutlet pressure to balance it. The second contributor todroop is diaphragm effect. As with spring effect, the wideopen valve position (fully down) is again the problematicdiaphragm position. The effective diaphragm areaincreases giving the outlet pressure more area to actupon, which in turn means a lower outlet pressure isneeded to balance the “see-saw.”

Acting together, both spring and diaphragm effect cancause a considerable mount of droop. To combat thisperformance deficiency, raising the outlet pressure tocompensate is necessary. Boost is a method that usesgas velocity at high flow to create a low pressure underthe diaphragm relative to the downstream pressure. Thislower “sensed” pressure aids in the lowering of thediaphragm causing the valve to open and elevating theoutlet pressure. The boosting needed to overcome thenegative droop can be attained by various methods;angled valve seats, pitot tubes, and loading rings to namea few. Droop is not the only factor which effects the outletpressure adversely. Varying inlet pressure can be anuisance as well.

CHANGING IN INLET PRESSURE VS. MECHANICALADVANTAGE

As discussed earlier, one of the forces in the balancingact is the inlet pressure acting on the valve seat throughthe orifice opening. As the inlet pressure drops, the forcetrying to push open the valve also declines. This in turnallows the valve seat to reposition closer to the orificethus decreasing the flow rate and downstream pressure.Again, the regulator needs to overcome this obstacle.

Mechanical advantage can answer this dilemma. Thisadvantage, know as power ratio, directly relates to thediaphragm size and linkage ratio or the amount of travelof the diaphragm with respect to the valve seat. The largerdiaphragm gives more force at a given pressure andcombined with the lever ratio will equalize the outletpressure over a range of changing inlet pressures.

Figure 4 shows a typical example of the effects ofchanging inlet pressure. Raising the inlet pressure fromset pressure (curve “set”) tends to increase the outletpressure (curve “↑”). A reduction in inlet is observed inthe bottom curve (“↓”).

SAFETY MECHANISMS

Under normal operation a regulator will deliver servicearound the clock. But in the event of a regulator failure,what type of safety features exist?

Even though a regulator is built to give a constant outletpressure, what happens if the valve fails to make thepressure reduction? An internal relief valve is incorporatedinto most self-operated regulators that have no othermeans of pressure relief.

In the instance that a foreign object gets trapped betweenthe valve seat and the orifice preventing a positive shutoff, pressure can continually build exceeding the setpressure.

If this pressure is allowed to increase, a dangerouslyelevated pressure situation will occur. This is where aninternal relief valve steps in. The relief valve (Figure 5),usually a spring-loaded device, will allow the outlet tobuild to a small pre-determined amount over the setpressure before it opens, relieving gas through thediaphragm to the atmosphere. This relief pressure istypically not adjustable. The “odorous” gas smell will alertthe customer and a service call will be made to repairthe malfunction.

FIGURE 5. Internal Relief


Another built-in safety feature is the restricting orifice.As gas passes through the reduction in area there is anatural pressure drop. This drop varies with the area ofthe orifice opening. The smaller the opening, the largerthe pressure drop. This translates to a smaller pressurebuildup downstream in a failure situation. Safety is thekey when dealing with a device that regulates the deliveryof volatile gases to business and homes around the world.

PRINCIPLES OF PILOT REGULATORS

Why Use A Pilot Regulator?

Self-operated regulators may take a variety of forms.They can be single ported, double ported, balanced orunbalanced. All types have one common limitation whichis “spring effect,” resulting in some control inaccuracy.Self-operated regulators are exceptional devices for theirsimplicity, performance and cost. However, when moreaccurate regulation is desired, spring effect along withother mechanical factors will cause the self-operatedstyle to fall short of expectations and needs. These short-comings can be solved, by replacing the spring with amore constant diaphragm loading force. The result is asignificant accuracy increase throughout the entire rangeof flow as shown in Figure 6.

CONSTANT-LOADED REGULATORS

A constant force from gas pressure can be used in placeof a simple loading spring to eliminate this “spring effect.”Constant-loaded regulators were designed with “straightline” regulation in mind (Figure 6).

FIGURE 6. Spring vs. Pilot Regulator

Constant-loaded regulators use the outlet pressure froman additional pilot regulator as the constant loading forceneeded (Figure 7). The pilot regulator can supply avirtually constant pressure to the diaphragm throughoutits range of movement. This constant force adds to theusable capacity range of a regulator by eliminating any“droop” caused by the loading spring. Outlet pressurewill remain stable for essentially the full valve travel ofthe main regulator, achieving accuracy of ±1% of theabsolute outlet pressure or better. Increased flow ratesare also expected results with the constant-loadedregulators. The increased capacity does not occurwithout some disadvantages, however.

FIGURE 7. Pilot Loaded Regulator

Constant-loaded regulators respond poorly to fast off or“shock” loads. The nature of the design causes slowresponse to sudden load decreases. When the regulatoris asked to shut off flow quickly, it takes time for thispressure difference to be communicated to the pilotthrough the bleed hole in the main diaphragm. The pilotrelies on its pressure sensing through a bleed hole in thediaphragm and thus is not directly sensing andresponding to the downstream pressure.

Constant-loaded regulators also require higher lock-uppressures under “no flow” conditions. The increasedlock-up is needed to fully close the main valve as well asthe pilot valve. This adds to the additional lock-up forceneeded. In special applications these characteristics cannot be tolerated, therefore the need for a spring-loadeddesign is required.

PILOT-OPERATED REGULATORS

To obtain the advantage of both the spring-loaded andconstant-loaded regulator performance, one mustexplore the use of “pilot-operated regulators.” With theuse of a more complex pilot and pressure sensing lines,the pilot-operated regulator can make rapid changes todownstream demand since the pilot is continuouslysensing the downstream pressure. However, this designis still slower than the spring-loaded design due to thenumber of mechanical steps the regulator must undergoin order to open or close. Due to the gain and accuracyof the pilot, these regulators can control pressure withflat line response without many of the shortcomings ofthe constant-loaded style.

There are two basic types of pilot-operated regulatorsthat will be discussed in the following:

1. Dual-path control2. Pressure Unloading system


DUAL-PATH CONTROL

The first style of pilot-operated is a dual-path controlsystem (Figure 8). The dual-path system incorporates afixed and a variable orifice within the control system.The loading pressure that is generated between the fixedand variable orifice gives the pilot its “gain” and abilityto quickly position the main valve with accuracy. With asmall downstream pressure change, a high gain pilot cancompletely position the main operating valve from closedto fully open. This quick re-positioning of the valve MUSTbe controlled to insure stability.

As the load varies, the main valve will tend to respondquickly to re-position its valve; however, it is uncontrolledand can become unstable. The solution to this problemis to combine a high gain pilot with a slightly dampenedmain valve to give accurate and stable control. Thesensing of the downstream pressure coupled with theloading pressure in the main case provide what mightbe called a shock absorbing quality. However, at higherinlet pressures some regulators of this design can stillbecome unstable due to the high gain of the regulator.In these cases a pilot supply regulator must be employedin order to reduce the inlet pressure supplied to the pilotand, thus, desensitize the unit.

PRESSURE UNLOADING SYSTEM

The unloading regulator system is used mainly in theflexible element valves (Figure 9). The pilot functions tocontrol the differential pressure across the flexibleelement causing it to open or close depending on thedemand. This system also employs a fixed restrictionand variable restriction. The fixed restriction is typically“tuned” to achieve a particular rate of response.

Upon a downstream demand increase, the pressure dropis sensed through the static line in the control chamber.Next, the pilot adjustment spring pushes the seat off the

bleed orifice allowing the loading chamber pressure toescape downstream. The bleed orifice now exhaustsloading pressure faster than it can be replenished dueto the fixed restriction. The bleed orifice is then controlledby the pressure difference across the measuring elementbetween the loading chamber and the control chamber.Finally, the reduced pressure behind the flexible elementcauses the flow to increase around the main valve’sannular seal restoring the set pressure.

Upon a no flow condition, outlet pressure rises forcingthe pilot seat to close the variable pilot orifice. Thepressure in the loading chamber equalizes with the inletpressure across the fix restriction. Finally, the resiliencyof the rubber forms a positive no flow condition in themain valve.

The unloading pilot can supply exceptional accuracy onlywhen the system is properly tuned to its application.Certain situations will cause stability problems that canbe addressed by a few methods. One method is toprovide an adjustable inlet restriction to control the speedof pilot response. Another method regulates the speedof response through an independent proportional bandchamber. Both methods use pilot control to eliminatethe source of instability.

KEYS TO SELECTION

Matching the regulator to the application is mostimportant. Accurate knowledge of the system in whichthe regulator will be placed is very crucial. The followingare the variables that must be understood in order toselect an appropriate regulator:

• Minimum and Maximum Inlet Pressure• Minimum and Maximum Capacity• Rate of Load change• Load Diversity

FIGURE 8. Dual-Path RegulatorFIGURE 9. Pressure Unloaded Regulator


Values for the minimum and maximum inlet pressures,as well as the maximum capacity for the regulator, needto be clear and accurate. Other factors such as rate ofload change and load diversity must be take intoconsideration. Rate of load change refers to the on/offnature of the equipment downstream. For example, someboilers are equipped with “snap-acting” burner valvesthat open and close very quickly which then requiresthat regulator be able to respond at the same rate.Loaded Diversity refers to the probability that allequipment downstream will be operating simultaneously.If this probability is low, sizing the regulator for the “totalconnected load” may cause over-sizing of the regulatorand ultimately instability at low flow rates.

The following “rules of thumb” should be applied whensizing any regulator:

• Use minimum inlet pressure expected andmaximum flow required when consideringcapacity needs

• Choose the smallest orifice available that canmeet capacity requirements

• Use maximum inlet pressure expected whenconsidering “lock-up” and relief performance

• Choose the lightest adjustment spring availablethat can meet outlet pressure requirements (i.e.,choose set point at upper end of spring range)

The selection procedure then dictates the smallest orificesize to be implemented while allowing required flowcapacity at minimum inlet pressures. The smallest orificegives two benefits. First, it allows the smallest pressurebuildup downstream in the event of a failure when usingan internal relief type regulator. Secondly, it optimizesmechanical advantage by allowing less fluctuation inoutlet pressure for a given inlet pressure change.

Another good practice is to use the lightest adjustmentspring that will give you the outlet will minimize springeffect, which causes “droop.” In addition, the internalrelief valve has to work against the main adjustmentspring to open. The relief valve will open at a lower outletbuildup given the lighter adjustment spring.

For spring-loaded regulator, the power ratio of a regulatoris another selection criterion. This ratio, as previouslydiscussed, is based upon diaphragm size and linkageor lever ratio. It is a measure of how well the regulatorutilizes the downstream pressure to close against theupstream pressure. The larger the ratio the lower the lock-up pressure will be.

However, with increasing power ratios come costhindrances. A larger power ratio means a physically largerregulator and a higher price.

When considering rate of loaded change, beware of theorder of rate of response of various regulator types asshown in Figure 10.

CONCLUSIONS

Every type of regulator represents a compromiseinvolving factors such as price, capacity, accuracy,stability, simplicity, safety and speed of response. Table1 below summarizes the advantages and disadvantagesof the various regulator designs available. It is important

Parameter SpringLoaded

ConstantLoaded

PilotOperated

PressureUnloading

TypicalApplication

ServiceRegulator

ServiceRegulator

ServiceRegulator/

District

District/Gate

StationRegulator

Stability Good Good Can beunstable

Can beunstable

MinimumDifferentialRequired

No Very little Very little Yes

Sensing LineRequired No No Some Yes

Sensitive togascontaminants

No No No Yes

Variable Gain No No Some YesSpeed ofResponse Fast Relatively

FastRelatively

Fast Slow

Susceptible toinlet pressureswing

Yes Yes Yes Yes

Product Cost Low Cost Moderate Expensive Expensive

Table 1 — Regulator Summary Comparison

Spring

Pilot (Constant) Loaded

Pilot Operated (two-path)

Flexible Element (boot)

FASTEST

SLOWEST

FIGURE 10. Regulator Rates of Response


Kevin Shaw

that manufacturers, utility engineers, and field servicepersonnel alike understand the workings of self-operatedand pilot regulators. A correct analysis of the applicationwill give the customer clean, safe, and accurately

controlled gas. Working closely with the manufacturerand matching the right regulator to the right applicationwill also lead to a problem free life.


FUNDAMENTALS OF EGM — ELECTRICAL INSTALLATIONSMichael D. Price

CenterPoint Energy Field Services525 Milam Street, Shreveport, LA 71101

INTRODUCTION

The areas of gas measurement and communicationshave seen substantial changes in the last few years asthe natural gas industry adapts to effects of the economy,fluctuating gas prices, warm winters and governmentderegulation. Every energy delivery company hasstudied, debated, hired consultants, and finallydetermined how gas flow data is to be measured andcollected. All gas companies have hundreds and eventhousands of points which must be accurately measured.Data is retrieved from very remote and rugged locations.Climate conditions can range from humid off-shoreplatforms to desert conditions with both temperatureextremes included. No commercial power is available,allowed or even desired at these locations making thesolar-powered electronic gas measurement equipmentthe ideal method of gathering flow data.

Companies have standardized on one or more modelsof EGM equipment that meets its particular requirements.Its associated components usually include the computer,differential pressure transmitter, pressure transmitter,solar-panel, battery, and possibly a communicationssystem of some type. This could be a modem for atelephone line, radio data system, or cellular datacollection unit.

The manufacturers of the solar-powered computersprovide recommended installation procedures basedupon agency approvals that they applied for. There aremany other techniques that can be utilized to insure thatoptimum performance is realized from the EGMequipment. Solar-powered EGM equipment must use alow amount of power from the battery source in order tokeep the PV system small enough to remain costeffective. Low power consumption makes the system ahigh impedance device. Solar-powered EGM equipmentrequires special attention to installation details for properoperation under most conditions.

This paper will look at the installation of solar-poweredflow computers of the most common configurationsavailable. There are several recommendations presentedto improve upon the manufacturer's installationprocedure based upon isolation, insulation, andgrounding.

ENCLOSURE MOUNTING

All solar-powered EGM equipment is designed to bemounted out in the open, exposed to the elements. It is

constantly exposed to the daily intense ultra-violet raysand infrared heat from the sun. The best of paintedsurfaces becomes faded and chalky after only a fewyears of exposure. The temperature extremes that occurwithin the enclosure literally “bakes” the semi-conductorcomponents and shortens the life-expectancy of allelectronic parts. Ideally, the EGM should be located in awalk-in meter house or mounted in a small metalenclosure mounted on the meter run, but realistically, itwill be mounted completely exposed. At minimum, aneffective sun-shield protecting the enclosure should beutilized. This raises the cost of each installation but thelonger life expectancy of the equipment will offset theinvestment.

Most enclosures have mounting hardware for attachingthe enclosure, to a vertical 2” diameter pipe usuallymounted over or to the side of the orifice fitting or thepositive displacement meter. This will insure that allmechanical and electrical lines will be as short aspossible. The 2" diameter pipe is the first area of concernbecause it will be a path between the EGM enclosureand the pipeline for electrical surges that will surely occur.

Insulating the enclosure from the pipeline can beaccomplished several ways. The first method to consideris to install an insulating union at the bottom of the 2”pipe where it attaches to the leveling saddle. The cost ofthe union is high and it's electrical integrity will eventuallydegrade. Other installations use a 2”, schedule 80 CPVCpipe to mount lighter weight devices, such as processtransmitters. It is not recommended for heavy enclosuresdue to its weakness where the threads are made andexposure to the elements will cause it to become brittle.A third way is the 2” pipe can be covered with a strongheat-shrink material used by the electrical industry toinsulate high voltage splices. It is RAYCHEM WCSM 68/22 EU 3140, a black tubing that will slip over a 2” pipeand then after being symmetrically heated by a rose-bud torch, will make a very strong insulated pipe stand.It must be acknowledged that if the manufacturer designsthe EGM to not use the case in the ground system(floating ground), there is no need to insulate theenclosure from the pipeline. It is still a good practice toisolate an EGM as much as possible. This provides anextra measure of protection above the manufacturer'srequirements.

If the EGM is to be mounted at a power generating plantor near over-head, high-voltage power lines, even moreprotection is needed. Ground currents around theselocations during an electrical storm are potentially


harmful. In some instances, EGM with an enclosure madeof heavy steel works better in these locations becauseof steel's ability to shield EMI (electromagneticInterference). Aluminum or a composite material may notshield EMI even though it may work well against RFI(radio frequency interference). Think of each wire enteringthe enclosure as an antenna.

The maximum humidity rating for any EGM is 95%condensing. The enclosure should have desiccantmaterial in the enclosure to absorb moisture. There is anew product called “Humidisorb” that replaces thecommon desiccant known as silica gel. It regeneratesitself and does not have to be “dried out” or replaced asoften as silica gel. If there is corrosion inside theenclosure, an inhibitor package should be replaced every2 years. Seal all openings and keep the enclosure doorsecure.

Use a carpenter's level to mount the enclosure straightand keep it at a good level to work with. The way aninstallation looks is a reflection on the technician(s) thatinstalled it and will be a source of pride when it looksprofessional. When an installation looks good it is usuallygoing to operate well.

PRESSURE TRANSMITTERS INSTALLATION

The installation of the differential pressure transmitterand pressure transmitter requires special attention if theyare not an integral assembly inside the EGM enclosure.Mount the transmitters on an insulated pipe stand toinsure that the transmitter mounting brackets are not apath for electrical surges from the pipeline. The pressuretransmitter is normally connected to the down-streampressure of the orifice fitting or to the case of a pulse,type meter. The pressure transmitter can also be mountedjust above the differential transmitter to minimize thetubing connection length and reduce the number ofisolating valves required.

A direct connection type manifold has the ability toconnect the transmitters directly to the orifice fittingwithout the use of tubing. It presents several advantagesdespite it's higher cost over an installation using tubing,fittings and valves. The direct connect manifold optionprovides a mounting for the transmitter(s), isolationvalves, unrestricted bore from the taps of the orifice fittingto the transmitters bellows, valving for testing, anddielectric isolation parts.

TUBING AND FITTINGS

The differential transmitter and the pressure transmitterwill require block valves, tubing and a five-valve manifoldfor testing and operation of the measurement equipment.Just as the 2” pipe used for mounting the EGM enclosurecan be an electrical path to the pipe line, so can thegauge lines from the transmitter to the orifice fitting. Evenif the meter run has flange insulating kits installed oneach end, they will deteriorate in time. To isolate thegauge lines between the transmitters and the meter-run

piping, di-electric fittings should be installed. For 3/8”stainless steel gauge lines, a common one to use is theCAJON SS-&DE-6, 3/8” tube to tube connector or theImperial Eastman P/N 962-DC-06xO6.

After installation of the transmitter, tubing, manifold anddielectric fittings, measure with an ohm-meter acrossthe dielectric fitting to insure that isolation has beenachieved.

TEMPERATURE PROBE AND TRANSMITTER

The most popular method of temperature measurementis to mount a 100-ohm or a 500-ohm temperature sensorin a thermowell downstream of the primary device andthe EGM electronics will convert the resistance changesfor the, corresponding temperature value. A cost savingsis realized because there is no need for a processtransmitter to convert the probe's resistance value to astandard voltage range that some EGM computersrequire.

Temperature Probe

The 100-ohm temperature sensor means that theresistance of the probe will measure 100.0 ohms whenit is exposed to 0 deg. C or 32 deg. F. The 500-ohmtemperature sensor will measure 500.0 ohms resistanceat 0 deg. C or 32 deg. F. They are not inter-changeable.The temperature sensor is also known as a ResistanceThermal Detector or just “RTD.”

The construction of temperature probes has the sensingpoint located at the tip of the probe. It is isolated by apotting material to keep it from contacting the protectivemetal sheath. The actual spacing between the sensorand the metal sheath is very small and could breakdownif the voltage surge is high enough. This may be theweakest point for isolation of the EGM from voltagesurges.

The probe is mounted in a thermowell and the tip of theprobe should touch the bottom of the thermowell. Thiswill insure good transfer of heat from the gas stream tothe probe tip. The interconnecting wiring from the probeto the EGM is surrounded by a flexible armoredprotection to prevent damage from normal use. Whenthe armored cable enters the EGM enclosure, it stopsinside an insulated compression connection. This is tosecure the cable and insulate the enclosure from thepipeline via the metal armored cable. Surround thearmored cable with a piece of plastic tubing from theenclosure to a point past any isolating devices to insurethe metal cable does not compromise the isolation fromthe pipeline. Make a resistance measurement betweenthe armored cable and the case to insure that theinsulation integrity is maintained.

Temperature Transmitter

If the EGM does not have a direct resistance input butrequires a voltage signal that is designated as the


temperature value, a process transmitter will be requiredto convert the resistance to a voltage range of 0.8 to 3.2volts or 1.0 to 5.0 volts.

Some temperature transmitters are mounted directly ontothe temperature probe and thermowell. This puts thetransmitter case in direct contact with the pipeline and itnow becomes an area of concern when installing EGMequipment. The connection size between the transmitteris normally 1/2” NPT threads. CPVC plastic collars andshort nipples can be installed for insulating the transmitterhousing from the pipeline (See Figure 1). The plasticCPVC parts are not as strong as steel and can break asthey age.

Solar Panels

The more sunshine the solar panel is exposed to, themore power they will produce to charge the battery.Locate the panel where there is nothing to shade itespecially during the peak sunlight hours of 9 a.m. to 4p.m. Even the thin shadow of an electric wire cansignificantly reduce the panel’s output.

In northern latitudes solar panels should face south andin southern latitudes should face directly north. Don’tguess at the direction to point the panel, use a compassto correctly orient the panel. Tilt the panels for maximumexposure. Mounting the panel at the best angle meanthat the panel will get maximum exposure to the sun.The best possible tilt angle puts the flat surface at rightangles to the midday sun.

If the panel is remote from the EGM and is to be clampedto a pipe that is connected to the pipeline, install a pieceof the heat shrink wrap around the pipe to insulate thepanel. Attach an insulated ground wire to the panel’smetal frame and connect it to the grounding system. Donot ‘daisy chain’ any ground system wiring.

Batteries

The most common type of battery used in solar powersystems is the deep-cycle, lead-acid battery. They aredesigned to be repeatedly cycled (charged anddischarged). An automobile starting battery is different.It is designed to provide a powerful, short burst of currentto get an engine started. At that point, the alternator takesover and recharges the battery. An automobile battery isnot designed for deep-cycle applications and willprematurely fail.1

The battery of choice for EGM solar-power systems isthe “gel cell”. This is a lead-acid battery with additivesthat turns the electrolyte into a non-spillable gel. Thesebatteries can be mounted sideways or even upside downif needed because they are sealed and with geltechnology, they will not spill. They are more expensivebut the savings in maintenance can justify the additionalcost.

Temperature will have the strongest influence on batterybehavior and life. High temperatures will increase thedeterioration in a cell due to accelerated chemicalreactions. On the other extreme, freezing of theelectrolyte can cause damage to the internal plates. Thebattery life is generally estimated to be between 3-5 yearsunder normal conditions.

Charging Systems

The charging system will prevent the over-charging ofthe battery during the intense summer sunshine and alsodisconnect the load when the battery approaches it’sminimum state of charge. When the battery is below 20%of it’s state of charge for a long period of time, the batterywill not recover to it’s previous capacity or it may fail.

FIGURE 1. Temperature Transmitter With CPVC Coupling

As with the enclosure and the pressure transmitters, thetemperature transmitter can be mounted on an insulatingpipe stand and the wiring between the temperature probeand the transmitter can be insulated by using all-plasticflexible conduit or by rugged inter-connecting wireaccepted for mounting intrinsically safe devices in ahazardous area.

SOLAR PANEL AND BATTERY

The power required to operate remote EGM equipmentcan be obtained from different sources but the mostcommon is the PV (photo-voltaic) system consisting ofa solar-panel, deep-cycle battery and a charging system.The size of the panel and battery is determined by severalfactors:

• Geographical location• Average current consumption• Number of days autonomy required• Voltage requirements

Many EGM manufacturers size their PV systems toprovide power to their equipment over an averagenumber of days without sunlight to recharge the battery.The is referred to as ‘autonomy.’ This can be seven, tenor even up to fifteen days. During January of 1992,northwest Louisiana and east Texas went 22 days withoutsunshine. PV systems failed and batteries were beingreplaced, recharged, or supplemented to insure that theEGM continued to operate. Size your systems with theworst case conditions and the extra cost of a larger paneland battery will be worth the investment.


Preventive Maintenance

Here is a check list of suggested preventive maintenanceitems to keep solar power systems in good condition:

1. Keep a record of the date of installation of the PVsystem and measure the voltage after five days ofoperation.

2. Summer time will cover up the performance of weakbatteries because of the intense sunshine and fewercloudy days. In the fall, begin to record the batteryvoltages just before sunrise. This will give you yourfirst clue to which batteries have low capacities.

3. Start a battery refurbishment program that willreplace batteries at a maximum of five yearsinstallation life.

4. Have an appropriate amount of spare batteryassemblies that are fully charged and ready for use.

5. Check all battery and solar panel connections forcorrosion. Repair and water-proof.

6. Keep the solar panel clean.

GROUNDING SYSTEMS

A good grounding system is vital to any EGM installation.All equipment must be connected to a low impedanceground system to provide transient dissipation. Earth/ground resistance testers are used to determine theresistance of earth surrounding a ground rod. It shouldbe less than 25 ohms. Some of the devices used to testthe ground resistance are:

1. “Biddle-250260”; James G. Biddle Co.2. “Vibrogound”; Associated Research, Inc.3. “Groundmeter”; Industrial Instruments, Inc.4. “Ground-ohmmeter”; Herman H. Sticht Co., Inc.

Resistance from all pieces of equipment to the earthground connection should be less than 1 ohm asmeasured with a quality DVM (digital volt meter).

The actual ground system may consist of a single groundrod, multiple ground rods, or an elaborate, groundinggrid system.

Ground Conductors

Every conductor has measurable inductance (resistanceto the flow of AC current not DC current). Solidconductors exhibit more inductance than stranded wireconductors but the most effective conductor is a copperstrap.4 The best practical compromise is to use #10 AWGcopper stranded wire with a green colored insulation forproper identity. If you use a bare copper wire, it maycontact piping and bypass all of your isolation efforts.

Soil Doping

The earth is a conductor due to the large amount of ionicsalts that exist in soil. Conductivity can be increased byadding more ions to the soil especially if the soil is rocky.

Augmentation back-fill material that is used by cathodicprotection systems is one method of increasing theconductivity but a simpler way is to add Epsom salt tothe soil around the ground radial.1 Depending upon howwet the soil is due to rainfall and water drainage, dopingthe soil with about 4 pounds of Epsom salt around theground rod should last about two years.

Ground Radials

Large gauge copper wire and copper-clad steel rodsmakes copper the most commonly used groundingmaterial. Joints between copper wire and the groundrods should be made by exothermic welds or by usingan anti-oxidant, joint compound such as the Ideal“Noalox,” in high-compression clamps. After theconnection is made, it should be wrapped and madewater tight with 3-M Mastic tape to prevent corrosion.

INTER-CONNECTING WIRING

Most EGM enclosures that are solar powered are locatednear the primary measurement device and the wiringlength required for transmitter connections are minimal.Wiring requirements will also be minimal if the pressureand differential pressure transmitter electronics areinternally mounted in the enclosure. The temperatureprobe/cable assembly is with a fixed length and nospecial connections are required.

If an extra process signal is to be included in theinstallation, a good quality wire with the appropriateamount of conductors, surrounded with a strong shieldshould be used. The wire size should be #16 AWGstranded and the shield should have a stranded “drain”wire for connection. If it is to be exposed to the elementsthe outer covering should be ultraviolet (sun light)resistant and have a direct burial rating. When terminatingthis type of wire under compression connectors, the wirewill eventually conform to the area of the connector andbecome loose due to temperature cycling. It is a goodpractice to moderately re-tighten all connections everysix months. Terminate the shield's drain wire in theenclosure at the recommended termination point anddo not connect it in the transmitter. This will preventground loop currents from occurring during a transientsurge. Wrap a piece of electrical tape around the cableat the cut-off drain wire. Remember to leave some extrawire neatly routed in the enclosure before it is terminated.You will need that extra length of wire in the future. If thetransmitter is fully isolated from the pipeline connect a#10 AWG ground wire from the transmitter case to theground radial. Do not “daisy chain” any ground systemand keep the resistance below 1 ohm from the case tothe ground system.

Solar panel wiring should be as large as possible toprevent any voltage drops from occurring. If the panelhas to be located away from the battery or EGMenclosure, make sure the wire splices are solid to preventcorrosion from occurring. Connectors between the solar


panel and the battery have been known to have problemsand should be avoided. Soldering the connections is thebest method but a good quality butt-splice, crimping tooland the use of 3-M Mastic Tape will make the connectionwater-tight. The wire should be sun-light resistant andthe shield should be connected to ground in theenclosure only. As with the transmitter cases, if the solarpanel is isolated from the pipeline connect a #10 AWGground wire, from the metal frame to the ground radial.Keep the resistance below 1 ohm from frame to theground system.

Put a “drip loop” curl in any wire or cable before enteringan enclosure to prevent rain water from running alongthe wire. Protect the wire from any sharp edges that willeventually cut it's way through the protective insulation.

CONDUIT

Your company may require that conduit be used toprotect all inter-connect wiring. Special precautionsshould be utilized to maintain the integrity of the isolationbetween the pipeline and the grounding system. CPVCcouplings and fittings are available to isolate metal andflexible conduit runs. Chemical resistant flexible conduitis a viable alternative.

If seals are required, follow your company's installationguidelines to properly install the fittings. A standard,multi-conductor cable that passes through a seal mustbe separated and the sealing compound poured aroundit.2 The multi-conductor “gel-filled” cable will not let avapor pass through its wrap. if the enclosure is in ahazardous area and a conduit will be run to a safe areafor connections to other equipment, a seal must beinstalled as it leaves the hazardous area. Each companyfollows certain standards concerning the classificationof the operating areas and have specific requirementsfor hazardous area installations.

Long conduit runs that are buried at a new installationwill eventually move due to settling of the soil. It will puta strain on equipment and CPVC couplings if they areutilized. Reinforce the conduit as it comes above groundlevel by using stable supports such as Uni-strut bracinganchored separately from the piping. Before installation,check inside of each joint of rigid galvanized conduit tolook for any galvanizing residue left over from processing.This could cut into wiring as it is being pulled through it.Do not use pipe-dope or thread tape to water proofunderground connections of conduit.

TRANSIENT PROTECTION

Each transmitter should have some type of transientprotection at each transmitter. Transient protection willalso be built into the EGM's point of termination for alloutside connections. These items usually consist ofdevices such as MOVs (metaloxide varistors), transorbs,resistors, zener diodes, fuses or a combination of these.Install transient protection at each external device. The

temperature probe cable that terminates directly in theEGM's enclosure has transient protection only at thepoint of termination.

Some of the transient protection devices are sacrificialand have to be replaced after they perform their designedtask. The MOV (Metal Oxide Varistor) can deteriorate andbecome less effective after each surge. MOVs willpartially fail and begin to interfere with transmitter signalscausing incorrect values to be measured. The bestprotection is a non-sacrificial protector that win continueto operate correctly after repeated suppressions ofsurges.

LIGHTNING PROTECTION

There are two areas of thought on just how to providethe best lightning protection system for differentinstallations. One area of engineering thought is todissipate the static charge build-up into the earth beforea strike occurs and the other is to control the strikecurrent by spreading the strike's charge in the earth atsurvivable levels.5 By taking the best practices from each,an effective lightning protection system can be installedon the EGM equipment. Installation of EGM for lightningprotection should be based on isolation, prevention ordissipation, and control.

According to Underwriters Laboratories, Inc., centralFlorida has the most annual days per year of electricalstorms in the United States of approximately 1006. Theutility companies in Florida are experienced in lightningprotection for its EGM equipment and have utilizedseveral techniques to prevent lightning surge damage.Consider some of these when installing any type ofequipment in the field:

1. Ground all fences at each comer or angle changewith a deep ground rod.

2. Ground any metal buildings.3. Provide surge protection on the cathodic rectifier

equipment and its power source.4. Install surge protection on the pipeline insulator

connections to prevent arc over.5. Have a good grounding system.6. Install transient protection devices on data

circuits, telephone lines and field devices.

CONCLUSION

Electronic gas measurement equipment is rapidlyreplacing the mechanical chart recorders to respond tothe changes that are occurring in the way that naturalgas companies are doing business. Careful installationof the computers and attention to details will result in adependable and accurate measurement of gas flow overa long period of time.


REFERENCES

1. Integrated Power Corporation. The PhotovoltaicBattery: Introduction to the PV Battery. TechnicalBulletin No. 504, Rev. A. N.P., Rockville, MD. Oct,1992.

2. National Electrical Code Handbook 1993. Article5015(e)(1). Cable Seals Class 1, Division 1.

3 . PolyPhaser Corporation. Lightning/EMP and GroundSolutions. Catalog 93. N.P. Minden, NV. 1993.

4. PolyPhaser Corporation. Striking News. Volume 3,No. 1, N.P., Minden, NV. Feb, 1994.

5. PolyPhaser Corporation. Striking News. Volume 3,No. 2, N.P. Minden, NV. May, 1994.

6. Underwriters Laboratory, Inc. UL LightningProtection. N.P., Northbrook, IL 1993.

Michael Price


AN OVERVIEW OF THE McCROMETER V-CONE METER AMERICANSCHOOL OF GAS MEASUREMENT TECHNOLOGY 2003

Dr. RJW Peters and Dr. R. StevenMcCrometer

3255 W, Stetson Ave., Hemet, CA 92545-7799

INTRODUCTION

The V-Cone meter is a differential pressure (DP) typemeter patented by McCrometer Inc. a subsidiary ofDanaher Corporation. The V-Cone meter is in manyrespects a classical DP meter using the physical laws ofthe conservation of mass and energy as its principle ofoperation. However, there are important differencesbetween the V-Cone meter design and other DP metertypes. These differences give the V-Cone meter importantperformance advantages. These advantages nclude theability of the V-Cone meter to operate with very shortupstream and downstream straight pipe lengths, tocreate a low total pressure (or “head loss”), to create avery stable DP, to give a large turn down, to createrelatively low signal noise and to cope well with liquidand particulates in the gas stream. The aim of this paperis to discuss the design of the V-Cone meter and explainwhy this design gives these advantages over traditionalDP meters.

1. The V-Cone Meter is a DP Meter

The V-Cone meter like several other popular meters is adifferential pressure (or “DP”) meter. These meters allwork according to the same principle. That is anobstruction in the pipe (i.e., a reduction in the crosssectional area available to the flow) causes an increasein flow velocity and a corresponding reduction inpressure. Hence be measuring the upstream pressure,the temperature and the difference in the static pressurebetween the upstream and the minimum cross sectionalareas the flowrate can be determined as long as the fluidproperties are known. The flow rate determination is doneby applying the laws of conservation of mass and energy.

The difference in the DP meter designs is the particulargeometry of the primary element (in particular theobstruction in the pipe creating the pressure and velocitychange). With the traditional DP meters the obstructionhas consisted of blocking the outside of the flow area invarious ways thus forcing the fluid to accelerate througha reduced opening in the centre of the pipe line. Examplesof this are the Orifice Plate, Venturi and Nozzle metersas shown in Figure 1.

The Orifice Plate is a plate with a hole in the centerinserted between two pipe flanges. The Nozzle is acontoured converging section (or “nozzle’) discharginguncontrolled into the full pipe area and the Venturi(sometimes called “Venturi Tube”) has a converging

FIGURE 1. Traditional DP meters.

section (or “nozzle”), a length of straight pipe of reducedarea comparative to the upstream pipe (or “throat”) anda diverging section (or “diffuser”) to allow a controlleddischarge. Clearly these meters all accelerate the flowtowards the center line of the pipe. The V-Cone meterdoes the opposite. It accelerates the flow to the outsideof the pipe by having a cone positioned pointing up thecenter line of the on coming flow. Figure 2 shows thegeometry of the V-Cone.

FIGURE 2. V-Cone meter sketches.

From Figure 2 it can be seen that the V-Cone meterupstream tapping (denoted “H” to in indicate the highpressure port) is on the wall of the pipe but thedownstream tapping (denoted “L” to in indicate the lowpressure port) is not on the wall of the pipe but on theback face of the cone on the pipe center line. This is aradical departure from the traditional DP meter designsand this design gives some significant performanceimprovements which will be discussed in chapter two.

The different geometry can cause some engineers towonder if the V-Cone meter operates according to adifferent principle than the traditional DP meters but thisis not so. The governing flow equation for the V-Cone isidentical to that of all other DP meters. From the userspoint of view all secondary equipment used is exactlythe same as would be chosen for any DP meter. That isthe same manifolds, DP transmitters and FlowComputers are used as are used with other DP meters


√

√

(with the flow computer requiring the V-Cone equationjust as it requires the unique equation for any DP meter).Therefore use of a V-Cone meter requires no furtherunderstanding than that knowledge required to use anyother DP meter. Figure 3 illustrates the typical set up ofany DP meter.

FIGURE 3. A typical DP meter set up.

Now that it is known that the V-Cone meter is similar inset up and use to all other DP meters the differencescaused by having a cone in the primary element can bediscussed.

2. The Differences (and Advantages) Between V-ConeMeters and Traditional DP Meters

2.1) The Flow Equation

All DP meters work according to the principles of theconservation of mass and energy. This results in equation1.

m = EYCdAd √2ρ∆P (1)

where m is the mass flowrateE is the “Velocity of Approach Factor” √

___

Y is the expansibility factor.Cd is the discharge coefficient (i.e. the ratio of

the actual to theoretically predicted flow-rates).

Ad is the minimum cross sectional area throughthe meter.

ρ is the fluid density.∆P is the measured differential pressure.β is the square root of the ratio of the minimum

flow cross sectional area through the meterto the inlet cross sectional area.

Depending on specific DP meter geometries some ofthe above parameters are calculated in different ways.For the V-Cone the following should be noted.

a) Due to the different geometry the V-Cone meter betaratio is not calculated by the standard equation 2 but byequation 3.

Ad dβTraditional = A D

βV-Cone =Ad dc √

2

A D ↵

where A is the upstream pipe cross sectional area, d idthe traditional throat diameter, D is the upstream pipediameter and dc is the cone diameter. Figure 4 illustratesthe origins of these equations.

FIGURE 4. The Difference in beta calculations betweenthe V-Cone and traditonal DP meters.

b) Clearly the calculation of the minimum cross sectionalarea for the V-Cone meter will be different to thetraditional DP meters. The comparisons are shown inequations 4 & 5.

AdTraditional =π d2 (4)4

AdV-Cone =π (D2 - d

2c) (5)

4

c) The Expansibility (Y) Equation

For all DP meters including the V-Cone meter if the fluidflowing is liquid the expansibility factor is effectively unityas the expansion factor corrects for changes in densitythrough the DP meter and liquids are effectivelyincompressible. If the fluid is a gas however the flow iscompressible and the correction factor is required. The

= (2)

(3)√= 1-

βTraditionaldA

A

d

D= =

βV Coned cA

A

d

D− = = −

1

2

Orifice Plate V-Cone


Orifice Plate meter has an experimental expansibilityequation as shown in equation 6. The V-Cone meter hashad its expansibility equation found by experiments atNEL in the UK [1]. The V-Cone meter expansibilityequation is shown in equation 7.

Orifice Plate Meter: (6)

V-Cone Meter: (7)

κ is the isentropic exponent and P1 is the upstreampressure. Again it is clear that the V-Cone meter behavessimilarly to other DP meter devices.

d) The Discharge Coefficient, Cd and MeterCalibration

For all DP meters the discharge coefficient is defined asshown in equation 8.

(8)

Where mactual is the actual flowrate and is the theoreticalflowrate calculated by consideration of the masscontinuity equation and Bernoulli (energy) equation.

V-Cone meters need to be “calibrated” to find the valueof the discharge coefficient. Calibration for all DP metersconsists of using a reference meter of extremely lowuncertainty (which gives a flowrate value as close to theactual flowrate as can practically be managed) upstreamof the DP meter in question and fitting the results toequation 8. Over a large turn down (a wide range offlowrates) V-Cone meters are found to have a dischargecoefficient which is a function of the upstream pipe flowReynolds number. For small turndowns an averageddischarge coefficient is often sufficiently accurate formany applications. Figure 5 shows a typical V-Conemeter calibration.

Note that in Figure 5 the discharge coefficient line fit isquoted to have an uncertainty of 0.5%. This is the quoteduncertainty for all V-Cone meters. To achieve thisMcCrometer calibrates all meters over the operationalrange of the flow.

The latest version of ISO 5167 [2] recommends thatVenturi meters should be calibrated for flows where theReynolds number is greater than one million. Thus, theVenturi and V-Cone meters have similar calibration

requirements. Orifice Plate meters (and Venturi metersat Reynolds numbers lower than one million) do not needto be calibrated as long as the upstream and downstreamflow conditions are met as then the discharge coefficientcan be read off a table [2]. However, these upstreamand downstream straight lengths are considerable andas the V-Cone meter does not need these extra pipelengths (see section 2.2) the weight and space saved bya V-Cone meter can match the extra cost of the pipework or flow conditioner required making the V-Conemeter extremely competitive.

For any given V-Cone meter (or any other DP meter) if aconstant discharge coefficient is used then with a readupstream pressure and differential pressure and with thefluid properties known equation 1 can be directly applied.If any given V-Cone meter (or any other DP meter) has aline fit describing the discharge coefficient to theReynolds number relationship found by calibration thenan iterative solution is required. The Reynolds number iscalculated by equation 9. The iteration for a V-Cone meter(or any other DP meter) is given by equation 10. A startingflowrate for the iteration is recommended to be that foundby initially considering the discharge coefficient to beunity. This ensures a quick convergence of the iteration.

4mRe = (9)πµD

where Re is the flows upstream Reynolds numberm is the mass flowrateD is the upstream pipe diameterµ is the fluid viscosity

4mm - EYAdf 2π∆P = 0 (10)πµD

Y = 1 - (0.41+0.35β4) ∆P κP1

Y = 1 - (0.649+0.696β4) ∆P κP1

mactual mactual

Cd = mtheoretical = EYAd √_____2ρ∆P

FIGURE 5. A typical calibration graph for a V-Cone meter.

√___

√√√↵

√√√↵

m EYA fm

DPd

..

−

=42 0

πµρ∆


4mwhere Cd = f(Re) = f πµD

2.2) Flow Conditioning

An advantage the V-Cone meter has over other metersis that it is a flow conditioner as well as a flow meter.This results in the meter having the ability to operate ininstallations where other meters in the market (DP andnon-DP meters) can not. With space and weight in manyindustrial applications being important the savings inspace, weight and cost of avoiding the requirement forlarge straight lengths of pipe upstream and downstreamof the metering station by using a V-cone meter can beconsiderable.

The V-Cone meter will continue to operate as theupstream length is reduced to less than the minimumallowed by other meter designs. In fact V-Cone metershave been installed successfully directly onto thedownstream flange of various pipe disturbances (e.g. 900

bends, out of plane double bends, valves etc) andcontinued to give reliable readings. Figures 6 and 7 showthe difference in using an Orifice Plate meter and a V-Cone meter in a confined space. The test graphs arecourteously of NIST.

FIGURE 6. An Orifice Plate meters performance atvarious distances downstream of a double out of plane

bend.

Figures 6 and 7 where published in the ASME FluidsEngineering Conference 1993 [3]. Further independentverification of the V-Cone meters ability to continue tooperate in poor installations that have a-symmetricalvelocity profiles and heavy swirl is given by Chevron andStatoil who presented results of swirl tests on the V-Conemeter at the North Sea Flow Measurement Workshop in1995 [4]. Chevron and Statoil concluded “Since the

accuracy of V-Cone meters are not significantly affectedby swirl flow, they are better suited for applications incramped quarters than Orifice Plate meters.”

2.3 A Stable Differential Pressure

With the V-Cone meters unique positioning of thedownstream pressure port on the back face of the conethe differential pressure is very stabile. In fact, with thedownstream port not in direct contact with the flow pastthe minimum cross sectional area, what is in reality beingread is the pressure created by the vortices generatedbehind the cone. The strength of these vortices (andhence the magnitude of the DP read) is dictated by theshear force of the jet coming off the end of the cone.Hence, the vortices strength is a direct indication of theflowrate.

Figure 8 shows a sample result of experimentsconducted to show the relative fluctuation in the DPreadings of a 1/2” Orifice meter and a 1/2” V-Cone meter.The paramerer δ is the standard deviation of the DParound its mean.

Clearly the V-Cone has less pressure fluctuations thanthe Orifice Plate meter. Figure 8 is from an internalMcCrometer report [5].

2.4) The V-Cone Meter Turn Down

The V-Cone meters turndown (i.e., the ratio of highest tolowest readable flow with in the quoted 0.5% uncertainty)is 10:1 for a single DP transmitter. This is considerablyhigher than other DP meters (e.g., an Orifice Plateturndown for a single DP transmitter is typically 3:1). Thisperformance is due to the positioning of the downstreampressure port on the back end of the cone and thestability of the DP signal produced.

√√√↵C f f

m

Dd = ( ) =

Re

.

4

πµ

FIGURE 7. An V-Cone meters performance at variousdistances downstream of a double out of plane bend.


FIGURE 8. A comparison of the DP signal fluctuations ofOrifice Plate and V-Cone meters.

Figure 9 is shows a recently calibrated V-Cone meterholding an uncertainty of 0.54% across a turndown of18:1 for a single DP transmitter.

FIGURE 9. A calibrated V-Cone meter showing highturndown.

2.5) The Total Pressure (or “Head”) Loss

The total head loss of the V-Cone meter is relatively lowcompared to other DP meters and many other type of flowmeters. In fact the V-Cone meter head loss is comparableto the Venturi meter (i.e., the DP meter traditionallydesigned to be low head loss meter). Note that both theV-Cone meter and the Orifice Plate meter do not measurethe pressure at a point in contact with the streamlines ofthe main flow but rather at a point in contact with thevortices. (see Figure 10a and Figure10b).

If it is also noted that the typical discharge coefficient ofthe Orifice Plate meter is 0.6 [2] and the typical dischargecoefficient of the V-Cone meter is 0.8 then for a givenflow rate and equivalent meter geometries Equation 1shows that the differential pressure of the V-Cone meteris less than for the Orifice meter. As the upstreampressure is the pipe pressure before it is influenced bythe meter in question then the higher differential pressure

produced by the Orifice Plate indicates a lowerdownstream pressure which indicates stronger vortices.This therefore is the reason why the V-Cone meter has alower total head loss than the Orifice Plate meter asstronger vortices means greater energy losses.

Miller [6] gives the head loss calculations for differentmeters. Figure 11 shows a comparison of many metersthat have had their total head loss predicted according tothe equations given by Miller. Water in a 3” pipe line wasused with a flow rate of 1145 liters per minute flow rate.Clearly the V-Cone meter is similar to the Venturi metersand better than the other meters in this comparison.

2.6 Contaminated Flows

The V-Cone meter is sturdy and copes with particulatesin the flowing fluid well. The cone is protected by theboundary layer that exists around it so most small solid

FIGURE 10a. The Orifice Plate meter pressure portposition relative to the vortices.

FIGURE 10b. The V-Cone meter pressure port positionrelative to the vortices.


particles tend not to strike the cones edge and thereforethe cone survives in abrasive flows longer than manyother meters. It does not wear quickly unlike an OrificePlates sharp edge. When damage does occur tests haveshown that even moderate damage to the cones edgearound the minimum cross sectional flow area has hadonly a small effect on the meters performance. Due tothe geometry of the V-Cone meter there is no place forcontaminates to collect and therefore the meter tendsto be self cleaning.

2.7 Wet Gas Flow Metering

Wet gas metering is becoming increasingly essential inmany different industries. DP meters are considered tobe one of the better single phase meters for use in wetgas flows (i.e., gas flows with a relatively small quantityof liquid present). All DP meters tend to give a positiveerror (called an “over-reading”) when liquid is present ina gas flow. The V-Cone meter allows most of the liquidphase in a wet gas to pass by undisturbed along thepipe wall meaning it tends to have a relatively smallpositive error compared to other DP meters.

Figure 12 shows a comparative plot of the effect of liquidin a gas flow on a geometrically equivalent (0.55 betaratio) V-Cone meter and Venturi meter. The quantity ofliquid is shown in terms of a parameter called theLockhart-Martinelli parameter denoted by “X.” Equation11 defines X.

m1 ρgX = (11)mg ρ1

where mg and m1 are the gas and liquid mass flowratesrespectively.

ρg and ρ1 are the gas and liquid densitiesrespectively.

Thus in Figure 11 with a constant pressure of 60 Bar(i.e., set fluid densities) the parameter X is indicating theratio of liquid to gas mass flowrates in the wet gas. TheV-Cone meter has a slightly smaller liquid induced errorthan the Venturi meter. The spread visible for both metertypes at the higher relative liquid to gas ratios is partiallydue to a relationship between over-reading and pressureamong other effects but this is beyond the scope of thispaper. What is clear however, is that the V-Cone metercontinues to operate when liquid is present in the gasflows and for a given liquid to gas ratio it reads with arepeatable and therefore predictable error.

FIGURE 12. A plot showing the comparativeperformance of a V-Cone meter and aVenturi meter used with wet gas flow.

STANDARDS

The V-Cone meter is covered by the API 5.7 documentfor differential pressure meters [7]. It also has custodytransfer approval in Canada.

CONCLUSIONS

The V-Cone meter is a relatively new meter on the marketthat operates according to the same principles as othertraditional DP meters but with significant advantages.These are the elimination of large pipe runs, a stabledifferential pressure is produced, a high turndown isobtained using a single DP transmitter, a low head lossis produced, the meter is more resistant to wear thanmany other designs and handles wet gas flows betterthan any other DP meter. These advantages are slowlybecoming known through out the many industries andthe V-Cone meters market share is continuing to increasethrough out the world.

REFERENCES

1. Reader-Harris M. et al. North Sea Flow MeasurementWorkshop 2001, “Derivation of an ExpansibilityFactor for the V-Cone Meter.”

FIGURE 11. A plot of Millers total head loss calculations[6] for 3” meters with 1145 liters per minute flow rate.

Xm

m

l

g

g

l

=

.

.

ρρ


2. ISO 5167 Rev. 4, “Measurement of Fluid Flow byMeans of Pressure Differential Devices Inserted inCircular Cross-Section Conduits Running Full.”

3. Ifft S. ASME Fluids Engineering Conference 1993,“Pipe Elbow Effects on the V-Cone Flowmeter.”

4. Shen J. et al. North Sea Flow MeasurementWorkshop, 1995 “A Performance Study of a V-ConeMeter in Swirling Flow.”

5. Ifft S. McCrometer Internal Report “Signal NoiseRatio Comparison with V-Cone and Orifice Plate.

6. Miller R. “Flow Measurement EngineeringHandbook” 3rd Ed. Published by McGraw Hill.

7. API 5.7 “Testing Protocol for Differential PressureFlow Measurement Devices.”


BASIC ELECTRONICS FOR FIELD MEASUREMENTRick Heuer

Emerson Process ManagementHouston, Texas

INTRODUCTION

Try this at home. We are professionals. If you are able toinstall your own TV dish satellite system and a wirelesshome network for PCs, you have a head start on installingand maintaining electronic field measurement equipment.

Today’s Measurement Technicians and Engineers arerequired to operate and maintain a variety of Hi-Techfield measurement equipment. Most of the fieldinstrumentation is tightly integrated in a complete systemfunctional environment. The larger the metering station,the more complex the system.

The goal of this paper is to provide an introductory toelectronics based on current field measurementtechnology.

Basic field duties would include:

calibration, data collection, configuration, orifice platechanges, communication checks and maintenance.

Some of the basic types of wire connections includepower, analog signals and digital signals. Each of theseplay unique circuitry roles and require different knowledgeof their intended operation.

POWER

Current flows from one terminal to another only whenthere is a completed circuit. The device being poweredmust have enough current and voltage to insure properoperation. Most Field electronics operate off of 6, 12 or24 volts DC.

ELECTRONICS FORMULAS

V = I * RI = V / RR = V / IWhere V is Voltage (Volts), I is current (Amps), R isResistance (Ohms Ω)P = Power (Watts) where as:P = I2 * RP= E * I

FIGURE 1

Antenna

Flow Computer

SolarPanel

Batteries& Radio

ProcessSensor

FIGURE 2

Resistor/Load

Switch

DC SupplyCurrent

FIGURE 3

V

I R

V = I • R I = V R = V R I

P = I • V P = I2 • R P = V2

R

I = Current (A)

P = Power (W)

R = Resistance (Ω)V = Voltage (V)


ANALOG SIGNALS

Some of the more common things encountered iscalculating 4 to 20 mA instrument loops. Loop loadingcan be seen on the Input Burden Chart below. This chartshows the maximum number of 250Ω loads a devicecan power, based on loop voltage levels. This applies toboth analog inputs and analog outputs.

CURRENT TO VOLTAGE BASICS

Most RTUs, PLCs and flow computers convert currentstyle transmitters into voltage for use in the analog todigital process. Current output devices are still commondue to only two or three connection wires and their abilityto locate thousands of feet between devices.

TYPICAL RESISTOR

4mA x 250Ω = 1 V20mA x 250Ω = 5 V

The 4 – 20mA example shows how a single transmittersignal can go to two RTUs (Remote Terminal Unit)

FREQUENCY/PULSE

Turbine meters and gas ultrasonic meters have pulse orfrequency type outputs. They are best measured by usingan oscilloscope. This allows you to view the amplitude andfrequency of the signal. Un-amplified signals are usuallysmall amplitude sine waves. Amplified device outputs aremore than likely a square wave. The chart below gives anexample of the millivolt input to frequency ratio on the inputof a preamp. This shows that a high frequency (Noise) smallamplitude signal would be filtered out.

FIGURE 7Preamp Frequency Input Graph

RELAYS

Relay type circuits are magnetically operated switches.A wetting voltage usually DC is applied to the coilwindings which produce a magnetic field used to openor close the relay switch. Relay terminals are referred toas “Normally Open,” or “Normally Closed.” Relays canbe used to turn on or off different actuations such as:Solenoid valves, gas samplers, and radios.

FIGURE 4

FIGURE 5

FIGURE 6 FIGURE 8

250 + 250 = 500 ohm

4–20 mA Example

TemperatureTransmitter

24 VDCPower

250 ohm

RTU 1

250 ohm

RTU 2

Last in the loop


OPEN COLLECTOR

Some signal inputs and outputs may require a “pull-up”resistor to source the voltage. These are mostly referredto as “Open Collector”. Check the specifications to insureproper resistor sizing. The resistor has to be sized toallow only enough current to flow through it. The drawingbelow shows a turbine meter preamp using a “OC”,termination.

DIGITAL SIGNALS

Digital signals are normally low level, low voltage inputtypes. They are typically looking for a voltage transitionbetween 0 and 5 volts DC. Some digital signal types areas follows:

• Digital Inputs are normally used as status typeinputs that sense a voltage level transition. Thiscould be from a valve limit switch. The signal isa low to high voltage level transition. Sometimesthey can be “OC” type for use with long distanceconnections.

• Digital Outputs are normally used to trigger anevent like a smart gas sampler. This output couldalso be “OC” type. The signal is a low to highvoltage level transition.

• Communications are digital in nature becausethey carry the binary (Computer Talk) information

needed for data transfer. While there are manydifferent hardware formats for carrying data,RS232 is still the most common device interfacetoday. The table shows the connector pin outfor two types of common cables.

ETHERNET

Ethernet connectivity is being used for most plant typeoperations involving line power flow computers. This isa very common network medium that is spreading tomore complex field measurement scenarios. Most RTUs,Gas Chromatographs and Gas Ultrasonic meters haveEthernet ports. Simply follow the instructions needed forsetting up your laptop’s Ethernet port.

FIELDBUS

Connecting a set of measurement transducers to a flowcomputer has been traditionally an analog world. Thereis the low power 1 to 5vdc method and the 4 to 20 mAmethod for line powered applications. Fieldbus changesthis in two ways. A standard set of wires and connectorsare used between devices and a commoncommunication protocol is used. Less field wiring isneeded and device interchangeability is added. Deviceinterrogation is done digitally now instead of analog.Process variables along with performance diagnosticsare now available at the device level.

Smart Multivariable ProcessTransmitter

PROTOCOLS

The de facto standard for flow computer protocol is EFM(Daniel/ENRON) Modbus ASCII or RTU. All major vendorsoffer some variation of this. This was the first step towardvendor interchangeability. The next step is the use ofhigher level protocols based from the internet standards.

FIGURE 9

FIGURE 10

RS-232DTE (Computer) DB9Pin# DB9 RS-232 Signal Names

#1 Carrier Detector (DCE)

#2 Receive Data (Rx)

#3 Transmit Data (Tx)

#4 DTE Ready/Data Terminal Ready (DTR)

#5 Signal Ground/Common (SG)

#6 DCE Ready/Data Set Ready (DSR)

#7 Request to Send (RTS)

#8 Clear to Send (CTS)

#9 Ring Indicator (RI)

DTE (Computer) DB25Pin# DB25 RS-232 Signal Names

#1 Shield to Frame Ground

#2 Transmit Data (Tx)

#3 Receive Data (Rx)

#4 Request to Send (RTS)

#5 Clear to Send (CTS)

#6 DCE Ready/Data Set Ready (DSR)

#7 Signal Ground/Common (SG)

#8 Carrier Detector (DCE)

#20 DTE Ready/Data Terminal Ready (DTR)

#22 Ring Indicator (RI)


TROUBLESHOOTING & CHECKUPS

In a perfect world there would be: As-built drawings,Make & Model of equipment lists and of course, themanuals. For each measurement station!

Most manuals can be found on the vendors website.Other useful information as to support contacts, softwareupgrades/patches and FAQs (Frequently AskedQuestions) is also online.

FIELD CHECKS

1. Are there Lights, LCD, or other health indications? 2. How’s the Power? 3. Check the Fuses. 4. How’s the system grounding? (Another lengthy

discussion)

The best way to verify correct operation is to interrogatethe device by using it’s diagnostic modes. A laptop PCor PDA is needed. Some new technology even haspredictive diagnostics, integrated documentation,calibration management and device configuration.

FIELD ELECTRONIC EXAMPLES

Smart Transducers, Digital Valve Controllers, Ultrasonicmeters, Turbine meters, radios, gas chromatographs,odorizers, sample systems, radios and power systems.

Digital Valve Controller

SOME FIELD TECHNICIAN TOOLS

• Portable Laptop Computer with device software• Cables• Digital Voltmeter• Transducer Calibration Test equipment

FIGURE 11

Rick Heuer


PRINCIPLES OF OPERATION FOR ULTRASONICGAS FLOW METERS

John LansingDaniel Measurement and Control, Inc.

9270 Old Katy Rd, Houston, Texas 77055

ABSTRACT

This paper discusses fundamental issues relative toultrasonic gas flow meters used for measurement ofnatural gas. A basic review of an ultrasonic meter’soperation is presented to understand the typicaloperation of today’s Ultrasonic Gas Flow Meter (USM).The USM’s diagnostic data, in conjunction with gascomposition, pressure and temperature, will be reviewedto show how this technology provides diagnostic benefitsbeyond that of other primary measurement devices. Thebasic requirements for obtaining good meterperformance, when installed in the field, will be discussedwith test results. Finally, recommendations for installationwill be provided, including an example of a good pipingdesign.

INTRODUCTION

During the past several years, the use of ultrasonic flowmeters for natural gas custody transfer applications hasgrown significantly. The publication of AGA Report No.9, Measurement of Gas by Multipath Ultrasonic Meters[Ref 1] in June 1998, has further accelerated theinstallation of ultrasonic flow meters (USMs). Todayvirtually every transmission and many distributioncompanies are using this technology fiscal or foroperational applications.

Since the mid-1990s the installed base of USMs hasgrown by approximately 50% per year. There are manyreasons why ultrasonic metering is enjoying such healthysales. Some of the benefits of this technology includethe following:

• Accuracy: Can be calibrated to <0.1%.• Large Turndown: Typically >50:1.• Naturally Bi-directional: Measures volumes in

both directions with comparable performance.• Tolerant of Wet Gas: Important for production

applications.• Non-Intrusive: No pressure drop.• Low Maintenance: No moving parts means

reduced maintenance.• Fault Tolerance: Meters remain relatively

accurate even if sensor(s) should fail.• Integral Diagnostics: Data for determining a

meter’s health is readily available.

It is clear that there are many benefits to using USMs.Although the first several benefits are important, the most

significant may turn out to be the ability to diagnose themeter’s health. The primary purpose of this paper is todiscuss basic gas ultrasonic meter operation,diagnostics, review the fundamentals of fieldmaintenance, discuss some test results and provide thereader with an examples of good and not-so-good pipingdesigns.

ULTRASONIC METER BASICS

Before looking at the main topic of integral diagnostics,it is important to review the basics of ultrasonic transittime flow measurement. In order to diagnose any device,a relatively thorough understanding is generally required.If the technician doesn’t understand the basics ofoperation when performing maintenance, at best theycan only be considered a “parts changer.” In today’sworld of increasingly complex devices, and productivitydemands on everyone, companies can no longer affordthis type of service.

The basic operation of an ultrasonic meter is relativelysimple. Consider the meter design shown in Figure 1.Even though there are several designs of ultrasonicmeters on the market today, the principle of operationremains the same.

FIGURE 1. Ultrasonic Flow Meter

Ultrasonic meters are velocity meters by nature. That is,they measure the velocity of the gas within the meterbody. By knowing the velocity and the cross-sectionalarea, uncorrected volume can be computed. Let usreview the equations needed to compute flow.

The transit time (T12) of an ultrasonic signal traveling withthe flow is measured from Transducer 1 to Transducer2. When this measurement is completed, the transit time(T21) of an ultrasonic signal traveling against the flow ismeasured (from Transducer 2 to Transducer 1). Thetransit time of the signal traveling with the flow will be


less than that of the signal traveling against the flow dueto the velocity of the gas within the meter.

Let’s review the basic equations needed to computevolume. Assume L and X are the direct and lateral (alongthe pipe axis and in the flowing gas) distances betweenthe two transducers, C is the Speed of Sound (SOS) ofthe gas, V the gas velocity, and T12 and T21 are transittimes in each direction. The following two equationswould then apply for each path.

of sound (Equation (4)), gas velocity is not required. Thisis true because the transit time measurements T12 andT21 are measured within a few milliseconds of each other,and gas composition does not change significantlyduring this time. Also, note the simplicity of Equations(3) and (4). Only the dimensions X and L, and the transittimes T12 and T21, are required to yield both the gasvelocity and speed of sound along a path.

These equations look relatively simple, and they are. Theprimary difference between computing gas velocity andspeed of sound is the difference in transit times is usedfor computing velocity, where as the sum of the transittimes is used for computing speed of sound.

Unfortunately, determining the correct flow rate withinthe meter is a bit more difficult than it appears. Thevelocity shown in Equation (3) refers to the velocity ofeach individual path. The velocity needed for computingvolume flow rate, also know as bulk mean velocity, isthe average gas velocity across the meter’s area. In thepipeline, gas velocity profiles are not always uniform,and often there is some swirl and asymmetrical flowprofile within the meter. This makes computing theaverage velocity a bit more challenging.

Meter manufactures have differing methodologies forcomputing this average velocity. Some derive the answerby using proprietary algorithms. Others rely on a designthat does not require “hidden” computations. Regardlessof how the meter determines the bulk average velocity,the following equation is used to compute theuncorrected flow rate.

Q = V * A (5)

This output (Q) is actually a flow rate based on volume-per-hour, and is used to provide input to the flowcomputer. A is the cross-sectional area of the meter.

In summary, some key points to keep in mind about theoperation of an ultrasonic meter are:

• The measurement of transit time, both upstreamand downstream, is the primary function of theelectronics.

• All path velocities are averaged to provide a “bulkmean” velocity that is used to compute themeter’s output (Q).

• Because the electronics can determine whichtransit time is longer (T21 or T12), the meter candetermine direction of flow.

• Speed of sound is computed from the samemeasurements as gas velocity (X is not required).

Transit time is the most significant aspect of the meter’soperation, and all other inputs to determine gas velocityand speed of sound are essentially fixed geometric(programmed) constants.

and

T12 = C+V • XL

L(1)

T21 = C–V • XL

L(2)

Solving for gas velocity yields the following:

L2 T21 – T12V = 2X T21 • T12( ) (3)

Solving for the speed of sound (C) in the meter yieldsthe following equation:

L T21 – T12C = 2 T21 • T12( ) (4)

Thus, by measuring dimensions X & L, and transit timesT12 & T21, we can also compute the gas velocity and speedof sound (SOS) along each path. The speed of sound foreach path will be discussed later and shown to be avery useful parameter in verifying good overall meterperformance.

The average transit time, with no gas flowing, is a functionof meter size and the speed of sound through the gas(pressure, temperature and gas composition). Considera 12-inch meter for this example. Typical transit times,in each direction, are on the order of one millisecond(and equal) when there is no flow. The difference in transittime during periods of flow, however, is significantly less,and is on the order of several nanoseconds (at low flowrates). Thus, accurate measurement of the transit timesis critical if an ultrasonic meter is to meet performancecriteria established in AGA Report No. 9.

It is interesting to note in Equation (3) that gas velocity isindependent of speed of sound, and to compute speed


INTEGRAL DIAGNOSTICS

One of the principal attributes of modern ultrasonicmeters is their ability to monitor their own health, and todiagnose any problems that may occur. Multipath metersare unique in this regard, as they can compare certainmeasurements between different paths, as well aschecking each path individually.

Measures that can be used in this online “healthchecking” can be classed as either internal or externaldiagnostics. Internal diagnostics are those indicatorsderived only from internal measurements of the meter.External diagnostics are those methods in whichmeasurements from the meter are combined withparameters derived from independent sources to detectand identify fault conditions. Some of the commoninternal meter diagnostics used are as follows.

Gain

One of the simplest indicators of a meter’s health is thepresence of strong signals on all paths. Today’s multipathUSMs have automatic gain control on all receiverchannels. Any increase in gain on any channel indicatesa weaker signal, perhaps due to transducer deterioration,fouling of the transducer ports, or liquids in the line.However, caution must be exercised to account for otherfactors that affect signal strength, such as pressure andflow velocity.

Gain numbers vary from manufacturer to manufacturer.Thus, recommendations may also differ. However,regardless of design or methodology for reporting gain,it is important to obtain readings on all paths undersomewhat similar conditions. The significant conditionsto duplicate are metering pressure and gas flow rate.

Gain readings are generally proportional to meteringpressure (and to a much lesser extent, temperature). Thatis, when pressure increases, the amount of gain(amplification) required is reduced. If an initial gain readingwere taken at 600 psig, when the meter was placed intoservice, and subsequent readings taken at 900 psig, onewould expect to see a change. This change in reading(assuming gain values are linear, not in dB) woulddecrease by the ratio of pressures (600/900).Understanding that pressure affects gain readings helpsguard against making the false assumption somethingis wrong.

Fortunately, most applications do not experience asignificant variation in metering pressure. If pressure doesvary, the observed gain value can be adjusted relativelyeasily to allow for comparison with baseline values. Thismethod of adjustment varies with manufacturer, so nodiscussion will be incorporated here.

Gas velocity can also impact the gain level for each path.As the gas velocity increases, the increased turbulenceof the gas causes an increase in signal attenuation. This

reduction in signal strength will be seen immediately byincreased gain readings. These increases are generallysmall compared to the amount of gain required. Typicalincreases might be on the order of 10-50%, dependingupon meter size and design. Thus, it is always better to“baseline” gain readings when gas velocities are below30 fps. Using velocities in excess may provide goodresults, but it is safe to say that lower velocities providemore consistent, repeatable results.

So, what else causes reductions in signal strength(increased gain)? There are many sources other than gasvelocity and pressure. For instance, contamination of thetransducers (buildup of material on the face) will attenuatethe transmitted (and received) signals. One might assumethat this buildup would cause the meter to fail (inabilityto receive a pulse). However, this is not generally thecase. Even with excessive buildup of more than 0.050of an inch of an oily, greasy, and/or gritty substance,today’s USMs will continue to operate.

One question often asked is “What impact on transit timeaccuracy could be attributed to transducer facecontamination?” It is true the speed of sound will bedifferent through the contaminated area when comparedto the gas. Let’s assume a build-up is 0.025 of an inchon each face, and the path length is 16 inches. Alsoassume the speed of sound through the contaminationis twice that of the typical gas application (2,600 fps vs.1,300 fps). With no buildup on the transducer, and atzero flow, the average transit time would be 1.025641milliseconds. With buildup the average transit time wouldbe 1.024038 milliseconds, or a difference of 0.16%. Thiswould be reflected in the meter’s reported speed of sound(more on that later). However, it is the difference in transittimes that determines gas velocity (thus volume). This isthe affect that needs to be quantified.

Maybe the easiest way to analyze this is assume thetransit time measurements in both directions are reducedby 0.16% (from the previous example). Remembering inEquation (3) that gas velocity is proportional to a constant(L2/2X) multiplied by the difference in transit times, alldivided by the product of transit times. The decrease intransit times will occur for both directions, and this effectappears to be negated in the numerator. That is, the Dtwill remain the same. However, the error in both T12 andT21 will cause the denominator value to decrease, thusproducing an error that is twice the percentage of transittime (0.16%), or 0.32%. Thus, the meter’s output willincrease by 0.32%. However, this amount of buildup isabnormal, and not typical of most meter installations.

Concluding the discussion on gain readings, USMs allhave more than adequate amplification (gain) toovercome even the most severe reductions in signalstrength. The amount of buildup required to fail today’shigh-performance transducers and electronics generallyexceeds pipeline operational conditions. Periodicmonitoring of this parameter, however, will help insuregood performance throughout the life of the meter.


Metering accuracy (differences in transit time velocitycomputation) can be affected, but only when significantbuildup of contamination occurs.

Signal Quality

This expression is often referred to as performance (butshould not be confused with meter accuracy). Allultrasonic meter designs send multiple pulses across themeter to another transducer before updating the output.Ideally, all the pulses sent would be received and used.However, in the real world, sometimes the signal isdistorted, too weak, or otherwise the received pulse doesnot meet certain criteria established by the manufacturer.When this happens the electronics rejects the pulserather than use something that might distort the results.The level of acceptance (or rejection) for each path isgenerally considered as a measure of performance, andis often referred to as signal quality. Meters provide avalue describing how good signal detection is for eachultrasonic path.

As mentioned above, there are several reasons whypulses can be rejected. Additional causes may includeextraneous ultrasonic noise in the same region thetransducer operates, distorted waveforms caused byexcessive gas velocity, and to some degree,contamination on the face of the transducer.

Typically, the value of acceptance for each path, undernormal operating conditions, will be 100%. As gasvelocity increases to near the meter’s rating, thispercentage may begin to decrease. Depending upondesign, this percentage may decrease to below 50%.Generally, this reduction in performance will have littleimpact on meter accuracy. However, if the percentageof accepted pulses is this low, it is safe to say the meteris not operating at top performance, and investigationmay be warranted (assuming the meter isn’t operatingat 110+% of rated capacity).

Concluding the discussion on performance, thisparameter should be monitored periodically as poorperformance on a path may be an indication of possibleimpending failure. Lower than expected performance canbe caused by several factors. Besides excessive gasvelocity, contamination on the transducer face andexcessive extraneous ultrasonic noise can reduce signalquality. However, by monitoring gains, this condition canbe easily identified before it becomes a problem.

Signal-to-Noise Ratio

This parameter is another variable that providesinformation valuable in verifying the meter’s health, oralert of possible impending problems. Each transduceris capable of receiving noise information from extraneoussources (rather than its mated transducer). In the intervalbetween receiving pulses, meters monitor this noise toprovide an indication of the “background” noise. Thisnoise can be in the same ultrasonic frequency spectrumas that transmitted from the transducer itself.

Noise levels can become excessive if a control valve isplaced too close and the pressure differential is too high.In this scenario the meter may have difficulty indifferentiating the signal from the noise. By monitoringthe level of noise, when no pulse is anticipated, the metercan provide information to the user, warning that meterperformance (signal quality) may become reduced. Inextreme cases, noise from control valves can “swamp”the signal to the point that the meter becomesinoperative.

All meters can handle some degree of noise created fromthis condition. Some USM designs can handle more thanothers can. The important thing to remember is the besttime to deal with control valve noise is during the design.Today’s technology has improved significantly in dealingwith extraneous noise. Reducing it in piping design isalways the best choice (more on this later).

Other sources can cause reduced signal to noise values.Typically they are poor grounding, bad electricalconnections between electronics and transducers,extraneous EMI and RFI, cathodic protectioninterference, transducer contamination and in someinstances, the meter’s electronic components. However,the major reason for decreased signal to noise ratiosremains pressure drop from flow control or pressurereducing valves.

Concluding this discussion on signal to noise, the mostimportant thing to remember is high-pressure drop(generally in excess of 200 psig) across a control valvecan cause interference with the meter’s operation. If thenoise is isolated to a transducer or pair of transducers,the cause is generally not control valve related. Hereprobable causes are poor component connections or apotential failing component. Control valve noise usuallycauses lower signal to noise levels on the transducersthat face the noise source (all would be affected).

Velocity Profile

Monitoring the velocity profile is possibly one of the mostoverlooked features of today’s ultrasonic meter. It canprovide many clues as to the condition of the meteringsystem, not just as a monitor of the meter. AGA ReportNo. 9 requires a multipath meter to provide individualpath velocities. As mentioned previously, the output usedby the flow computer is an average of these individualreadings.

Once the USM is placed in service, it is important tocollect a baseline (log file) of the meter. That is, recordthe path velocities over some reasonable operatingrange, if possible. Good meter station designs producea relatively uniform velocity profile within the meter. Thebaseline log file may be helpful in the event the meter’sperformance is questioned later.

Many customers choose to use a “high performance flowconditioner” with their meter. This conditioner is intended


to isolate any upstream piping effects on gas profile. Inreality, they don’t totally isolate the disturbance, but doprovide a reasonably repeatable profile. The importantissue here is the velocity profile is relatively repeatable.Once a baseline has been established, should somethinghappen to the flow conditioner, it can be identified quicklyby comparing path velocities with the baseline. Manythings can happen to impact the original velocity profile.Changes can be caused by such things as:

• partial blockage of the flow conditioner,• damage to the flow conditioner,• or upstream piping affects, such as a change in

a valve position.

Of course, something could have also occurred with themeter to cause a significant profile change. Generallyspeaking, this is unlikely as all components are securelymounted. However, the velocity of a given path couldbe affected by other problems. When considering thatonly X and L dimensions, and transit times, impact pathvelocity, it is relatively easy to eliminate these. If a problemdevelops within the meter that impacts only one or morepaths, other performance indicators, such as gain, pathperformance, and speed of sound will also be indicatingproblems.

One of the major benefits of analyzing path velocities isthe ability to determine if the meter assembly is becomingcontaminated with any pipeline debris. Surfaceroughness changes in the upstream piping will changethe velocity profile the meter sees. A profile change canbe observed by analyzing the different path velocitiesrelative to the meter’s reported average. Typically thevelocity profile becomes more “pointed” as the surfacefinish becomes rougher. This is a very important featuresince contamination on the inside of a meter will impactthe meter’s accuracy [Ref 2].

Different manufacturers utilize different path velocityintegration techniques. The ability to monitor profilechanges, and thus predict the significance of this effect,may vary by design. Thus, it may not be possible for allUSM designs to provide this diagnostic information.

Concluding this discussion on path velocities, most goodinstallations produce somewhat symmetrical velocitieswithin the meter. Comparing each path’s velocity withthe average, and sometimes to other paths, dependingupon the USM design, can give the user confidence theprofile has not significantly changed. Today’s USM canhandle some relatively high levels of asymmetry withinthe meter. It should not be assumed that the meter’saccuracy is significantly impacted just because thevelocity profile has changed. It is usually an indication,however, that something within the meter set, other thanthe meter itself, is probably causing the effect. Carefulreview of other diagnostic parameters can determine ifthe meter is at fault, or not. Identifying changes in pathvelocities are very helpful in determining if contamination

has occurred on the inside of the piping. Contaminationmay have an impact on the meter’s accuracy.Speed Of Sound

Probably the most discussed and used diagnostic toolis the meter’s speed of sound (SOS). The reader mayrecall that speed of sound is basically the sum of thetransit times divided by their product, all then multipliedby the path length (Equation (4)). As was discussedearlier, the primary measurement an ultrasonic meterperforms to determine velocity is transit time. If the transittime measurement is incorrect, the meter’s output willbe incorrect, and so will the speed of sound. Thus, it isimportant to periodically verify that the meter’s reportedspeed of sound is within some reasonable agreement toan independently computed value.

Modern USMs use high frequency clocks to accuratelyperform transit time measurements. In a typical 12-inchmeter, the average transit time may be on the order ofone millisecond. To obtain a perspective on thisdifferential time, values start out in the 10’s ofnanoseconds and typically increase to maybe 100microseconds at the highest velocities.

Obviously accurate meter performance requiresconsistent, repeatable transit time measurements.Comparing the SOS to computed values is one methodof verifying this timing. This procedure would beconsidered an external diagnostic technique. Let’sexamine the affects (or uncertainties) on computingspeed of sound in the field.

Pressure & Temperature Effects

The speed of sound in gas can be easily computed inthe field. There are several programs used for thispurpose. Most are based upon the equation of stateprovided in AGA Report No. 8, Compressibility andSupercompressibility for Natural Gas and OtherHydrocarbon Gases [Ref 3]. When computing speed ofsound, there is always some uncertainty associated withthis operation. It is important to realize that the speed ofsound is more sensitive to temperature and gascomposition than pressure. For example, a one degreeF error in temperature at 750 psig, with typical pipelinegas, can create an error of 0.13%, or about 1.7 fps. Anerror of five psig at 750 psig and 60 degrees F onlycontributes 0.01% error. Thus, it is very important toobtain accurate temperature information.

Knowing the temperature measurement error contributessignificant error in computing SOS is important. However,if the temperature is in error by one degree F, a moresignificant question might be “what error is this causingin the volumetric measurement?” A quick calculationshows a one degree F error will cause the correctedvolumetric calculation to be incorrect by 0.28%. Havinga history of calculated SOS vs. measured may actuallybe a good “health check” on the stations temperaturemeasurement!


Gas Composition Effects

Sensitivity to gas composition is a bit more difficult toquantify as there is an infinite number of sample analysesto draw from. Let’s assume a typical Amarillo gascomposition with about 90% methane. If thechromatograph were in error on methane by 0.5%, andthe remaining components were normalized to accountfor this error, the resulting effect on speed of sound wouldbe 0.03%. Thus, minor errors in gas composition, forrelatively lean samples, may not contribute significantlyto the uncertainty.

However, lets look at another example of a Gulf Coastgas with approximately 95% methane. Suppose themethane reading is low by 0.5%, and this time thepropane reading was high by that amount, the error incomputed speed of sound would be 0.67% (8.7 fps!).Certainly one could argue this may not be a “typical”error. There are many scenarios that can be discussedand each one would have a different effect on the result.The uncertainty that gas composition contributes to thespeed of sound calculation remains the most elusive toquantify, and, depending upon gas composition, mayprove to be the most significant.

A typical question is “what difference can be expectedbetween that determined by the meter, and onecomputed by independent means?” It has been shown[Ref 4] that the expected uncertainties (two standarddeviations) in speed of sound, for a typical pipeline gasoperating below 1,480 psig, are:

• USM measurement: ± 0.17%• Calculated (AGA 8): ± 0.12%

Since the USM’s output is independent of the calculationprocess, a root-mean-square (RMS) method can be usedto determine the system uncertainty. Thus, when usinglean natural gas below 1,480 psig, it is expected that95% of readings agree within 0.21% (or about 2.7 fps).Therefore, it may be somewhat unrealistic to assumethe meter will agree within 1 fps under typical operatingconditions.

Concluding this discussion on speed of sound, this“integral diagnostic” feature may be the most powerfultool for the technician. Using the meter’s individual pathspeed of sound output, and comparing it to not only thecomputed values, but also comparing within the meteritself, is a very important maintenance tool. Cautionshould be taken when collecting the data to helpminimize any uncertainty due to gas composition,pressure and temperature. Additionally, it is extremelyimportant to obtain data only during periods of flow astemperature stratification can cause significantcomparison errors. By developing a history of meter SOS,and comparing with computed values, it can also be usedas a “health check” for the temperature measurementused to determine corrected volumes.

IMPORTANCE OF SOS VERIFICATION

As was discussed earlier, SOS verification helps insurethe meter is operating correctly. However, what otherchanges in a meter can affect the reading? From theprevious discussion on gain, buildup on the face of atransducer will affect the speed of sound. Thus, if a pairof transducers has a different value, when compared tothe average (or to other paths, depending upon meter’sdesign), this might be an indication of contamination.

One thing to remember is that the percent change inspeed of sound, given the same buildup, will be greaterfor a smaller meter than a larger one. As path lengthincreases from say 10 inches to 30 inches (or more), abuildup of 0.025 inches will affect the transit time less.By utilizing gain information with SOS data for a givenpath, it can be quickly determined if the change in SOSis due to contamination, or other causes.

Another benefit in monitoring path SOS is to verify properidentification of reception pulses. In the section on signalto noise, extraneous noise was noted to potentiallyinterfere with normal meter operation. That is, if ultrasonicnoise within the meter (caused by outside sources)becomes too great, meter performance will be impacted.As the noise level increases, there is the possibility thatthe circuit detecting the correct pulse will have difficulty.Good meter designs protect against this and rejectreceived pulses that have increased uncertaintyregarding their validity. If this scenario occurs, it is unlikelyall paths will be affected simultaneously, and by the sameamount. Monitoring variations in SOS from path to pathwill identify this problem and help insure the meter’shealth is satisfactory.

Typical Speed of Sound Field Results

This section provides actual data from two differentmeters. Figures 2 and 3 show trended vs. time. Data isshown for an eight-inch meter in Figure 2 [Ref 3]. Itcompares the average speed of sound over the fourpaths with the AGA 8 calculated value.

FIGURE 2. 8-INCH METER MEASURED VS.CALCULATED SOS


At each measurement point, ten successive values ofthe ultrasonic meter’s SOS were logged. The two curvesthat show the minimum and maximum values in Figure2 demonstrates repeatability in SOS measurements ofbetter than 0.03%. The difference in the meter’s speedof sound vs. computed values are also, for most points,less than 0.3%.

Figures 3 shows the AGA 8 calculated speed of soundtrended against the individual SOS readings from thefour paths. Note that in each case the agreement on allchords is roughly as expected (better than 0.3%). In thearea where speed of sound deviations exceeded 0.3%,(Figure 3) low flow temperature stratification was likelythe cause. In the event of significant contamination onone or more pairs of transducers, this graph would haveshown the impact.

Basic Piping Issues

Ultrasonic meters require adhering to basic installationguidelines just as with any other technology. Primarymetering elements, such as orifice and turbine, haveadopted recommendations for installation long ago.These are provided through a variety of standards (API,AGA, etc.) to insure accurate performance (within someuncertainty guidelines) when installed. The reason forthese guidelines is the meter’s accuracy can be affectedby profile distortions caused by upstream piping. Oneof the benefits of today’s USM is that they can handle avariety of upstream piping designs with less impact onaccuracy then other primary devices.

Installation effects have been studied in much more detailthan ever before. This is due in part to the availabletechnology needed for evaluation. Reducing uncertaintyfor pipeline companies has also become a higher prioritytoday due to the increasing cost of natural gas. Let’slook at a typical velocity profile downstream of a singleelbow.

From this mathematical velocity profile model it isapparent the velocity profile at 10D from the elbow is farfrom being fully symmetrical. What isn’t apparent in thismodel is the amount of swirl generated by the elbow.According to research work performed at SouthwestResearch Institute (SwRI) by Terry Grimley, it would takeon the order of 100D for the profile to return to a fullysymmetrical, fully developed, non-swirling velocity profile[Ref 5]. More complex upstream piping, such as twoelbows out of plane, create even more non-symmetryand swirl than this model shows. Today’s USM musthandle profile distortion and swirl in order to be accurateand cost-effective. However, just as with orifice andturbine meters, installation guidelines should be followed.

In 1998 AGA released the Transmission MeasurementCommittee Report No. 9 entitled Measurement of Gasby Multipath Ultrasonic Meters. This document discussesmany aspects and requirements for installation and useof ultrasonic meters. Section 7.2.2 specifically discussthe USMs required performance relative to a flowcalibration. It states the manufacturer must “Recommendupstream and downstream piping configuration inminimum length – one without and flow conditioner andone with a flow conditioner - that will not create anadditional flow rate measurement error of more than±0.3% due to the installation configuration.” In otherwords, assuming the meter were calibrated with idealflow profile conditions, the manufacturer must then beable to recommend an installation which will not causethe meter’s accuracy to deviate more than ±0.3% fromthe calibration once the meter is installed in the field.

During the past several years a significant amount oftests were conducted at SwRI in San Antonio, Texas todetermine installation affects on USMs. Funding for thesetests has come from the Gas Research Institute (GRI).

FIGURE 3. 10-inch Meter SOS with 4 Chords

Concluding this discussion on external calculations, theresults demonstrate multi-path ultrasonic meters showgood correlation between the computed speed of soundand the meter’s reported speed of sound. Even thoughthere are differences between computed and reportedvalues, these remain relatively constant though out thetest period. This also suggests that when performing anon-line comparison of speed of sound, an alarm limit ofabout ± 0.3% between the meter and computed values,as recommended earlier, is reasonable. However, asshown in Figure 3, for a short interval the error exceeded0.3% (during periods of low (or no) flow and temperaturestratification). Since this situation can occur in the field,safeguards should be implemented to insure gas velocityis above some minimum value, and for a specified time,before alarming occurs. Thus, the use of independentestimates of gas speed of sound, derived from ananalysis of the gas composition, can be an effectivemethod of understanding how well an ultrasonic meteris performing.

BASICS OF USM INSTALLATIONS

When installing ultrasonic flow meters, many factorsshould be taken into consideration to insure accurateand trouble-free performance. Before discussing theseissues, let’s review the basics of a good installation.


Much of the testing was directed at determining howmuch error is introduced in today’s USMs when a varietyof upstream installation conditions are present. This waspresented in a report entitled Ultrasonic Meter InstallationConfiguration Testing at the 2000 AGA OperationsConference in Denver, Colorado. Following is an excerptfrom this report that shows the impact of upstreameffects on an ultrasonic meter.

Installation Effect: One Elbow Elbows Out Elbows In

Meter Orientation: 0˚ 90˚ 0˚ 90˚ 0˚ 90˚

No Conditioner, 10D 0.07 0.02 0.53 0.04 0.02 0.24

No Conditioner, 20D 0.13 0.11 0.05 0.10 0.11 0.12

19-Tube Bundle 0.35 0.37 0.13 0.37 0.15 0.22

Flow Conditioner #1 0.10 0.03 0.02 0.02 0.07 0.14



Meets AGA 9 Doesn’t meet AGA 9

Table No. 1 – GRI Installation Test Results

The preceding table presents metering accuracy resultsfrom a 4-path meter with a variety of upstream effects(single elbow, two elbows in plane and two elbows outof plane). Tests were conducted with no upstream flowconditioner and four brands of flow conditioners, alllocated at their manufactures recommended position.One thing to note is the 19-tube bundle did not performvery well. Also of importance is the meter met the AGA 9installation requirement test producing less than ±0.3%shift with no flow conditioner when located a minimumof 20D from the upstream effect.

In conclusion, for basic piping issues, upstream pipingdoes have an effect on the meter’s performance. Manycustomers choose to use a flow conditioner in order toreduce potential upstream effects. The use of 19-tubebundles is not recommended by most manufacturerstoday as the results are not consistent and are generallynot as good as with other flow conditioners. Flowconditioners are not always required. As can be seen inline two of the table, this meter passed the installationaffects test with no flow conditioner when located 20Dfrom the effect, and passed all but one test when locatedat 10D.

Other Piping Issues

One aspect to keep in mind when designing an ultrasonicmeter station is the use of control valves (regulators).Ultrasonic meters rely on being able to communicatebetween transducers at frequencies in excess of 100kHz. Control valves can generate ultrasonic noise in thisregion. How much depends upon several factorsincluding the type of valve, flow rate and differentialacross the valve.

Manufacturers have different methods for dealing withcontrol valve noise. Whenever an ultrasonic meter is usedin conjunction with a control valve, the manufacturershould be contacted prior to design. Following is adiagram of meter set with a flow conditioner and controlvalve.

FIGURE 4. Poor USM Piping Diagram

In this design pressure reduction occurs at about 27Dfrom the meter. There are two elbows between the meterand valve. At low flow rates this design would probablywork fine. However, as the flow rate increases, so doesthe amount of energy generated by the pressurereduction. The amount of noise generated is roughlyproportional to the square-root of the flow rate times thedifferential pressure. Thus, as flow rate (or differentialpressure) increases, so will the amount of noisegenerated. At some higher flow rate the meter will beunable to identify the signal, and measurement will cease.Following are two sets of waveforms. The first is a typicalsignal received by a pair of transducers when there is noextraneous ultrasonic noise. The second is an exampleof a meter experiencing noise from a control valve. Inorder to continue operation, the meter must be able tohandle this type of noise (Graph No. 2).

GRAPH NO. 1 – Typical USM Waveform

A better design would be to locate the meter furtherupstream (see Figure 6 following). By installing two teesbetween the meter and control valve, much of theultrasonic noise is reflected back downstream, helpingisolate the meter from the noise source. Also, in thisdesign, the meter has been located more than 70D fromthe valve. Ultrasonic noise, just like audible sound,becomes attenuated the further you get from the source.


FIGURE 5 – Better USM Piping Diagram

In conclusion, control valves can create enough noisethat will over-power the USMs signal. Control valvesshould be located away from the meter. Install the meterupstream of the control valve whenever possible as morenoise propagates downstream than upstream. Also, withthe higher pressure upstream, the USM will obtain astronger signal from the transducers, making it easier todetect the signal when in the presence of noise. Teesbetween the meter and the control valve are moreeffective than elbows at reducing noise. (about twice thenoise attenuation). Probably the most significant thingto remember is consult with the manufacturer during thedesign phase. Testing of USMs with control valve noiseis ongoing with all manufacturers, and better methodsof handling noise are constantly being developed. Somemanufacturers have internal meter digital signalprocessing that can handle increased levels of controlvalve noise.

FLOW CALIBRATION BASICS

The primary use for USMs today is in custodymeasurement applications. As was discussed earlier, theintroduction of AGA Report No. 9 has helped spur thisgrowth. Section 5 (of AGA 9) discusses performancerequirements, including flow calibration. It does notrequire meters be calibrated before use. However,paraphrasing, it does require ..“the manufacturer toprovide sufficient test data confirming that each metershall meet these requirements.” The basic accuracyrequirement is that 12-inch and larger meters be within

±0.7%, and 10-inch and smaller meters to be within±1.0%. Again, these maximum error values are “prior”to flow calibration.

Most customers feel their applications deserve, andrequire, less uncertainty than the minimum requirementsof AGA 9. Thus, a for virtually all USM custodyapplications, users are flow calibrating their meters.

In a majority of applications today customers are usingflow conditioners. USMs are designed to be installedwithout a flow conditioner. Part of the benefit of anultrasonic meter is there is no pressure drop. However,many feel that using a “high performance” flowconditioner (not a 19-tube bundle) further enhancesperformance. Even though data exists to support someUSMs perform quite well without flow conditioners, theadded pressure drop and cost is often justified byassuming uncertainty is reduced. One thing that everyonedoes agree on is that if a flow conditioner is used with ameter, the entire system should be calibrated together.Most companies have standard designs for their meters.They typically specify piping upstream and downstreamof the flow conditioner(s) and meter. Thus, USMs aretypically calibrated with either 3 or 4 piping spools.Calibrating as a unit helps insure the accuracy of themeter, once installed in the field, is as close to the resultsprovided by the lab as possible.

There are several flow calibration labs in North Americathat provide calibration services. Labs usually test metersthroughout the range of operation. Once all the “as-found” data points have been determined, an adjustmentfactor is computed. The adjustment is uploaded to themeter and either one or two verification points are usedto verify the “as-left” performance.

PERIODIC FLOW CALIBRATION

AGA 9 does not currently require an ultrasonic meter bere-calibrated. In the next update it is expected that allcustody applications will require flow calibration. AsUSMs have no moving parts, and provide a variety ofdiagnostic information, many feel the performance of themeter can be field verified. That is, if the meter isoperating correctly, its accuracy should not change, andif it does change, it can be detected. This, however,remains to be proven.

The use of USMs for custody began increasing rapidlyin 1996. Thus, with less than 5 years of installed base, itis difficult to prove USMs don’t require re-calibration.Many companies are not certain as to whether or notthey will retest their meters in the future. They are waitingfor additional data to support their decision.Manufacturers are also trying to show the technologyshould not require re-calibration.

The benefits of flow calibrating USMs have been welldocumented over the past few years [Ref 6]. Not onlydoes flow calibration reduce the meter’s uncertainty, it

GRAPH NO. 2- USM Waveform With Valve Noise


is often used to extend the rangeability of a meter toextremely low flow rates. This expanded rangeability canoften permit one meter to have a flow range of greaterthan 100-1 and a measured accuracy on the order of±0.1% (relative to the calibration facility). Flow calibrationalso has been used to validate a meter’ performancewhen less than the full compliment of transducers isoperating. This is beneficial for those times when atransducer is removed for inspection, but the meter mustremain in service.

During the next several years many meters will requirere-calibration in Canada. Their governmental agency,Measurement Canada, requires USMs be re-tested every6 years. Many meters will be due for re-testing in 2003and 2004. Once data is obtained from these tests, fromrandom re-testing by customer, and long-term data frommeters at calibration labs is analyzed, customers canbetter determine if they should re-calibrate their USMsin the future.

CONCLUSIONS

During the past several years ultrasonic meters havebecome one of the fastest growing new technologies inthe natural gas arena. The popularity of these deviceshas increased because they provide significant value tothe customer by reducing the cost of doing business.One of the most significant benefits is the reduction inmaintenance over other technologies.

There are several factors that can be attributed to thisincreased usage. First, as there are no moving parts towear out, reliability is increased. Since USMs create nodifferential pressure, any sudden over-range will notdamage the meter. If the meter encounters excessiveliquids, it may cease operation momentarily, but nophysical damage will occur, and the meter will return tonormal operation once the liquid has cleared. Mostimportantly, ultrasonic meters provide a significantamount of diagnostic information within their electronics.Most of an ultrasonic meter’s diagnostic data is used todirectly interpret its “health.” Additional diagnostics canbe performed by using external devices. This diagnosticdata is available on a real-time basis and can bemonitored and trended in many of today’s remoteterminal units (RTUs). USMs support remote access andmonitoring in the event the RTU can’t provide this feature.There are five commonly used diagnostic features beingmonitored today. These include speed of sound by path(and the meter’s average value), path gain levels, pathvelocities, path performance values (percentage ofaccepted pulses), and signal to noise ratios. By utilizingthis information, the user can help insure the proper meteroperation.

Probably the most commonly used tools are path speedof sound and gains. Speed of sound is significant sinceit helps validate transit time measurement, and gains helpverify clean transducer surfaces. When computing speedof sound in the field, care should be taken to collect

data only during periods of flow in the pipeline astemperature gradients will distort comparison results.Additionally, as shown in one of the graphical examples,low-flow limits should be implemented to insure pipelinetemperature is uniform and stable before comparingmeter speed of sound with computed values from gascomposition, pressure and temperature.

One significant benefit in performing online comparisonsbetween the meter’s speed of sound and a computedvalue is to provide a “health check” for the entire system.If a variation outside acceptable limits develops, theprobable cause will be temperature, pressure, or gascomposition measurement error rather than the USM. Inthis regard, the USM is actually providing a “health check”on the measurement system!

Monitoring path velocities is gaining in popularity daily.It has been shown that velocity information can helppredict if the inside of a meter is becoming contaminatedwith pipeline buildup [Ref 2]. In the past it was believedthat buildup inside of a meter would be detected by anincrease in gains. However, recent analysis of metershas shown this may not be the case [Ref 2, 7 & 8]. Thus,path velocity information will probably remain as thesingle most important tool for identifying if a meter isdirty internally.

Control valve applications are much better understoodtoday than a few years ago. All manufacturers havemethods to deal with this issue, and it varies dependingupon design. The manufacturer should be consulted priorto design to help insure accurate and long-term properoperation.

Today’s USM is a robust and very reliable device withmany fault-tolerant capabilities. It is capable of handlinga variety of pipeline conditions including contaminantsin the natural gas stream. In the event of transducerfailure, the meter will continue to operate, and some USMdesigns maintain excellent accuracy during this situation.When encountering contamination such as oil, valvegrease, and other pipeline contaminants, today’s USMwill continue working and, at the same time, provideenough diagnostic data to alert the operator of possibleimpending problems.

The issue of re-calibration of meters, after a number ofyears of service, has been discussed for a number ofyears. Most users are flow calibrating their USM prior toinstallation. Whether to re-calibrate after a number ofyears still remains a question to be answered. Somedesigners have opted to install a secondary in-situtransfer standard in the field to verify performance on aperiodic basis [Ref 8]. However, most users feel thismethod is too expensive and does not provide thenecessary traceable certification that might be neededshould the buyer of the gas question the accuracy ofthe primary meter. Thus, if a user is concerned, they haveopted to remove a sample and return it to the calibrationlab for checking.


As ultrasonic metering technology advances, so will thediagnostic features. In the near future USM diagnosticdata will become even more useful (and user friendly) asmore intelligence is placed within the meter. They willnot only provide diagnostic data, but will identify whatthe problem is. When this happens, ultrasonic metersmay be considered “maintenance free.”

REFERENCES:

1. AGA Report No. 9, Measurement of Gas by MultipathUltrasonic Meters, June 1998

2. John Lansing, Dirty vs. Clean Ultrasonic Flow MeterPerformance, AGA Operations Conference, 2002

3. AGA Report No. 8, Compressibility andSupercompressibility for Natural Gas and OtherHydrocarbon Gases, July 1994

4. Letton, W., Pettigrew, D.J., Renwick, B., and Watson,J., An Ultrasonic Gas Flow Measurement System withIntegral Self Checking, North Sea Flow MeasurementWorkshop, 1998.

5. T. A. Grimley, Ultrasonic Meter InstallationConfiguration Testing, AGA Operations Conference,2000.

6. John Lansing, Benefits of Flow Calibrating UltrasonicMeters, AGA Operations Conference, 2002.

7. John Stuart, Rick Wilsack, Re-Calibration of a 3-YearOld, Dirty, Ultrasonic Meter, AGA OperationsConference, 2001

8. James N. Witte, Ultrasonic Gas Meters from FlowLab to Field: A Case Study, AGA OperationsConference, 2002

John Lansing


FUNDAMENTALS OF NATURAL GAS CHEMISTRYSteve Whitman

Coastal Flow Measurement, Inc.Address

In order to understand the chemistry of natural gas, it isimportant to be familiar with some basic concepts ofgeneral chemistry. Here are some definitions you shouldknow:

Matter — anything that has mass and occupies space.

Energy — the capacity to do work or transfer heat.

Elements — substances that cannot be decomposedinto simpler substances by chemical changes. There areapproximately 112 known elements. Examples: carbon,oxygen, and nitrogen.

Atom — the smallest unit in which an element can exist.Atoms are composed of electrons, protons, andneutrons.

Compounds — pure substances consisting of two ormore different elements in a fixed ratio. Examples: waterand methane.

Molecule — the smallest unit in which a compound canexist or the normal form in which an element exists.Example: One molecule of water consist of two hydrogenatoms and one oxygen atom. One molecule of nitrogenconsist of two atoms of nitrogen.

Mixture — combination of two or more pure substancesin which each substance maintains its own compositionand properties. Examples: natural gas, gasoline, and air.

BONDS

The attractive forces that hold atoms together incompounds are called chemical bonds. There are twomajor classes of bonds — ionic bonds and covalentbonds. Most of the bonds involved in natural gascomponents are single covalent bonds. A single covalentbond consists of a pair of electrons shared by two atoms.A double bond is two pair of electrons shared betweentwo atoms. Some minor components of natural gas maycontain both single and double bonds. Aromaticmolecules, such as benzene, contain covalent bondswhere multiple electrons are shared among more thantwo atoms.

STATES OF MATTER

Matter exists mainly in three physical states — gas, liquid,and solid. Solids are rigid and have definite shapes. Thevolume a solid occupies does not vary much with

changes in temperature and pressure. In liquids, theindividual particles are confined to a given volume. Aliquid flows and assumes the shape of its container.Liquids cannot be easily compressed. Gases are muchless dense than liquids and solids. They occupy all partsof any vessel in which they are confined. Gases arecapable of infinite expansion and are easily compressed.They consist primarily of empty space because theindividual particles are so far apart.

UNITS OF MEASUREMENT

Quantities of matter can be expressed in a variety ofways depending on the nature of the substance beingmeasured. For example, solids are generally measuredby weight while liquids are measured by volume orweight. Gas is most commonly measured in units ofvolume but can also be measured by weight. Thestandard unit of volume used in natural gas measurementis a cubic foot corrected to a standard (stated) pressureand temperature. Large volumes of natural gas areusually expressed in units of one thousand cubic feet(Mcf). A million cubic feet is indicated by MMcf. However,when a gas sample is analyzed, the composition isusually expressed in mole percent — the percent (bynumber) of moles of the particular substance out of thetotal molecules of the gas. This is roughly equivalent tovolume percent. A mole is defined as the amount of asubstance that contains the same number of units asthe number of atoms in 12 grams of Carbon12 which is6.022 x 10-23, otherwise known as Avogadro’s number.Therefore, a mole of any element or compound is simplythe molecular weight of that substance in grams. Forexample, the molecular weight of water is 18, so a moleof water is 18 grams and contains 6.022 x 10-23

molecules. Moles are often used in chemistry becausethey make it easier to keep track of quantities ofsubstances involved in chemical reactions. One mole ofoxygen will react with 2 moles of hydrogen to form onemole of water. However, there would be an excess ofhydrogen if one gram of oxygen were reacted with 2grams of hydrogen.

NATURAL GAS

Natural gas is a mixture of many compounds which canbe classified into three major groups — hydrocarbons,inerts, and miscellaneous trace compounds.Hydrocarbons are compounds which contain hydrogenand carbon. Most of the hydrocarbons in natural gasare saturated, meaning that each carbon atom is bondedto four other atoms while each hydrogen atom is bonded


to only one carbon atom. This group of compounds isalso known as alkanes, paraffins, and aliphatics. Themost abundant alkane in natural gas is methane,commonly referred to as C1 because it contains onecarbon atom. Next is ethane (C2) with two carbons,followed by propane (C3), iso-butane and normal butane(C4), iso-pentane and normal pentane (C5), and hexanesand heavier hydrocarbons (C6+). The C6+ fraction cancontain up to 100 or more compounds includingaromatics such as benzene, toluene, ethylbenzene andxylenes (BTEX).

Deposits of natural gas are usually found with petroleumdeposits. Most of the natural gas and petroleum beingproduced today was formed by the decay of plants andanimals buried deep within the earth millions of yearsago. A small fraction of natural gas is also being producedby the decomposition of residential and commercialwaste in landfills.

Petroleum products are complex mixtures of aliphaticand aromatic compounds including sulfur and nitrogencompounds. Some of the most common refinedpetroleum products include methane gas, natural gasliquid or NGL (the ethane and heavier gas components),liquefied petroleum gas or LPG (mostly propane andbutane), gasoline (C5-C10), kerosene (C11-C12), dieselfuel (C13-C17), heating oil (C18-C25), and lubricants(C26-C38). Heavier residues of refining are referred toas asphalts.

In addition to hydrocarbons, natural gas componentsinclude the inerts, nitrogen and carbon dioxide, whichdo not combust. Therefore, they do not contribute tothe heating value of the gas, making them undesirable.Trace components commonly include hydrogen sulfide,water vapor, and helium. Hydrogen sulfide is highly toxicand is characterized by a very strong foul odor. It is alsoreferred to as sour gas. Helium is a true inert gas that isnonreactive with other compounds. Most of the nation’shelium production comes from natural gas deposits inTexas. Other less common trace components includeoxygen, hydrogen, and carbon monoxide.

PHYSICAL PROPERTIES

Physical properties are those that can be observedwithout causing any change in the chemical compositionof the specimen. One of the most important physicalproperties used in the petroleum industry is density.Density is the mass (weight) per unit volume of asubstance and may be expressed as specific gravity orrelative density. In liquids, the specific gravity is the ratioof the density of the liquid to the density of water. Ingases, it is the ratio of the density of the gas to the densityof air. The specific gravity of natural gas varies from lessthan .600 for gas containing mostly methane to greaterthan 1.000 for gas containing a high percentage of heavyend components.

Other important physical properties of gases includeboiling point, hydrocarbon dew point, odor, toxicity, andthermal conductivity. Boiling point is the temperature atwhich the vapor pressure of a liquid equals theatmospheric pressure. Hydrocarbon dew point is thetemperature at which hydrocarbons start to condensefrom a gas stream. This is important in gas productionand transmission because condensation in a natural gasline will lower the capacity of the line to carry gas.Consequently, there will be problems with compressors,dehydrators, and other processing equipment. Moreimportantly, liquids in a gas line make it impossible toaccurately measure the gas. The dew point also allowsthe heavier gases to be liquefied by processing. Theyare generally more valuable as liquids than gas.

Odorizing is important in gas processing andtransportation as a relatively inexpensive way ofdetermining the location of leaks. Unless it contains highconcentrations of hydrogen sulfide or othercontaminants, natural gas is normally odorless andnontoxic when it comes out of the ground. Nontoxicodorants, such as mercaptans, are added duringprocessing to make it detectable by sense of smell.

Thermal conductivity is the property that enables thedetector on a chromatograph to quantify the amount ofeach component in a gas mixture. Simply stated, it isthe ability of a substance to conduct heat. Thermalconductivity usually decreases with increasing particlesize. For this reason, helium makes a good carrier gasfor gas chromatographs. Its molecules are very small,allowing it to effectively draw heat away from the detector.

HEATING VALUE

The Btu, or British thermal unit, is a measure of the energyproduced by burning natural gas. A Btu is equal to theamount of heat required to raise the temperature of onepound of water one degree Fahrenheit at 62º F. The Btumay be expressed as dry, wet, or as delivered. The dryBtu calculation assumes that there is no water vapor inthe gas. As might be expected, the wet Btu is calculatedon the assumption that the gas is saturated with watervapor at standard conditions (60º F and atmosphericpressure). Hence, the wet Btu is less than the dry Btu.The as delivered or actual Btu is calculated by accountingfor the actual amount of water in the gas based ondelivery conditions. The Btu factors of the individualcomponents in natural gas increase with the number ofcarbon atoms. The table in Figure 1 illustrates the Btu ofthe most common components in natural gas.

The inerts commonly found in gas, carbon dioxide andnitrogen, do not participate in combustion and contributeno heating value or Btu to a gas.

Water vapor, though it does not burn, has a heating valueas defined by this industry. Water vapor has a heatingvalue of 50.4 Btu per standard cubic foot.


COMPRESSIBILITY

Compressibility is a measure of a real gas’s deviationfrom ideal gas behavior. Real gases behave like idealgases at low pressures and high temperatures. Becausethey deviate significantly at high pressures and lowtemperatures, the deviation should be accounted forwhen calculating Btu.

An ideal gas is one that obeys the Ideal Gas Law — itoccupies half its initial volume at twice the initial pressureand twice its volume at twice its (absolute) temperature.

REACTIONS AND PROCESSES INVOLVING NATURALGAS

Under normal conditions, the hydrocarbon componentsin natural gas are unreactive because saturated carbonsform chemically stable molecules. However, at hightemperatures and in the presence of oxygen, alkaneswill undergo combustion, giving off heat in the process.This capacity to produce heat, otherwise known asheating value, is what makes natural gas and crude oiltwo of the most precious energy sources in the world.The two major by-products of combustion, water vaporand carbon dioxide, also occur naturally as individualcompounds.

Other reactions and processes include corrosion, hydrateformation, and condensation. Corrosion is the processin which metals are oxidized in the presence of moistureor some other corrosive agent such as carbon dioxideor hydrogen sulfide. The degree of corrosion in pipelinescan be significantly reduced by removing thesecomponents from the gas during production. Theremoval of water vapor (dehydration) is commonlyaccomplished by bubbling the gas stream through avessel containing a chemical, such as glycol, which hasan affinity for water. Similarly, carbon dioxide andhydrogen sulfide are removed by passing the gas throughchemicals, such as amine, that have an affinity for eachof them.

Natural gas hydrates are solids that form from naturalgas molecules and water. The water molecules form ahoneycombed structure with a molecule of one of thenatural gas components occupying each void. Becausethese solids are denser than water ice, their formation isfavored at higher pressures. They are favored so muchto the extent that natural gas hydrates may form attemperatures up to 70º F. Like liquids, they interfere withthe flow of gas and cause operational problems. One ofthe places hydrates often form is in the sensing lines tothe orifice recorder. When formed in these lines, hydratescause incorrect recording of the differential pressure. Thisis what is often called meter freeze and is a major causeof measurement errors.

GAS LAWS

Accurate measurement of natural gas must take intoaccount a variety of gas laws involving relationshipsbetween temperature, pressure, and volume. Pressureis defined as force per unit and is commonly expressedin a variety of ways including psi, atmospheres, bars,mm of Hg, inches of water, and pascals. The quantitativerelationship between volume and pressure is summarizedby Boyle’s Law which states that the volume of a sampleof gas varies inversely with the pressure under which itis measured, given a constant temperature. If thepressure is doubled, the volume is reduced by half. Ifthe pressure is cut in half, the volume is doubled. Charles’Law states that the volume of a sample of gas variesdirectly with the absolute temperature, given a constantpressure. Absolute temperature is measured in degreesKelvin and starts with absolute zero, representing acomplete absence of heat. Thus, on the Kelvin scale,the freezing point of water is 273ºK and on the Celsiusscale, absolute zero is -273ºC. If the absolutetemperature of a gas is doubled, its volume will double.Combining Boyle’s Law and Charles’ Law results in asingle expression known as the Combined Gas LawEquation:

Boyle’s Law P1V1 = P2V2Charles’ Law V1/T1 = V2/T2Combined Gas Law P1V1/T1 = P2V2/T2

Based on Avogadro’s Law, equal volumes of gases, atthe same temperature and pressure, contain the samenumber of molecules. A standard molar volume of anideal gas is 22.4 liters per mole at STP (standardtemperature and pressure, 0ºC and 1 atm.).

According to the kinetic theory, gas pressure is causedby molecular collisions with the walls of the container.Therefore, the larger the number of molecules per unitvolume, the greater the number of collisions and thehigher the pressure. The average kinetic energy of themolecules is also proportional to the absolutetemperature.

While the molecules are at rest at absolute zero, at hightemperatures, the molecules move at increasing speedsresulting in higher pressures. Combining these conceptsinto one equation produces what is known as the idealgas equation:

PV = nRT

where P = pressure (atmospheres)V = volume (liters)N = number of molesR = the ideal gas constant (.08206 liter atm/ºK mol)T = temperature (degrees Kelvin)


CONCLUSION

Natural gas measurement and processing is basedlargely upon the fundamental principles of chemistry andphysics. The intent of this presentation was to bring forthsome of the basic concepts helpful to personnel involvedin the industry.

Carbon Number Name Btu @14.696 psia and 60ºF

1 Methane 1010.0

2 Ethane 1769.7

3 Propane 2516.1

4 Iso-butane 3251.9

5 N-butane 3262.3

6 Iso-pentane 4000.9

7 N-pentane 4008.9

8 Hexane 4755.9

FIGURE 1

Steve Whitman


TECHNIQUES FOR NATURAL GAS COMPOSITE SAMPLINGKris Kimmel

YZ Systems, Inc.3101 Pollok Drive, Conroe, TX 77303

INTRODUCTION

Since a gas sampling system can be referred to as a“cash register” it is very important that the correctsampling method be selected and the appropriateindustry standard be followed. Methods reviewed by thispaper will include spot sampling, composite sampling,and on-line chromatography. In addition, Gas ProcessorsAssociation (GPA) 2166-86 and American PetroleumInstitute (API) 14.1 will be described.

Natural gas is sampled to determine quality for custodytransfer applications, balance a plant process, or gatheringsystem. In the late 1970’s most natural gas custodytransfer contracts used gas volume (MCF) for the units ofmeasure. In 1978 Congress passed the Natural Gas PolicyAct in an attempt to deregulate the natural gas industry.This act dictated that natural gas should be purchased orsold based on energy content. Today natural gas ispurchased based on the amount of energy delivered. Thequantity of energy delivered is calculated by multiplyingthe gas volume per unit time by the energy value (BTU)per unit volume. A gas chromatograph is typically used toidentify individual components of the sample and theirquantity, thus determining quality of the sample taken.Correctly obtaining, transporting and analyzing the sampleis crucial to the accuracy required for custody transfer ofthis product.

Natural gas sampling can be accomplished using severaldifferent methods. The most common method, spotsampling (depicted in GPA 2166-86), involves manuallyfilling a sample cylinder connected to the center of thepipeline by a probe and separator. The entire volume ofgas that has flowed between sample periods is evaluatedfor its quality with that one sample or “snapshot in time”.It is obvious why this method is used only in gas streamsthat are extremely constant in volumetric rate andcomposition.

In locations where high volumes of gas are transferred,it is common for on-line chromatographs to be utilizedin determining the composition of the flowing gas streamin a real-time manner. Special temperature compensatingregulators are utilized to supply a constant representativesample from the center of the pipeline to thechromatograph.

However, many other applications have volumetric flowrates, which cannot justify the cost of an on-line analyzer,but require the same accuracy for custody transfer. Theseapplications require the use of an automatic composite

sampling system. These systems take periodic samples,based on time or volumetric flow rate, to create acomposite sample which best represents the truecomposition of the gas stream during the sample period.The sample is stored in a sample cylinder and transportedto a laboratory where its composition and quality aredetermined by gas chromatography. The advantages ofthese systems are their accuracy, low cost, simplicity indesign and minimum amount of maintenance requiredto keep them operating.

Our task is to determine “What technique of samplingbest fits the application, and what is the properinstallation and operation of said equipment for anaccurate sampling program?” Improperly applied orinstalled equipment and poor sampling methods cancause BTU variances worth thousands of dollars.

SPOT SAMPLING

In 1986 the Gas Processors Association (GPA) issued arevision of the standard 2166-86 “Obtaining Natural GasSamples for Analysis by Gas Chromatography” A totalof 1050 samples were collected and analyzed resultingin nearly 70,000 data points. Eight sampling methodswere evaluated for their impact on the components ofcommon natural gas.

The standard determined that “good samples” could beobtained using any of the eight methods, providedextreme care is taken while the samples are beingobtained. A brief description of each method…

Fill & Empty: requires sample line separator, samplecylinder, extension tube, appropriate valves and gauges.This method involves purging the sampling apparatus,then repeatedly filling and emptying the apparatus. Thenumber of cycles is line pressure dependent.

Controlled Rate Method: requires the same apparatusas Fill & Empty method with the addition of a flow tubeplug (orifice) in the extension tube. This method allowsfor a controlled flow of sample through the apparatusfor a specific length of time. The amount of time variesdepending on line pressure at the sample source.

Evacuated Container Method: requires an evacuatedsample container with a pressure of 1 mm Hg or less.Instead of an extension tube this method requires a venttube installed between the sample source and the samplecylinder. After a careful purge procedure the cylinder isfilled with natural gas.


Reduced Pressure Method: similar to Evacuated Cylindermethod. This method is not recommended when linepressure is below 100 psi.

Helium “Pop” Method: Begin with an evacuated cylinder(1 mm Hg). Carefully fill sample cylinder with Helium(approximately 5 psi). Use the same samplingconfiguration as the Evacuated Cylinder method. Afterpurging the sampling apparatus, the cylinder is filled toline pressure. This method requires a separatedetermination for helium to calculate the correct BTUcontent.

Glycol or Water Displacement Method: A sample cylinderfilled with clean glycol or water and a vessel to measurethe displaced liquid are required for this method. Carefullypurge the sampling apparatus. Open sample source,slowly open outlet valve allowing displacement of fluid.Close outlet valve as soon as gas can be seen or heard.The source gas must displace all fluid during sampling.

Floating Piston Cylinder Method: This sample cylinderhas a piston that creates a precharge chamber, andsample chamber. Fill the precharge chamber with an inertgas (slightly above line pressure). Carefully purge thesampling apparatus. Slowly open the outlet valveallowing the sample to displace the piston and fill thecylinder. The amount sampled should not exceed 80%of the cylinders capacity.

ON-LINE CHROMATOGRAPHY

An on-line gas chromatograph (GC) can provide almostreal time measurement of the natural gas. Unlikecalorimeters, which only give us energy value, a GC alsogives us the compositional makeup of the gas. Thisadditional information is fed into a flow computer andused in the AGA 8 supercompressibility equation, whichis then used in the volumetric calculation.

Because the GC is on-line it is able to run an analysis ofthe flowing gas every few minutes and supply the flowcomputer with up to date data. Most on-line analyzerscan also provide other important information, such asarchived data (hourly, daily, or monthly averages for BTUor gas compositions etc.), and limit alarms for componentconcentrations that may go out of a specified range. On-Line Gas Chromatographs (GC’s) have been widely usedwhen rigid custody transfer standards are needed fornatural gas trading, and when gas specifications needto be monitored closely.

Gas chromatography is a scientific method in which a gassample is separated into its component parts formeasurement. The gas chromatograph consists ofsubsystems that inject the sample, separate the sample,detect the components, integrate the peaks, and reportthe results. The injection, separation, and detection all occurin the heart of the GC known as the GC oven. The integrationand calculation of results are done in the controller, whichcan be considered the brains of the system.

COMPOSITE SAMPLING SYSTEMS

These are the major considerations of a compositesampling system installation and its operations:

• Locating the right sample point / sample probe.• Choosing the proper sample system

configuration.• Components of the sampling system.• Controlling the sampling system.• Sample cylinder selection.• Operational training.

Each of the components must receive proper attentionor the results of the sample may be compromised.

LOCATING THE RIGHT SAMPLE POINT

In all sampling applications, no matter what the product,it is important that the sample probe be in a locationthat best represents the product to be sampled. In naturalgas sampling, we are only interested in the compositionof the flowing stream that is in a gas phase. If the sampleprobe is in an improper location, the sample enteringthe system may not be representative and the endanalysis in the lab may be inaccurate. The sample pointshould be top center mounted in a straight and horizontalpipe section. It should be five pipe diameters away fromorifice plates, pipe bends, fittings, valves or otherrestricting devices that create turbulence.

When liquids are present in the piping, turbulence cancreate aerosols that are a mixture of liquids and gasesflowing together. As mentioned before our only interestis to sample the flowing stream in a gas phase. The liquidsthat create the aerosol can be present along the bottomor upper circumference of the pipe wall. Areas ofturbulence cast these liquids into flight where they mixwith the flowing gas to create a mist or aerosol. If oursample point is located in or close to this region, theaerosol will be sampled and stored in the sample cylinder.If these mists are included with the sample, the analysismay be inaccurate.

CHOOSING THE RIGHT SAMPLE PROBE

Once the best sample point has been selected, the nextconsideration is the type of probe to be used. Two typesof probes are commonly used, the single flow and dualflow probes. The inlet of any probe must be located inthe center 1/3 of the pipe. This will ensure that the sampletaken will be from a region of representative flow, andavoid sampling any liquids that may be present alongthe bottom or upper circumference of the pipe wall.

The single flow probe consists of a 316 stainless steelprobe, cut at a 45 or 90 angle on the end, and welded ormachined integrally with a male threaded nut. Incomposite sampling, a means must be provided forpurging the volume of gas residing in the single flowprobe before the sampling cycle begins. To achieve this,


some sampling systems either bleed gas continually tothe atmosphere, or to a low pressure source, ensuringthe probe is purged before the sampling cycle begins.

PROPER SAMPLE SYSTEM CONFIGURATION

Composite sampling systems are normally configuredin one of two ways. They may be mounted on a levelingsaddle, or if areas of vibration are a concern, mountedon a freestanding pole next to the pipeline. Thisconfiguration utilizes a dual flow probe. Or the samplingsystems can “probe mounted.”

“Probe Mounted” sampling systems are close coupledwith the single flow probe, and are designed to use theentire volume of gas in the probe as an actuation sourceto power the sampling system. When the actuation gasis exhausted, the next sample is removed from thepurged probe to reside in the pump, awaiting the nextactuation. This process is called “self purging”; meaningthe entire volume of gas in the probe is used and replacedbefore each sample is taken.

The dual flow probe is designed to create a slipstreamloop for a continuous supply of representative samplegas from the middle 1/3 of the pipeline. It requires onlyone threaded coupling on the pipeline for the inlet andreturn of the probe. The dual flow probe consists of astainless steel probe that extends to the center 1/3 ofthe pipeline, and a return port located at the bottom ofthe probe body. The dual flow probe works by using thedifferential pressure created between the sample pointlocated in the center 1/3 of the pipeline, and the returnport at the bottom of the probe body inside of the treadedcoupling. This pressure imbalance creates a slipstream,which provides a constantly moving supply of freshrepresentative sample gas, from the middle 1/3 of thepipeline to the sampling systems manifold. Tests haveshown that the probe penetrating the center 1/3 of thepipeline, if bent 90 degrees and facing the flowing gas,may increase the velocity of the slipstream and ensureits direction of flow.

“Probe mounted” composite sampling systems areconfigured to mount directly on the pipeline. Thiseliminates the need for a dual flow probe because thesystem comes complete with a single flow sample probeand isolation valve as integral components. Because thesample pump and pipeline are closely coupled, only asmall volume of gas occupies the integral probe thatneeds to be purged. Additionally, no external tubingexists for the sample gas to flow through to the samplingsystem. There is concern that the exposed tubing, usedto interconnect a dual flow probe with the samplingsystem, may provide a means for heavy components inthe slipstream to condense due to chilling of the tubingunder specific ambient temperature conditions. Loss ofthese components could have an effect on the integrityof the sample.

It is important to mount the sampling system verticallyabove the sample point to prevent the flow of liquidsinto the system. Also, if the sampling system is to bemounted on a leveling saddle or pole, it should bemounted as close to the sampling point as possible withno traps or right angles in the slipstream tubing.

COMPOSITE SAMPLING SYSTEM COMPONENTS

Most composite sampling systems contain thesecomponents:

• Manifold / Probe Body• Sample Pump• Filter / Regulator• System Controller

The sample pump is the component that pumps thesample in specific volumes into the sample cylinder. Thesample volume can be changed to accommodatedifferent size sample cylinders and sampling periods. Ina probe mounted automatic sampler, the sample pumpis powered by natural gas supplied through a lowpowered solenoid valve that is normally closed. Whenenergized, natural gas flows from the probe body, throughthe regulator and solenoid valve, to actuate the pump.

A filter/regulator is integral with the system to regulatethe pump actuation gas. The regulator has a filter andpot to remove any particulate and moisture from thepump actuation gas. It is important for the regulator tobe located downstream of the pump sample inlet on themanifold, so the pressure drop will not affect the integrityof the sample.

The manifold and probe body each provide points ofentry to the sample pump, and for sample discharge tothe sample cylinder. A purge valve is integral with themanifold and probe body that provides communicationbetween the sample cylinder and the inlet of the sampleprobe. The purge valve provides a means for purgingthe ambient air and stagnant gas in the sample probeand related tubing, manifold, and sample cylinder beforethe start of a sample period.

The enclosure that houses the sampling system must fitthe extremes of the environment it will be placed in.Offshore and marine applications should have anenclosure that is resilient to corrosion. This is also trueof the sampling system components. Not only shouldthey be compatible with the gas they are sampling, butalso with the environment they are operating in.

CONTROLLING THE SAMPLING SYSTEM

A composite sampling system controller determineswhen a sample should be taken, and activates the systembased on pre-set parameters. The type of control utilizedis dependent on the flow conditions of the stream to besampled. In all flowing condition, the sampling systemmust be “proportional-to-flow”. Four typical flowconditions are:


1. Continuous flow / constant volumetric rate.2. Intermittent flow / constant volumetric rate.3. Fluctuating volumetric rate w/electronic

measurement.4. Fluctuating volumetric rate w/out electronic

measurement.

In applications where the volumetric rate does notchange, and the flow is constant, a sampling systemwith a simple re-cycling timer is sufficient to beproportional-to-flow. In this mode, the controller willactivate the sample pump at predetermined time intervalsover the entire sampling period. Based on the pumpdisplacement and sample cylinder volume, a “timebetween strokes” can be selected and programmed intothe controller which will bring the sample cylinder to linepressure at the end of the sampling period.

In applications where the volumetric rate is constant,but the flow starts and stops, a sampling system with asimple re-cycling timer and an “on-off” switch, issufficient to be proportional-to-flow. The “on-off” switchsenses differential pressure at the orifice plate in themeter. When differential pressure is not present, thecontroller is interrupted. When differential pressure exists,the controller is allowed to operate the sampling system.Operating in this manner is considered to beproportional-to-flow, and not sampling a stagnant non-moving stream.

Some applications have a volumetric flow rate thatfluctuates the entire range of the meter setting, whilethe flow starts and stops. In this scenario, the samplingsystem will need to change the rate at which it samples,and stop when flow ceases in order to be proportional-to-flow. This type of sampling system has a controllerthat will accept a proportional-to-flow signal input froma flow computer. A simple calculation can be used todetermine the sample rate needed to bring a samplecylinder to line pressure during the sample period.

If there is no flow computer to provide the flow signal, orthe sampling system cannot be interfaced with the flowcomputer, some sampling systems can provide their owndifferential pressure transmitter as a part of the samplingsystem. The differential pressure is measured andconverted into an analog flow signal to drive the samplingsystem proportional-to flow. The advantage to thesesystems is that they stand alone, and require no interfacewith a flow computer or external power.

SAMPLE CYLINDER SELECTION

One of the most critical aspects involved in compositesampling is the proper use and selection of the samplecylinder. There are two types of sample cylinderscommonly used. The constant volume and the floatingpiston sample cylinders. The constant volume cylinderis typically used with “dry gas”, and the floating pistoncylinder is used with gas having a high BTU content or“wet gas”. The floating piston sample cylinder keeps the

sampled gas at line pressure until it is analyzed. This willensure that all constituents present in the sample, in agas phase, will remain in the sample and not condensedue to retrograde condensation.

The floating piston sample cylinder is equipped withrupture discs and purge valves. The interconnectingtubing between the sampling system and any samplecylinder should be as short as possible to eliminate deadspace. The Gas Processors Association (GPA), standard2166-86 contains procedures for purging and cleaningthe cylinders. Department of Transportation (DOT)regulations govern the design, manufacture, andtransportation of these cylinders. Every effort should bemade to comply with these regulations.

API 14.1

This standard details requirements and proceduresrequired to correctly collect and handle natural gassamples for custody transfer. A particular emphasis isplaced on the impact of hydrocarbon dew point to theoverall accuracy and success of your sampling program.

Accurate sampling from gas streams with temperatureat the hydrocarbon dew point temperature is moredifficult than sampling from streams with temperaturesabove the hydrocarbon dew point temperature.

If any part of the sampling process causes the sampleto fall below the hydrocarbon dew point, scattered andbiased analytical results and non-representative samplesare likely to result. In order to avoid this problem, thesample gas temperature must remain above the gashydrocarbon dew point during sampling. This can beaccomplished by heating sample probes and by heattracing lines, regulators and sample cylinders or byemploying some other means of delivering heat to thefluid in the sampling system.

Due to the uncertainty in measuring or calculating thehydrocarbon dew point, it is recommended that the gasbeing sampled be maintained at 20-50oF (11-28oC) abovethe expected hydrocarbon dew point throughout thesampling system. If ambient temperatures are above thehydrocarbon dew point, heating may not be required.When the sampling process involves a pressurereduction, provide sufficient heat at or prior to, the pointof pressure reduction to offset the Joule-Thomson effect(approximately 7oF (3.9oC) per 100 psi of pressurereduction.

OPERATIONAL TRAINING

The scope of any sampling program should be welldefined. Technicians responsible for installing thesampling system, gathering the samples and conductingthe analysis should be well versed in their tasks. A lapsein any step of the process could skew the results. If anydoubt exists regarding the proper use of compositesampling systems and their ancillary equipment, most


manufacturers offer operational training and technicalassistance.

CONCLUSION

Obtaining a sample is easy. However, obtaining anaccurate sample that is representative of the flowingstream is a challenge. The process requires use of theproper method, selecting the correct equipment, andattention to detail.

REFERENCES

Garrett Lalli, “Techniques for Composite Sampling”,Proceedings of the International HydrocarbonMeasurement Short Course, 1999

David Anders, “Energy Measurement Utilizing On-LineChromatographs”, Proceedings of the InternationalHydrocarbon Measurement Short Course, 2001.

Gas Processors Association, 2166-86. Obtaining NaturalGas Samples for Analysis by Gas Chromatography.

American Petroleum Institute, Chapter 14 section 1.Collecting and Handling of Natural Gas Samples forCustody Transfer, Fifth Edition, June 2001.

Kris Kimmel


TECHNIQUES OF COMPOSITE GAS SAMPLINGDavid J. Fish

Welker Engineering CompanyP.O. Box 138, Sugar Land, TX 77487

INTRODUCTION

The level of interest in effective and accurate gassampling techniques is currently at a very high prioritywithin the natural gas industry. With the fluctuating rangesin natural gas prices, exploration interests, profitability,deregulation and consolidation of the work force,recoverable revenue must be accounted for. At largevolume delivery points, a 3-5 BTU error in energydetermination can cost companies tens of thousands ofdollars within a very short time period. Accurate samplingtechniques must be implemented with equal interest asis given to accurate volume measurement. MMBTU isthe total of volume and energy. Sampling is the energydetermination delivery system for this equation.

Proper knowledge and correct implementation of theprimary methods of sampling will provide a qualityminded sampling program for the measurementpersonnel in natural gas companies.

The amount of hydrocarbon product that is transportedbetween producer, processor, distributor and user issignificant. To be able to verify the exact composition ofthe product is important from an economic and producttreatment standpoint. A small percentage of savingsmade by correctly determining composition will quicklyrecoup the investment made in the purchase of a systemdesigned to obtain an optimum sample. In addition, ifthe best sampling procedures are followed, the potentialfor disputes between supplier and customer will begreatly reduced. The importance of properly determininghydrocarbon gas composition, benefits all partiesinvolved and will achieve greater significance as thisprecious commodity becomes less plentiful and moreexpensive.

Sampling is an art. It begins with the proper selection ofthe method and procedure for a given location. Thencomes the correct selection of the equipment to be usedfor that installation. Proper installation and operation ofthe chosen apparatus will determine whether arepresentative sample is taken. The technician shouldbe trained in the procedures that will be used and howto carry out the entire sampling operation. A solidunderstanding of the rules and regulations surroundingthe transportation of the collected sample, will assurecompliance with rules. The end result will be an accurateand repeatable sampling procedure to provide arepresentative sample to the laboratory and an analysisthat can be trusted as correct.

GAS SAMPLING

The definition for sampling from the Gas ProcessorsAssociation publication GPA 2166-86, is as follows, “Theobject of any sampling procedure is to obtain arepresentative sample of hydrocarbon from the systemunder investigation. Any subsequent analysis of thesample regardless of the test, is inaccurate unless arepresentative sample is obtained.” And, from ISO-10715, a representative sample is, “A sample having thesame composition as the material sampled, when thelatter is considered as a homogeneous whole.” API 14.1offers a similar statement in the latest revision, “arepresentative sample is compositionally identical or asnear to identical as possible, to the sample sourcestream.” These standards are the most commonreferenced on Gas Sampling procedures, along with theAGA Gas Measurement Manual, Part No. 11, Section11.3 and ASTM 5287-97.

Proper sampling is fundamental to the correctdetermination of the product composition. In a majorityof cases, the sample is also the source for thedetermination of the specific gravity of the gas. This figureis a critical component of the flow formula, from whichwe derive the product quantity. An error in samplingeffects both quality and quantity, and ultimately,profitability.

Natural Gas sampling has been performed for years withtechniques handed down from generation to generation.Many of the older methods are not sufficient to meettoday’s requirements of accuracy and repeatability;however, standards have been developed to reachtoward these demands. The most widely knownstandards are GPA-2166-86 and ISO-10715. API hasproduced a revised API 14.1, which was published inJune, 2001. This new standard has already generatedsignificant interest in proper sampling techniques, dueto a large volume of data produced during the revisionwork.

Proper maintenance of all sampling equipment is vital tothe operations of all sampling methods. A review ofrelative sampling standards and the manufacturer’soperation, installation and maintenance manuals, is animportant step the total accurate sampling process. Dirtyor poorly maintained sampling apparatus will adverselyaffect the final results and profitability of the gascompany’s operation.


SAMPLING COMPONENTS

Sampling can be accomplished by primarily threetechniques; spot, continuous composite or continuouson-line sampling systems. Our concern in this paper iswith the composite sampler system. The variouscomponents of a composite sampling system deserveindividual consideration before the sampling system isinstalled and initiated.

Regulators - Composite Samplers use instrumentregulators for instrument supply only. A compositesampler main body should be designed to operate underfull line pressure so that the process is not regulated inpressure prior to the sample being taken. The regulatorshould be capable of handling the full line pressure and,the gas composition and contaminants

Relief valves - Regulators should have a relief valveinstalled downstream, if the equipment downstream isnot able to withstand full upstream pressure. Mostsampler actuators need a regulated and protectedinstrument supply. Regulators will not always give aguaranteed shut-off and their lock-up pressure will climbto a dangerous level should there be failure to attain agood shut-off as a result of seal damage, diaphragmdamage or impurity build-up on working parts andsensing lines.

Valves - If shut-off/isolation valves present a restrictionthat causes a pressure drop, it is possible thatcondensation could occur. When used with a collectioncylinder it is important that there are no leaks from thegland. Light ends will be the first to leak off, therebycausing the sample to be over-represented with heavyends. It is wise to use valves with soft seals to give apositive shut-off. Large orifice valves should be used,as restrictive valve paths can cause fractionization ofthe sampled gas.

Filters - Filters should be carefully chosen, if used at all.It is not wise to filter the pipeline gas prior to taking thesample, as alteration of the gas is a possibility. A filteron the instrument supply, downstream of the sample grabpoint, is acceptable.

Pipework - Should be as short and as small in diameteras possible. It will help maintain the sample integrity. Probeand lead lines to continuous samplers should slope backtowards the pipeline. All connections must be leak tight!

Heating Elements - There is sufficient evidence to showthat heating all components of a sampling system isfrequently a prudent step in having a reliable and accuratesampling system. The hydrocarbon dew point of a naturalgas stream is a critical issue in obtaining a representativegas sample.

Probes - The use of a probe is imperative in propersampling. Without the use of a probe, an accurate sampleis not likely to be taken. The probe may have a bevel or

be cut flat across the end. The probe itself is typically 3/8 inch tubing, but can be larger in some cases. The probeshould be equipped with a full open ball valve or gatevalve, or large ported valve. Do not use a metering valveor small orifice valve. The probe design must be checkedfor wake frequency and bending moment calculations,to insure that it is structurally sound in design. The correctplacement is at the top of the pipe, into the center onethird or at least 200 mm (8 inches) for larger diameterpipes; in an area of minimum turbulence, that is to say,away from headers, bends, blow-down stacks, elbows,traps, valves, etc. Turbulence will stir up the contaminatesthat usually reside at the bottom of the pipeline and aretherefore not normally part of the gas stream. By havingthe probe at a point of turbulence these contaminateswill be taken into the sample, thereby providing a samplethat is not representative. Probe placement should be ina flowing and non-turbulent spot in the pipeline.

Sample Pump - These pumps are, of course, needed toextract the sample from the line and transfer the sampleto the analyzer or collection cylinder. They should havethe capability to be able to extract the sample underflowing conditions, maintain a consistent discrete sizeof sample, take a fresh purged sample every time andhave the ability to be controlled by a timer or proportional-to-flow controller. This forms the heart of the continuousgas sampling system. If the pump or sampler is unableto perform all these functions, a representative samplewill not be taken and the sampling exercise will be flawed.

Pumps can be either pneumatic or electric. The safetyrequirements of the electrical components such asmotors and solenoid valves and the environmentalprotection rating, dictate careful selection andcompliance with applicable codes. The selection optionsmay well be limited if electrical components haverequirements which are incompatible with the use ofstandard components elsewhere in the system.

Sample Cylinders - Used for the collection of gasesand light liquid hydrocarbons, sometimes called “samplebombs”. The cylinders come in two forms; one is a plainsingle cavity cylinder with a valve at each end, and theother is known as a Constant Pressure Sample Cylinder,which takes the form of a closed end cylinder with aninternal piston. Using the Constant Pressure Cylinder thesample can be collected at a pressure above the vaporpressure of the light ends. By having the piston at theend of the cylinder, the need for excessive purging iseliminated. Pulling a vacuum in the sample cylinder(which is often destroyed by technicians) or using thewater outage method is not necessary. The hook-up issimple and straightforward making the operation easierfor technicians and minimizing the possibility of anincorrect sample being taken.

The need for maintaining the gas at full line pressurefrom beginning to end has been recognized as a positivefeature for several years. Any reduction in pressure andchange in temperature from the line condition at the time


of sample, was deemed to alter the gas analysis in almostevery case. Only low BTU gas (1000 BTU and below)seemed to possibly escape alteration.

Then, once the cylinder is in the laboratory, a gas supplycan be connected to the pre-charge side equal to thepipeline pressure. As the sampled gas is injected intothe chromatograph, the piston is being pushed by thepre-charge gas. While the cylinder is being emptied, fullpressure is being maintained and the gas compositionis not being altered as a result of pressure reduction.The cylinder can be stored, or sent to another laboratoryfor confirmation, and when the remaining gas is analyzed,it will give repeatable results, because the condition ofthe gas is maintained by the constant pressure cylinder.

The cylinder is equipped with valves, safety reliefs andgauges on both ends, and thus the pressure can becontrolled and monitored at all times on both ends. Thetemperature is maintained just as with Standard Cylindersi.e. heating blankets, ovens, or water baths.

The Constant Pressure Cylinder also brings with it,additional safety in handling the sample. No longer doyou have to purge the cylinder and vent large amountsof gas to the atmosphere. A brief purge of the sampleline up to the cylinder is all that is required. The piston isat the sample end of the cylinder when you commenceto fill, so there is no “dead volume” to purge.

Also, because of the design of the cylinder, with sealson the end of caps, it cannot be over pressured to thepoint of exploding. If the cylinder is over pressured, thesafety reliefs will allow the pressure to escape. In therare event that they fail to work the cylinder will swelland the seals will stop sealing, allowing the product toescape safely.

Constant Pressure Cylinders provide for accuratesampling procedures, better sampling systems,repeatability, safer handling, accurate analysis andstorage of samples as well as storage of gas and liquidstandards for the laboratory.

All updated ISO, GPA, ASTM and API standards andcommittee reports, address the proper usage of Standardand Constant Pressure Cylinders for the gas and liquidsindustry.

Sample cylinders should be constructed with a materialthat is compatible with the gas. For instance, H2S canbe absorbed into the structure of 316 stainless steel.This will necessitate coating the inside of the cylinder,otherwise the resultant sample will not be trulyrepresentative.

Sample cylinders are normally protected with burstingdiscs. They are less expensive and are lighter weightthan relief valves, though their proper selection andreplacement should have more importance than issometimes given them.

With all of the notes on the various components shouldgo the comment which is one of the basic rules ofsampling. The materials of construction of the samplingequipment that come into contact with the sample areto be compatible with the product being sampled. It isnormally reasonably safe to use 316 stainless steel andViton elastomeric components. One should look for thesematerials in selecting equipment, and ask questions ofsuppliers about material selections.

HYDROCARBON DEW POINT KNOWLEDGE

An additional major factor in correct sampling proceduresis an awareness of the hydrocarbon dew point of thegas stream being sampled. The importance of knowingthe HCDP is related to 1). The ambient temperature; 2).The temperature of the equipment being used to collectthe sample; and 3). The temperature of the flowingstream. The creation of liquids due to equipment designand equipment temperature must be avoided.Determination of the HCDP of the gas stream can bedone by the chilled mirror method or by the use of anumber of equation of state models for hydrocarbon dewpoint determination. There are several programs availablesuch as Peng-Robinson or SRK. The variations of thecalculated results between different equations of stateare so wide, that it is strongly recommended to add 20°to 50°F (11° to 28°C) to the answers. This is to assurethe operator that he is designing his sampling systemtemperature requirements above the actual hydrocarbondew point.

COMPOSITE SAMPLING

Composite sampling is the proven middle groundbetween spot sampling and the continuous on-lineanalytical gas chromatographs.

A composite gas sampler or gas sampling systemconsists of a probe, a sample collection pump, aninstrumentation supply system, a timing system (timedor proportional-to-flow) and a collection cylinder forsample transportation. Its sole objective is to collect andstore over a period of time or volume of flow, arepresentative composite sample at line conditions,allowing it to be transported to the laboratory forrepeatable analysis, without changing the chemicalcomposition, heating value, or physical characteristicsof the products being sampled.

This package will mount on a pipeline and collectsamples over a desired sample period unattended. Forthe sake of illustration, a description of a common systemis provided here.

A probe should be installed which extends into the middle1/3 of the flowing stream. This location should be chosento provide a representative sample of the gas stream,thus devoid of stagnant gas, i.e. blowdown stacks, anddevoid of free liquids and aerosols, i.e. downstream ofpiping elbows or orifice fittings which cause turbulent


flow. The probe should have a large ported outlet valveto prevent fractionation, resulting in compositionalchanges in the gas.

A self-purging sample collection pump designed tooperate under pipeline conditions should be locatedabove and as close to the probe as is practical andpossible. Filters, drip pots, screens, regulators and suchconditioning equipment shall not be placed between theprobe and the sampler, as this will affect therepresentative nature of the sample which is taken. Inletcheck valves can also cause the gas to fractionate, dueto the restriction it causes in the line.

The sampler instrumentation source can be from thepipeline itself (the most common installation) or from anauxiliary instrument supply.

The timing system can be a simple function timer andsolenoid, a proportional-to-flow signal conditioner andsolenoid, or simply, a solenoid ready to be connected tofield RTU’s or other electronic devices capable ofproviding the desired signal.

The sample collection cylinder can be either aconventional single cavity sample cylinder or the morecontemporary piston style, constant pressure samplecylinder. As these cylinders will be transported, theyshould meet design criteria such as ASME Section 8 orcarry approvals from recognized agencies such as D.O.T.,DNV, Lloyds, etc. A typical system would include a 500mlcylinder which would be used on a monthly basis tocontain 2200+ bites of .2 cc size during the sampleperiod.

Using the grab sampler, it is possible to obtain arepresentative sample over a pre-determined period. Itis the only practical method for collecting a continuoussample. The grab sampler will introduce a set volume,taken in equal amounts to the collection cylinder over aset period and is the preferred method when arepresentative sample has to be taken over time.

It has the advantage of being able to measure preciselya predictable amount over a given period when using atimer, and can also take samples proportional-to-flowwhen taking a modified signal from a flow meter.

In addition, the sample is taken from the flowing streamat the system pressure and can be fed into the sampleror sample cylinder at the flowing pressure; thus anychanges in composition is avoided.

Another feature required of any sampler is that it shouldnot have areas or pockets where residue of previoussamples can accumulate and, must take a fresh grab orbite of gas each time it samples.

This then describes a typical continuous compositesampling system, which has been proven to provide arepresentative sample for analysis. Such systems have

been tested against continuous on-line gas calorimetersand gas chromatographs with + 1 BTU accuracy for thetotal sample period, at considerably less cost andmaintenance than on-line GC’s.

TRANSPORTATION

The transportation of natural gas samples is a veryimportant issue for both the companies that are involvedand the individual personnel who are transporting thesamples. The United States Department ofTransportation (DOT), covers the transportation ofsamples in CFR-49. Everyone involved in transportingsample cylinders and other sampling apparatus, both toand from sample collection locations, should be familiarwith the rules and regulations set forth in CFR-49.

As well as the safety issues, markings and forms thatare to be filled out for DOT purposes, otherconsiderations should be addressed as well. Amongthese are:

• Proper tagging of the cylinder for time, date,location of the sample

• Pressure and temperature of the pipeline source• Technician who took the sample• Method used to obtain the sample• Plugging of the valves and checking for leaks

prior to transport• Protection of the cylinder and sample apparatus

during transport, both to and from the samplelocation

• Temperature concerns during transport, both toand from the sample location - if necessary orrequired

• Other company procedures that will assist in thesuccess of a quality sample being delivered tothe laboratory for an accurate analysis.

CONCLUSION

The methods, techniques, and designs of today’ssampling systems should be considered by everyproducer, shipper, buyer and end-user. Regardless ofthe application or installation, there is a system whichmeets your needs, and will effect your company in theprofit and loss column. Sampling and metering are thecash register of your company. Sampling is an art!Examine your methods, procedures and needs closely.


REFERENCES

“Proper Sampling of Light Hydrocarbons”, O. Broussard,Oil and Gas Journal, September 1977

“Standard Cylinder vs. Constant Pressure Cylinders”, D.J. Fish, Gas Industries, January 1994

“Analyzing Heating Value”, T. F. Welker, Pipe Line Industry,October 1990

“DOT Requirements for the Transportation of SampleCylinders”, D. J. Fish, Presented at American School ofGas Measurement Technology, September 2001

“Natural Gas Sampling”, T. F. Welker, Presented at AGAAnnual Meeting, Anaheim, California, 1981

“Methods, Equipment & Installation of CompositeHydrocarbon Sampling Systems”, D. J. Fish, Presentedat Belgian Institute for Regulation and Automation,Brussels, Belgium, 1993

“Practical Considerations of Gas Sampling and GasSampling Systems”, D. J. Fish, Pipeline and Gas Journal,July 1997

“Selection and Installation of Hydrocarbon SamplingSystems”, D. A. Dobbs & D. J. Fish, Presented atAustralian International Oil & Gas Conference,Melbourne, Australia, 1991

Various Standards of AGA, GPA, API, ASTM and ISO

David Fish


OPERATIONS OF ON-LINE CHROMATOGRAPHYCharlie Cook

Daniel Measurement and ControlHouston, Texas 77055

INTRODUCTION

Since the early eighties it has become common in theUnited States, and elsewhere in the world, for naturalgas to be bought and sold based on the amount of energydelivered. The quantity of energy delivered is calculatedby multiplying the gas volume per unit time by the heatingvalue (BTU) per unit volume.

Gas chromatography is normally employed for thecalculation of the heating value. However, when the gaschromatograph runs an analysis we not only get theheating value, but other valuable information; such asgas composition and relative density. This additionalinformation is used in the gas volume calculations. GasChromatographs (G.C.s) have been widely used inhydrocarbon processing facilities when high qualityproduct specifications are required, or when rigid custodytransfer standards are needed for natural gas trading.

Gas chromatography is a scientific method in which agas sample is separated into its component parts formeasurement. The gas chromatograph consists ofsubsystems that inject the sample, separate the sample,detect the components, integrate the peaks, and reportthe results. The injection, separation, and detection alloccur in the heart of the GC known as the GC oven. Theintegration and calculation of results are done in thecontroller which can be considered the brains of thesystem.

This paper describes the basics of how a BTU analyzerworks. In a typical natural gas application the gas isseparated into eleven different components. Hexane’sand heavier components such as heptane, octane, andnonane are combined together to form a single peakknown as C6+. Because we separate up to thesecomponents, we call this a C6+ application. Typicalranges are as follows:

Component RangeC6+ Hexanes and heavier (0-0.7%)C3 Propane (0 -5%)IC4 Isobutane (0-1%)NC4 Normal Butane (0-1%)NeoC5 Neopentane (0-1%)IC5 Isopentane (0-1%)NC5 Normal Pentane (0-1%)N2 Nitrogen (0-15%)C1 Methane (0-100%)CO2 Carbon Dioxide (0-15%)C2 Ethane (0-15%)

The following section describes and illustrates what takesplace in the GC oven.

GC OVEN FOR C6+ NATURAL GAS APPLICATION

The GC oven is heated at a constant temperature (around80C), and has a constant flow of carrier gas (usually highpurity helium) flowing through it. The carrier gas is usedfor transporting the sample through the oven during theseparation process. The oven consists of three valvesand four columns, and a set of balanced thermalconductivity detectors.

The first valve is the sample valve which is used forinjecting the sample into the system. The second valveis called a backflush valve. The backflush valve is usedfor backing out the C6+ heavies so that they can bedetected first without having to go through columns twoand three. The third valve is a dual column valve whichis used to trap the light components and allow themedium components to go around column three.

Of the four columns the first three are made up of tubingwith special material inside called packing that isdesigned to separate the gas into its differentcomponents. The fourth column acts as a buffer to keepthe flow steady when the valves are switched. Thefollowing steps illustrate the process taking place in theGC oven.

GC OVEN ILLUSTRATION

Step 1: Start of an analysis. Sample valve (V-1) OFF,backflush valve (V-2) ON, and dual column valve (V-3)ON. The sample purging system maintains a sample inthe gaseous phase and passes the sample throughtransport tubing to the sample valve and through thesample loop.


Step 2: The sample valve is turned ON to capture aprecise volume of sample, and to allow the carrier gas(helium) to sweep the sample into column 1 to start theanalysis.

Step 3: The sample valve turns back OFF to start purgingthe next stream. The backflush valve turns OFF reversingthe flow through column 1 so that C6 and heaviercomponents elute first (all combined as a single peak).Note: By switching the direction of flow through column1 we bypass columns 2 and 3. This helps to expeditethe analysis.

Step 4: C6+ (heavy) is on its way to the detector. Column2 separates the mediums C3, C4’s and C5’s while thelightest components N2, C1, CO2, and C2 continuetraveling through column 2 into column 3

Step 5: Dual column turns OFF after trapping the lightestcomponents N2, C1, CO2, and C2. The mediumcomponents bypass column 3 by going through therestrictor column and follow C6+ to the detector. Onceagain this helps speed up the analysis.

Step 6: After the heavy and medium components eluteto the detector the dual column valve is turned back onfreeing the light components and allowing them to crossthe detector next.

This ends the analysis and the next one is ready to begin.Typical analysis time is 4 minutes.

QUALITATIVE AND QUANTITATIVE INFORMATION

In a gas chromatograph it is vital that three things neverchange.

1. Oven temperature2. Flow of carrier gas3. Sample size

By keeping these three things constant we are able torun continuous analysis that repeat within ±.5 BTU / 1000BTU’s (±.25 BTU / 1000 BTU’s when installed in atemperature controlled environment). By knowing thistime (Retention Time) we can program the controller andhave it identify the peaks as they come out. Thisidentifying of the peaks is known as qualitativeinformation.


To find out quantitative information we use a set ofmatched thermal conductivity detectors (TCD’s)integrated in a bridge circuit. One detector is thereference and is only exposed to the carrier gas. Theother is called the measure detector and is exposed toboth the carrier and the separated sample components.When carrier gas is flowing across both the referenceand measure detectors at the same time the detectorbridge is balanced and no signal is generated. When acomponent, other than carrier gas, is going across themeasure detector the bridge is not balanced and a signalis generated, amplified, and sent to the controller. Thisraw signal is integrated by the controller which thencalculates a raw area. A raw area is calculated for eachpeak (see Figure 1).

FIGURE 1. Chromatogram

Raw areas are calculated by integrating the area undereach peak. Each raw area is directly proportional to thecomponent concentration

During a calibration raw areas are divided by the knowncalibration concentrations and response factors (RF) arecalculated. Each response factor is only changed duringa calibration. The following formula shows how aresponse factor is used for calculating the mole percentof a sample stream component.

Response Factor = Raw Area / CalibrationConcentration

Mole % of component = Raw Area / Response Factor

The following is a simplified example of how the responsefactor is calculated and then used in the stream analysis.

Given: Calibration was run and a raw area for ethanecalculated by the controller is 100.

Given: The calibration gas had a molecular concentrationof 5% ethane in the bottle (this was programmed intothe controller).RF = Raw Area / Calibration ConcentrationRF = 100 / 5 = 20

Given: Sample stream gas is run, and the raw areacalculated by the controller is 200.

What is the quantity of ethane in the sample stream?

Mole % of Sample = Raw Area / RFMole % ethane = 200 / 20 = 10%

Note: Realistically the Raw Area for the differentcomponents will typically be in the thousands.

BTU CALCULATIONS FROM MOLE PERCENT

After the controller has calculated the mole percent ofeach component, it normalizes the components so thatthe summation of all mole percentages equal onehundred. A simplified example of normalization using onlyfour components instead of eleven is as follows:

Measured component Normalized ComponentN2 0 .99% 1.0%C1 89.1% 90%CO2 0.99% 1.0%C2 7.92 8.0%Totals 99% 100%

To normalize mole %:CONCNn = (CONCn / CONCt) X 100

Where:CONCNn = Normalized concentration of component nCONCn = Measured concentration of component nCONCt = Measured total concentration of components

After the mole percentages have been normalized thecontroller multiplies each of these percentages by acorresponding BTU value often taken from an internaltable (GPA 2145). From the individual BTU calculationsfor each component the controller can then do asummation and calculate a total BTU per cubic foot ofgas. This is an uncorrected or Ideal BTU value. To get acorrected or Real BTU value the controller multiplies theIdeal BTU by the compressibility factor. Figure 2 showsthe results from an analysis report.

FIGURE 2. Analysis Report

The analysis report gives the BTU values as well as gascomposition, and relative density which can be used forvolumetric calculations.


CONCLUSION

Gas Chromatographs have been used for process andBTU measurement for many years. They have proven tobe an accurate and reliable source for gas energymeasurement needs. During the gas analysis thechromatograph uses columns to separate the gas intoits constituents, and uses thermal conductivity detectorsto quantify the mole percentage of each componentpresent. A BTU value is then calculated for eachconstituent based on this mole percentage. By summingup these individual values we can find the total BTU valueof the gas.

GLOSSARY:

1. Carrier Gas – the gas supply, regulated to a specificpressure, which carries the sample through thecolumns to the detector. The most common carriergas used for natural gas applications is pure helium.

2. Columns – various tubes or capillaries mounted in achromatograph oven that are packed with solidmaterial, combinations of liquid and solid material,or liquid material.

3. Analytical valves are installed in the chromatographoven both to inject a sample for analysis and to divertit or its separated components as per the applicationschedule (timed events).

4. Detectors are devices or assemblies of instrumentsthat transmit electronic signals proportional to the

concentration of a component. The ThermalConductivity Detector (TCD) is the most widely useddetector for natural gas analysis.

5. A Calibration cylinder is used as the basis ofcomparison for the each of the componentsmeasured on a stream(s). The components andranges blended in the cylinder are roughlycomparable to the stream.

6. Retention time — the amount of time elapsed fromthe start of an analysis to the maximum moment ofinfluence a component makes on the detector.

7. The chromatogram is the graphic depiction of thechromatograph analysis. It is used to allow thetechnician to view the baseline integrity and peakseparation as a part of quality assurance.

8. The Molecular percentage (mole %) is the volumetricmeasurement of the percentage of a component ina sample (also known as the percentage by volume.)

9. Response factor — used for quantitativemeasurement. RF = RA / CCWhere:RF = Response FactorRA = Raw AreaCC = Calibration Concentration from calibration gas.


CALIBRATION OF STANDARD GASESRonald C. Geib

Matheson Tri-Gas, Inc.166 Keystone Drive, Montgomeryville, PA 18936

INTRODUCTION

Calibration standard gases are essential to quantitativeanalytical measurements in petrochemical processes,natural gas, environmental compliance, and health andsafety programs. The calibration gas standardestablishes a known analyzer response to a certifiedchemical component concentration which enables theconversion of sample responses to a concentration witha determinable accuracy. In consideration of the criticalityof calibration standard gases to valid measurements inchemical processes and monitoring programs, theobjective of this paper will be to provide an in-depthreview of how calibration standard gases aremanufactured, certified, and properly maintained.

ORDERING CALIBRATION STANDARD GASES

The specialty gas industry are specialists in the provisionof calibration standard gases and in the support gasesessential to analytical measurement processes. Certaininformation is essential for the specialty gas companyto provide the calibration standard that is required whichis summarized as follows: minor components,concentrations, units, balance gas, cylinder size, blendtolerance, and analytical accuracy. All specialty gascompanies will offer standard product grades that havedefined blend tolerance and analytical accuracy.Customers must be aware that grade names, blendtolerances, and accuracy are not uniform in the industryso that careful evaluation of suppliers and specificationsis essential to ensuring value and technical satisfaction.

Blend ToleranceBlend tolerance is the concentration range within whichwill be the mixture minor component concentrations. Forexample, a supplier's primary standard grade may claim1% accuracy and 2% blend tolerance. For a 100 ppmconcentration, the blend tolerance says that thecomponent certified concentration will be 98 - 102 ppm.The end user must be careful in defining the blendtolerance that is acceptable for his analyzer andcalibration practice. Should there be a need for acalibration gas that is at the top of the instrument range(span gas), then the mixture should be ordered with amaximum concentration specified because there is nouse for a calibration gas standard that exceeds full scale.If a minimum of calibration adjustments are desired fromstandard to standard, then the end user should considera homogeneous batch or pursuing a supplier that hasthe capability to blend with very narrow blend tolerancessuch as less than 1% relative.

AccuracyAccuracy is defined as the agreement of a measuredvalue with its true value. Common synonyms for accuracyare analytical accuracy, analytical uncertainty andcertification accuracy. All calibration standard gasesshould have a certified component concentration alongwith an expression of the accuracy of the certified value.Analytical accuracy is a statistically derived value, andthe generally accepted formula for calculating accuracyare in International Standard Organization (ISO) and U.S.institute papers. Basically accuracy is calculated usinga propagation of error model which is commonly referredto as the square root of the sum of the squares of thecommon error factors. In measurements, the commonerror factors for supplier certification of a calibration gasstandard are the reference standard error, the imprecisionof the measurements, and the instability of the gasmixture. The mathematical expression is as follows:

Accuracy = [(std error)2 + (precision)2 + (stability)2]1/2Example: Standard Error = 1%

Precision = 1%Instability = 1%

The result of this calculation is 1.7% which a supplierwould round to +/-2% accuracy. A word of caution hereis that the specialty gas industry does not uniformly applyaccuracy calculations - some suppliers either do notreport accuracy or base it on other calculations such asgravimetric additions. The end user of the calibrationstandard gas must determine the measurement accuracythat is needed for the specific application. If there is aprocess control chart, the user could assess the impactof having deviations up to twice the reported accuracy(remember that +/- 2% on individual certified standardscould result in a range of cylinder values -2 to +2% or4% for repeat orders). Often there is a clause in contractsthat specifies the accuracy required of reported values.Finally, environmental regulations frequently specify theaccuracy required of measurements and even the correctcertification procedure (example EPA Protocol gases forenvironmental instrument calibration). The accuracypropagation of error calculation is also fundamental toany process or laboratory analyst understanding themeasurement accuracy of his process.

NIST TraceabilityISO Guide 25 defined traceability as:"the property of a measurement result whereby it canbe related to appropriate standards, generallyinternational or national standards, through an unbrokenchain of comparisons. Organizations that have achieved


ISO 9000 certification must demonstrate that theirmeasurements comply with the preceding definition. Inthe U.S., national standards are provided by the NationalInstitute of Standards and Technology, NIST. In the fieldof chemical measurements, NIST traceability iscomplicated by the fact that only about 20 chemicalsare addressed by NIST. See Table 1.

TABLE 1: NIST SRM COMPONENTSMINOR BALANCE CONCENTRATIONCOMPONENT GAS RANGESCarbon Dioxide N2 300 ppm - 16%Carbon Dioxide Air 330 - 380 ppmCarbon Monoxide N2 10 ppm - 16%Carbon Monoxide Air 10-45 ppmHydrogen Sulfide N2 5 - 20 ppmMethane Air 1 - 10 ppmMethane N2 50 - 100 ppmNitric Oxide N2 5 - 3000 ppmNitrogen Dioxide Air 100 - 2500 ppmNitrous Oxide Air 300 - 330 ppbOxygen N2 2 - 21%Propane Air 0.25 - 500 ppmPropane N2 100 ppm - 2%Sulfur Dioxide N2 50 - 3500 ppm

For many U.S. environmental regulations, traceability ofthe above minor components through measurementversus a NIST traceable gas standard is essential. Wherepossible, the end user should request NIST gastraceability, plus the traceability information must bereported on the certificate of analysis as proof of theNIST traceability. For the hundreds of chemical speciesthat may be needed as gas standard minor components,NIST traceability is achieved through gravimetric weightcalibrations and other mechanisms which should alsobe reported on the certificate of analysis. For non-NISTSRM traceable chemicals, the end user should requestinformation on how the supplier guarantees the accuracyof the minor component as this will ensure that thecalibration gas standards will be consistent over timeplus meet the ISO 9000 certification requirements.

Storage and Shelf-life ConsiderationsFor any calibration gas standard with hydrocarbon minorcomponents with mole% or higher concentration, thehydrocarbon dewpoint becomes important. Thehydrocarbon dewpoint is that temperature where the mostvapor pressure restricted minor component condensesfrom the gas phase into liquid. Condensation invalidatesthe certification. Specialty gas companies usually have astandard dewpoint temperature, and an example wouldbe 32 deg F. This is satisfactory in some geographicregions for most of the year, but in the winter in Minnesota,a complex gas mixture calculated at 32 degrees F willexperience significant condensation. The supplier and enduser must anticipate shipping and storage conditions andnegotiate the proper dewpoint temperature for thegeographic region; and this should be reported on thecertificate of analysis as well. In some situations, unopenedcylinders can be warmed to room temperature and acondensed component(s) can be re-volatilized and re-mixed. However, data and a procedure from the supplier

should be available to ensure that this will work. Also, acomplex mixture requires software to do dewpointcalculations because manual calculations can not takeinto consideration the molecular interactions which oftenreduce the allowable cylinder fill pressures. Mixture storagewith respect to the mixture dewpoint has been discussed,the mixture shelf-life or stability should also be considered.For ISO 9000 compliance and also for gas standards madeto comply with environmental regulations, a shelf-life mustbe reported on the certificate of analysis. U.S. EPAspecifies EPA Protocol mixture shelf-lives, all othermixtures require the study and reporting by the individualsupplier. For many reactive gas species at concentrationsless than 1000 ppm, the mixture stability depends on thetechnology of the supplier, and the mixture stability canvary significantly.

SPECIALTY GAS INDUSTRY PROCESSES

Once the end user has resolved all of the gas mixturespecifications with the specialty gas supplier, an orderis entered into the supplier's manufacturing process. Ifthe mixture is a supplier catalog item, most of theproduction and laboratory procedures are routine and inmost cases, defined. Many gas and liquid mixtures arenon-catalog or custom; and many challenge thetechnology and know-how of the respective gas supplier.In the engineering of a catalog and custom item, the gasmanufacturer must address the following issues:

• Cylinder material such as steel vs aluminum• Cylinder preparation technology• Raw material grades and impurities• Chemical compatibility• Blending technology• Blending measurements validity• Laboratory instruments and procedures• Certificate of Analysis• Shipping

Role of the Quality System in Specialty Gas IndustryThe preceding paragraph identified nine key areas thatmust be addressed in the successful preparation andcertification of a calibration gas standard. All of theseactivities must be systematized to ensure the end userthat a reliable calibration standard gas will be provided.The suppliers quality system provides the organizationand control mechanism to provide the assurance of areliable supplier. A common quality system is ISO 9002,but there are other programs such as LaboratoryAccreditation processes. The end user should verify thatthe prospective supplier has sufficient quality systemsto meet the end users reliability expectations. In addition,the supplier's understanding and use of technology andmeasurement science should be demonstrated.

Specialty Gas Manufacturing ProcessSuccessful gas mixture manufacturing requires cylinderpreparation and cylinder surface treatment technology.Different suppliers will apply unique terminology to theircylinder treatment processes, and the proof of their utilityis whether the supplier can back up the treatments withshelf-life studies of challenging mixtures. Reactive


chemicals at concentrations less than 10 ppm remain asreal challenges to the suppliers' technology. Some of thesechallenging chemicals and classes are listed in Table 2.

TABLE 2: REACTIVE CHEMICAL SPECIESH2S NH3NO NO2SO2 HClPhosgene AminesDienes EthersAldehydes Organosulfursethylene propylene oxide

Once the proper cylinder and cylinder treatment havebeen identified, the supplier will forward the cylinder forblending. Depending on the grade of mixture as definedby the blend and analytical tolerances, the manufacturermay select from a number of blending options:

• Gravimetric• Volumetric• Dynamic blending

among other options. Gravimetric blending is quitecommon for high accuracy and tight blend tolerances.The purchaser of a calibration standard gas must beaware that some suppliers may base their certificationson the gravimetric quantity weighed into the cylinderalone.

Gravimetric certifications have the following limitations:• Loss of reactive minor components due to

surface absorption (example chlorine in a steelcylinder)

• Presence or addition of impurities inmulticomponent mixtures (example isobutaneimpurity in propane adding to isobutane weightin a Propane/isobutane mixture)

• Reaction of impurities in the mixture with certifiedminor components (example O2 impurity withNitric Oxide)

• Blunders which may include addition of thewrong component, an erroneous weight, or evenno component

For these reasons, most end user measurements shouldbe done with mixtures that have been verified usinglaboratory measurements, and the end user shouldconfirm that the supplier includes laboratory certificationin the mixture grade and pricing.

The reliability of gravimetric concentrations is also subjectto the purity of the raw materials, plus the stability of themixture is frequently dependent on the raw materialimpurities. Some raw materials do not include majorimpurities in their specifications and the individual suppliermay not be aware that they are present. For example,pure nitrogen frequently has several hundred ppm ofargon, pure helium may have 20 ppm or more of neon,and carbon monoxide may have 0.7% argon. Theseexamples also point out the value added of laboratorycertifications by technically proficient laboratories.

Laboratory CertificationOnce the gas mixture has been blended, better grade

gas mixtures and the better specialty gas suppliers willrequire laboratory certification of the components'concentrations. The laboratory must address the order'sand advertised traceability requirements by conductingmeasurements against an appropriate referencestandard, or reference standard mechanism. The highestNIST traceable standard is the NIST SRM. NIST andprivate industry also have a NIST Traceable ReferenceMaterial (NTRM) program that allows the copying of eitherthe SRM or other NIST primary standards. The NTRMprograms are essential for specialty gas suppliers tomaintain directly NIST traceable certification programs.

For the hundreds of minor components that must becertified where there are no NIST SRM, the supplier hasthe responsibility to develop certification practices thatcan statistically validate the concentration of minorcomponents in the cylinder. Some technology that isused includes wet titrations after quantitative collectionof minor components of the gas phase into scrubbingsolutions, multiple preparations of gravimetric standardsand conducting calibration curve studies, and workingwith regulatory and industry sources to prepare andcertify mutually acceptable standards.

The Role of Quality Assurance ProcessesQuality Assurance functions typically sponsor and assessthe quality system within the supplier and end userorganization. In that the supplier - customer relationshipinvolves quantitative measurements, the qualityassurance programs must include the support andevaluation of the measurement programs. In this regard,the application of the best available, NIST traceablereference standards by the supplier is essential. Boththe supplier and customer should participate in industryround robins that confirm the conformance ofmeasurements to the industry. The supplier also needsto organize internal round robins to challenge theuniformity of multiple locations. The most importantquality assurance function is organizing datainterpretation, and corrective action processes.

CONCLUSIONS

In consideration of the criticality of calibration standardgases to leading industrial measurement processes, thispaper has gone into detail identifying the key variablesfor both the user of calibration standard gases, and thesuppliers of calibration standard gases. If the end useradheres to stringent application of the principles that need tobe addressed, then the end userwill be able to identify areliable supplier.

Ronald C. Geib


INSTRUMENTS FOR THE DETERMINATION OFSPECIFIC GRAVITY / RELATIVE DENSITY OF GAS

Myles J. McDonough, Jr.Chandler Engineering Company LLC

P.O. Box 470710, Tulsa, OK 74147-0710

HISTORY

The terms Specific Gravity and Relative Density havebeen used for a number of years. Yet there seems to besome confusion over what exactly they mean.

Specific Gravity is formally defined as the ratio of gasdensity to air density when both are at standardconditions of 0 Degree C and 760 mm. Over the yearsthe definition evolved to become the ratio of gas densityto air density at the same temperature and pressure,“Relative” to each other. Hence, the term “RelativeDensity.” This is the most commonly used definitiontoday. The fixed or “Specific” requirement of temperatureand pressure, (0 degree C and 760 mm), had beenremoved over the years.

The American Gas Association in 1985 officially replacedthe term Specific Gravity with Relative Density. However,both terms are still used synonymously.

TERMINOLOGY

Density of gas sampleRelative Density =

Density of Air(Both at same temperature and pressure)

Density of gas sampleSpecific Gravity =

Density of Air(Both at Standard Conditions)

Density = weight per unit volumeUnits: lbs./ft.3, grams/cubic centimeter, etc.Example: Pure water has a density of 62.4 lbs/ft3

Notice that both the Specific Gravity and Relative Densityterms are unit less. Also, it should be noted that both aretemperature and pressure dependent. Therefore, if thegases are not at identical conditions, some type ofcompensation would have to be applied.

APPLICATION OF SPECIFIC GRAVITY IN FLOWMEASUREMENT

Specific gravity values are primarily used in the orificemeter flow correction calculations and for custodytransfer. The fundamental orifice flow equation is a massflow calculation based upon the physical laws ofconservation of energy and mass. Differential pressureand density are the two main variables in the equationonce the orifice meter is installed. Both variables are

under the square root meaning that flows varies by thesquare root of the differential pressure and density. Inpractice density is not directly measured but calculatedfrom the temperature, pressure, compressibility andspecific gravity values. Specific gravity is also used indetermining the super-compressibility multipliers fordisplacement meters when correcting volume.

METHODS FOR DETERMINING SPECIFIC GRAVITY

Many different types of instruments are used in the oil andgas industry for determining specific gravity. Including:

• Effusion and weighing methods (Gravity Balance)• Direct weighting• Kinetic energy• Vibrating element• Gas chromatography

EFFUSION AND WEIGHING TECHNOLOGY

The best known effusion and weighing instrument is thegravity balance. Gravity balances operate on the principleof measuring the pressure exerted by a gas of a givendensity and the pressure exerted by air of the samedensity, the specific gravity being determined from theratio of these pressure measurements.

In this instrument, the means for measuring the buoyantforce exerted upon a body suspended in gas or air areprovided by a balance beam having a sealed float onone end and a balance weight and a graduated scale onthe other end. When the buoyant force of the gas sampleand air taken on successive tests is equal, the densitiesare equal. The pressures at which this equality occursare measured by a mercury manometer, and the pressureratio is used in calculating specific gravity.

This instrument calculated specific gravity by thefollowing equation:

Pair TgasSpecific Gravity = ×Pgas Tair

Where

Pair is air absolute pressure (barometric + manometer)

Pgas is gas absolute pressure (barometric + manometer)

Tgas is air absolute temperature (temperature F + 460)

Tair is gas absolute temperature (temperature F + 460)


The gravity balance is a primary measurement device.The accuracy obtained is dependent upon the care usedin construction of the gravity balance and the care usedin its operation.

The gravity balance is a very accurate device, howeverit does required a fair amount of time to obtain a reading.

As the gas industry was making its initial movementtowards an on-line continuous specific gravitymeasurement two other devices appeared. The directweighing and the kinetic energy devices.

DIRECT WEIGHING TECHNOLOGY

One of the more popularly used direct weighinginstruments is the UGC Gravitometer. This gravitometeruses a simple balance beam system to measure specificgravity. Identical tanks, referred to as reference andsample, are hung on each end of the balance beam atequal distance from the pivotal axis.

Pressure, temperature and other ambient changes areexperienced equally by both tanks and thus nullify thesechanges. For example the effects of ambient temperaturefluctuations are eliminated by having the reference tankgas pressure regulate the sample tank gas pressurethrough the use of an especially adapted pressureregulator. The pressure in the sample tank is not keptconstant but is maintained equal to the reference tankpressure which in turn is a function of the ambienttemperature as predicted by the perfect gas law for aconfined gas. The ambient temperature which governsthe pressure in the sealed referenced tank is allowed tocontrol the pressure in the sample tank. Only the genuinespecific gravity changes are recorded and not theapparent changes due to a variation in the ambienttemperature.

Calibration is performed by the use of calibration discsthat are equal to the weight difference of “full scale”gravity and “minimum scale” gravity.

KINETIC ENERGY TECHNOLOGY

The most popular kinetic energy type gravitometer is theRanarex gravitometer. This device became very popularand is still widely used, especially the portable version,by field technicians. This instrument’s popularity grewbecause for the first time an operator could take a spot(field) specific gravity measurement in a matter ofminutes. There are no additional readings to record ormathematical equations to remember. The fieldtechnician simply connected the portable gravitometerto a pipeline probe and within a few minutes had therequired reading for the spot sample.

This gravitometer consists of two cylindrical gas-tightmeasuring chambers, each having separate inlet andoutlet connections. Each chamber contains an impellerand an impulse wheel, both with axial vanes. These wheelsare mounted on separate shafts facing each other but

not touching. An electric motor and drive belt rotate theimpellers at the same speed and in the same direction.

The impellers draw continuous flows of the gas sampleinto the upper chamber and dry reference air into the lowerchamber, and spin the gas and air against the vanes oftheir corresponding impulse wheels which are proportionalto the density of the gas and of the air. The other variablesof pressure, temperature and speed are equal for boththe air and gas sample and therefore cancel each other.These torques are transmitted from the chambers by theimpulse wheel pivot shafts to two external measuringwheels. These external measuring wheels are shapeddifferently for the gas and air samples. A flexible tape iswrapped over the measuring wheels rims in the crosseddirection so that the torque creates two opposing forces.The measuring wheels are thus restrained from continuousrotation, but a difference between the torques permitslimited motion of the entire system.

Calibration involves using either calibration pulleys orreference gases of known specific gravity. The calibrationis performed by adjusting the effective distance betweenthe impulse wheel and the front of the cylindrical chambermaking the pointer indicate the appropriate specific gravity.The zero adjustment is re-aligned after each change.

VIBRATING ELEMENT TECHNOLOGY

The vibrating element technology has been employedsince the 1960’s. This type of gas gravitometer is usuallyelectronic and produces a frequency signal that isproportional to the specific gravity of the gas.

The principle of resonant frequency measurement issimply that if a very stable element is maintained inresonance, only two factors have any major effect onthe natural frequency, these being a change in mass ofthe fluid surrounding the element or the stressing of theelement. The element is made to vibrate at it naturalresonant frequency by means of electromagneticinduction. A solid state amplifier is used to maintain theconditions of vibration and to provide the output signalfor measurement. This signal is linearized using thequadratic equation shown below:

Specific gravity = K0 + K2 T2

where, K0 and K2 are calibration coefficients unique toeach gravitometer. T2 is the output time period (theinverse of frequency)

The gravitometer consists of a vibrating cylinder gasdensity transducer surrounded by a constant volumereference chamber. This gas reference chamber has afixed volume which is initially pressurized with the samplegas. It is then sealed by closing the reference chambervalve, thus retaining a fixed quantity of gas now knownas the reference gas.

The sample gas enters the instrument at the base plateand passes through a filter, followed by a pressure


reducing orifice. The sample gas is then fed through aspiral heat exchanger (wound around the referencechamber) so that it enters the gas density transducer atthe equilibrium temperature. the output of thegravitometer is generated from the frequency of vibrationof a thin-walled metal cylinder contained within this gasdensity transducer.

Following the gas density transducer, the gas flows downto a pressure control valve chamber. The reference gaspressure acts through a separator diaphragm on thepressure control valve chamber so that the gas pressureson both sides of the diaphragm are equal, ie. the gaspressures within the gas density transducer and thereference chamber are equal.

With a fixed quantity of pipeline gas contained in thereference chamber and with pressure and temperatureequalization established, the density changes beingmeasured by the density transducer will be causedentirely by changes in the molecular weight of the flowinggas. As the ambient temperature changes, the pressureof the fixed volume of reference gas will change asdefined by the gas laws. This change in pressure willaffect the sample gas pressure within the gas densitytransducer such that the temperature and pressurechanges are self-compensatory.

Calibration involves charging the reference chamber withthe pipeline gas to a defined pressure in the range of 30to 90 psig and then calibrating the output signal usingtwo gas samples of known specific gravity. Typically puregases such as methane and nitrogen.

GAS CHROMATOGRAPHY TECHNOLOGY

Gas chromatography is a technique by which a gasmixture is physically separated into it’s individualcomponents and quantified. Gas chromatographsinclude the following major components:

• Sampler• Column Oven• Packed or Capillary Columns• Sample Loop• Column Switching Valve(s)• Detector and Associated Electronics• Carrier Gas• Computer used for Data Acquisition, Calibration

and Communications

In the natural gas application, the gas chromatographseparates the gas sample into its individual componentsby injecting a precisely measured volume of the mixtureinto the separation columns.

The carrier gas, usually helium, transports the samplethrough the columns and is also used to actuate thevalve(s) in the system. The columns are packed with amaterial that selectively retards (adsorption) the passageof the hydrocarbon molecules based on the number of

carbon atoms (molecular weight) in the molecule. Themolecules with fewer carbon atoms (light components)will exit (elude) the columns first. The molecules withmore carbon atoms (heavy components) will elude later.Therefore, the components are carried through thecolumns at different rates and elude separately.

A detector usually Thermal Conductivity Detector (TCD)is used to sense the elution of the components from thecolumn. The time at which the component appears atthe detector identifies the component. The detector alsois utilized to determine the concentration of thecomponents. The detector’s output is a DC voltage thatchanges when any gas other than helium is present atthe detector. When this output is taken to a recorder,each component forms a separate “peak” which can beidentified and quantified. The physical recording of thesepeaks is a “chromatogram.” After all the peaks have beenidentified and quantified, the BTU (heating value), specificgravity, etc. are computed.

The specific gravity of the gas is calculated by thefollowing equation.

Theoretical specific gravitySpecific Gravity =

Compressibility of the sample

where, Theoretical specific gravity = sum of the specificgravity of each component multiplied by the mole % ofthat component.

MAJOR COMPONENTS OF A GASCHROMATOGRAPH

Sampler

The sampling system must extract a representativegaseous sample from the pipeline, reduce the pressure,filter the sample, prevent sample condensation andefficiently deliver the sample to the chromatograph. Moremeasurement inaccuracies occur due to impropersampling systems than any other reason.

Column Oven

The chromatograph oven usually contains the columns,the switching valve(s), the sample loop, and the detector.The temperature of the oven must be precisely controlledsince the performance of the columns and detector isaffected by changes in temperature.

Packed or Capillary Columns

The columns separates the sample into its individualcomponents. They consists of a length of 1/16” diameterstainless steel tubing which is micro-packed with agranular liquid coated porous support material. Thispacking is termed the “stationary phase” of the analysis.A tiny mesh screen is installed at the tip of the columnsto hold the packing in place. The column length variesdepending on the components to be separated.


Sample Loop

Fixed gas sample volume that is injected into the columnsto be analyzed.

Valve(s)

The sample valve(s) seizes and injects the gas sampleinto the carrier gas ahead of the column.

Detector and Associated Electronics

The detector is used to identify the individual componentsof the gas sample. The TCD is the most commonly useddetector for natural gas chromatographs. The TCDconsists of a reference side and a measurement side of aWheatstone Bridge circuit. As the components elude fromthe columns, they flow across a thermistor (measurementside) where heat is removed in direct proportion to thethermal conductivity of the component. This produces animbalance in the circuit which is amplified and sent to thecomputer for calculation. This signal is represented in thechromatogram as a separate peak.

Carrier Gas

A carrier gas usually helium, is used to transport the gassample through the columns. This carrier gas must bepure (99.995%) or a reduction in the sensitivity of thedetector will occur.

Computer

An internal computer provides all of the instrumentcontrol and the integration of the chromatogram peakareas. Once the compositional analysis of the sample iscompleted the computer calculates the heating value,specific gravity an compressibility of the gas using ASTM,GPA, AGA, and ISO methods. The computer is also usedto automatically transfer data to a central computersystem using a data acquisition network known as aSCADA system.

The computer may also be used to perform automaticfunctions such as daily calibrations and calculate dailyaverages of stored analyses.

SUMMARY

Specific gravity is an important physical measurement.Even as the instruments for measuring gas compositionbecome more accurate, reliable and rugged for field usethe usefulness of the specific gravity value will remain.It is an important variable in correcting the flowmeasurement. Experience has shown accounting for thegas by mass is more accurate than accounting byvolume.

Myles J. McDonough, Jr.


DEVICES FOR FIELD DETERMINTION OF H2OIN NATURAL GAS

Borys J. MychajliwMEECO Inc.

250 Titus Avenue, Warrington, PA 18976

INTRODUCTION

Natural gas is rapidly becoming the fuel of choice foreverything from home heating to power generation. Thedetermination of the water vapor content in natural gasis one of several key factors in determining the ultimatequality of the gas. With economic conditions as they existtoday, many companies have been forced to cutpersonnel in order to maintain a reasonable balancesheet. The loss of experienced measurement techniciansplaces a heavy burden on instrument manufacturers toprovide an accurate and reliable means of determiningthe water vapor content of natural gas. This paper willreview the different sensor technologies in use today andalso address key issues and proper procedures inassembling a sampling system to provide a clean,representative gas sample to the sensing device.

IMPACT OF WATER VAPOR ON GAS QUALITY

As removed from the ground, natural gas is typicallydescribed as being either rich or lean (wet or dry) asrelates to the heavier hydrocarbon components in thestream. In addition to these hydrocarbons, natural gaswill almost always contain water vapor as well asnitrogen, helium, carbon dioxide and, in some cases,hydrogen sulfide. Our focus here will be on one of themost undesirable components, water, both as free waterand water vapor.

Natural gas, as removed from the ground, can besaturated with water vapor and must go through adehydration process prior to transportation in the vastnetwork of underground transmission pipelines. Thedehydration process can be accomplished throughvarious means. Typical dehydration technique is to passthe gas through a TEG (tri-ethylene glycol) contactor orpassing the gas through a molecular sieve desiccant bed.Both of these processes will add up front cost to thegas, but are necessary for the following reasons:

1. Water vapor can combine with other tracecontaminants in the gas stream, namely H2S andCO2, and form potentially corrosive acids in thepipeline. The presence of these acids in the pipelinewill decrease the life expectancy of the pipeline andit’s components while increasing long-termmaintenance costs. The risk potential forcatastrophic failure of the pipeline or any one of itscomponents will also increase if the water vaporcontent remains unchecked.

2. As the gas passes through meter stations,compressor stations, orifice plates and regulators,rapid expansion of the compressed natural gas willcool the gas based on the Joules Thompson effect.The interaction of water vapor with heavierhydrocarbons in the gas stream at this point canresult in the formation of hydrates in the pipeline.These hydrates can cause complete or partialblockage in valves, regulators or measurementdevices in the pipeline resulting in significantdowntime while the hydrate is thawed or removedfrom the pipeline.

3. Contractual obligations at custody transfer pointsrequire that the gas not exceed a specified maximummoisture level. Typical contracts call for thismaximum to be 7 lbs/mmscf, although this numbermay change depending upon the geographic areaand normal ambient conditions.

4. The BTU or heating value of the gas will be reducedas the amount of water vapor increases. Ifcontractual obligations are not met, the producer canbe shut in and will face a loss of revenue until theproblems are corrected.

SENSOR TECHNOLOGIES

Today, there are many manufacturers of moistureanalyzers and sensors competing for a slice of the salespie. Most of these manufacturers will use a sensor thatfalls into one of the four basic technologies used todetermine the water vapor content of the gas. They are:

1. Electrolytic2. Capacitance3. Chilled Mirror4. Vibrating Crystal

In addition to these types of sensors, length of stain tubesare widely used to determine an approximate water vaporcontent, and new laser based devices are being utilizedin certain applications where traditional technologies maynot operate properly.

ABSOLUTE VS. RELATIVE

All sensor types will fall into one of two categories –Absolute or Relative. Absolute sensors are based onprimary laws of physics and do not require periodiccalibration against known moisture standards, nor canthe sensors actually be calibrated. Relative sensors arebased on a comparative measurement and will require aknown, certified moisture source for calibration.


Calibration frequency will vary depending on theindividual sensor characteristics.

ELECTROLYTIC SENSORS

Electrolytic sensors are unique in that they can be eitherabsolute or relative measurement devices. In either case,the basic operation of the sensor is the same.

The typical electrolytic sensor consists of two preciousmetal electrodes, wound around a support mandrel orimbedded in a hollow glass tube. These electrodes arethen coated with a thin layer of phosphorus pentoxide(P2O5). In operation, a controlled amount of gas is allowedto constantly flow through or across the sensor allowingsufficient time for the P2O5 coating to adsorb all of themoisture from the gas stream. A voltage potential isapplied across the electrodes, splitting (or electrolyzing)the water molecules that have been collected on thecoating. Once equilibrium conditions are attained, therate at which moisture molecules enter the cell will exactlymatch the rate at which the molecules are electrolyzed.Each electrolyzed molecule causes two electrons to bedisplaced from the anode to the cathode. The electrolysiscurrent (Amps) gives the electrical charge (coulombs)discharged per second. Since the elementary charge ofan electron is known, by measuring the current, we candetermine the rate at which water molecules are enteringthe sensor. Combined with a known flow rate throughthe sensor, the moisture content of the gas can bedetermined. The critical aspect here is the controlled flowrate, as this is the lone variable in the absolutemeasurement version of electrolytic based analyzers.

Electrolytic sensor based moisture analyzers are availableas portable and stationary devices with flow controlmechanisms built into the unit. As with all analyzers, it isimportant that the proper sampling techniques andsample conditioning criteria be strictly adhered to. Failureto do so may cause premature failure of the sensor. Liquidintrusion or the presence of conductive particles will alsocause premature sensor failure.

CHILLED MIRROR

Commonly known as the Bureau of Mines Tester, thechilled mirror hygrometer is a very simple and basicmoisture measurement device that has been used formany years as a primary, absolute moisturemeasurement tool. In operation, the sample gas streamflows across a temperature controlled polished surfaceor “mirror”. As the temperature of the mirror is slowlylowered, the water vapor will begin to condense or formdew on the mirror. The temperature at which the dewfirst appears is considered the dew point. Once thetemperature dew point is attained, a simple conversionusing existing charts or tables will provide the actualwater vapor content of the gas. This type of analyzercan also be utilized to determine hydrocarbon dew pointsin the gas stream. Usually, an iridescent ring will form onthe mirror surface indicating you have reached thehydrocarbon dew point temperature. Moisture dew point

will appear as a cloudy, opaque spot in the center of themirror.

The critical components of the chilled mirror analyzerare the pressure chamber with valves to control the gasflow and pressure, a small mirror or polished surface, athermometer or thermocouple to measure the mirrortemperature, a chilling device (typically propane or CO2)and a view port to allow for observation of the mirror. Inapplications where the ambient temperature is below thedew point of the gas, it may be necessary to heat thesample line and analyzer to prevent condensation in thesampling system. Care should also be taken so as notto confuse dew point with frost point. Significant errorscan occur when converting dew/frost point values toactual moisture content (i.e. a reported dewpoint of -40°C at atmospheric pressure is equivalent to 8.8 lbs/mmscf while the same frost point temperature translatesto 6 lbs/mmscf).

Since this is a direct measurement device, no calibrationof the system against a known moisture standard isrequired. Care should be taken to collect a representativesample of the gas using proper sampling and filtrationtechniques. Experienced operators can make highlyaccurate and reproducible measurements, butinexperienced operators may have some problemsassociated with interpreting the visual results. It is notuncommon to get different results from differentoperators on the same gas line.

CAPACITANCE SENSORS

Capacitance sensors fall into the category of relativemeasurement devices. Periodic calibration of the sensoragainst a known moisture standard is an absoluterequirement. Of all the different sensor technologies,capacitance sensors are the choice of a majority ofmanufacturers. There are numerous types of capacitancesensors, including aluminum oxide, silicon oxide,polymer base and thin film, but they all share the samebasic principle. Regardless of the sensor type, the coreof the sensor consists of two electrodes and a dielectricmaterial that absorbs the water vapor in the gas streamand achieves an equilibrium condition based on thepartial pressure of water vapor in the specific gas stream.

In operation, moisture in the sample stream is absorbedinto the dielectric material creating impedance within thesensor. An excitation voltage is applied to the electrodesand a return signal, proportional to the water vaporcontent, is transmitted back to the base electronics.

Capacitance sensor based analyzers are available in bothportable and stationary configurations. Although thesensor probe can be mounted directly in the processline, glycol and other contaminants in the gas streamcan cause false readings and premature sensor failure.The sensor should be mounted in an independent sampleconditioning system adjacent to the sample point in orderto protect against sensor contamination and to facilitatecalibration and maintenance.


VIBRATING CRYSTAL SENSORS

Piezo-Electric sensors are commonly referred to asvibrating quartz crystal sensors. The crystal is coatedwith a hygroscopic material in order to allow absorptionof the water vapor from the gas stream. The analyzerelectronics monitor the vibration frequency change ofthe crystal as water vapor is absorbed onto the coating.During operation, the crystal is alternately exposed tothe sample gas stream and a dry reference stream. Thereference stream is the actual sample gas, passedthrough an on-board gas dryer.

In operation, the sample gas flows across the crystal fora fixed time period. During this time, moisture in the gasstream is absorbed onto the coating of the crystalcausing a change in the vibration frequency. Thisfrequency is read, stored and compared against a sealedcrystal. The sample gas is then diverted through the onboard dryer and again is passed across the crystal. Onceagain, the frequency is read, stored and comparedagainst the sealed crystal. Using the differential in thevibration frequency of the sample gas and the driedsample gas, one can determine the water vapor contentof the sample gas.

These analyzers are strictly stationary devices and arenot suitable for portable applications. They are alsoconsidered relative measurement devices that requireperiodic calibration against a known moisture source.

STAIN TUBES

For those of us who might only be interested in ballparknumbers on an infrequent basis, the length of stain tubesmay be the best alternative. Encased in a glass tube, achemically treated compound will change color whenexposed to moisture in the sample gas. This is a quickand rough method to get a visual indication of themoisture content of the gas stream. Accuracy of thesetubes is only ± 25% and is dependent upon the operator.Sampling is accomplished by breaking off both ends ofthe glass tube and inserting it into a hand pump. Caremust be taken to avoid contaminating the reading withatmospheric moisture. A fixed number of pump strokesis required to achieve the final moisture value.

Although this technology is not suited for custody transfermeasurements, it is a useful tool when looking for a quick,rough measurement. If you suspect high moisturecontent at your sample point, use the stain tube as thefirst pass before exposing the sensitive analyzer topotentially high water content gas. Moisture analyzersare traditionally slow to respond after being saturatedwith water vapor.

LASER BASED ANALYZERS

Laser based analyzers are the new wave of moisturemeasurement technology. The basic principle is rathersimple. Every molecule has a unique electronic signatureor “fingerprint”. By tuning a laser beam to a specific

wavelength, the light energy is absorbed by themolecules in the gas stream that have an electronicsignature at the selected wavelength. The absorption ofthe laser energy is directly related to the concentrationof the selected molecule. The advantage of a laser is theability to detect different molecules, assuming there is alaser diode available within the specified absorptionwavelength. Current systems have the ability to measuremoisture and carbon dioxide with individual, dedicatedlaser diodes and sample chambers.

The basic concept is simple, but the reality is less so.Even though every molecule has a unique fingerprint,other molecules may have a lesser absorption efficiencyat the same wavelength, leading to false values. It iscritical to select a wavelength where no other moleculewithin your specific sample stream can also absorb thelight energy. Not all applications are suitable for laserbased systems.

One of the key advantages of laser based measurementis that the sensor itself is not in direct contact with thegas sample. This avoids the traditional problem of sensorcontamination, but does not preclude the need for propersample conditioning. There are mirrors and tubingsurfaces that are exposed to the sample gas and canalso be contaminated resulting in a loss of sensitivity inthe measurement.

SAMPLING SYSTEMS

All of these sensing technologies have their ownadvantages and shortcomings, but one issue is never indoubt. That is to say, we must use proper samplingtechniques and sample conditioning to bring arepresentative gas sample to the sensing element.

The design of the sampling system should adhere tosome basic principles. These principles include thefollowing:

1. Use high quality stainless steel components2. Low volume components3. Avoid dead space4. Use a sample probe5. Install and maintain proper filters

The first and most important component of the samplingsystem is the sample probe. The sample probe shouldbe installed in the pipeline such that you are extractinggas from the center third of the flowing gas stream. Thereason for this is to avoid pulling all of the collected liquidsoff the walls of the pipe and putting them directly intoyour analyzer. What we are trying to do is to sample thegas phase, not the liquid phase.

Probably the single largest cause of failure of a moisturesensor is contamination. Although the sample probe iscritical, this does not eliminate the need for additionalfiltration. There is always the potential for entrainedliquids in the sample stream and these liquids must beremoved. These liquids include glycol, methanol,


hydrocarbon liquids and compressor oil to name just afew. It is imperative that a filtration system be installedupstream of your analyzer. Ideally, any sampleconditioning should be designed specifically for theindividual application, but most sample systems willconsist of a liquid removal filter along with a glycol or oilvapor removal filter. Liquid removal can be accomplishedwith a coalescing filter or a membrane separator or acombination of both. Glycol vapor filters are also availablein a variety of sizes and configurations. I suggestdiscussing your application with the analyzer supplier todetermine to what extent you will need filtration. Alsoremember that filters require periodic maintenance. Don’tinstall a system and forget about it, rather, maintain itaccording to the supplier’s guidelines.

Several other points to consider in the sampling systemare:

1. Always use stainless steel sample lines andcomponents to minimize corrosive effects and tomaintain good sensor performance. Too many times,the inclination is to use any old rubber hose thathappens to be lying around to deliver the sample tothe analyzer and then blame the analyzer when thereadings are not correct.

2. If you choose to heat your sample lines, make surethat you also heat your entire system, including theanalyzer, to the same temperature. Hot gas hittingthe cold metal of the analyzer is a prime catalyst forcondensing liquids out of your gas stream.

Borys Mychajliw

3. Select a sample point that is free from swirling gasflow. Try to select a sample point that is in the middleof straight, unimpeded run of pipe. Elbows, orificeplates and control valves can create swirl effects,which can sweep liquids from the wall of the pipeand right into the analyzer.

4. Design your sample system with minimal deadvolume components to help the analyzer respondquickly to changes in moisture. Keep thecomponents to an absolute minimum and place youranalyzer as close as possible to the sample point.The more complex your sampling system is, the moretime it will take to both purge it out and get arepresentative gas sample to your analyzer.

Using these basic guidelines, you can design a systemthat will give you protection from contamination while atthe same time bringing a representative gas sample toyour analyzer.

CONCLUSION

Today, we have many choices when selecting a moistureanalyzer. This overview touched on several of the choicescurrently available. Do your research and talk to themanufacturers to determine what will work best in yourapplication. Remember that the analyzer will provideaccurate results only if you provide it with a clean,representative gas sample.


FUNDAMENTALS OF ENERGY DETERMINATIONDavid Hailey

Instromet, Incorporated12650 Directors Drive Suite 100, Stafford, TX 77477

ABSTRACT

This paper presents fundamental information necessaryto understand and appreciate the concept of total gasenergy in a natural gas pipeline. That is, to be able toconverse with peers within the natural gas industry andunderstand basic concepts and terminology.

Discussed is the historical transition from volumetricmeasurement to total gas energy including some of thebasic terminology, physics, measurement, as well as thereasons for changes in methodologies. Included isindustry acceptance of new concepts and regulationsinvolving custody transfer as well as the instrumentationand systems involved in traditional and newer, moreprogressive forms of gas measurement.

GAS ENERGY VS. VOLUME

It has been the traditional practice for gas companies toconduct business on a cost per volume basis, (volumetriccustody transfer). Typically in the United States, this hasbeen done on a cost per standard cubic foot of naturalgas. This practice began to change, however, whenCongress passed the Natural Gas Policy Act of 1978that came as a result of price differentials between stateand federally regulated natural gas. The Act sought torestore consumer confidence and instill a sense offairness in the gas system. One of its provisions specifiedthat natural gas be bought and sold on the basis ofenergy content. Another milestone was FERC 436 thateffectively opened interstate pipelines for use astransportation conduits for companies independentlybuying and selling natural gas. In many cases, interstatepipelines had restricted access to anyone other than theirown organization. Under revised laws and regulations,end users were free to purchase gas on a commoditybasis and were encouraged to bargain for best pricingthat would directly benefit end use consumers. The costbenefits were initially impressive, but not to all customers.Compared to previous experience, it soon becameapparent that changes in gas quality resulted in increasedconsumption at some sites while others metered less.Unknown qualities had an unsettling effect on the entiresystem.

Gas flow is effected by factors such as temperature,pressure, composition, and physical dimensions of theprimary element and supporting meter run. As gaspressure increases, more gas molecules can beessentially packed within a given space — a pipeline.Likewise, as gas temperature decreases, more gas

molecules can be accommodated. In general, flow hasan inverse relationship to heating value, that is, for higherquality gas, less volume is required to support a givenload than is required by a lower quality gas.

There are industry procedures, reports and standards,(e.g. AGA-3, AGA-7, AGA-9, ASTM D4891, GPA-2172,etc.) which account for the impact that these variousfactors have on volumetric and energy flows.

Gas energy is volumetric flow at base (reference)conditions multiplied by the heating value (overall quality)of the gas. The buying and selling of natural gas canonly be truly effective and fair if total energy is the basisof measurement.

COMMON TERMINOLOGY

A common problem when discussing energy flow is thedifference in each persons understanding of terminology,reference conditions and the standards in place. It isimportant to agree on the terms that define the piecesof the energy puzzle. Terminology, definitions, andreference conditions must be understood before anymeaningful discussion can take place.

Gas Energy:

Gas volume or flow adjusted for gas quality, that is,volume at base conditions multiplied times heating value.Total gas energy is the capacity for doing the work naturalgas is targeted for. The quality of the gas has an inverserelationship to volumetric flow. The higher the gas qualitythe lower the flow requirements for a defined load andvisa-versa.

Base Conditions:

Base, standard or normal conditions are the referencepoints for measurement. These define the referencetemperature, (usually 600) Fahrenheit and the referencepressure. The most common pressure base used is 14.73psia but other pressures may be encountered such as14.60, 14.65 or 14.70 psia. Which ever figure is used itmust agree with other gases under comparison.

Standard Cubic Foot (SCF):

The quantity or volume of gas occupying a cubic foot ofspace at a specified pressure and temperature.


Gross Heating Value:

The heating value (Btu) produced by combustion atconstant pressure with the following conditions:

(a) a volume of one cubic foot.(b) 60° Fahrenheit.(c) reference base pressure.(d) with air and gas having the same temperature

and pressure.(e) recovered heat from the water vapor formed by

combustion.

Gross heating value is the most commonly usedcondition/definition of heating value in the natural gasindustry, it is also referred to as the superior heating value.

Net Heating Value:

The heating value produced under conditions similar togross heating value conditions excepting the amount ofheat potentially recovered from the water vapor producedat combustion. Net heating value is always less thangross heating values. It is sometimes referred to as theinferior heating value.

Saturated Heating Value:

The heating value of a standard cubic foot of gassaturated with water vapor. The amount of water presentper given volume is defined by contract.

Dry Heating Value:

The heating value of a standard cubic foot of gas withoutwater vapor.

The saturated heating value of a standard cubic foot ofgas is always less than the dry heating value per SCFbecause of the displacement of hydrocarbon moleculesby water molecules. For example, the saturated value ofmethane (CH4) is 997 BTU at 14.73 psia and 60°F. butthe dry value of the same gas under the same conditionsis 1014.3 BTU.

Relative Density:

The ratio of the density a gas to the density of dry airunder the same pressure and temperature conditions,(it sometimes referred to as specific gravity).

Wobbe Index:

The ratio of the gross heating value of a gas to the squareroot of the relative density of the gas, (WI = Hv /(√RD).Wobbe index is a measure of the amount of energydelivered to a burner via an injector (orifice). The energyinput is a linear function of Wobbe index. Two gasesdiffering in composition but having the same Wobbeindex will deliver the same amount of energy for any giveninjector/orifice under the same injector pressure.

VALUE COMPARISONS

Today’s energy reports are commonly made by volumewith energy corrector factors, BTU per standard cubicfoot or in dekatherms (1,000 scf of 1,000 Btu of gas).Energy accounting is commonly referred to as “thermbilling” as opposed to volume billing.

There are many factors involved with total energymeasurement. Some are fixed such as pipe sizing orwhether a differential or pulse output primary element isused. Others may be variable such as static anddifferential pressures, temperature, or gas composition.These factors are used in combination to arrive at thestandard base conditions which is the reference forcustody transfer.

The most common base conditions are 60°F and 14.73PSIA but it is possible that your associate may be usingdifferent conditions which can lead to misunderstandings.For example, if methane (CH4) is valued at 1014.3 BTU(dry) at 60°F and 14.73 PSIA, the same gas will convert to1006.8 at 60° and 14.65 PSIA. The higher the basepressure, the higher the Btu.

Often conflicts and miscommunications arise simplybecause the parties are not comparing gases at the samebase conditions, proverbially not comparing apples withapples!

FLOW MEASUREMENT

There are two major methods for measuring gas flow,inferential and positive displacement. Although there aremany types of meters falling into each category, we areprimarily concerned with five, orifice plates, turbinemeters, ultrasonic, diaphragm and rotary meters.

Orifice Meters:

An orifice plate is a restrictive element in the pipelinewith a precise hole in its center, Figure 1.

FIGURE 1.

An orifice plate essentially functions as a meter from whicha differential pressure signal is developed as a function ofgas flowing through the orifice bore. Flow is proportionalto the square of the pressure differential (DP or ∆P) betweenan upstream tap, across the orifice plate to the downstreamtap, hence the term, “square law devices.” This is aninferential signal developed under flowing conditions anddoes not take into consideration the various factors

∆P

Flow


necessary to produce an accurate representative number.Correct flow at pipeline conditions or flow at base referenceconditions is derived from the rules, conditions andequations set forth in AGA Report Number 3.

Also known as ANSI/API 2530 and GPA 8185, AGA-3describes and defines the mathematics, equipment andprocedures involved with the measurement of natural gasflow using flange tap and pipe tap orifice meters. Itprovides standards of construction and installation oforifice plates, meter tubes, and associated fittings as wellas computations providing basic factors to adjust for theprimary element expansion, flowing fluid density, flowingfluid expansion, Reynolds number, temperature, pressure,relative density and compressibility. That is, to correct forall the variations effecting measurement and to arrive at atrue and accurate measurement relative to flow at pipelineconditions and or flow at base reference conditions.

Turbine Meters:

Turbine meters are velocity measuring (inferential) deviceswhereby a rotor ideally spins at a speed proportional tothe gas flow rate — an analogy is a wind vane. However,rotational speed is a function of passageway size andshape and rotor design. It is also dependent upon theload imposed due to mechanical friction, fluid drag,external loading and gas density. Rotor revolutions arecounted mechanically or electrically, Figures 2a & 2b.

FIGURE 2A.

FIGURE 2B.

Counts or pulses are converted to a continuouslytotalized volumetric registration but since this flow ismeasured at pipeline conditions, it must be corrected tospecified base conditions for true reference and billingpurposes. Rules, conditions and corrective factors aredetailed in AGA Report Number 7.

AGA-7 defines installation, calibration, operation andcalculation methods applied to axial-flow-type turbinemeters measuring volumetric and mass flow.

In Europe, turbine meters are now the preferred meteringmethod for custody transfer.

Ultrasonic Flow Meters:

Ultrasonic flow meters are available in two fundamentaldesign categories, doppler and transit time or time of

flight. The preferred version used in natural gasapplications is the latter, time of flight approach. Theyoperate by bouncing a high frequency energy pulseagainst the piping sidewall to a second receiver ortransducer, Figure 3. By subtracting the time of flight ofalternating pulse transmissions downstream, thenupstream, a transit time differential is calculated. Oncethe meter has been properly calibrated, this differencein time is proportional to the gas flow rate,

Q = ∆T = ta – tb.

As with all meters, pressure, temperature, gas compositionand meter factors must be accounted for to correct actualcubic feet flow to flow at base conditions.

FIGURE 3.

Ultrasonic flow meters subscribe to Report AGA-9 for basicstandards of operation and installation and AGA-10 forSpeed of sound calculation (not finalized at this writing).

Diaphragm Meters:

A diaphragm meter is a positive displacement devicewhereby the flexing of an internal diaphragm indicates aspecific volume of gas that has been transferred. Thisdisplacement action is actuated by the differentialpressure developed by gas demand on the downstreamside of the meter. Diaphram meters are in wide use inthe low pressure gas distribution business although highpressure meters are also on use. (Figure 4).

FIGURE 4.

Flow

Flow


Rotary Piston Meters:

Commonly known as rotary meters these positivedisplacement devices feature two counter rotating “figure8” or lobed impellers. These impellers, or rotors, moveas a function of the force of flowing gas and areunaffected by gas density, turbulence and pulsation inthe line. The rotor lobes form a defined space betweenthe impeller and the meter wall, hence the term, “positivedisplacement meter.” For each complete internal rotationof the meter four defined volumes of gas are displaced.By knowing the volume of these spaces very precisemeasurements can be accomplished. (Figure 5).

FIGURE 5.

GAS QUALITY

Heating value generally notes quality of natural gas. Inthe USA, our standard of measurement is in BritishThermal Units or “Btu.” There are two general methodsfor determining BTU, direct and indirect. Two approachesrepresentative of the methods are gas chromatographyand calorimetry. Both have advantages anddisadvantages as compared to one another.

Gas Chromatography:

Indirect measurement (calculated) is most commonlyrepresented by gas chromatographs. These instrumentsrun the full range accuracy, repeatability andreproducibility. At one end are laboratory chromatographswhich are quite sophisticated, require close support butprovide extensive information on a detailed basis. At theother extreme are process instruments that typicallyoperate faster and do not require as much operatorattention. However, they lack the detailed analysisproduced by the more complex laboratory instruments.The number of gas components to be identified typicallydetermines the level of sophistication.

Gas chromatographs consist of three main sections, thesample conditioning system, the oven and the controller.Each has its own identity and must be carefully integratedinto the final instrument designed to best target theapplication.

A precise amount of sample gas is injected into a flowingstream of carrier gas which is usually ultra high purityhelium. The mixture then flows into and through a column

packed with a coated granular material unique to theproperties of the sample gas. There are literally hundredsof different columns available which requires the operatorto specify the correct column for the particularapplication. As gas travels through the column andaround the coated material, its components begin toseparate as a function of time and temperature. Gas flowsout and by a detector which measures each identifiedconstituent or group of constituents. A graphicrepresentation called a chromatogram is produced as aresponse to the detector output, Figure 6. Thechromatogram is raw data listing of component peaknumbers, retention time, peak area and height. Acomputer uses this information to calculate and producea report detailing the mole percent of components,heating value, relative density and compressibility factor.

Calorimetry:

Known as direct measurement because of its method ofdetermining heating value, combustion calorimetry isrepresented by two main groups of instruments, heattransfer and stoichiometric combustion.

The recording calorimeter known as the Cutler Hammerwas the primary direct measurement standard datingfrom the 1930’s. This instrument uses a fixed ratio tomix gas and air resulting in optimum combustion. Heatof combustion is transferred to an exchanger that in turntransfers heat to a stream of air. Air temperature isrecorded before and after passing around the exchanger.The difference of temperatures before and after passingaround the exchanger is proportional to heating value(Figure 7).

FIGURE 6.

FIGURE 7.

Gas

Air

Air

∆t=Hv

t+

t–


The second direct method utilizes principles ofstoichiometric or optimum combustion. When aparticular air/gas mixture is burned at its maximum flametemperature the hydrocarbon gases are converted to H2Oand CO2 (Figure 8). Btu can be accurately measured asan inverse function of gas flow rates when the instrumentis tuned to the optimum stoichiometric air/fuel ratio andcompared to the flow rate of a standard reference gassuch as methane.

TOTAL ENERGY MEASUREMENT SYSTEMS

To calculate total energy of natural gas the followingvalues are needed; pressure, temperature, compositionand flow rate. Supervisory systems can interface withflow computers, gas chromatographs, ultrasonic meters,turbine meters, pressure transducers and temperaturetransducers (Figure 9). These inputs combined with othersupervisory control functions such as actuating controlvalves, monitoring alarms, performing metering

diagnostics, graphical presentation of data to theoperator, and generating extensive reports constitutesa total energy measurement system. Operation becomessimple using its graphic screen with mimics, trends,extensive alarms and events (Figure 10).

FIGURE 10.

SSCADA & DCS

For pipeline control systems a real-time overview of thecomplete status of the pipeline is important. Figure 10.Since the metering sites are often unmanned, control ofthe metering sites and the block valve operation is oftencentralized in the gas control center. Traditionally, pipelineSCADA (supervisory control and data acquisition)systems are used in conjunction with telemetry overtelephone lines, cellular, radio or satellite. Connectivityover fiber optic cables, are becoming more cost effective,local area networks and intranet technologies arebecoming more popular.

As typical SCADA system consists of a central host ormaster MTU (master terminal unit), connected to severalRTU’s (remote station units) for data gathering andcontrol. These systems have open-loop control andutilize mostly long distance communications, but someparts of these systems utilize closed-loop controls andwill use short distance communications for local controland data gathering.

A DCS (distributed control system) is somewhat similarto a SCADA system in functionality, but the datagathering and control functions are usually co-located.Communication connections are made generally via aLAN (local area network) because of its reliability andhigh speed. A DCS system usually employs a largenumber of local closed loop controls.

FIGURE 9.

Air I/P

M

Gas

Hv = K(Qr/Qs)

where:K is the Btu reference valueQr is the reference gas flow rateQs is the sample gas flow rate

FIGURE 8.


David Hailey

CONCLUSION

Because of the dynamic nature of today’s natural gasquality, total natural gas energy is paramount to accuratemeasurement and fair custody transfer. There are avariety of hardware and software systems and devicesto accomplish accurate measurement. As long as thevarious parties agree to what constitutes accuratemeasurement disputes will be minimized. It is vital thatall parties agree on the technical terms, language andbase references of total energy measurement.


EFFECTS OF ENTRAINED LIQUID ON ORIFICE MEASUREMENTSWilliam Johansen

Colorado Engineering Experiment Station, Inc.54043 WCR 37, Nunn, Colorado 80648

INTRODUCTION

Natural gas often has some liquid content. The liquidmay be water, hydrocarbons, or compressor oil. As thegas flows through an orifice meter is the gas beingmeasured correctly? The measurement methods andcalculations described in ANSI/API 2530 are for dry gas.

Many researchers have studied the effect of entrainedliquids on orifice measurement. The existing literaturecan provide much information about orifice flowmetererrors. This information can be used to determine thecourse of future orifice plate research efforts.

This paper will discuss four test programs that wereconducted to examine the effects of entrained liquids onorifice meter performance. The results of these programswill be discussed as well as some simple flow models.The flow models will be used to explain why research intothis area has been so difficult. The flow models are notintended as a guide for flow measurement correction.

WET GAS

The liquid in natural gas may be in the form of very smalldroplets or there may be enough liquid to completely fillthe pipe at times. The behavior of the gas/liquid mixturepasses through phases as the liquid content and flowingvelocities change. These phases are referred to as flowregimes. Three simple flow regimes will be introduced.

Flow Regime 1

The first regime has the least amount of liquid in theflow. When the amount of liquid in the flow is low, theliquid is in the form of drops being carried in the flowstream. The drops are continually colliding with the pipewall while new drops are formed when the gas flowscours the liquid from the pipe surface. As the liquidcontent increases in the flow, more and bigger drops hitthe wall resulting in a film of liquid coating the entirepipe surface. As the film thickens, the roughness of thepipe wall begins to change due to the liquid filling in lowspots on the pipe surface. The pipe seems to becomevery smooth. The film moves down the surface of thepipe and is pushed onto the orifice plate by the flowinggas. As the liquid streams across the plate the platesurface conditions also begin to change. Streaks onorifice plates suggest that the fluid is drawn across thesurface as lines of fluid stretching from the outer edgeof the plate into the bore. Streaks make the orifice platesurface rough. The liquid also makes other changes to

the meter geometry. Liquid near the edge of the boremakes the sharp edge seem less sharp to theapproaching flow and liquid coating the inside surfaceof the bore make the bore make the bore diametereffectively smaller.

Flow Regime 2

Let the amount of liquid in the pipe increase so that thereis more than enough liquid to completely coat the pipe.If the gas velocity is low enough, a thicker layer of liquidmay form on the bottom of the pipe. The thick layer offluid may acquire a rough surface if the liquid surfacebecomes rippled or wavy, or the surface may be smooth.The thick layer of fluid is flattened against the pipe bottomby the flow. If the gas velocity increases, the thick layerof fluid can be flattened so much that it becomes auniform coating on the inside surface of the pipe. As theflow approaches the orifice plate a puddle of liquid mayform just in front of the plate. Large drops can be carriedfrom this puddle through the orifice bore. As the largedrops pass through the orifice plate bore the area of theplate effectively decreases, restricting the much fastergas flow. Liquid still moves across the plate surface instreaks and collects near the bore and on the insidesurface of the bore. The liquid behavior can change quitea bit in this flow regime with small changes in gas velocityand liquid load.

Flow Regime 3

If the liquid in the flow continues to increase, a permanentthick layer of fluid forms at the bottom of the pipe. Thesurface of this thick layer is wavy or rippled making thesurface seem rough to the gas flowing through the pipe.The liquid surface condition does not change as theamount of liquid increases and the layer begins to fill thepipe. Liquid is dammed up in front of the orifice plateand is drawn through the orifice bore at a constant rateby the gas flow. Much of the flow area of the orifice boreis now occupied by liquid which is moving at a muchslower rate down the pipe than the gas.

TEST METHODS AND TERMINOLOGY

Two-phase or multi-phase flow are terms that may beused to describe wet gas flow. Wet gas is two-phaseflow because there is a gas phase and a liquid phasepresent in the pipe. Wet gas researchers can measurethe gas flow and liquid flows separately before combiningthem in the test section. The liquid and gas flowratescan be changed allowing researchers to use a wide range


of liquid loads. A common way of describing the amountof liquid in the flow is to use a mass ratio. A commonexpression for mass ratio is:

Mass flowrate of liquidMass Ratio = —————————————— (1)

Mass flowrate of gas

The mass ratio may also be called liquid load and is oftenexpressed as a percent.

Wet gas studies are performed using systems similar tothe one shown in Figure 1. Dry gas enters the system onthe left side. An orifice plate or some other flowmeter isused to accurately measure the flow of dry gas into thetest section. Liquid flow enters the test system from thebottom of the figure. The flowrate of liquid is accuratelymeasured prior to being injected into the test system.The liquid is injected downstream of the flowmeter usedto measure the dry gas flow and upstream of the orificeflowmeter being tested. The total flowrate passingthrough the meter being tested is the sum of the dry airand the injected liquid. As flow passes through the orificeflowmeter being tested the differential pressure, staticline pressure, and temperature are measured. Thesevalues allow a calculation of flow which can be comparedto the known flowrate of the combined gas and liquid.

using steam/water, air/water, natural gas/water, naturalgas/distillate, and natural gas/salt water combinations.Testing was conducted at pressures ranging from nearatmospheric to over 900 psia. Testing was performedusing 21/2”, 3”, and 4” piping and beta ratios of 0.25 to0.5. The mass ratio was varied from about 0.02 to over 8.A plot of some of Murdock’s data is shown in Figure 2.

Note that as the mass ratio increases the percent of over-measurement increases. Also note that the errorsincrease more rapidly when the system pressure is higher.The data shown is grouped along the two straight linesshown on the plot. The trend of increasing flowmeasurement error as pressure increases appears to beclear. The clear trend suggests flowing conditions arestable. The flow in Regime 3 is stable. The large increasein differential pressure due to the load of the liquidpassing through the orifice bore dominates any othereffects and clearly correlates with pressure.

Flow Regime 2

McConaghy et. al. performed tests to determine theeffects of liquid and solid contaminants on orifice plateperformance. The testing was performed using highpressure natural gas and a system composed of 10”piping with beta ratios of 0.2 and 0.6. The mass ratioswere varied from 0.004 to 0.0004.

Orifice plate surface conditions were varied by machininggrooves into the plate surface and by putting grease onthe orifice plate surface. Two different groove depths andmany different grease applications were examined. Testresults showed that plate roughness caused under-measurement of flow but only when the height of theroughness became large enough to disturb the flowacross the plate.

Two-phase flow testing was performed by injecting glycolor oil into the natural gas flow stream upstream of theorifice plate. After each of the test runs the orifice platewas removed for inspection. Residue on the platesindicated that some puddling of liquids was taking place

FIGURE 2. Some of Murdock’s Experimental Data

FIGURE 1. Typical Two-Phase Test Setup

PREVIOUS RESEARCH

There has been a considerable amount of research onthe effect of entrained liquids on the performance oforifice plates. A few of the papers will be discussed interms of the flow regimes described above.

Flow Regime 3

Murdock performed a number of tests using orificemeters in two-phase flow. The testing was performed

Orifice meter or some otheraccurate meter to measure thedry gas flow Orifice meter being tested with

two-phase flow

Two phase flow leaving testsection

Compressed dry gassupply

Liquid is injected downstream of dry gasmeasurement and upstream of orificemeter being tested

P PT TdP dP

t e low regimes described above.

Figure 2. Some of Murdock’s Experimental Data Flow Regime 3 Murdock performed a number of tests using orifice meters in two-phase flow. The testing was performed using steam/water, air/water, natural gas/water, natural gas/distillate, and natural gas/salt water combinations.

0

2

4

6

8

10

12

14

0 0.5 1 1.5 2 2.5 3 3.5

Mass Ratio

Per

cent

Ove

r-M

easu

rem

ent

Low Pressure Data

High Pressure Data


just in front of the orifice plate and radial streaks indicatedthat the liquids were moving across the plate surface. Fromthese descriptions it appears that these tests wereperformed in the second flow regime discussed above. Theliquid behavior in this flow regime can vary considerably.

The test results showed considerable variations. Testingconducted using a low beta plate and high injection ratesproduced a smaller effect on flow measurement thantesting with a high beta orifice plate and lower injectionrates. The authors suggest that there were many potentialmechanisms which could affect flow measurement. Someof the mechanisms produce under-measurement of flowand some of the effects produce over-measurement offlow. Over-measurement of flow was caused by theeffective decrease in orifice bore area as large dropletspassed through the hole. Over-measurement was alsocaused by the roughness of the orifice inlet pipingchanging as the liquid layer made the surface smooth.There were several potential mechanisms that could causeunder-measurement of flow. Increasing orifice platesurface roughness, puddling of liquid in front of the plate,liquid on the plate surface close to the orifice bore, andincreasing pipe roughness as the surface of the liquid filmbecomes rippled or wavy all cause under-measurementof flow. The authors felt that many of these mechanismswere present at the orifice plate all of the time. Othermechanisms, such as the puddling, large droplets passingthrough the orifice bore, or varying roughness of the filmsurface may appear and disappear with small variationsin flow. Depending on how these mechanisms interact itwas found that the flow measurement errors could be verysmall or the mechanisms could produce significant under-measurement errors.

Flow Regime 1

Ting and Corpron performed testing using compressedair and water. The pipe size used was 8 inch and thebeta ratios were 0.5 and 0.7. The injected water producedmass ratios of 0.02 down to 0.0002. The test resultsshowed a negligible shift with the 0.5 beta plate butshowed large shifts with the 0.7 beta plate. Under-measurement of flow by as much as 1.7% with a 0.7beta plate occurred at high Reynolds numbers.

Clear trends of increasing under-measurement as massratio increased and as Reynolds number increasedshowed that the testing was performed in the first flowregime described. An even coating of liquid covers all ofthe pipe internal surfaces. The injection rates were highenough to maintain some level of coating but it is notclear whether or not there was damming of liquids in frontof the plate or large droplets of liquid passing through theorifice bore. In fact, this is one of the problems with testingconducted by injecting liquids into a gas stream. It is neverclear whether or not damming of liquids in front of theplate occurred. With these test results it was not clearwhether or not the under-measurement was caused bythe film of liquid alone or if some other mechanism suchas puddling contributed to the measurement error.

Work by the present author attempted to answer the questionabout the effects of thin films of liquids in orifice flowmeters.Testing was performed for the Gas Research Institute usingcompressor oil in systems flowing dry compressed air. Testingwas performed in 2 inch, 6 inch, and 10 inch pipe. Betaratios tested were 0.25, 0.5, and 0.7.

This testing was unique as it did not use liquid injection.The orifice plates and piping immediately upstream of theorifice meter were coated with oil. After allowing the excessoil to drain away the piping system was assembled andtesting was started. Figures 3 and 4 show some of thetest results. The plots show the percent change indischarge coefficient. A positive shift in dischargecoefficient corresponds to under-measurement by thesame percentage. Each test was started by taking dataat low flowrates and then proceeding to higher flowrates.Once the maximum flowrate had been achieved data wastaken as the flowrate decreased back to the originalflowrate. The magnitude of the shifts tended to decreaseas the test proceeded. This change corresponded to the

FIGURE 3.6” Pipe, β = 0.67 Thin Oil Film on Orifice Plate

and Inlet Piping

b

Figure 4. 10” Pipe, b = 0.66 Thin Oil Film on Orifice Plate and Inlet Piping

This testing was unique as it did not use liquid injection. The orifice plates and piping immediately upstream of the orifice meter were coated with oil. After allowing the excess oil to drain away the piping system was assembled and testing was started. Figures 3 and 4 show some test results. The plots show the percent change in discharge coefficient. A positive shift in discharge coefficient corresponds to under-measurement by the same percentage. Each test was started by taking data at

0

0.5

1

1.5

2

2.5

0 0.5 1 1.5 2 2.5 3 3.5

Bore Reynolds Number (Millions )

Perc

ent

Chang

e in D

ischarg

e C

oeff

icie

nt

Ascending Flowrates Descending Flowrates

FIGURE 4.10” Pipe, β = 0.66 Thin Oil Film on Orifice Plate

and Inlet Piping

Per

cent

Cha

nge

in D

isch

arg

e C

oef

ficie

nt

Bore Reynolds Number (Millions)


diminishing oil film as the air swept the oil out of the system.After each test the piping was taken apart and the plate andpiping were inspected. These inspections revealed that nearlyall of the oil had been scoured from the pipe wall by theflowing air during each test. The data in Figure 1 show thatwhen the test ended there was still a large under-measurement of flow. A very small amount of oil couldtherefore produce under-measurement of flow. Consistentunder-measurement errors occurred when both the orificeplate and upstream piping were coated with a thin film of oil.

CONCLUSIONS

Three two-phase flow regimes were described. The firstflow regime had very low liquid content. The liquidcontent was just enough to produce a thin coating ofliquid on all of the orifice meter internal surfaces.Research performed with low injection rates has shownthat flow is under-measured. The under-measurementis consistent and in some cases can be over 1%. Thesecond flow regime had a higher liquid content than thefirst regime. The liquid content was high enough that athicker layer of fluid could form at the bottom of the pipe.This thicker layer of fluid could become redistributedaround the pipe surface if fluid velocities increase.Puddling of liquid in front of the orifice plate could alsooccur. In this flow regime there is a lot of variability in thebehavior of the liquid and as a result the flowmeasurement may not be affected or there may besubstantial under-measurement of flow. The third flowregime has enough liquid in the pipe so that there is apermanent thick layer of fluid on the bottom of the pipe.The surface of the liquid has a fairly constant roughness.The liquid is dammed up in front of the orifice plate andis drawn through the orifice bore at a steady rate. Orificeplates used with liquid loads this high consistently over-measure flow. The amount of over-measurement isaffected by the flowing gas pressure.

Future Research

Figure 5 shows the general trends in flow measurementerror from the papers discussed. The plot shows thepercent error in flow measurement as the liquid to air massratio increases. The values shown are included only toprovide some idea of the magnitude of the measurementerrors. This plot and the previous discussion can providesome insight into the course of future research.

In flow regime 3 the orifice meter clearly over-measures flow.Future testing must be performed with natural gas and liquidsthat are representative of liquids found in pipelines. Murdock’swork clearly showed that the flow measurement errorschanged as line pressure changed. Those changes must beexamined in detail. Any other effects such as how beta ratioand Reynolds number change the over-measurement needto be identified and studied.

In flow regime 2 the performance of the orifice plateexperiences a great deal of variability. The research resultsof McConaghy et. al. illustrate the potential variability inflow measurement. If future research also shows variability

then the research results may be used to establish someboundaries of the possible errors in flow measurement.These boundaries could extend from 0.0% or slightlypositive flow measurement error to as much as 1% under-measurement. Test results should also define the upperand lower mass ratio boundaries of this region.

The performance of the orifice plate in flow regime 1 isnot as variable as in flow regime 2. The data suggest thatthe orifice meter under-measures flow when thin films ofliquid coat the internal surfaces of the pipe. Although itappears that Reynolds number does not strongly affectmeasurement errors future research needs to verify thisas well as the effect of beta ratio, line size, and pressure.

REFERENCES

Murdock, J.W., “Two-Phase Flow Measurement WithOrifices,” ASME Journal of Basic Engineering, Dec. 1962.

McConaghy, B.J., Bell, D.G., Studzinski, W., 1989, “HowOrifice Plate Condition Affects Measurement Accuracy,”Pipeline Industry, Dec Issue.

Ting, V.C. and Corpron, G.P., “Effect of Liquid Entrainment onthe Accuracy of Orifice Meters for Gas Flow Measurement”

Johansen, W.R., “Effects of Thin Films of Liquid CoatingOrifice Plate Surfaces on Orifice FlowmeterPerformance,” Report GRI-96/0375, Gas ResearchInstitute, Dec 1996.

William Johansen

FIGURE 5. Predicted Trends in Flow Measurement Error

Per

cent

Err

or

in F

low

Mea

sure

men

t


GAS FLOW CONDITIONING — DEVICES USED TO PRE-CONDITIONTHE GAS FLOW PROFILE PRIOR TO MEASUREMENT

B.D. Sawchuk P. Eng. and D.P. Sawchuk C.E.T.Canada Pipeline Accessories Company ltd.6710-30 St. S.E. Calgary, Alberta, Canada T2C 1N9

INTRODUCTION

A flow Measurement Engineer’s presentation of FlowConditioning for Orifice, Turbine and Ultrasonic Meters.The latest references are sited to provide an overview ofthe state of flow technology today and how thesetechnologies affect various popular meters used inNatural Gas measurement.

PIPE FLOW CONDITIONS

The most important—and most difficult to measure—aspects of flow measurement are flow conditions withina pipe upstream of a meter. Flow conditions refer to: thegas velocity profile, irregularities in the profile, varyingturbulence levels within the velocity or turbulenceintensity profile, swirl and any other fluid flowcharacteristics which will cause the meter to register flowdifferent than that expected. This will cause the meter todiffer from the original Calibration State referred to asreference conditions that are free of installation effects.

Installation effects which cause flow conditions withinthe pipe to vary from reference conditions are: insufficientstraight pipe, exceptional pipe roughness or smoothness,elbows, valves, tees and reducers, just to name a few.Certainly, a common understanding of how theseinstallation effects impact the meter is important sincedevices which create upstream installation effects arecommon components of any standard metering design.Flow Conditioning refers to the process of artificiallygenerating a reference, fully-developed flow profile andis essential to enable accurate measurement whilemaintaining a cost-competitive meter standard design.

Industry-accepted nomenclature and discussions arepresented which explain commonly referred to flowconditions.

The most commonly used description of flow conditionswithin the pipe is the velocity flow profile. For generalfluid dynamic background Miller (1996) offers a thoroughtextbook description of velocity profiles and distortionsof the profile due to upstream piping effects. The mostcommon method used to describe velocity flow profilesfor natural gas measurement is shown in Figure 1,Velocity Flow Profile.

Equation 1 describes the shape of the velocity flowprofile. The value of n determines the shape of thevelocity flow profile. Karnik (1993) and others use

Equation 1 to determine the flow profile’s shape withinthe pipe by fitting a curve to experimentally measuredvelocity data. Karnik (1993) was the first to actuallymeasure transverse velocities within the high-pressurenatural gas environment using hot wire technology toaccomplish the data fit.

(1)

A fully developed flow profile is used as the ReferenceState for meter calibration and determination ofCoefficient of Discharge (Cd). For Reynolds Number 105

to 106 n is approximately 7.5; for Re of 106, n isapproximately 10.0 where a fully developed profile in asmooth pipe is assumed.

Since n is a function of Reynolds Number and frictionfactor, more accurate values of n can be estimated byusing

(3)

where f is the friction factor. It is not the intent here toprovide detailed instructions for determining frictionfactors. The Colebrook (1939) equation or Moody (1944)diagram can be utilized as illustrated and detailed byKarnik (1993).

FIGURE 1. Velocity Flow Profile

Uy y

Umax = R )1

n

(

n = ___ ,1

√f

Umax

yR

Uy


A good estimate of a fully developed velocity flow profilecan be used for those without adequate equipment toactually measure the velocities within the pipe. White(1986) and Karnik (1993) utilize the following straight-pipe-equivalent length to ensure a fully developed flowprofile exists.

PipeDiameters ≈ 4.4D(Re) (2)

As one can see, the pipe lengths required by equation(2) are significant, hence the need for devices thatcondition the flow over a shorter pipe length allowingmetering packages to be cost competitive and accurate.

It is important to point out that the velocity flow profile isgenerally three-dimensional. For simplicity, normally thedescription requires no axial orientation indication if theprofile is asymmetric. If asymmetry exists, then axialorientation with respect to some suitable plane ofreference is required. Asymmetry exists downstream ofinstallation effects such as elbows or tees.

Normally, the velocity flow profile is described on twoplanes 90˚ apart. With today’s inexpensive computer andsoftware technology a full pipe cross sectional descriptionof the velocity profile is possible (if sufficient data pointsare provided of course).

The second description of the flow-field state within thepipe can be made using turbulence intensity. Karnik,Jungowski and Botros (1994) showed that metering errorsmay exist even when the velocity flow profile is fullydeveloped and pipe flow conditions seem perfect.Conversely, they found zero metering error at times whenthe velocity profile was not fully developed. Theyattributed this behavior to the turbulence intensity of thegas flow that can cause metering bias error. This behavioraccounts in part for the less than adequate performanceof the conventional tube bundle.

Delving into the mechanisms governing the effects ofturbulence intensity is not within the scope of this paper.It is therefore highly recommended that the state oftechnology be pursued via the references provided,namely (Karnik, Jungowski and Botros).

The third description of the flow field’s state is swirl. Swirlis the tangential flow component of the velocity vector(The velocity profile should be referred to as the axialvelocity profile. Recall that the velocity vector can beresolved into three mutually orthogonal components, thevelocity profile only represents the axial component ofvelocity).

Figure 2, Swirl Angle illustrates the definition of flow swirland swirl angle. Note that swirl is usually referenced tofull body rotation (that which the full pipeline flow followsone axis of swirl). In real pipeline conditions, such asdownstream of elbows two or more mechanisms of swirlmay be present. Miller (1996) provides additional detailspertaining to the effects of installation effects such asone and two elbows in and out of plane.

FIGURE 2. Swirl Angle

ORIFICE METERING

Orifice metering is one of the most commonly usedmetering technology in the gas production andtransmission industry. This is due to the low installationand maintenance cost, and the low uncertainty whichcan be achieved using these meters.

Another significant advantage orifice meters have overmost other technologies is the ability to use the meterswithout having them proved or calibrated (Note that thereis a requirement to calibrate the instrumentationmeasuring pressure, differential pressure andtemperature, but not to flow prove the meters themselves).This feature is achieved by precisely defining the geometryof the meter, so that each meter reacts similarly to theflow of gas. This is referred to as dynamic similarity.

In order to ensure that the meters have the property ofdynamic similarity, orifice meters should conform to theindustry standard for these devices, Orifice Metering ofNatural Gas and Other Related Hydrocarbon Fluids,referred to as A.G.A Report No. 3, or API 14.3.

A key factor in maintaining dynamic similarity is thevelocity profile of the gas as it enters the meter. Theorifice equation was developed using a fully developedvelocity profile. This is considered to be the profile ofthe gas after travelling through a long length of uniformstraight pipe. Unfortunately, it is generally not possibleto arrange the piping upstream of a meter so that thereis enough straight pipe to achieve fully developed flow.The 1999 version of the standard results from testingdone at a number of recognized facilities. The intent ofthese tests was to provide installation requirements,including the piping upstream of the meter so that theresults are the same as those obtained from a longuniform upstream pipe.

A key aspect of the upstream installation is a flowconditioner, which modifies the flow prior to passingthrough the meter. A.G.A 3/API 14.3 provides threeoptions for flow conditioners: 1)The user of the metercan opt for not using a flow conditioner (a bare tube); 2)A tube bundle flow straightener of specific design canbe used (the details for this can be found in the standard);or 3) A flow conditioner, which passes a series of tests.

16

Axial Component

Tangentialcomponent

Radial

component

Swirl Angle


These tests are outlined in the standard. This third optionis the subject of this paper.

The reason that someone using an orifice meter woulduse this third option is that it there are several flowconditioners on the market that easily outperform thetube bundle flow straightener. The testing for the 1999version of the standard has shown that the installationrequirements for the tube bundle, in the previous versionof the standard, do not allow sufficient flow developmentto reproduce the long uniform length of pipe used indeveloping the standard. Because of this, many existingmeter installations do not meet the new standard. It ispossible, however, by simply replacing the tube bundlewith another flow conditioner that the meter will againmeet the standard, and be capable of measuring flowaccurately.

A listing of the various flow conditioners available to themarket place are listed below

Flow Conditioners Available

• 19 Tube• Akashi• AMCA• Bellinga• Bosch & Hebrard• Etoile• Gallagher• ISO• K-Lab• Kinghorn• Laws• K-Lab/NOVA 50E• PG&E• Sens & Teule• Spearman• Sprenkle• Stuart C-3• Zanker

The reader is STRONGLY cautioned to note that all theflow conditioners are not equal in terms of performancecharacteristics, pressure drop or approvals.

From a commercial standpoint two of the listed flowconditioner brands are generally accepted through outthe Unites States. They are the; Gallagher or GFC FlowConditioner and the CPA 50E flow Conditioner.

Gallagher Flow Conditioner

Below, as an example, the CPA 50E Flow Conditioneris used to describe A.G.A. – 3 2000 compliance. Thereader should ask for this information whenspecifying a flow conditioner.

The CPA 50E Flow Conditioner consists of a steel plate0.125D to 0.15D in thickness with a central circular holeand two rings of circular holes in concentric circlesaround the central hole. The spacing and design of theflow conditioner has been developed to produce a fullydeveloped profile downstream of the conditioner. Theholes account for approximately 50% of the area of theplate, and are sized such that the velocity profile a shortdistance downstream will be the same as the fullydeveloped profile. The thickness of the plate alsoeliminates swirl which may exist upstream of the meter.

A diagram of the CPA 50E flow conditioner is shownin Figure 3.

FIGURE 3. CPA 50E Flow Conditioner

TEST REQUIREMENTS

Detailed test requirements are outlined in the 1999version of API 14.3/AGA 3 Appendix 2D (2). ThisAppendix outlines the range of pipe sizes, b ratios, andRenolds Numbers to be used (users are advised to referto the published standard). The test requirements forapproval of type are presented here in abbreviated form. 1. Baseline Calibration using a bare meter tube with

a minimum of 70D upstream of the orifice plate, withless than 2 degrees of swirl upstream of the 70Dtube. This test should be performed with the sameorifice plates and b ratios that will be used for tests2 to 5.

2. Good Flow Conditions with the flow conditionerdownstream of the meter tube used in Test 1. Thistest is done to show how the flow conditioner affectsthe meter baseline.

3. Two 90° Elbows in Perpendicular Planes — Thistest will show how the flow conditioner handlesnormal amounts of swirl.

4. Gate Valve Closed 50% — This test shows howthe flow conditioner handles highly asymmetricvelocity profiles.


5. High Swirl — This test shows how the flowconditioner handles high swirl (min. 24° at 17D). Thedesign of a device to generate this swirl is given inthe standard.

In addition, there are requirements for testing differentsizes to ensure scalability of the flow conditioner, andtesting at multiple flow rates. These last two tests mustbe performed on the baseline calibration and at leastone of the disturbance tests.

The performance that the flow conditioners must meet topass these tests is very stringent. The threshold foracceptance is half the 95% confidence level in the RGequation at infinite Reynolds Number. This is slightly morethan twice the repeatability for the SwRI facility, whichmeans that it is close to the detectable limit for that facility.

TEST RESULTS

Southwest Research Institute — This testing wassponsored by GRI at the GRI Metering Research Facility(GRI MRF) at Southwest Research in San Antonio. Thepurpose was to provide data to support the developmentof the tests which later became part of API 14.3/AGA 3.Only a small portion of the data is included in this paper.For a complete report, please contact GRI.

Test 1 — Baseline Calibration - Southwest Researchhas done a baseline calibration on the meter used in theother tests. This calibration falls within the uncertaintyof the overall data set used to develop the flow equation.

Test 2 — Good Flow Conditions - This test is performedto ensure that The flow conditioner does not alter themeter indication when place in a line with good flowconditions (already fully developed). Note that the resultspresented here are a subsection of the results presentedin the full report developed by Southwest Research.These results are those for b = 0.67. The results for Test2 are shown in Figure 4. Note that when the flowconditioner is very close to the orifice plate the readingis outside the limits set forth. However, after the flowconditioner is moved 5D or so from the plate, themeasurement is within the specification.

Test 3 — Two 90° Elbows in Perpendicular Planes —This test demonstrates the performance of the flow

conditioner in conditions of swirl. The results for a 17Dmeter run are shown in Figure 5. In this case, 7D is requiredso the reading is within the tolerance allowed. 29D and45D meter runs were also tested, with similar results.

Test 4 — Gate Valve Closed 50% — The results areshown in Figure 6. Again, 7D is required to meet thespecification.

FIGURE 4. Nova 50E Performance in Good FlowConditions GRI MRF Data - Beta 0.67

FIGURE 7. Nova 50E Performance — High Swirl GRI MRF Data - Beta 0.67

FIGURE 6. Nova 50E Performance — Partly Closed ValveGRI MRF Data - Beta 0.67

Test 5 — High Swirl — The results are shown in Figure 7.The flow conditioner meets the requirements 7Dupstream of the orifice plate.

FIGURE 5. Nova 50E Performance — 2 Elbows Out ofPlane GRI MRF Data - Beta 0.67

Figure 5 - Nova 50E Performance - 2 Elbows Out of Plane GRI MRF Data - Beta 0.67

-1.00%-0.80%-0.60%-0.40%-0.20%0.00%0.20%0.40%0.60%0.80%1.00%

1 3 5 7 9 11 13

x/D

Dev

iati

on

fro

m B

asel

ine

Cd Tap 1

Tap 2

Figure 6 - Nova 50E Performance in Good Flow Conditions

GRI MRF Data - Beta 0.67

-1.00%

-0.80%

-0.60%

-0.40%

-0.20%

0.00%

0.20%

0.40%

0.60%

0.80%

1.00%

0 5 10 15 20 25 30 35 40 45 50

x/D

Dev

iati

on

fro

m B

asel

ine

Cd

Tap 1Tap 2

Figure 7 - Nova 50E Performance - Partly Closed Valve

GRI MRF Data - Beta 0.67

-1.00%

-0.80%

-0.60%

-0.40%

-0.20%

0.00%

0.20%

0.40%

0.60%

0.80%

1.00%

1 3 5 7 9 11 13

x/D

Dev

iatio

n fr

om B

asel

ine

Cd Tap 1

Tap 2

Figure 8 - Nova 50E Performance - High Swirl GRI MRF Data - Beta 0.67

-1.00%

-0.80%

-0.60%

-0.40%

-0.20%

0.00%

0.20%

0.40%

0.60%

0.80%

1.00%

1 3 5 7 9 11 13

x/D

Dev

iati

on

fro

m B

asel

ine

Cd

Tap 1Tap 2


Scaling — Tests were carried out by the Nova Research& Technical Centre at TransCanada’s Didsbury Testfacility (testing sponsored by TransCanada Pipelines).As per the requirements of the standard, this testing wascarried out on the baseline condition and on one otherof the proscribed tests. In this case, Test 3 – Two 90°Elbows in Perpendicular Planes, was repeated. Theresults are reported in detail in Reference 5. Results areshown in Figure 8.

Reynolds Number Sensitivity — Results reported bySwRI (3) for testing on 0.67b plates are at Re 2.7 x 106.Testing performed by NRTC (6) on air at 0.68 b plates atRe 1.7 x 105 and 1.0 x 105 give results as follows:

NTRC Testing on Two Elbows our of Planeon Air at Low Pressure

Beta Reynolds Deviation fromRatio Number Baseline Cd

0.68 170,000 –0.24

0.68 100,000 –0.08

This meets the requirements of the standard over a muchlarger range of Re than specified. This testing wasconducted on 4 inch flow conditioners, as per therecommendation of the standard.

enough development downstream of the flow conditionerto meet the requirements of the new standard. Eventhough these installations may be grandfathered underthe current revision some additional uncertainty is present.

For example, the 1991 version of API 14.3 Part 2, usingfigure 2.5 Partly Closed Valve Upstream of a Meter Tube,an overall length of 17D from the orifice plate upstreamto the partly closed valve is required (b = 0.75). 7D isrequired from the straightening vane outlet. The workdone by Southwest Research at the GRI MRF (3) showsthat this configuration results in much higheruncertainties than described in the previous version ofthe standard (almost 2%).

Under all of these installations using tube bundles andthe minimum lengths specified in API 14.3, the situationcan be corrected by installing a CPA 50E flow conditionerin place of the tube bundle.

Another example is a meter with no straightening vaneor flow conditioner, downstream of two in-plane elbows.In this configuration, the previous version of the standardrequires 17D between the last elbow and the orifice plate.Again, testing done by SwRI shows that this length ofupstream pipe is not adequate to avoid high uncertainty.A flow conditioner such as the CPA 50E can also beused in this case, instead of substantially increasing theupstream length required.

These examples show how a CPA 50E flow conditionercan be used to retrofit existing meter runs to meet thenew standard.

New Installations — When new meters are beinginstalled, using a CPA 50E flow conditioner can reducethe amount of upstream pipe required to keep uncertaintyas low as possible. Of course, flow conditioning is onlyone of many factors which affect overall meterperformance.

ULTRASONIC FLOW METERING

One of the many benefits of ultrasonic metering is thattypically, ultrasonic meters measure high volumes ofnatural gas. As a result, efforts to provide small increasesin accuracy and/or repeatability result in large benefitsto the customer. Installation experience, testing andcomputational exercises have been conducted whichshow how small efforts to provide correct velocity profilescan have considerable payback. Strategies forminimizing the influences of installation effects andprovision of correcting velocity profiles on the ultrasonicmeters are presented. An almost zero cost method isalso presented to drastically increase the measurementconfidence from the primary measurement meter.

Ultrasonic meters have gained rapid acceptance in themeasurement of large volumes of natural gas in the lastseveral years. There has been a concentrated effort bymany parties, AGA taking a lead role, to improve

FIGURE 8. Nova 50E Scability Test NRTC Data

FLOW CONDITIONER APPLICATION

The CPA 50E flow conditioner can be used as a retrofitin existing orifice meter applications, or as part of a newinstallation. Both of these uses of the flow conditionerwill result in a good installation when some care is takenin the design of the facility. In general, the upstreamportion of the piping should be such that strong swirl orhighly asymmetric flow is not present. When these areunavoidable some additional meter tube between theflow conditioner and orifice plate is suggested.

Retrofit Installations — There are many meters installedin accordance with the former version of the standard.In these cases, the minimum length does not always give

Figure 10 - Nova 50E Scalibility TestNRTC Data

-1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

0 5 10 15 20 25

x/D

Dev

iati

on

fro

m B

asel

ine

Cd

B 0.31B 0.41B 0.50B 0.60B 0.67B 0.75


understanding of these meters so that they can be usedproperly. If we examine the reasons for this acceptancethere are several advantages that ultrasonic meters haveover other technology. The main advantages are:

1. Installed cost at high volume facilities 2. Low pressure loss through the installation 3. Inherent bi-directional capability 4. Low maintenance 5. High tolerance for liquids or other foreign material

The work conducted to date has led to the publicationof AGA Report No. 9 [1], and further work is underwayto provide continued revision to AGA-9 as required.

Ignoring all other benefits of ultrasonic metering, in termsof accuracy, it could be argued that other technology,such as turbine or orifice meters, have a 25 to 70 yearadvantage in terms of optimized installation knowledge.This point may never be fully resolved, but it is safe tosay that current testing and constant productimprovement by manufacturers has shown a clearadvantage for ultrasonic metering.

Facility designers are looking for direction to best installthese meters to take advantage of the benefits thatultrasonic meters offer, while ensuring that accuracy isoptimized. Testing has been done by Grimley atSouthwest Research [2], as well as other locations,however, this testing is expensive, and it is difficult totest all the possible factors which could lead to increasedmeasurement accuracy. Since there are many factorsthat have not been tested, users should employconservative design practices that will meet current andlikely future requirements. These design practices taketesting which has been completed into account, as wellas an understanding of ultrasonic metering.

The AGA task group on ultrasonic metering, inconjunction with test facilities, manufacturers and users,is undertaking to develop new testing regimes which willprovide additional information on how the ultrasonicmeters work under various conditions.

Note that there is a difference in requirements betweenthose users who need the most conservative design toensure accuracy, and those who have limited space (suchas production platforms), who need solutions thataccount for their constraints. This paper is focused moreon exploring conservative design practices.

SENSITIVITY TO VELOCITY PROFILE

Ultrasonic meters are an alternative for turbine meters,especially for high flow (high capital cost) applications. Themeters are similar in how the flow rate is presented to theflow computer, and both have a relatively high turndownratio. Ultrasonic meters have the additional advantages ofbeing bi-directional, requiring less maintenance, having lesspressure loss across the meter and being more tolerant offoreign material in the gas stream.

There are significant differences, however. The turbinemeter is an intrusive device, as such, the turbine meter(Figure 9) forces the gas into an annular ring around theoutside of the pipe, possibly through a built in flowstraightener, then through the turbine rotor. The rotor itselfwill help to balance any skewed velocity profile. As aresult these meters are less sensitive to upstreamconditions, and a flow conditioning plate - five to sevendiameters upstream will produce excellent results.

The ultrasonic meter, however, is generally a continuationof the pipe, and does not condition the flow at all withinthe meter. The meter is therefore unable to alteraberrations in the velocity profile and is exposed to worstcase installation effects.

Figure10 shows a diagram of a single path ultrasonicmeter. There is one measurement taken (per path), - thedifference in the transit time. Because the ultrasonic pulsedoes not differentiate where it is in the pipe, it cannotcorrect for the difference in volume in the outer 50% ofthe diameter and the inner 50%. In order to correct forthis, manufacturers incorporate a correction factor basedon an assumed velocity profile.

FIGURE 9. Turbine Meter

FIGURE 10. Ultrasonic Meter

Since the transit time ultrasonic meters are measuringthe time difference of a sound pulse traveling on apredetermined path in the gas stream, the meter onlyhas one measurement per path. This measurement canonly determine the average velocity along the path. Thereare a number of factors which affect this measurement:

1. The velocity profile of the fully developed flow stream,which depends on the Reynolds number, roughness,imperfections in the pipe such as welds and flanges,and other factors.

Flow

Gearbox

Processor

Centre 50% of Path covers 25% of area and 35 to 40% of flow.Flow


2. How much the velocity profile is distorted (includingthe effects of swirl) resulting from the upstreamconfiguration, and the orientation of the meter to thisskewed velocity profile.

3. Changes to the velocity profile as it is passing themeter. Since the currently available meters areroughly one or more diameter long, there is thepossibility of changes to the velocity profile as itpasses through the meter.

FULLY DEVELOPED PROFILE

First we will examine the effect of fully developed velocityprofile, considering small changes and calculating theeffect of these changes.

The value of n is dependant on the Reynolds number andfriction factor. As such it will vary with gas composition,pipe roughness, diameter and imperfections such aswelds, gaskets and flanges. Values of 10 or 11 arecommon in the pipe sizes and conditions common withlarge diameter ultrasonic meters. This is described in moredetail by Karnik [3].

The meter sensitivity was compared to two profiles thatare very close to each other. These were calculated usingthe power law profiles for N=10 and N=11. The N=11profile is slightly flatter, with a centre line velocity of a bitover one percent lower than the N=10 profile (Figure 11).In practical terms, it would be difficult to differentiate thevalue of N from experimental data from these two profiles.

Calculations using these slightly varying axial velocityprofiles show sensitivity to the profile. For these twoprofiles the transit time difference is 0.42% for the sameflow rate.

The reason for this difference in transit time is that thevelocity in the centre of the pipe, is treated the same asthe velocity closer to the wall, even though the bulk flowis closer to the wall. Over 60% of the bulk flow is travelingin the outer 50% of the diameter.

This result is not sensitive to the model used to calculatevelocity profile, because all the models for velocity profilehave similar characteristics, being empirical curve fits tothe same data. Even though this simple model breaksdown at the centre line, other models will produce verysimilar results.

ASYMMETRIC VELOCITY PROFILE

Yeh and Mattingly have conducted CFD studies onultrasonic meter performance in ideal and non-ideal pipeflows [5]. They concluded that single path meters havea high degree of flow profile sensitivity and thatperformance is strongly dependant upon path orientationto a disturbance (as high as 35% close to an elbow).These results lead to the further conclusions thatinstallation location and orientation are critical, and thatmulti-path meters are desirable for high accuracy.

Certainly, a 35% shift in meter reading is not acceptablein current meter applications. The issue for custodytransfer meters is whether the specifications in AGA-9[1] are appropriate, and whether a particular installationmeets this specification.

In order to reduce the effect of this asymmetry, Yeh andMattingly recommend using multi-path meters. Thesemeters will significantly improve the response toasymmetric profiles, but may not totally compensate forthe profile. Results from meter testing by Grimley showthat some conditions will result in shifts of over onepercent when compared to good flow conditions.

FIGURE 11. Fully Developed Profile

There are several models for the fully developed velocityprofile. These are empirical approximations, and aresomewhat interchangeable. When measurements aretaken on flow that is fully developed, there is somedifference between the theoretical profile and themeasured profile. This is normal, and reflects somemeasurement uncertainty as well as theoreticalassumptions. If the measured profiles are within one ortwo percent of the theoretical profiles, they can beconsidered fully developed. In ISO 5167 [4], the profileis considered to be fully developed if the measured profileis within of five percent of the calculated profile.

The Power Law is one of these empirical profiles. Theequation is shown as equation 1.

U y

Ub = R )1

n

(Where:U = local axial velocityUb = Bulk or average axial velocityy = distance from the pipe wallR = pipe radiusn = Empirical exponent

(1)


CHANGING VELOCITY PROFILE

If the meters are sensitive to small changes in velocityprofile, they are probably also sensitive to the profile itselfchanging within the meter. This is an issue with ultrasonicmeters since they are from one to three diameters long,and numerous tests have shown that the velocity profilecan change significantly within this distance.

In a straight pipe, a distorted velocity profile will transformitself, moving toward a ‘fully developed’ profile. If a profileis close to the fully developed state, the change will begradual. An example of a rapidly changing profile is theprofile at the exit of a perforated plate flow conditioner(Figure 12). The profile at the exit of the flow conditionerwill be a number of jets, corresponding to the holes inthe plate. These jets quickly form into one profile. Testinghas shown the distance to form one profile is about fivepipe diameters.

conditioner and spool piece(s), to ensure that it is thesame as at the calibration facility. Even if this care istaken, the reassembly may not be exact, since there isno comprehensive testing to show how exact thereassembly needs to be.

TESTING AT SOUTHWEST RESEARCH

Grimley [2] tested the meters with a variety of upstreamflow conditions, including single elbow, elbows out ofplane, and elbows in plane. These conditions are similarto those found in field locations, but of course, are notrepresentative of everything found in the field.

Testing done on several ultrasonic meters with differentflow conditioners shows that the user must be careful inhow the meters are installed. Some combinations ofmeter, flow conditioner, and piping configuration wereclearly shown to be better at repeating baselinemeasurements than other combinations.

Grimley concluded that use of a high performance flowconditioner was preferable due to the difficulty intranslating ‘bare tube’ results (without flow conditioners)to a field location. These high performance flowconditioners isolate the meter, to a certain extent, fromthe upstream conditions.

Testing using high pressure gas on these large meters isexpensive, and the results of the testing complete todate are not conclusive. There are other questions yetto be addressed, such as: Are the results repeatable?What is the effect when other variables are altered? Howmuch depends on the spool pieces, the exact positionand orientation of the flow conditioner, and other specificinstallation conditions? Are the results the same at highergas velocity?

These results show some variation in performance ofthe ultrasonic meters tested which is not fully understood.While these deviations are generally small in percentageterms, they are of concern due to the high volume ofgas that is generally measured by these meters.

It should be noted that in the testing carried out atSouthwest Research Institute, the objective was todetermine the effect of changing the upstream profile,and the meter remained bolted to the flow conditionerand spool piece throughout the test. At custody transfermeter stations, the meter is generally unbolted from theflow conditioner after calibration, and re-assembled onsite. It will be difficult to re-assemble the meter exactlyas it was assembled at the calibration facility.

The CPA 50E flow conditioner is designed to recombineinto a fully developed profile within about five diameters,in order to minimize this effect. Even if the flowconditioner, spool piece, and meter is disassembled aftercalibration, the shift should not be expected to be large.If the opportunity to keep the meter run bolted togetheris available, it may be best to leave it this way until

FIGURE 12. Profile Downstream of a Perforated Plate

Perforated Plate flow conditioners accelerate the gas flowthrough holes in the plate. These individual gas streamsthen recombine to a single stream, which will then movetoward full development given enough straight, uniformpipe.

As the jets recombine, the profile will be changing quickly.If the meter is located too close to the flow conditioner,the indicated flow rate could be affected by these jets,and the bulk velocity may not be calculated correctly.While it may be possible to calibrate this effect out ofthe system, a slight change could significantly alter theindicated flow rate.

Meters are currently being installed with flow conditionersas close as three pipe diameters. There is some supportfor this, however in testing by Grimley, the meter wasnot unbolted from the flow conditioner when theassembly was placed in front of the flow disturbance.The testing was done to evaluate the effect of thedisturbance, not the effect of a rotated flow conditioner.The normal procedure that companies now employ, isto calibrate the flow conditioner, spool piece and metertogether at the calibration facility. The parts are thendisassembled for shipping to the site. At the site, careneeds to be taken to re-assemble the meter, flow

~ 5 D

CPA 50E FlowConditioner


installation. This option is of course not possible with aNPS 30” or 24” meter run.

VERIFICATION TECHNIQUES

There are techniques which will improve the usersconfidence that the metering installation is workingproperly. These include meter diagnostics and checkmeasurement.

METER DIAGNOSTICS

When the meter is calibrated, a log file from the meter isavailable which shows such things as the individual pathvelocities, speed of sound, gain, and other information.Users should acquire this data from the calibration facility,and gather some similar information when the meter isinstalled and measuring gas. Differences, especially inthe path velocity ratio, indicate that the velocity profile isnot the same as when calibrated. This likely means thatthere is a shift in meter performance.

CHECK MEASUREMENT

Conservative design suggests that there be a mechanismto verify or trend the measurement. Ultrasonic meterscan provide some redundancy if there is more than onepath, and there is sometimes other measurement to trendagainst the meter. However, in some cases operatorsmay want some other verification.

While installing an additional meter may not beconsidered cost effective, it is possible to use the flowconditioner as a check meter. Testing has been donewith the CPA 50E flow conditioner with differential tapsdrilled on the upstream face for stagnation pressure, andin the wall of one of the holes for downstream pressure.This measurement has been shown to be very stableand -repeatable.

The data can be collected using a standard flowcomputer setup. In order to reduce data handlingrequirements, the data can be trended and saved. It onlyneed be retrieved and analyzed if there is a changecompared to the ultrasonic meter.

RECOMMENDATIONS

Meter Run DesignThe testing conducted by SwRI indicates that there issome sensitivity to flow conditions upstream of the flowconditioners. This is clearly a result of high sensitivity tovelocity profile, even with multi-path meters. Usersshould take this into account when designing facilities.The following is recommended:

1. High performance flow conditioners should be used,at a distance from the meter which will allow the jetsto fully re-combine, so that the velocity profile at themeter will be stable.

2. Meters should be calibrated with the flowconditioners and spool pieces which will be used inthe meter station.

3. Manufacturers minimums for upstream lengthsshould be maintained or exceeded. A minimum of 8diameters should be between the flow conditionerand the meter.

4. Spool pieces should be finished with welds groundout and perhaps internally coated. Diameter of thespool piece should closely match the meter diameter.

5. Flow conditioner orientation at the calibration facilityshould be marked and reestablished.

6. Piping upstream of the meter run should be designednot to produce extreme velocity profiles, swirl ornoise

The installation might me slightly more expensive, butthere will be more confidence in the measurement.

DiagnosticsDiagnostics should be employed when the meteringinstallation is commissioned. These include log fileevaluation and check measurement (if available).

CONCLUSIONS

The CPA 50E flow conditioner meets the requirement forapproval of type in the 1999 version of API 14.3/AGA 3.

Existing meters designed to meet previous of thestandard can be retrofitted to meet the new standard, inmost cases at nominal cost, by installing a new flowconditioner.

Flow conditioning is an important aspect of orificemetering, but a flow conditioner cannot correct allproblems associated with metering. The condition of themeter tube and orifice plate, as well as the conditionand design of the secondary equipment are all factorswhich may also affect the quality and performance ofthe meter.

The most critical component of the meter station is themeter, flow conditioner, meter run and upstream pipingconfiguration. In terms of the overall capital cost of thefacility these components do not necessarily get theattention they deserve.

Small percentage cost savings efforts aimed at the meter,meter, flow conditioner, meter run and upstream pipingconfiguration are dangerous. Especially considering thelife span of a typical meter station is 20 years.

Measurement bias cost over this capital life span can besubstantially larger than the initial savings obtained.If there is no conclusive testing data to substantiate costsaving efforts in these critical areas it is best not tochance it.

1. Put thought into the upstream meter station pipingto minimize loading of the flow conditioner and meterby minimizing flow disturbing designs.


2. Ensure that the meter run is not allowed to becomethe last item of interest in the bill of materials.Roughness, roundness, length and overall fabricationquality is important until testing proves otherwise.

3. Work closely with your meter and flow conditionervendors.

4. If at all possible leave the meter run assembly boltedtogether.

5. Don’t change pipe roughness through the meter runand meter.

6. Use the diagnostics the meter manufacturers supply. 7. Use the CPA 50 E flow conditioner as a secondary

measurement device. It will drastically increase theconfidence level of the primary measurement.

REFERENCES

Measurement of Gas by Multipath Ultrasonic Meters,AGA Transmission Measurement Committee Report No.9, American Gas Association, June 1998

Grimley, T. A., “Ultrasonic Meter Installation ConfigurationTesting”, AGA 2000 Operations Conference, May 2000,Denver, Colorado

Karnik, U., “Measurements of the Turbulence StructureDownstream of a Tube Bundle at High ReynoldsNumbers”, Transactions of the ASME, Vol. 116,December 1994

ISO 5167, Measurement of Fluid Flow in Circular CrossSection Conduits Running Full

Yeh, T. T. and Mattingly, G. E., “Computer Simulations ofUltrasonic Flow Meter Performance in Ideal and Non-Ideal Pipeflows”, 1997 ASME Fluids Engineering DivisionSummer Meeting, June, 1997API 14.3/AGA 3 Part 2, 1991

API 14.3/AGA 3 Part 2, 1999

Morrow, T . 1997 Gas Research Institute/SouthwestResearch Institute Technical Memorandum“Development of a Flow Conditioner Performance Test”

Studzinski, W., Karnik, U., LaNasa, P., Morrow, T.,Goodson, D., Hussain, Z. and Gallagher, J. 1997 “GRIreport White Paper on Orifice Meter InstallationConfigurations with and without Flow Conditioners”

Karnik, U. and Kowch, R., 1999, “Scale Up Tests for theNova Flow Conditioner for Orifice Meter Applications”4th International Symposium of Fluid Flow Measurement,Denver, Colorado

Karnik, U., 1995, “A Compact Orifice Meter/FlowConditioner Package” 3rd International Symposium ofFluid Flow Measurement, San Antonio, TexasMiller, W. Richard, “Flow Measurement EngineeringHandbook”, McGraw-Hill, Third Edition, 1996, ISBN0-07-042366-0

Karnk, U., “Measurements of the Turbulence StructureDownstream of a Tube Bundle at High ReynoldsNumbers”, ASME Fluids Engineering Meeting,Washington D.C., June 1993

Colebrook, C.F., “Turbulent Flow in Pipes, with Particularreference to the Transition between the Smooth andRough Pipe Laws”, J. Inst. Civ. Eng., vol. 11, pp. 133-136, 1938-1939.

Moody, L.F. “Friction Factors for Pipe Flows”, Trans.ASME, vol. 66, p 671, 1944.

White M. Frank, “Fluids Mechanics”, Second Edition,McGraw-Hill, 1986, ISBN 0-07-069673-x

Karnik U., Jungowski W.M., Botros K.K., “Effect ofTurbulence on Orifice Meter Performance”, 11th

International Symposium and Exhibition on OffshoreMechanics and Arctic Engineering, ASME, May 1994,Vol. 116

American Gas Association Report No. 3, AmericanPertroleum Institute API 14.3, Gas ProcessorsAssociation GPA 8185-90, “Orifice Metering of NaturalGas and Other Related Hydrocarbon Fluids”, ThirdEdition, October 1990, A.G.A. Catalog No. XQ9017

The International Oragnization for Standardization, “ISO5167, Measurement of Fluid Flow by Means of OrificePlates, Nozzles and Venturi Tubes Inserted in CircularCross-Section Conduits Running Full”, first Edtion, 1980-02-01, Ref. No. ISO 5167-1980 (E)

Scott L.J., Brennan J. A., Blakeslee, NIST, U.S.Department of Commerce, National Institute ofStandards and Technology, “NIST DataBase 45 GRI/NISTOrifice Meter Discharge Ceofficient”, Version 1.0 NISTStandard Reference Data Program, Gaithersberg, MD(1994).

Karnik, U., “A compact Orifice Meter/Flow ConditionerPackage”, 3rd International Symposium of Fluid FlowMeasurement, San Antonio Tx., March, 1995

Morrow, T.B., “Orifice Meter Installation effects in the GRIMRF”, 3rd International Symposium of Fluid FlowMeasurement, San Antonio Tx., March, 1995

Morrow T. B., “Orifice Meter Installation Effects: ResearchUpdate for A’ = 29 D Meter Tubes”, AGA OperationsMeetings, Nashville, TN., May, 1997.

Morrow T. B., Metering Research Facility Program, “Orifice Meter Installations Effects, Development of a FlowConditioner Performance Test”, GRI-97/0207, Dec. 1997.

Gallagher J.E., LaNasa P.J., Beaty R.E., “The GallagherFlow Conditioner”, 1994 North Sea Flow MeasurementWorkshop.


B.D. Sawchuk

McBrien R.K., “Performance of a Single and Dual RotorTurbine Meter in Short and Close Coupled Installations”,AGA Operating Section Proceedings, Montreal, 1996,p. 586.

Park J.T., “Reynolds Number and Installation Effects onTurbine Meters”, Fluid Flow Measurement 3rd

International Symposium, Mar, 1995.

Micklos J.P., “Fundamentals of Gas Turbine Meters”,American School of Gas Measurement Technology 1997Proceedings p. 35.

Dijstelbergen H.H., Bergervoet J.T.M., “OptimalStaightening Vanes for Turbine Meters”, Fluid FlowMeasurement 3rd International Symposium, Mar, 1995.

The International Oragnization for Standardization, “ISO9951, Measurement of Gas Flow in Closed Conduits -Turbine Metersl”, first Edtion, 1993.

Stuart J.S., “New A.G.A. Report No. 9, Measurement ofGas by Multipath Ultrasonic Gas Meters”, 1997Operating Section Proceedings, Nashville, TN., May,1997.

Karnik U., Studzinski W., Geerligs J., Rogi M., “Effect ofFlow Conditioning and Pulsation on the 8” MultipathUltrasonic Meter”, International Pipeline Conference,June 1998 Calgary (to be published)

Karnik U., Studzinski W., Geerligs J., Rogi M.,“Performance Evaluation of 8 Inch Mutipath UltrasonicMeters”, A.G.A. operating Section Operations Confernce,May, 1997, Nashville TN.


VERIFYING GAS CHROMATOGRAPH ATCUSTODY TRANSFER LOCATIONS

Mark F. MaxwellInstromet Inc.

12650 Directors Drive, Suite 100, Stafford TX 77477

INTRODUCTION

On-line Chromatographs are commonly used todetermine the individual components of a natural gasstream. The individual components are then used tocalculate Btu and Specific Gravity. The chromatographicdata is combined with flow rate data to calculate a totalenergy value. The energy value is then used for custodytransfer. Accurate chromatograph data is a critical partof the gas measurement process. Understanding theoperation and principles of the chromatograph allowssystem optimization for individual field conditions.Planning is an essential element of a successfulchromatograph installation and continued operationalaccuracy.

PRINCIPALS OF OPERATION

An On-line gas chromatograph typically contains severalkey elements to accomplish separation of the individualgas components so that energy (Btu), relative densityand the compressibility of the gas may be calculated.The heated oven provides a stable climate controlledcondition for the separation to occur in a reproduciblemanner. Only a slight variation in oven temperature willgive non-reproducible results. In a non-temperaturecontrolled environment the gas components will elutefaster at higher temperatures and slower at lowertemperatures. The columns are contained within theheated oven area so that their temperature may beaccurately maintained to optimize repeatable separationof the components of the gas mixture. The columns arewhere the separation of the gas mixture occurs. Theheavier components move slower through the columnsthan the lighter components. A gas-sampling valve isalso located in the heated oven compartment. The naturalgas is injected into the columns by this valve. The gas-sampling valve is driven by a solenoid that is activatedat a specific time during the analysis run based on thecomputer program that operates the chromatograph. Theactivation fluid is typically helium. Helium is also usedas a carrier gas that flushes the gas mixture beinganalyzed through the columns towards the TCD (thermalconductivity detector). The TCD is also located in theheated oven compartment. The gas mixture is firstinjected into the columns by the gas-sampling valve,separated into individual components in the columns andthen detected by the TCD. The detector response to theunknown sample of natural gas is compared to a certifiedgas standard using ratios and then the gas quality of thesample being analyzed is calculated. Verification of thechromatographs operational conditions must be done

on a routine basis to insure accuracy of the calculatedresults. The sample system external to thechromatograph must also be monitored to insure arepresentative sample of the natural gas is directed tothe chromatograph for analysis.

SAMPLE CONDITIONING SYSTEM

The sample system is of primary importance in obtainingthe most accurate analysis. The objective is to obtain arepresentative sample of gas and maintain the integrityof that sample until it reaches the analyzer. The naturalgas found in most pipelines is relatively stable becauseof the individual hydrocarbon component mix. If the gascontains higher concentrations of heavy hydrocarbons(pentanes and above) special care should be taken. Toobtain a representative sample of natural gas, a sample-probe that extends down into the center of the gas streamshould be used. Do not use any device that will allowcontaminants near the pipeline wall to enter the sampleline. Even in the best pipelines there will be particulatematter and/or liquids moving along the pipeline walls.The sample tap should be located in an area of the pipethat is least likely to generate aerosols and is convenientlylocated for maintenance and adjustment. Select theshortest route for the sample line from the pipeline tothe chromatograph. If the chromatograph is located veryclose to the sample tap this is easy to accomplish. Insome cases other conditions dictate the equipmentlocation. Location of the analyzer and sample tap mayrequire additional planning on the part of the user in multi-stream systems. A central location between the sampletaps is recommended. If one of the sample points is moreimportant than the others the user may wish to optimizethe equipment location for that point. Speed loops maybe considered if the sample points are located somedistance from the equipment. Most gas people arefamiliar with the term “line pack”. In pipeline operationsit is usually desirable to maximize the volume of gas inthe pipeline. The exact opposite is true in the line fromthe sample tap to the analyzer. Keep the line pressureas low as possible to obtain a real time sample. Usesmall diameter stainless steel tubing (1/8 inch is best).Insulate or heat trace sample lines if required. Traceamounts (low ppm) of hydrogen sulfide, mercaptans, andwater will not damage most systems. Eliminate thesecomponents if possible by placement of the sample tapup stream of odorant injection or other sources ofcontaminants. Most sample systems incorporate liquidfilters and particulate filters to protect the chromatograph.These filters should be checked on a routine basis toinsure contamination free operation. The design of most


chromatographic sample systems requires the sampleto be at or close to, atmospheric pressure when injectedinto the chromatograph. When taking a sample at highpressure, considerable pressure reduction is required. Acertain amount of cooling is associated with this pressurereduction (Joule-Thompson effect). If the sample gascomposition contains higher concentrations of theheavier hydrocarbons the cooling may cause thesehydrocarbons to condense and form liquid droplets.These droplets will remain in the regulator or sample line.When this occurs, the sample injected into thechromatograph is not representative of the gas in thepipeline. To prevent this problem the pressure reducerwill require some form of heat to keep the gas in a singlephase. A commercially available probe is recommended.A good selection is the combination probe and regulatorspecifically designed for sampling a high-pressure gasstream. This combination device has a probe thatextends into the center of the gas stream and anadjustable regulator to reduce the outlet pressure in theportion of the probe that extends into the gas stream.The flowing gas warms the pressure reducer andprevents liquid formation without external heat. Severalvendors offer a probe of this type. The user should beaware that a small amount of liquid in the sample systemmight cause numerous problems. These problems canrange from bad analysis to physical damage to thesample system or analyzer.

CALIBRATION AND CARRIER GAS

It is important to use a high quality calibration gas. Awell-tuned gas chromatograph is only as accurate as itscalibration gas. Remember the chromatograph is not adirect type of measurement but depends on its calibrationgas for reference to accurate results. The calibration gasshould be blended with pure components and becertified as a primary standard traceable to NIST(National Institute of Standards and Technology). Thecalibration gas will be supplied in a cylinder. This cylindershould be located as close to the sample system aspossible. Connections and pressure regulation shouldbe treated the same as any other sample. The calibrationgas should be protected from exposure to temperaturesthat will cause the heavy ends to fall out. Consult yourcalibration gas vendor for specifics. The carrier gasshould be helium with a 99.995% purity level. Thechromatograph will not operate properly unless heliumof this purity is used.

VISUAL/WRITTEN CHECKS

The on-line chromatographs in use by the natural gasindustry are very reliable. Most of the systems aredesigned to execute an automatic calibration every 24hours. This calibration not only insures system accuracybut also provides a type of system diagnostic. If thereare any problems the system should output some typeof alarm condition. A daily calibration using this auto calfeature is highly recommended. The user should alsodevelop a check procedure that can be followed based

on internal company requirements or gas contractrequirements. Depending on the location of theequipment, and the time between visits to the equipmentsite, the user should develop a visual and written checkof the system. The following are some of the moreimportant items to observe, however, it is not to beconsidered as the only equipment to be checked. A visualinspection of the equipment including signs of any leaks,excessive dirt (especially in filters), water or oil damage,or crimped or damaged tubing. Observe and record thecarrier gas pressure on the bottle and output setting fromthe regulator mounted on the helium cylinder. This willgive the operator a history of helium consumption.Increased consumption of helium could indicate a leakin the system and may require a more in depth check forleakage. Observe and record the calibration gas pressureat the primary gauge and also the output pressure to thechromatograph. Developing a history will help theoperator diagnose leaks, which could lead to inaccurateresults and or calibration gas contamination. Observeand record each sample streams pressure and flow ratethrough the chromatograph. Each sample stream andcalibration gas stream should be maintained at aconstant and equal flow rate through the chromatograph.Utilities in the computer program or manual switchesare available to set and maintain equal flow rates for eachstream. These simple checks will spot a majority of theproblems that can arise in a chromatograph system. Iwould suggest performing this routine check at leastonce per month. Developing a “history” on a system isone of the most important tools to have when a realproblem develops. The flow rate of the sample vent andmeasurement vent should be checked also. Insure thatthe tubing ends are protected from water, ice and bugs.Plugged vent lines can give erroneous results andemulate other mechanical problems in thechromatograph hardware.

MANUAL/MECHANICAL CHECKS

Each vendor supplies diagnostics for checking thesystem. The diagnostics are dependent on the overallsystem design and therefore will vary greatly betweendifferent manufacturers. The user should becomeacquainted with the diagnostics for his machine andunderstand the limitations. No set of diagnostics will beable to identify every problem. This is especially true ofprocess chromatographs. Barring a major componentfailure many problems are associated with incorrectanalysis or they are intermittent in nature. Resolving thistype of problem requires a basic understanding ofchromatography and the system. Most of the effort inresolving a system problem is isolating the problem area.The complexity of modern electronic circuits preventsall but the most experienced user from trouble shootinga system to the component level. Many circuits requirethe use of a special computerized tester to diagnosecomponent failures. If the user can isolate the problemto the module level most vendors recommend replacingthe module. The defective module then can be returnedto the vendor for repair. This technique also minimizes


the system down time. The stream select and sampleinject solenoids should be inspected for leakage. Manualoperation of the solenoids using program utilities orswitches and the use of a leak detector will insure thatthere are no leaks present. Each sample stream and thecalibration gas should be set to a flow rate equal to allother streams. This will ensure that the exact volume ofsample gas is injected into the columns for analysis.Helium may be run through the sample stream lines tocheck for contamination or sample carryover from onestream to the next. When helium is connected andanalyzed as a gas sample, there should be no peaksfound and the baseline should be relatively flat with theexception of any valve firing upsets. The bridge circuitshould be checked to make sure it is balanced to thecorrect readings that the manufacturer suggests.Incorrect or drifting detector voltage may adversely affectthe analysis results. If the detector voltage is adjusted,then it is recommended the chromatograph be re-calibrated. The most important tool in trouble shootingis the chromatogram. An experienced user can look at achromatogram and ascertain information regarding peakidentification, valve timing, leaks, heater problems, bridgebalance and proper integration of the peaks. Savechromatograms when the unit is functioning properly tocompare to chromatograms when trouble is suspected.Valve timing is important to understand and a large partof obtaining accurate results. Valves are used to decreasethe systems analysis time. This is common to mostmanufacturers. The term back flush to measure is awidely used function. Back flush to measure involveselution of the light components into the main columnwhile not allowing the hexane plus peak to enter the maincolumn. Instead the hexane plus peak is reversed by avalve change and sent directly to the detector allowingfor a much faster analysis and allowing smallconcentrations of heavy components to be combinedas one peak for better detection. If the valve timing is offjust a little then, the hexane plus peak and the normalpentane peak could be affected.

AUDIT SAMPLE

An audit sample or two should be used to verify accuracyof the chromatograph. The calibration gas is used to

calibrate the chromatograph, however problems withchromatograph valve timing or incorrect certified valuesof the calibration gas will not be found unless a secondcertified standard is used to verify accuracy. Rememberthe chromatograph only knows what the operator tellsit. Utilizing a secondary certified sample as a test gaswill allow the operator to confirm the accuracy of thechromatograph as well as the operation of thechromatograph. A third certified sample might be usedto cover the entire range of the chromatographinstallation if the chromatograph is a multi stream unitanalyzing very different gas qualities (plant inlets andoutlets). Repeatability may be determined be runningseveral consecutive analysis of the same gas. Most on-line chromatographs are repeatable to within + or - .5Btu and .25 is not uncommon. When running an auditsample or comparing the chromatograph results to acertified standard report, the operator must make surethat the data being compared is at the same pressurebase and calculated at an ideal or real value. Calculatedvalues such as Btu and relative density must becompared using the same conditions. When determiningsystem balances it is imperative that the on-linechromatographs and any computer database used forcalculating data from continuous or spot samples usethe same Btu and relative density factors. The entire gasquality system should employ the same factors.

SUMMARY

Successful installation and operation of achromatographic system requires the user to consider anumber of variables. The manufacturer should beconsulted if questions arise on any aspect of the systeminstallation or operation. Proper planning will not onlyease the tasks of installation and training but will providebetter long term operation and results. Althoughmaintenance is not a major problem, user personnel mustbe trained and some simple check routines developed.When a system problem develops try to isolate theproblem to a replaceable module. Rely on the vendor torepair defective modules. Accurate results and reliableoperation are easy to verify using any number of processchromatograph systems presently being marketed to thenatural gas industry.

Mark E. Maxwell


ADVANCES IN NATURAL GAS SAMPLING TECHNOLOGYPaula Lanoux

A+ Corporation, LLC41041 Black Bayou Road, Gonzales, LA 70737

INTRODUCTION

The analysis of natural gas plays an important role indetermining its monetary value. Natural gas is bought andsold based on its energy content and volume. The energycontent or heating value is computed directly from theanalysis. Physical constants of the gas, which arenecessary to accurately determine its volume, are alsocomputed from the analysis. Therefore the correctassessment of the monetary value of natural gas isdependant to a large extent on overall analytical accuracy.

The largest source of analytical error in natural gas isdistortion of the composition during sampling. Samplingclean, dry natural gas, which is well above itsHydrocarbon Dew Point (HCDP) temperature is arelatively simple task. However, sampling natural gas thatis at or near its HCDP temperature is challenging. Forthese reasons, much attention has been focused onproper methods for sampling natural gas which have ahigh HCDP temperature.

This presentation will address several problemsassociated with sampling natural gas which is at or nearits HCDP temperature. Industry standards or guidelinesfor sampling natural gas will also be discussed.

DEFINING THE PROBLEMS

The following is a list of some major problems associatedwith proper sampling of natural gas:a. Lack of consensus in the industry as to what

constitutes a “representative” sample.b. Lack of distinction between the sources of liquid,

which may be present in some natural gas samples(i.e.-liquid entrained in the source gas versus liquidresulting from condensation in the sample system).

c. Many key personnel involved in sampling of naturalgas lack understanding of “Phase Behavior”.

LACK OF INDUSTRY CONSENSUS

The key difference of opinion between producers andtransporters of natural gas lies in the treatment of liquidentrained in some natural gas streams.

Producers typically would like for natural gas samplesto represent in some way the presence of the entrainedliquid. They contend that to exclude the liquid is notproper since it was present when the gas volume wasmeasured and represents a source of high energy (andmonetary) value.

The transporters typically contend that the entrainedliquid presents operational problems that result inincreased expenses instead of increased profits.

Industry standards or guidelines do not totally agree onproper methods of sampling natural gas containingentrained liquid. They are not to be blamed because it isa very complex issue requiring substantial research. Tothe author’s knowledge, sampling gas having entrainedliquid is not even in the scope of the three “IndustryStandards” quoted below.

The following is a direct quotation from Appendix B.3 ofthe American Petroleum Institute (API) Manual ofPetroleum Measurement Standards, Chapter 14-NaturalGas Fluids Measurement, Section 1-Collecting andHandling of Natural Gas Samples for Custody Transfer,Fifth edition, June 2001, which addresses multi-phaseflow of natural gas.

“Sampling of multiphase flow is outside the scope of thisstandard. Sampling of multiphase (gas and liquid)mixtures is not recommended and should be avoided ifat all possible. In the multiphase flow, the ideal systemwould mix the gas and liquid flows uniformly and collecta sample of the true mixture flowing in the line by using aproperly designed sample probe and an isokineticsampling system. Current technology of natural gassampling is not sufficiently advanced to accomplish thiswith reasonable accuracy. When sampling a multiphaseliquid-gas flow, the recommended procedure is toeliminate the liquid from the sample. The liquid productthat flows through the line should be determined byanother method. The liquid fraction of the multiphaseflow may contain water and hydrocarbons. Thehydrocarbons can contribute significantly to the energy(measured in British thermal units) content of the gasand their presence in the gas line must not beoverlooked.”

The following is a quotation from the American Societyfor Testing and Materials (ASTM) Designation: D 5503-94 titled Standard Practice for Natural Gas Sample-Handling and Conditioning Systems for PipelineInstrumentation, Section 1.2.

“This practice is intended for single phase mixtures thatvary in composition. A representative sample cannot beobtained from a two phase stream.”

Section 6.4 of the Gas Processors Association (GPA)standard 2166-86, titled Obtaining Natural Gas Samples


for Analysis by Gas Chromatography, has the followingrecommendation for handling entrained liquids.

“Sample Line Separator- When entrained liquids arepresent at the sample point, a liquid separator must beinstalled between the source and the sample container.”

The GPA standard is currently being rewritten. Therevised standard is expected to be published sometimein 2003. Some changes relating to the sampling of naturalgas at its HCDP or containing entrained liquid isanticipated. The author also anticipates changes and/orclarification of that section of the standard.

The lack of consensus by producers and transporters isevident. The reason that they differ in their opinion shouldalso be evident. The reason why the various industry“standards” organizations differ in their approach tosampling natural gas containing entrained liquid may notbe as evident.

The scopes of the standards can vary considerably fromeach other. Additionally, the standards are usuallyinvolved in a continuous evolutionary process. Thenatural gas industry’s needs are also in a constant stateof fluctuation making it difficult for the standards to keepcurrent. At any given time, the current publication datesfor the standards may differ by several years from eachother. For example, the current GPA standard 2166-86was last revised in 1986 and the current API 14.1standard was revised in 2001. The ASTM DesignationD5503-94 was revised in 1994.

LACK OF DISTINCTION BETWEEN THE TWOSOURCES OF LIQUID

The origin of liquid found in any natural gas samplesystem, sample container, or analyzer may be either theresult of liquid which was entrained in the source gas orcondensation that occurred during the sampling process.Before the liquid is eliminated, precautions should betaken to correctly identify its origin. Otherwise seriousdistortion of the sample’s composition could result if theliquid is eliminated incorrectly. In order to fully appreciatewhy this is so, the relationship between a gas and liquidmixture which are in intimate contact must be reviewed.

VAPOR/LIQUID EQUILIBRIUM

When a gas and a liquid consisting of a single componentare in intimate contact at a given pressure andtemperature for a period of time, equilibrium is attained(See figure 1). When that occurs, the gas phase issaturated with the liquid vapor. The gas phase is also atits dew point temperature. An increase in temperatureor a decrease in pressure will result in additional liquidvaporizing thereby increasing its concentration in the gasphase. If the temperature is decreased or the pressureis increased, some of the vapor will condense and returnto the liquid phase causing a decrease in itsconcentration in the gas phase.

FIGURE 1 Gas and a liquid equilibrium

A second example is when a gas and a liquid consistingof multiple components are in intimate contact andequilibrated at a given pressure and temperature. Eachcomponent of the liquid mixture will vaporize to somedegree into the gas phase (See figure 2). The amount ofvapor contributed by each liquid at that conditiondepends upon its concentration in the liquid phase andits volatility.

FIGURE 2 Gases and a mixture of liquids in equilibrium

If the gas and liquid mixture temperature is increasedand/or its pressure is decreased, some of the liquidmixture will vaporize thereby increasing the concentrationof each liquid component in the gas phase (See figure#3). Since the volatility of each component varies, theirconcentrations in the gas phase will not increase in directproportion to their concentration in the liquid phase. Thenet result is that the vapor concentrations of all liquidcomponents will increase in the gas phase. The morevolatile components will increase in greater proportionthen the less volatile components. Therefore, not onlywill there be an increase of all liquid componentconcentrations in the gas phase, but their ratios will alsoshift.


FIGURE 3 Effect of Decreasing Pressure orIncreasing Temperature

If the gas and liquid temperature is decreased and/orthe pressure is increased, the opposite occurs. Some ofthe liquid vapor components of the gas phase willcondense thereby decreasing their concentration in thegas phase (See figure 4). Components will not condensein equal proportions; therefore their ratio in the gas phasewill also shift.

Liquid removed downstream of pressure regulationshould only be for protection of an analyzer since thegas phase composition will not be the same as that ofthe source gas.

The gas phase composition will also be distorted ifhardware is utilized to separate liquids from natural gassample at a temperature other than that of the pipeline.

In situations where the source gas does not containentrained liquid, the traditional methods of preventingcondensation are applicable. These methods arepressure reduction to lower the HCDP and/or heat tracingto maintain sample conditioning hardware above thesample’s HCDP.

LACK OF UNDERSTANDING OF “PHASE BEHAVIOR”

From the preceding discussion, it can be seen that havinga working knowledge of “Phase Behavior” is a must forthose involved in the sampling of natural gas. This isparticularly important when sampling sources which haveeither entrained hydrocarbon liquids, or when the sourcegas is near its HCDP temperature, or takes place in lowambient temperature environments.It can also be seen from the preceding discussion thatsome methods and/or hardware designed for liquid freegas sources will likely distort the sample compositionwhen liquid is present in the source gas.

In many cases it is difficult to determine if liquid isentrained in the pipeline gas. Therefore, one is advisedto assume that liquid may be present in designing oroperating sample systems where the gas is at or near itsHCDP temperature.

It should be noted that traditional open bore sampleprobes are not suited for eliminating suspended liquidaerosol droplets from entering the gas sample system.

Probe regulators that do not exclude liquid beforeregulating the gas pressure can distort the gascomposition when liquid is entrained in the source gas.Sintered metal or glass fiber filters are typically utilizedto coalesce liquid droplets as the gas flows through theelement. What is required however is rejection, notcoalescing, of liquid from the gas at the upstream elementsurface before gas flows through the element. To theauthor’s knowledge, only a phase separation membranecan reject liquid at its upstream surface. Hardwareemploying phase separating membrane is commerciallyavailable to properly precondition gas entering thesample system from a source gas containing liquids.

FIGURE 4 Effect of Increasing Pressure orDecreasing Temperature

It can therefore be stated that when liquid is present in anatural gas sample, a change in either the temperatureand/or the pressure will result in a change of the gasphase composition.

For this reason, it is imperative that liquid is removed atthe prevailing pressure and temperature condition of thesource gas. This is best accomplished inside of thepipeline before the gas enters the sample system.Sampling in that manner will provide a samplerepresenting the gas phase composition of the sourceat its prevailing pressure and temperature.


Paula Lanoux

CONCLUSION

Proper Sampling of natural gas is an important elementin determining its monetary value. Significant analysiserrors can occur when hydrocarbon liquids are entrainedin the source, or when the source gas is at or near itsHCDP temperature, or when sampling occurs at lowambient temperatures.

The traditional uses of heat tracing and/or gas pressurereduction are appropriate means of preventingcondensation. However, applying these traditionalsolutions for preventing condensation when hydrocarbonliquid is present in the source gas will distort the gasphase.

It is very important to determine the source of liquid tobe eliminated before implementing a means foreliminating it. Liquid removed after either the pressureor temperature is changed will alter the gas phasecomposition. When the source gas contains entrainedhydrocarbon liquid, an ideal solution is to utilize hardwaredesigned to remove it inside of the pipeline beforepressure or temperature changes occur.


HYDROGEN SULFIDE MEASUREMENT AND DETECTIONPatrick J. Moore and Rodney W. Spitler

Thermo Electron Corporation9303 W. Sam Houston Parkway S., Houston, Texas 77099

INTRODUCTION

The impetus for measuring and detecting hydrogensulfide, H2S, as it relates to the production anddistribution of natural gas, is rooted in two primaryconcerns. The first concern deals with protectingpersonnel from the lethal effects of H2S. Typically, themaximum pipeline H2S concentration is around 0.25grains per 100 SCF, nominally 4 ppm/volume. At theseconcentrations H2S is not lethal and its presence can bedetected by the sense of smell with its characteristicrotten egg odor. At the higher lethal H2S concentrations,typically found at production and acid gas removalinstallations, the nose becomes desensitized. Unable tosmell the H2S, a worker breathing such an atmosphereis oblivious to the life threatening danger.

The second concern deals with preventing hydrogenembrittlement of the natural gas transmission lines. Whenhydrogen sulfide reacts with the metal in the transmissionline to form a metallic sulfide, the released hydrogen isthen free to migrate within the molecular structure of thepipe. Transmission lines weakened by embrittlement aresusceptible to rupture failure allowing for large clouds ofgasses to escape and accumulate in the atmosphere.Not only are such release clouds lethal by the depravationof life sustaining oxygen, a single spark can set off adevastating explosion.

Various technologies are available for measuring anddetecting H2S. These technologies include on-linecontinuous analyzers, area monitors, test methods, andpersonal monitors. Proper use of these technologies cancontribute to the safe delivery of natural gas from thewell head to the consumer.

H2S

Though H2S is a flammable gas, the flammable limits of4.3% (43,000 ppm) to 46%, (46,0000 ppm) far exceedthe concentrations of concern for personnel protection,nominally 10 ppm, and pipeline transmission, 4 ppm.Because H2S is heavier than air, it will tend to accumulatenear the ground when leaked into the atmosphere. Astanding individual overcome by H2S will most likelycollapse to where the H2S concentration is even greater.The H2S toxicity danger is a function of the concentrationand the time of exposure. Concentrations on the orderof 500 ppm can result in rapid collapse, unconsciousnessand death. Prolonged exposure to lower concentrationscan also lead to hemorrhaging and death. At low

concentration H2S has a characteristic rotten egg odorthough, with sufficient exposure time, low H2Sconcentrations can also deaden the sense of smell. HighH2S concentrations rapidly deaden the sense of smell.Disturbed respiration, throat and eye irritation, sleepiness,headache, and pain in the eyes are all symptoms ofhydrogen sulfide exposure.

There is plenty of available information regarding H2Ssafety. Some of the web sites containing information onH2S are listed in Table 1 – H2S Web References.

Table 1 – H2S Web References

Agency Web AddressUSEPA www.epa.gov/swercepp/ehs/profile/7783064.txt

www.atsdr.cdc.gov/tfacts114.htmlOSHA www.osha-s lc .gov/SLTC/etools/o i landgas/

general_safety/h2s _monitoring.htmlNIOSH www.cdc.gov/niosh/npg/npgd0337.html

H2S DETECTION FOR PERSONNEL PROTECTION

Personnel protection devices provide information to aworker regarding a contaminating componentconcentration in the air so that appropriate actions canbe taken in the event an undesirable contaminantconcentration is detected. Such actions include, but arenot necessarily limited to, evacuation of the area, usinga self-contained breathing apparatus, turning onemergency ventilation system, and eliminating the sourceof the leak. H2S monitors for personnel protection canbe carried by the individual or fixed mounted for areamonitoring. H2S monitors typically use colorimetric orelectrical sensors.

An example of a colorimetric H2S sensor is an encasedroll of lead acetate impregnated paper tape, an exposurewindow, and a color chart. When moistened, the leadacetate impregnated paper tape will change color fromwhite to brown when exposed to H2S. Tape moisturizationis achieved by exhaling on the exposure window. Theamount of H2S present dictates how fast the colorchanges. The rate of tape color change and the staindarkness is directly proportional to the H2Sconcentration. The advantage of the colorimetric H2Ssensor is that electrical power is not required foroperation.


Electrical H2S sensor technologies include metal oxidesensors, sometimes called solid ceramic-metallic(cermet) film devices, and electrochemical cells. Eachof these devices depends on the migration of H2S fromits source to the sensor and can be incorporated into afixed-point detection system or carried by the individual.A personal protection monitor, sometimes referred to asa multimeter, will typically contain a flammable gasdetector and an oxygen deficiency sensor in addition toa device to detect H2S. The migrated H2S reacts withthe surface of the metal oxide sensor, or the reagentchemical in the electrochemical cell, to create anelectronic signal. In order for the generated signal to haveany meaning, the response of the device must becalibrated with a gas containing a known concentrationof H2S. The surface of the metal oxide sensor regeneratesitself in the presence of air. Electrochemical cells have afixed quantity of reagent that must be replaced whenconsumed.

Fixed mounted electrical H2S sensors, powered by abattery or a fixed electrical source, sound an alarm whenthe H2S concentration reaches a predetermined level.By using relays, the electrical H2S sensors can alsoactivate emergency control apparatus, such as,ventilation and alarm systems, when an alarm conditionis detected.

H2S MEASUREMENT IN GASES

Various organizations, such as ASTM, UOP, and GPA, toname three, publish test methods for measuring not onlyhydrogen sulfide but also a wide range of componentsand properties. Published test methods provide a wayto standardize test procedures so that the results fromone location are comparable to the results from anotherlocation. Some of the available test methods formeasuring and detecting hydrogen sulfide are listed inTables 2-4.

Table 2 – ASTM Methods for Measurement of H2S

ASTMMethods Description

D2420-91 Standard Test Method for Hydrogen Sulfide in(1996)e1 Liquefied Petroleum (LP) Gases (Lead Acetate Method)

D4084-94 Standard Test Method for Analysis of Hydrogen Sulfide1999 in Gaseous Fuels (Lead Acetate Reaction Rate Method)

D4323-84 Standard Test Method for Hydrogen Sulfide in the(1997)e1 Atmosphere by Rate of Change of Reflectance

D4913-00 Standard Practice for Determining Concentration ofHydrogen Sulfide by Direct Reading Length of Stain,Visual Chemical Detectors

D4952-02 Standard Test Method for Qualitative Analysis for ActiveSulfur Species in Fuels and Solvents (Doctor Test)

D5504-01 Standard Test Method for Determination of SulfurCompounds in Natural Gas and Gaseous Fuels by GasChromatography and Chemiluminescence

D6228-98 Standard Test Method for Determination of SulfurCompounds in Natural Gas and Gaseous Fuels by GasChromatography and Flame Photometric Detection

Table 3 – GPA Methods for Measurement of H2S

GPA Methods Description

Standard 2285 GPA Standard for Determination of Hydrogen Sulfideand Mercaptan Sulfur in Natural Gas (CadmiumSulfate-Iodometric Titration Method)

Standard 2377 Test for Hydrogen Sulfide and Carbon Dioxide inNatural Gas Using Length of Stain Tubes

Table 4 – UOP Methods for Measurement of H2S

UOP Methods Description

9-85 Hydrogen Sulfide in Gases by the Tutwiler Method

41-74 Doctor Test for Petroleum Distillates

Selecting a test method to detect and measure hydrogensulfide will depend on the desired degree of precision.Some test methods are designed to provide approximateresults while others are design to provide a greater degreeof precision.

An example of an approximate method is the Doctor Test.The Doctor Test involves passing a H2S containing gasacross a filter paper wetted with an aqueous lead oxidesolution and observing a color change. The lowerdetectable limit is about 0.03 grains per 100 SCF, or about0.5 ppm/v. A dark brown color is indicative ofconcentrations above 0.5 grains per 100 SCF, or about8 ppm/v (Pender, 1986).

Another approximate method involves using glass staintubes filled with lead acetate. As a known volume of gasis drawn through the tube, the lead acetate changes fromwhite to brown. The amount of color change dependson the amount of H2S in the gas. The greater the H2Sconcentration the greater the color change. Tubes forvarious concentration ranges are available.

Some methods to detect and measure H2S are basedon wet chemical procedures. An example of a wetchemical method is GPA Standard 2285 which isapplicable for determining the H2S content of natural gas.In this method an extracted sample is passed through acadmium sulfate (CdSO4) solution. The absorbed H2Sreacts to form cadmium sulfide (CdS) which is thenmeasured iodometrically. Another example of a wetchemical method is the portable titrator that uses anelectric current passing through a reagent electrolyticsolution, such as bromine, while the H2S containing gasis bubbled through the solution. The net generatingcurrent needed to maintain a slight excess of reagentsolution is directly proportional to the concentration ofthe reactant (H2S).

Some of the methods listed in Tables 2-4 are availableas on-line instrumentation. On-Line instruments offer theadvantage of providing analytical information twenty-fourhours a day seven days a week. Included in this categoryare gas chromatographs and lead acetate analyzers.


Each is an extractive type analyzer in that the sample isextracted from the process and is transported to theanalyzer for analysis.

Gas chromatography involves injecting a known volumeof sample into a carrier stream. The sample-containingcarrier then passes through a column where thecomponents of the sample are separated from eachother. Porous polymers, such as Poropack, are used asthe column material when analyzing for sulfurcompounds. Typically, the column is operated underconstant temperature and pressure conditions as thissimplifies the analyzer design. A fairly simple design ispreferred due to the potential for remote installations.As each component exits the column, it passes acrossa detector that, generates a Gaussian (peak) shapedsignal. The identity of a given component is determinedbased on the time it takes to pass through the column.The concentration of the component is determined byintegrating the area under the peak and comparing thisarea to the area obtained using a gas with a knownconcentration of the specific constituent. A distinctadvantage of the gas chromatograph is its ability toprovide a complete compositional analysis in additionto the determination of H2S concentration. The majordisadvantages of gas chromatographs include theirrelatively high purchase price and operational complexity.The major components of a simple gas chromatographysystem are shown in Figure 1.

FIGURE 1. Basic Gas Chromatograph

As the name implies, hydrogen sulfide analyzers usingthe lead acetate detector are based on the reaction ofhydrogen sulfide with lead acetate. The initially white leadacetate turns brown when exposed to hydrogen sulfidein the presence of water. The greater the amount of H2Sin the sample, the darker the stain that forms on the tape.Lead acetate is impregnated onto a roll of paper tapethat is installed in the analyzer. Only a small portion ofthe roll is exposed to the sample during an analysis cycle.This is accomplished by using a window in a samplechamber. Before exposure to the tape, the sample gaspasses through a humidification bubbler containing a

5% by volume acetic acid in water solution. Being acidic,the bubbler solution prevents the loss of H2S during thehumidification process. After passing past the exposedportion of the tape, the sample gas is vented from theanalyzer. At the start of each analysis cycle a fresh pieceof the tape is pulled in front of the window.

To be useful as an on-line analyzer, the color changecaused by the reaction of H2S with lead acetate must beconverted into an electrical signal. Earlier analyzerdesigns used a lens to focus a whitish light onto thereaction window. As the initially white surface changescolor it absorbs some of the incident light. Twophotocells, one measuring the other reference, initiallybalanced in a Wheatstone bridge arrangement, detectedthe drop in incident light. The now out of balancedWheatstone bridge creates a voltage signal that is thenprocessed by the electronic portion of the analyzer. Morerecent lead acetate analyzer designs utilize a tailoredincident light frequency, a bifurcated fiber optic cablefor light transmission to and from the reaction window, aphotodiode for detecting the drop in the incident lightintensity, and an analogue to digital converter formicroprocessor signal processing.

Tape transport on the first generation of lead acetateanalyzers was continuous. Because the tape was alwaysmoving, the motor used to transport the tape was subjectto mechanical wear necessitating its periodicreplacement. The H2S concentration was determined bythe difference between the reflected light intensity at thebeginning of the analysis cycle to that at the end of theanalysis cycle. This difference was proportional to theconcentration of H2S in the sample. On subsequentgenerations of the lead acetate analyzer the tape isstationary during the analysis cycle. By only having touse the motor for a short time between analysis cycles,it is subject to very little mechanical wear. Combinedwith employing periodic tape transport, a differentiatingcircuit was also added. Instead of looking at the changein reflected light intensity during an analysis cycle,differentiating the reflected light intensity meant that therate of change of reflected light intensity was beingmeasured. The rate of change is also proportional to theconcentration of H2S in the sample. Both means of H2Smeasurement yielded one result per analysis cycle, whichtypically lasted for 3 minutes.

With the advent of microprocessors it became possibleto slice the analysis cycle into smaller segments. Insteadof having one result per analysis cycle, multipleintermediate values of the rate of change could becalculated and then averaged at the end of the analysiscycle to report a result. Alarm outputs could bedetermined using the intermediate values, which meantthat the analyzer could respond to an upset conditionjust as fast as, and sometimes faster than, the earlierdesigns. By reporting an average of the intermediateresults at the end of the analysis cycle, the total analysistime could be increased. An increased analysis timemeant that the time between tape replacement was

Carrier Gas

Sample InInject Valve

Sample Out

Detector

GC Column

Thermostated Oven


increased from about two weeks to about a month. Figure2 illustrates a modern lead acetate tape analyzer fordetermination of H2S in natural gas.

FIGURE 2. Lead Acetate Tape Analyzer for Determinationof H2S in Natural Gas

Lead acetate analyzers cannot provide a completecompositional analysis as with gas chromatographs.These analyzers are typically configured to provide a H2Sand/or total sulfur analysis. Key advantages of leadacetate tape analyzers include their relatively lowpurchase price and operational simplicity.

CALIBRATION

Instrument calibration is required so that the electronicsignal generated by the analyzer can be interpretedintelligently. Calibration requirements are similar for alltypes of H2S analyzers. Basically, a span gas with aknown H2S concentration is introduced into the analyzerand the response is recorded. Many H2S detectors, suchas, the flame photometric detector used with gaschromatographs and the lead acetate tape can producea non-linear detector counts vs. H2S concentration curve.Therefore, it is often desirable to check the measurementlinearity in addition to calibrating with a span gas. Alinearity check can be performed with a gas containingat least 50% less H2S than the span gas.

The typical units of measure for H2S in gases is partsper million, ppm. For gases ppm is commonly expressedas a volume/volume or mole/mole basis. For gases theunits of ppm (volume/volume) and ppm (mole/mole) areequal. The concentration of H2S in ppm (v/v), that is,ppm by volume is calculated according to the equationshown in figure 3. Note that the calculation of ppm isvery similar to well known percentage calculations, forppm we multiply by 1,000,000, for percent we multiplyby 100. In fact, another name for percent is pph or partsper hundred.

Critical to obtaining a good calibration is the quality ofthe H2S standard gasses. These gasses can be obtainedin three ways: 1) purchase custom blends from a gassupplier, 2) using a permeation tube, and 3) preparing ablend by using a calibration kit.

A purchased calibration blend is typically under pressureallowing the user to have a long term supply of standardin a single cylinder. Low level blends, 1-10 ppm H2S(v/v), do require the cylinder to be stabilized which addsto the delivery time. Such blends most likely carry alimited validity time. Storage conditions for low levelblends also need to be seriously considered. A low levelcylinder exposed to extremes in cold or hot temperaturescan result in reduced or excess amounts of H2S actuallyin the gas phase within the cylinder. Under theseconditions, what comes out of the cylinder may not bethe concentration indicated or desired. The net result isa faulty calibration.

Permeation tubes contain a small amount of the purecompound of interest, for example H2S. The tube issealed at its ends with a membrane that allows the H2Sto permeate across the membrane at a known rate. Thepermeation rate is usually low enough compared to theamount of pure compound in the tube that the tube hasan extended lifetime. An H2S free carrier gas at a knownflow rate is required to utilize a permeation tube forcalibration. The carrier gas mixes with the permeatedH2S resulting in an H2S blend of a known composition.The permeation rate is characterized and certified at afixed temperature, which is typically 30 oC to 40 oC. Thismeans that a temperature controlled oven is required sothat the permeation rate obtained is the same as thepermeation rate expected.

Permeation tubes also require a stabilization time,sometimes as long as 4 to 10 hours after they are installedin a temperature controlled oven. It is important torecognize that a permeation tube permeates all of thetime until the material inside of the tube is exhausted.Even in standby mode a carrier gas flow rate is requiredto pass through the oven when a permeation tube isinstalled in it. Without a carrier gas flow in standby mode,the concentration of H2S accumulates. When the carriergas is finally turned on the H2S concentration in theensuing blend will far exceed what is expected. If thecarrier gas flow is off while the permeation tube is in theoven for an extended period of time, it may be necessaryto replace parts of the oven. These parts will be sosaturated with H2S that they are essentially useless forproducing low level calibration blends. Permeation tubesalso need to be stored at the recommended temperatureso as not to damage or alter the permeationcharacteristics of the tube.

Preparing calibration blends with a calibration kit requiresan H2S free blending gas and a source of either pureH2S or a cylinder blend with a known elevated H2Sconcentration, for example 1000 ppm/volume. A largeplexiglas cylinder with a piston with a typical volume of

FIGURE 3. Calculation of Parts Per Million, ppm

cm3 H2S

(cm3 H2S + cm3 diluent gas)1,000,000 ppm H2S (v/v)

BubblerHumidifier

SampleConditioning Rotometer

SensingTape

ExcessSample

Vent

=


10 liters is used to make the desired blend. The techniqueinvolves introducing a small amount of either pure H2Sor the H2S blend using a syringe into the 10 liter cylinderand diluting this to the correct volume with the blendinggas to obtain the desired H2S concentration. Becausethe blend is at atmospheric pressure, the amount of thestandard is limited and a pump is required to introduceit into the analyzer system. Some skill is required toachieve repeatable blends. Figure 4 depicts a typicalcalibration kit.

FIGURE 4. A Typical Calibration Kit

Regardless of the source of the calibration standard,attention must be paid to the composition of thebackground gas used to make the standard. This isespecially true when the sample flow rate is importantfor the successful operation of the analyzer. For example,a H2S analyzer configured for a methane sample shouldbe calibrated with a H2S in methane calibration standard.Another point to consider when using calibration blendsis where they are introduced to the analyzer. Processsample normally flows through a sampling system priorto being introduced into the analyzer. For purposes ofconsistency a calibration blend should be introduced intothe sampling system at the sample tap, if possible. Inthis way the calibration blend that arrives at the analyzerhas been exposed to the same conditions as the sample.A drifting analyzer response, under these conditions,could suggest the presence of a contaminant in thesampling system as well as an analyzer malfunction.Once a stable analyzer response to a calibration blendis achieved, the unit can then be calibrated.

SAMPLE CONDITIONING SYSTEMS

Independent of the type of extractive on-line analyzeremployed, the process sample must be transported fromthe sample point to the analyzer so as to maintain theintegrity of the sample. Sample integrity is obtained byinstalling a properly designed sample conditioning system.

Process variables taken into account when designing asampling system are the composition, temperature, andpressure at the sample point. One also must know theoutside ambient temperature as well as the temperatureinside of the analyzer shelter, if one is utilized. With thisinformation it is possible to calculate the hydrocarbonand water dew points of the sample as a function of thepressure along the sample delivery system. Dew pointsare important because the presence of liquid and gasphases impedes the ability to accurately monitor the flowrate of sample to the analyzer. When measuring hydrogensulfide, liquids can act as absorption sites thus reducingthe amount of H2S that arrives at the analyzer. Thus, theanalyzer can yield a legitimate result based on what isintroduced to the analyzer when in fact the actualconcentration in the process is higher. Since gaschromatographs use injection valves to introduce sampleto the column, two phase sampling leads to erraticresults.

Natural gas pipelines tend to operate at elevatedpressures, nominally on the order of 1000 psig. Whenthe pressure is reduced from 1000 psig to the typicalanalyzer inlet pressure of 50 psig, the gas cools. Thisphenomenon is known as the Joule-Thompson effect.Depending on the composition of the process and thetotal pressure drop, cold spots in the sampling systemcan cause liquid formation. Since H2S is soluble in liquidwater and liquid hydrocarbon, the amount of H2Sreaching the analyzer is reduced. Again, the analyzer canyield a legitimate result based on what is introduced tothe analyzer when in fact the actual concentration in theprocess is higher. Sample probes that take the pressuredrop inside of the process pipe are available. Theseoperate on the premise that the total heat flow acrossthe outside of the probe is very large compared to theheat flow requirement to reheat the cooled gas at thelower pressure inside the probe. Because probes canextend into the process pipe a third to 1/2 of its diameter,they may not be desirable in some installations. Underthese circumstances, taking the needed pressure dropin stages can mitigate the potential for cold spots and/or condensation.

If there is a chance for any particulate matter coexistingwith the sample, it needs to be removed prior to theanalyzer. Particulate matter can plug injection valves,pressure regulators, and flow control devices. Self-cleaning and cartridge type sample filters are available.Self-cleaning filters have a sweep line that is usuallyreturned to the process. In some natural gas installationsa sweep line is not practical because the needed pressuredrop in the process line is not available or venting to theatmosphere is not desirable. In these cases, two parallelcartridge filters with isolation and depressurizationprovisions can provide the needed filtration andmaintenance access.

Every sample system has a lag time, the time betweenwhen the sample exits the process and when the analyzerobserves it. Factors that impact the lag time for a singleanalyzer system include: 1) the analyzer location relative

Cal Kit

10 Liter

5 Liter

1 Liter

Bal.Gas

H2S or COSfilled syringe


analyzer system include: 1) the analyzer location relativeto the sampling point, 2) the sample line pressure, 3) thesample flow rate to the analyzer, and 4) line bulges suchas filter bodies. A long sample line at an elevated pressuremeans a large sample inventory between the sample tapand the analyzer. Calculating the lag time involvesdetermining the sample system volume. With the samplesystem volume, the total gas volume in the samplesystem can be calculated by correcting for the sampleline temperature and pressure. Dividing the samplesystem gas volume by the sample flow rate to theanalyzer, at the same conditions of temperature andpressure, yields the lag time to the analyzer. The totallag time of the system is the sum of the sample systemlag time and the lag time within the analyzer itself.

MAINTENANCE

Stated simply, routine maintenance on H2S analyzers,as well as any other analyzer, is a must. Without routinemaintenance, problems can develop that cancompromise the integrity of the results. Changes insample and/or carrier flow, as well as sample pressureand temperature can cause calibration drift. Failure ofan injection valve on a gas chromatograph or failure toreplace the tape when it is consumed and/or replenishingthe bubbler solution on a lead acetate H2S analyzer cancause the analyzer to read zero even though there isH2S in the sample. Malfunctioning alarms can preventdesired diversion or shutdown actions from occurring inthe event out of specification gas passes the samplingpoint. Sample system contamination and/or leakage canreduce, and possibly eliminate, the concentration of H2Sthat actually arrives at the analyzer. Plugging or

accumulation in vent lines can increase sample lag timesin addition to altering the response of the detector. Ingeneral, the recommended maintenance practices of themanufacturer should be followed and supplemented bysite and/or company specific requirements and practices.

CONCLUSION

The requirement for measuring and detecting H2S innatural gas is motivated by personal protection concernsand the need to maintain gas quality. Though thechallenges associated with measuring and detecting H2Scan seem daunting, the good news is that there aresuccessful H2S in natural gas analyzer applicationsinstalled in the field. Good maintenance practices, trainedpersonnel, well designed sample systems, propercalibration techniques, understanding the employedmethodology, and being cognizant of the dangers of H2Scombine to provide for a reliable and safe H2Smeasurement.

REFERENCES

Pender, A. B., Measurement Problems Dealing with andDetecting H2S, Gulf Coast Gas Measurement Society,Twenty First Annual School & Exhibit, Proceedings, pp.77-78, 1986.

ADDITIONAL READING

“Hydrogen Sulfide Analyzers”, Analytical Instrumentation,Chapter 19, Sherman, R. E., editor, Instrument Societyof America, Research Triangle Park: 1996, pp. 343-357.

Rodney W. Spitler


INTRODUCTION

Timely, diligent field testing and calibration of gas volumerecording and correcting instruments ensure thatmeasurement information fairly represents actualvolumes.

The instruments save a company capitol and operatingcosts because they can record or integrate volumes atpressures and temperature above the normal pressure-base conditions specified in contracts for volumecalculation. This allows the company to use smaller andfewer meters.

Recording and correcting instruments normally areconnected to positive displacement, rotary and turbinemeters in lieu of a direct reading/compensating index.The compensating instruments include:

• Volume and pressure/temperature recording gauges• Mechanical pressure/temperature volume correctors• Electronic pressure/temperature volume correctors• Electronic flow computer

TEST FREQUENCY

How often should recording and correctinginstrumentation be field inspected and calibrated?Generally, every time the meters are tested.

Company policies for test scheduling usually are basedon contractual obligations and average measuredvolumes. Large volume deliveries dictate monthly,quarterly and semiannual inspections. Devices thathandle smaller quantities might be scheduled annuallyor every four years.

Another scheduling factor is instrument type and age andthe type and age of the meter to which it is connected.Older mechanical devices require more frequent attentionthan electronic units. For this reason, high reliability meters,such as rotary and turbine, should not be paired with high-maintenance mechanical instruments. A 20-year-oldvolume and pressure recording gauge mounted on adiaphragm meter, might be scheduled for quarterlyinspections. Conversely, an electronic corrector,connected to a rotary meter, may require only semi-annualtesting. Whatever the case, when an inspection schedulehas been determined, it should be documented andmonitored to ensure timely and valid testing procedures.

FIELD INSPECTION AND CALIBRATIONOF MEASUREMENT INSTRUMENTS

George E. Brown IIIReliant Energy – Entex

210 W. Larissa, Jacksonville, Texas 75766

BASIC TEST PROCEDURE

Basic instrument proving procedure is simulating allconditions the device was designed for, and testing foraccuracy at several points. Pressure and temperatureresponses should be checked at a minimum, maximum,25%, 50% and 75% of instrument scale.

For indexes and counters, meter input responses shouldbe checked for value, coding and operation. When thedevice is a volume and pressure/temperature recordinggauge, volume cycle value and operation also should betested.

For electronic instrument tests, scaleable unitconfiguration values such as inputs, outputs, ranges,calculation parameter and audit trail definitions shouldbe verified during each inspection.

TEST EQUIPMENT

The most important piece of instrument test equipmentis the inspector’s eyes. Many measurement errors canbe avoided by a thorough visual inspection of aninstrument. Close attention to details will ensure utmostequipment accuracy and will find potential device failuresbefore they occur. The remaining test equipmentrequirements can vary, depending on whether theinstrument is mechanical or electronic. Following is aminimum list of typical items required for testing bothinstrument types.

MECHANICAL TEST DEVICES

• Deadweight tester• +/- 1/4% accuracy test gauge of suitable range• 0.5°F accuracy Yellowback measurement ther-

mometer• Device to simulate meter output

Electronic test units

• Pressure source—usually compressed air ornitrogen—and optional precision pressure regulator

• Electronic pressure gauge, 0.1% accuracy• Digital voltmeter, 41/2 digit• Digital thermometer, 0.1°F accuracy• Device to simulate and count meter output revolutions


CALIBRATING MECHANICAL DEVICES

Process inputs on most mechanical measurementinstruments consist of some form of the standard four-bar linkage (Fig. 1). Inputs include pressure, temperature,volume integration and volume cycle measurements.Basic linkage operation is the same for all types, withdifferences in the amount and direction of linkage travel.

Since mechanical wear is inevitable, linkage adjustmenteventually will be needed to maintain acceptable accuracy.These adjustments commonly called zero, span andlinearity. Zero is minimum scale, span is maximum scaleand linearity is set to ensure output equals input at allpoints between zero and span. Product manufacturersprovide maintenance manuals with diagrams andinstructions for adjusting their equipment (Fig. 2).

FIGURE 1. Mechanical linkage styles

Instruments should also be checked for repeatability sinceexcessive wear on mechanical parts will induce errorscaused by slack or slippage in the mechanism. Once theinputs are calibrated, the inspector must test the indexesor odometer counters for proper operation and, equallyimportant, proper coding of volume increments.

Final inspection is accomplished by applying test valuesto the inputs and simulating meter flow by rotating theinput drive equal to a sufficient test volume. The inspectorthen calculates what the instrument’s output should beand determines accuracy by making comparisons.Percentages of error will appear inflated at lower pressureand lower test volumes. In these circumstances, it is wiserto deal with larger test volumes rather than be confusedby percentages. Again, the manufacturer’s instructionsprovide best recommendations for dealing with thissituation.

The final inspection described above also is usuallyperformed as an initial inspection before calibrations areperformed. Initial inspection results generally are usedfor billing adjustments.

CALIBRATING ELECTRONIC DEVICES

Basic procedures for calibrating process inputs onelectronic instrumentation are similar to that formechanical devices except there is no four-bar linkage.

Instead, calibration changes are made with a voltmeterand adjusting, “tweaking,” zero/span potentiometers eithermanually, or electronically by using software in a hand-held or laptop computer connected to the unit’s serial

H-StyleBottom Driven

Driving Arm Driven Arm

H-StyleTop Driven

N-StyleOutput is opposite to input motion


communication port. Linearity usually is accomplished inthe computer or transmitter’s EPROM. A few transmittershave a potentiometer adjustment for this function.

Meter connection to an electronic volume corrector ismade directly to the hand-hole plate and these deviceshave a mechanical counter to back up the digitaluncorrected counter. Uncorrected counters should becompared to ensure that input pulses to the calculationcircuit are correct. If the y are not close to the same value,a pulse-switch problem exists. This leads to volume underrecording or under correcting.

Flow computers are connected to the meter via a remotepulser, which sometimes is backed up by an index orsandwiched between the meter output and a volume andpressure/temperature recording gauge. In either case,mechanical counters need to be inspected for operationand volume coding to make sure their setup matches thespecified measurement configuration for the site. Initialand final testing procedures are identical to those formechanical instruments. All “found” and “left” conditionsshould be properly documented on the appropriate testforms to validate the completed inspection.

CONCLUSION

Only the most common inspection and calibrationprocedures have been covered in this article becausespecial applications, such as volume controllednomination loops, tube switching and alarms, normallyare tailored to specific measurement requirements.

Acceptable accuracy limits are not mentioned becausecompany policies and equipment tolerances will varysomewhat. Measurement instrument inspectors mustobtain this information from company and manufacturersources. Whatever the instrument, the inspector’sprimary goal is to ensure that recorded volumes representactual volumes as accurately as is humanly possible. Astechnological advances bring new and more sophisticatedproducts to the industry, it is increasingly evident thatmeasurement perfection is a goal yet to be attained.

George E. Brown III


ONSITE PROVING OF GAS TURBINE METERSDaniel J. Rudroff

Flowline Meters Inc.1322 Foxwood, Houston, Texas 77008

INTRODUCTION

With the increased use of Natural Gas as a fuel, highernatural gas prices, and the new federal regulations,buyers and sellers of natural gas are seriously looking atways to improve their natural gas measurement andreduce the unaccounted for natural gas. An error inmeasurement of only one tenth of one percent (0.1%)on 100 MMSCF/D Natural Gas selling at $5.50/MCF willcause an over or under billing of $200,750.00 in one year.This will more than pay for a proving system.

The Btu in one barrel of oil for example is equivalent toapproximately 5,600 cubic feet of natural gas. At $5.50per thousand cubic feet, the natural gas equivalent ofone barrel of oil is $30.80 equal to, or more than the costof a barrel of oil. In the petroleum liquid industry nocustody transfer liquid measurement system would becomplete without a method to prove the meter, either aspart of the equipment or there would be connectionsprovided for a portable prover.

Under billing causes loss of revenue, and over billingcan cause a future correction that will cost the companymillions of dollars. For these reasons gas meter provingis important and necessary to insure precisemeasurement of natural gas that both the buyer and sellercan agree upon.

They have a long history, are well documented and theformulas, although a bit complicated, are in most flowcomputers.

In the ASME standard MFC-7M-1987 Reaffirmed 2001“Measurement of Gas Flow by Means of Critical FlowVenturi Nozzles” it is stated “The Venturi nozzles specifiedin this Standard are called primary devices.” Nozzleshave been used for many years to prove natural gasmeters.

REASONS FOR PROVING A FIELD METER

Although the installed field meter has been completelychecked and calibrated at the factory, and a performancecurve developed, there are many things that can affect ameter on site causing measurement errors.

Any meter system can have changes during operationthat will cause errors in measurement. Orifice plates canbecome damaged, bearings on Turbine Meters can wear,and dirt and trash can accumulate on flow conditionersand in piping. Trash can accumulate on straighteningvanes, and Ultrasonic meters and other meters canbecome dirty causing the meter to error.

The inputs into flow computers and other electronic devicescan be changed by mistake causing errors in measurement.

New meter installations can have debris fromconstruction caught on the straightening vanes or in themeter itself. Proving checks not only the meter, but alsothe complete meter system.

One of the more important reasons for meter proving inthe case of Custody Transfer is to give both the buyerand seller confidence the volumes they transfer areacceptable to both the buyer and seller, thus eliminatingdisputes


METER PROVING DEVICES

There are many different methods and devices availablethat can be used to prove a Natural Gas Meter.

Some of them are calibrated Master Meters, SonicNozzles, Bell Provers, and Volume Provers. Each devicehas its advantages and disadvantages.

These proving devices can be divided into twocategories, primary and secondary. A primarymeasurement device, such as a bell or volume prover isone that has had its volumetric flow rate measurementaccuracy checked and verified against measurementsfor which there are national or international standards(e.g., mass, time, length, etc.) This device can then beused to verify Master Meters. A Master Meter that hasbeen calibrated becomes a secondary standard and canthen be placed in series with a field meter to verify theaccuracy of the Field Meter. A secondary device is onethat has been checked against a primary device and isthen used to prove another meter. An example of asecondary device is an In-Situ Meter Prover using aMaster Meter. An In-Situ Meter Prover is the Master Meterand it’s associated piping that can be either taken to thesite of the meter in the field or is installed permanentlyon the meter skid.

Volume provers which are relatively new, are primarydevices and very accurate. They are however veryexpensive and work best at higher line pressures andlower volumetric flow rates.

Bell provers, which are also accurate, work only at lowpressures and cannot be used for large flow volumes.They can be used to calibrate a Master Meter, whichcan then be used to prove other devices at higherpressures and higher flows. Bell Provers are relativelyexpensive to build and maintain.

The sonic nozzle although it is precise and considered aprimary standard, (+/- 0.25% total measurementuncertainty) can only check a field meter at one flow rateand line pressure. The sonic nozzle also causes apermanent pressure drop in the system.

TRANSFER PROVING SYSTEMS

A Transfer Prover is a proving system that checks thefield meter against the Master Meter at AtmosphericPressure. The Master Meter and its associated piping isplaced in series with the field meter and an adjustableflow rate blower draws air through the two meters. Air atatmospheric pressure and at various flow rates is thenpulled through the two meters and a comparison is madebetween the Master Meter and the Field Meter. A TransferProver System can be used to calibrate field metersbrought into a shop or the Transfer Master Meter Provercan be portable and taken to the field. The atmosphericTransfer Prover is best suited for checking metersoperating at lower pressures. The disadvantage toTransfer Provers is that many Flow Meters exhibit linepressure sensitivity that can introduce a measurementbias if they are calibrated at a test pressure significantlydifferent than their field operating pressure.

The higher pressure In-Situ Master Meter Prover is aMaster Meter usually placed immediately downstreamand in series with the operating Field Meter. It isconnected on site and is used for proving at actualoperating conditions of flow, temperature, pressure anddensity. The Master Meter is either mounted permanentlyon the metering equipment skid or is portable andconnected to an existing three-valve manifold whenneeded.

A Portable High Pressure Proving System consists of anumber of different sized Master Meters in parallel meterruns with provisions for installing a field meter for testdownstream of the master meter.

A sonic nozzle is placed downstream of each of the twometers in series to limit the flow and to verify the accuracyof the Master Meter. Since the Sonic Nozzle is a precisionmass measurement device, it works very well todetermine the mass flow at a pressure and temperature.It can be used as a check for any variable flow MasterMeter.

Since proving with a Master Meter is currently one ofthe least expensive ways to calibrate an existing fieldmeter we will discuss the various Master Meter systems.It is noted that AGA Report No. 6 on Transfer MeterProving using Master Meters is presently being updatedand rewritten.

PROVISIONS NEEDED ON THE FIELD METERSTATION TO CONNECT THE MASTER METER

There are two ways to install a Master Meter on anexisting Field Meter skid. One way is to use a three-valve manifold downstream of the existing meter runshown in the P&ID above. Downstream of the field meteris the preferred location because the flow profile into theField Meter is not disturbed.


A Flow Conditioner should be installed upstream of theMaster Meter in the Master Meter prerun piping toeliminate any problems in the flow profile caused by theelbows and valves going into the Master Meter Run. If athree-valve manifold is used the inline valve must be aBlock and Bleed type valve to insure all the flow is goingthrough the Master Meter Run. As a cost saving, extravalves are not needed going into and out of the MasterMeter Run. Blind Flanges can be used to cover theconnections. However closing in the Field Meter andbleeding down the pressure to install a Master Meterwithout the valves is time consuming and costly.

The second method is to install a Master Meter in serieswith the Field Meter on the existing Field Meter Skid.With this arrangement it is necessary to provide spacedownstream of the Field Meter for an additional

Master Meter run. The advantage of this system is thatno additional valves are required. The disadvantage isthe Field Meter run has to be closed in and bled down toinstall the Master Meter.

EQUIPMENT REQUIRED FOR ONSITE PROVING OFGAS TURBINE METERS

Master Meter

The meter proving device used to check the field metershould have an accuracy that is at least several timesthe accuracy of the meter it is checking. Repeatability,the ability of a meter to give the same answer every time,is critical. Sonic Nozzles with no moving parts arerepeatable and accurate.

Flow Computers can take advantage of the repeatabilityof Turbine Meters by using linearization to improve theprecision of the meter.

Flow Computer

The engineering calculations performed in the FlowComputer should follow the procedures described in theexisting AGA and API specifications. The report from theFlow Computer must be presented in such a way thatthe calculations from raw information to finalcompensated volumes can be verified using handcalculations if necessary.

Using modern Flow Computers, the typical metercalibration curve of a precision master meter can belinearized. During the meter proof, the stated volume ofboth the calibrated Master Meter and the field meterbeing proved can be measured precisely during the provecycle by using Pulse Interpolation as described in theAPI Manual of Petroleum Measurement Standards,Chapter 4, Section 6, Pulse Interpolation.

Pressure and Temperature Transmitters

The gas pressure, temperature and differential pressureon DP meters, must be precisely measured at the MasterMeter and at the meter being proved. Because of thesmall pressure drop between the Master Meter and themeter being proved the pressure can best be measuredwith another differential pressure transmitter. By using aMulti-variable Transmitter between the meters, anypressure or calibration error of using two transmitters iseliminated. The lower the line pressure the more difficultit is to determine the difference in line pressure betweenthe two flow meters. With atmospheric Transfer MeterProver systems, a differential pressure transmitter canbe used to determine the gage pressure at the first meter.The high-pressure port of the Differential PressureTransmitter is left open to atmosphere and the low-pressure port is connected to the pressure port of themeter. Atmospheric pressure can also be a manual entrybased on weather bureau data or a precise atmosphericpressure transmitter can be used.

The gas temperature must be measured at a point in theline at least three pipe diameters downstream of eachmeter. There will normally be a slight difference in thegas temperature between the two meters. This difference

Some Meter Systems are designed in such a way thatthe Master Meter is permanently installed on the MeterSkid. The Master Meter can then be used as the sparemeter to be put on line if there are problems with thenormal operating meters or if the flow rate exceeds thelimits of flow through the installed meters.


will be due, in part, to the pressure drop between themeters. Since this pressure drop is very small, typicallyonly a small change in gas temperature will occur.Extremely precise and well calibrated temperaturesensors and transmitters must be used to measure thisprecise differential.

MAXIMUM GAS VELOCITY

Care must be taken not to exceed the safe maximumvelocity of the gas through the proving systems and thefield measurement system. Exceeding the maximumvelocity will cause erosion and wear of the piping andthe components resulting in an unsafe system.

The weight of 0.6 Sp.Gr. Gas at 1,440 psi is approximately5.727 lbs./cu.ft. The max velocity recommended by APIRP 14E is 78 ft/sec. Excess velocity will cause pipe wearespecially at elbows and bends.

PIPING PER AGA 7

Piping for both the master meter and field meter prerunsand post runs should be configured per the specificationsthat apply. Usually the axial lengths of the post-run ofthe first meter and the pre-run of the second meter (inseries) can be combined making the minimum axiallength separating the two meters a total of the requiredpipe diameters. If any piping manifolds, elbows or bendsare located upstream of either the Master Meter or thefield test meter, as is the case when connecting to athree-valve manifold, a high performance flowconditioner should be installed upstream of the meter inaccordance with the flow conditioner manufacturersrecommendations.

IN-SITU PROVING WITH A MASTER METER PLACEDIN SERIES WITH THE FIELD METER

The performance of the Master Meter and the Field Metershould be compared on a mass flow rate basis. For mostturbine flow meters, the measured flow rate is expressedin terms of volumetric flow rate. The line pressure (and,probably, the gas temperature) will be slightly differentat the two meter locations, so the actual volumetric flowrates measured by the two meters will not be equal. Asan alternative to comparing the field turbine meter andthe prover on a mass flow rate basis, the volumetric flowrates recorded by the two meters can be adjusted to‘standard’ conditions and then compared. Standardvolumetric flow rate is essentially a mass flow rate thathas been referenced to arbitrary temperature andpressure conditions (e.g., a predetermined pressure andtemperature, such as 14.73 psia and 68ºF, respectively)for the flowing gas composition. Standard volumetric flowrate is proportional to mass flow rate through theapplication of standard gas density and is, therefore,conserved from the upstream to the downstream meterlocations.

A calibrated Master Meter run of sufficient size andpressure rating is connected to a three-valve manifoldlocated upstream or downstream of the meter to betested. Tests have shown that some of the commonly-used turbine meters are relatively insensitive to upstreampiping effects. However since piping and field conditionsmay vary it would be good practice to follow the standardpiping configurations shown in AGA Report No. 7 forany turbine meter run. The use of a flow conditionerupstream of the Meter Run is recommended if there isthe possibility the Master Meter piping might have anadverse effect on the Master Meter. The Master Meterrun must be complete with a Flow Computer, pressureand temperature transmitter and differential pressure (DP)transmitter.

When needed the pressure, temperature and differentialpressure are available from one multivariable device. Itis possible for some resolution to be lost by the methodused to connect transmitters to the flow computer.Devices are available to convert the digital signal fromtransmitters directly to Flow Computers.

The Master Meter, it’s associated piping, and electronicsmust be calibrated as an assembly. Test at variouspressures can be done and correction factors establishedfor any shift in performance.

The field meter to be tested and the pressure,temperature, and DP transmitters associated with it areconnected to the proving Flow Computer in such a waythat the existing measurement is not affected.

Normally the signals from the meter come from a parallelconnection and the analog signals from the pressure andtemperature transmitters are connected in series. Adevice that can measure the gas density, either a GasChromatograph or a Correlative device also needed todetermine the mass flow. With a Gas Chromatographthe density of the gas is determined by its composition.A Correlative device uses other means such as speed ofsound, thermal conductivity etc. to determine the gasdensity.

HIGH PRESSURE PROVING WITH INLINE METERS

High Pressure proving at various flow rates and pressurescan be accomplished with a High Pressure Master MeterProving system. This system is located where there issufficient pressure and flow to check all the meters inthe system individually. For example the best locationfor High Pressure Proving in a distribution system iswhere gas enters the distribution pipeline. The systemhere can serve two functions. It can measure the gasbeing purchased and it can be used to prove all the highpressure meters used downstream in the distributionsystem.

Routing the flow through the other meters in the systemcan vary the flow rates through the meter being tested.


The pressure in a High Pressure Proving System can bevaried as long as the Master Meter System has beenproved at that pressure and it does not affect thedownstream system.

CONNECTING A PORTABLE MASTER METER RUNTO THE FIELD METER SKID

1. Connect the Master Meter Run to the three valvemanifold using either new gaskets or gaskets thatare suitable for multiple uses.

2. Connect the electrical cables from the Master Meter,Pressure Transmitter and Temperature Transmitterto either the existing Flow Computer or Connect theField Meter, Field Pressure Transmitter, andTemperature Transmitter signals to the FlowComputer on the Master Meter Run.

3. Slowly fill the Master Meter Run with gas by openinga small valve usually a needle valve or small ball valvethat connects the line pressure piping to the MasterMeter. As a rule of thumb pressurization of the lineshould not exceed one PSI per Second. It is also agood practice to check for leaks with an approvedleak detection method while the Master Meter Runis filling. Checking at a low pressure first before theline is completely filled and checked at line pressurecan save time and gas if a leak is detected.

4. Open the upstream valve connected to the MasterMeter Run.

5. Slowly open the downstream valve that connectsthe Master Meter Run to the Field Meter Run.

6. Slowly close the Block and Bleed valve downstreamof the Field Meter on the three valve manifold puttingthe Field Meter and Master Meter in series with oneanother.

7. Follow the Master Meter Proving Procedures usuallysupplied by the Company for Proving a Master Meter.

8. When the Proving has been completed open theBlock and Bleed Valve downstream of the FieldMeter.

9. Close the valves connecting the Master Meter Runto the Field Meter and slowly bleed down thepressure on the Master Meter Run. One PSI dropper second is always a safe pressure drop rate.

10. Disconnect the Master Meter Run from the FieldProver Skid.

ONSITE MASTER METER PROVING PROCEDURES

1. After the Master Meter has been properly installedon the three-valve manifold, the system must bechecked for leaks.

2. The precision of the Master Meter should be verifiedbefore each prove.

3. The manufacturer of the Master Meter will have aprocedure on how to verify the meter is still incalibration. When the Master Meter condition isverified its condition should be recorded on the provereport.

4. Enter the tracking information from the Master Meteror select a Master Meter whose information hasalready been entered into an existing database.

5. Enter all the information for the Field Meter to beproved or enter the serial number for the informationto be entered automatically if it already exists in adatabase.

6. Prove runs are normally done at 10%, 25%, 50%,75% and 95% of the maximum rated flow rate ofthe meter to be proved. These are recommendationsonly. The test flow rates requested by the end usershould be used wherever possible. If the requiredflow rates cannot be obtained the Field Meter mustbe proved at the available test flow rates.

Proves can be done based on time or on volume. Eithermethod if performed properly, can produce a good prove.In this example we will use proof runs based on time.

7. Set the time and number of proves for the proveruns. The number of pulses generated by the meterswill determine the time for each flow run. The timemust be long enough for a statistically significantnumber of pulses to be generated by the lowestfrequency output. If the Flow Computer is capableof Pulse Interpolation the time of each run can beshortened in accordance with the API Manual ofPetroleum Measurement Standards Chapter 4,Section 6, Pulse Interpolation.

8. The number of proofs at a specific flow rate dependon the client. However, it is recommended that aminimum of three runs be made at each test flowrate to determine if the meter being proved isrepeatable.

9. Start the prove. The Flow Computer will automaticallystop and start each run, based on the time enteredinto the Flow Computer. On the first run allow timefor stabilization of the flow rate through the metersbefore beginning the data acquisition.

10. When the prove runs for the specified time and therun is successful, the information is saved in the FlowComputer and/or can be printed out. If the flow rateof the station can be changed, it is recommendedthat the first test run be performed at maximum flowrate setting. If the meter repeats and is withinspecifications at the maximum flow it is an indicationthat it may prove at the lower flows. Proves at all


flow rates and operating pressures specified mustbe done to insure the field meter is operatingproperly.

11. At any time during a prove run or cycle, the provecan be stopped (Aborted). Aborting the prove stopsthe proving run. An aborted prove report should bestored or printed.

12. When the prove for all flow rates is complete, theflow control valve closes to the meter run closes,the prove report is printed, and all information issaved electronically in a file for that prove.

CONCLUSION

Using available knowledge, products and equipmentnatural gas meter proving in the field and in the shopcan be used to verify the accuracy and repeatability ofnatural gas measurement systems.

REFERENCES

ASME/ANSI MFC-7M-1987 Reaffirmed 2001“Measurement of Gas Flow by Means of Critical FlowVenturi Nozzles.

A.G.A. Report No 7, “Measurement of Fuel Gas byTurbine Meters.”

API Manual of Petroleum Measurement Standards,Chapter 4, Section 6, Pulse Interpolation.

AGA. Report No 6, “Methods of Testing LargeDisplacement Meters.”

Edgar B. Bowles, “Onsite Proving of Natural Gas TurbineMeters.” Southwest Research Institute, ISHM 2001.

Daniel J. Rudroff


LOCAL AND WIDE AREA NETWORKINGOF GAS FLOW COMPUTERS

King PoonThermo Flow Automation

9303 Sam Houston Parkway, Houston, Texas 77099

INTRODUCTION

Communication has been around ever since mandeveloped language and hand signs to exchange andshare ideas. Smoke signals were used in the ancientworld to send information from one place to another. Infact, a smoke signal is one form of wireless digitalcommunication. The advance in modern communication,network and computer technology has led to the growthof electronic forms of communication. Electronic datacan be transferred between workstations in the sameoffice and sometimes even between offices in differentcities.

Electronic flow computers are used widely by the gasindustry for measurement and control purposes.Measurement data is retrieved periodically and storedin the central computer system for accounting,engineering and administrative functions. Controlcommands are sent by the gas control department orthe field operators to the gas flow computers. Data arecollected from the field and shared by differentdepartments so that decisions can be made moreefficiently. Because of this, an effective communicationsystem between the electronic flow computers, thecentral computer system and remote offices need to bedeveloped.

Electronic flow measurement systems can now makeuse of this office interconnectivity to effectively movedata from the electronic flow computer to individualworkstations out in the field offices and the central office.Electronic flow measurement systems can utilize existinggeneral-purpose business equipment for communicatingmeasurement and control data across networks.

TERMS

Local Area Network (LAN): A Local Area Network is thenetworking or interconnection of multiple workstationscontained within a single building or facility.

Wide Area Network (WAN): A Wide Area Network is thenetworking or interconnection of multiple workstationsthat covers a broad geographical area such as a city,state or country.

TCP/IP: Transmission Control Protocol / Internet Protocol(TCP/IP) is a name given to the collection of protocolsthat have been used to construct the global Internet. Itis the typical protocol used for communications betweenworkstations across a LAN or WAN. It is the most widely

used protocol in the world and is now the de-facto setof protocols of the world wide internet.

IP Addressability: IP addressability is the ability toconfigure a device with an address that is unique to aparticular LAN/WAN.

Ethernet: Ethernet was developed in the mid-1970s byXerox Corporation. A cable connects all computers ordevices to an interface device. Carrier Sense Multiple/Collision Detection is used to avoid collision on thenetwork.

ELECTRONIC FLOW MEASUREMENT DATA

Measurement data stored in electronic flow computersconsists of vast amounts of hourly and daily records inaddition to alarm and event logs that must be retrievedon a periodic basis. Various departments use this datafor everything from accounting to engineering. Anexample of this would be retrieving historical tubing andcasing pressure information so that it can be analyzedby the engineering department.

A typical historical record consists of volume, energy,average differential pressure, average static pressure,average temperature, average square root extension,flow time, and possibly average gas quality values. Thisdata is stored for each hour for each meter run in eachflow computer in the system. Because of the nature ofhistorical data, retrieval is done at a less critical rate thancontrol data. For accounting or analysis purposes, thedata can be retrieved on a daily basis.

Control data is needed at a much faster rate. Decisionsmade by gas control personnel must be made using upto date information. When the commands are actuallymade to the flow computers, it must take placeimmediately, giving the operator immediate feedback.

Typical control data consists of instantaneous flow rates,set points for PID loops, nomination values, as well asvalve control commands and statuses. This is a typicallysmall amount of data retrieved and transmitted in realtime.

LAN/WAN

LAN is a computer network dedicated to share dataamong workstations, computers and devices (nodes)within an office. An interface device converts the userinformation (data) into a predetermined form (the


protocol, usually TCP/IP) on a communication medium(Ethernet). These types of networks have the ability totransfer data at rates of 100 Mega bits per second orhigher. A central server is used to store data and controlthe traffic within the network. Data, printers, storagemedia (such as hard drives) can be shared within thenetwork. With common protocol and medium, computersand devices can share information easily.

WANs are built to provide communication solutions foroffices that need to exchange digital information betweentwo distant places (in one or more different cities). Sincethe distance is long, the local telecommunicationcompany is involved to provide and maintain thenecessary service.

The main purpose of a WAN is to provide reliable, fastand safe communication between two or more places(nodes) with low delays and at low prices. WANs enablean organization to have one integral network betweenall its departments and offices, even if they are not all inthe same building or city, providing communicationbetween the organization and the rest of the world.

LAN/WAN FOR DATA SHARING

Most companies using gas flow computers are spreadacross a wide geographical area. The flow computersare typically in extremely remote locations. A field officeis usually setup in close proximity to the flow computersfor ease of maintenance and control. The accountingand engineering departments are located in a morecentralized location. Data from the flow computers mustmigrate from the remote flow computer all the way tothe central offices.

The existing LAN/WAN used for other business purposes(e.g. email, accounting, etc.) can be utilized for thesharing of this data. Host systems can use existingshared drives to move data from one office to the next.For example, a host system will communicate with thegas flow computer and store that data on the shareddrive. Workstations remote to the host system can thenaccess that data for reporting and analysis purposes.

Another method of data sharing is the use of TCP/IPbetween host programs. Host programs that are capableof communicating via TCP/IP can transmit data usingthe existing hardware to access data. For example, aprogram running on a workstation in a central office canrequest data from the host system residing in the fieldoffice. This type of system is typically used for data thatis more time critical such as control parameters that mustbe immediately relayed to the flow computer.

LAN/WAN FOR FLOW COMPUTERS

An alternate use of the existing LAN/WAN infrastructureis the direct communications of a host system with aflow computer. An IP addressable LAN/WAN interfacedevice can be connected to the host port of the flowcomputer. This makes it possible to connect a flowcomputer to any existing network on the LAN/WAN andallow a host system to retrieve its’ data. For example, aflow computer measuring gas on an offshore platformcan link into the existing microwave network used foremail, voice and network connectivity. A host systemthat resides anywhere on the LAN/WAN can thencommunicate with this device without the need of asatellite or cell phone system.

A communication system such as a radio network canalso be connected to the LAN/WAN using an existingnetwork interface device. A host system can requestinformation across the network and through theconnected radio network; data can then be transmittedto the flow computers in the field and back to the host.

CONCLUSION

Existing LAN/WAN infrastructure can be utilized to createa fast, reliable and cost effective communication networkbetween the flow computers, the host system in the fieldoffice and the computers in the central office. Differentdepartments and offices in different locations and citiescan then share electronic flow measurement data. Takenfurther, because of the use of TCP/IP and the LAN/WAN,data can now be shared on the internet.

King Poon


TRANSFER PROVINGLarry K. Wunderlich

Centerpoint Energy Entex4220 Laura Koppe, Houston, Texas 77016

INTRODUCTION

Transfer proving was initially developed to provide aneasier and more accurate field meter proving method.Because of the capacity capabilities of transfer provers(2000 CFH to 80,000 CFH) transfer provers are utilizedin meter shops where bell prover capacity is limited andallow for shop testing of the larger capacity meters.

TRANSFER TESTING SYSTEM

In the typical transfer testing system, air or gas passesthrough the meter under test (field meter) and thenthrough the “Master” meter. The vacuum method oftesting is basic to all commercially available provers.Atmospheric air is drawn through the field meter andprover by the blower system. Temperature and pressuredifferences are measured to enable correction of the datato a common base of comparison. The rotary metertransfer proving system is not affected by specific gravityor relative humidity of the flowing gas since under normalconditions no condensation or change of moisturecontact would occur as the gas passes from the fieldmeter to the master meter. It is a volumetric test in whichthe test time is not a critical variable. Automatic operationminimizes chances for human error, and built-in self-check features assure reliable system performance. TheROOTS Model 5 Transfer Prover System is typical of thecommercially available units and will be utilized in thisdiscussion. Other units may differ slightly in shape, size,or performance, but the base operating principles arethe same.

The prover System consists of one or more rotary positivedisplacement master meter(s) mounted on a wheeledcart. The master meter is calibrated over a flow range of100 to 10,000 CFH for a 10M master meter of 35 to 2,000CFH for a 2M master meter. This range covers the testingof larger diaphragm meters as well as most rotary meters.Blowers mounted downstream of the meters are usedwhen air is the testing medium. On air tests, the blowersdischarge to atmosphere through a muffler or silencerwhich minimizes noise when testing in public areas or inshops where noise could be objectionable.Comprehensive tests have demonstrated compliancewith OSHA regulations concerning acceptable noiselevels. The connection from the field meter to the proveris made with a 25˚ length of flexible hose equipped withquick disconnect fittings. In addition, a cable is requiredfor temperature, pressure, and pulser connections on afield meter.

All transfer provers are available to the gas industrycompute the following formulae:

Master Meter VolumeEquation 1 - % Uncorrected Proof = —————————— X 100 Field Meter Volume

ATM Press-Master Press Drop

Equation 2 - % Pressure Correction = —————————— X 100 ATM Press - Field Press Drop

Field Meter Temp ‘REquation 3 - % Temperature Correction = —————————1 X 100

Master Meter Temp ‘R

% Uncorr % Press % Temp Proof X Corr + 100 X Corr + 100

Equation 4 - % Corr Proof = ——————————————— 10,000

Testing can be done automatically or manually. Whentesting automatically, the index or instrument is removedfrom the field meter and a pulser unit installed to countthe output shaft revolutions. When testing a meter whoseindex cannot be removed, a retroreflective scanner maybe utilized or a remote start-stop switch in lieu of thepulser.

The standard transfer prover operates on a 115 VAC±10%, 47-60 Hz. Power consumption with blowers onhigh is approximately 1000 watts, but satisfactoryoperation can be obtained in the voltage range of 95 to130 VAC. Special provers have been designed foroperation on 230 VAC.

The proving system has been designed for operation byone man with a minimum of effort. At Centerpoint EnergyEntex a van is used for transporting the Model 5 prover.The accuracy and repeatability of the transfer proversystem is related to the permanent accuracycharacteristic of the rotary positive displacement mastermeter. For convenience, a direct readout of the proof ofthe field master is provided by using a simple but effectivemethod of obtaining a master meter curve of 100%accuracy over the full working range.

The master meter incorporates a pulse unit for generatingcontact closures representing the flow units from themaster meter. Provisions have been made in thecomputer software to add/subtract in extra countsrequired to produce a 100% accuracy curve for themaster meter.


The actual preset correction necessary to linearize eachspecific master meter curve is determined from a factorycalibration with a bell or piston prover. A chart is providedwith each prover, showing the calibration curve.

OPERATION

The Model 5 Proving System consists of the mastermeter(s), blowers, controller, and laptop computer.Operation of the Model 5 Transfer Proving System canbe broken down into five sections:

1. Powering up of the system.2. Making field meter connections.3. Purging the meter and leak testing.4. Selecting and starting the test(s).5. Running the test and saving the results.

Reasonability tests are run on the signal inputs to thecontroller to determine that the transducers are properlyconnected prior to starting a test run and thatmeasurements taken during a test run are withinreasonable limits.

POWERING UP THE SYSTEM

This step should be performed first as it allows time forthe controller to warm-up and stabilize prior to runninga test.

1. Plug the line card from an AC power source into thecontroller.

2. Connect the laptop computer AC cord to the laptopreceptacle on the controller.

3. Connect the RS232 cable between the laptopcomputer and controller.

4. Turn on the controller power switch and then thelaptop computer.

5. Go into the Model 5 laptop computer software andverify master meter serial numbers. This ensuresproper presets are loaded.

FIELD METER CONNECTION

Connections of the field meter sensors and pulser (ifused) should be done next.

1. Connect the field junction box cable to the controller.

2. Install the field meter temperature sensor in, or near,the inlet port for rotary meters and at the outlet portfor turbine and diaphragm meters. The tip of thesensor should be in the center of the flowing airstream. Plug the cable into the field junction box.

3. Connect the pressure lines from the field junctionbox to the inlet of the field meter (and outlet if rotarymeter differential is to be read during testing).

4. If the field meter pulser is being used rather then themanual start/stop button, install the field meter pulseron the instrument drive platen of the field meter.Ensure that the instrument drive properly engagesthe pulser and that the pulser shaft is centered overthe meter drive shaft. The pulser drive coupling mayrequire adjustment to engage the meter drive shaft.Plug the cable into the field junction box.

5. If the manual start/stop button is to be used, plugit’s cable into the field junction box instead of thepulser.

6. Connect one end of the 25 foot prover hose to theoutlet of the field meter.

PURGING AND LEAK TESTING

The Model 5 Transfer Prover can be set-up to requirepurging of the field meter or not require purging for shopuse. You can also require or not require a leak test. If apurge is required the blowers will not start for testinguntil the operator does a purge of the field meter.

1. Connect one end of the 25 foot prover hose to thecamlock on the Model 5 silencer. The other endshould be hooked to the field meter.

2. From the laptop software menu, select the purgeoption and run the blowers to purge the field meter.

3. Seal the inlet to the field meter, move hose fromsilencer to the inlet of the Master Meter and run leaktest. This takes approximately 30 seconds and apass or fail message will be isplayed.

4. Identify and eliminate any leaks as they will affectthe accuracy of the tests.

SELECT AND START TEST

Forty preconfigured tests can be run by simply selectingthe desired tests for a particular meter. Each test can berun at three different flows and each flow can be runtwice. The tests can be cascaded for more flow rates ifdesired. Flows are set and controlled by varying blowermotor speed which reduces noise. The following itemsmust be selected to configure for a manual test:

1. Prover Capacity (2M, 10M, 80M)2. Test Control Mode (ID, OPTO, Manual)3. Meter Output (UC, TC, PC, TCPC)4. Drive Rate (What each pulse represents.)5. Test Volume (Even multiple of drive rate.)6. Flow Rate (Up to three flows.)7. Base Temperature Correction (If Required.)8. Base Pressure Correction (If required.)


RUN TEST AND SAVE RESULTS

When the operator starts the test, the computer, via thecontroller, will start the blowers, stabilize flow as highestrate selected and begin test. At the end of the first run,results will be displayed and the blowers will stabilizefor the next run. At the completion of the tests, theoperator can view any particular run. The results can besaved and/or printed. The operator will be prompted tofill in some information for the report. The layoutappearance of the reports are customer configurable.

TESTING RESULTS

What is the accuracy of this testing device? Results fromextensive series of tests show that the ROOTS TransferProver duplicated a bell prover within ±0.1% on overallaverage results and had a standard deviation of the lessthan 0.5%. The test included a large number of differenttypes and sizes of diaphragm meters. The bell proverwas in a temperature controlled environment, but thetransfer prover operated in a room without temperaturecontrol. Repeatability has been found to be ±0.1%.

A warm-up period has been found to be desirable intesting diaphragm meters which have been inactive.Such meters are occasionally encountered in the metershop testing, and a running period of a few minutes willallow enough exercise of the moving parts to restorenormal meter friction and assure representative testresults.

Minor variations in blower operation due to supplyvoltage fluctuations may occur but will vary the actualflow rate only slightly. Tests have shown that fluctuationswhich might have invalidated the test run by other testmethods did not influence the transfer prover test results.

TEST SET-UP MODIFICATIONS

The transfer prover is an extremely versatile and flexibletest device and has been used to test meters with 1,000cubic foot drive rates, temperature compensatedoutputs, pressure compensated outputs, and intermittentoutputs. Information concerning special test techniquesor particular test problems can be obtained from theprover manufacturer.

CONCLUSION

The feasibility of using a rotary positive displacementmeter transfer testing system has been clearlyestablished. Tests have confirmed the accurateperformance of the equipment. The increase inproductivity, the speed of testing, the avoidance of testerrors, and the savings in shop test facilities maketransfer provers an attractive and economical additionto the array of meter test equipment.

The development of the portable transfer proving systemwith computer enhancements gives measurementpersonnel another useful and reliable tool for testing gasmeters. The ability to test meters in the field and spotinaccurate meters and adjust them without removing themeter saves not only time but money. Although theprimary usage of transfer proving has been for fieldtesting, increasing numbers are being utilized in metershops and provide a valuable supplement to existing testfacilities.

Larry K. Wunderlich


INSPECTION OF REGULATORS & RELIEF VALVESJohn Johnson

CenterPoint Energy4220 Laura Koppe, Houston, TX 77016

Regulators and Over Pressure Protection Devices (OPPD)must be inspected in accordance to Federal and StateLaw and Company policy. Over pressure protectiondevices are devices that protect the downstream pipingin the event of a regulator failure. These devices includea relief valve, a monitor regulator, or a positive pressureshut off. In Texas, inspection interval must be at leastonce per calendar year, at intervals of no more than 15months.

Prior to inspecting a regulator or OPPD:1. The person performing the inspection should have

some knowledge of the system downstream of theregulator. Is the system a one way feed or is it asystem that is fed by one or more stations.

2. Visually inspect all valves to make sure they are inthe correct position (opened or closed).

3. Install a gauge upstream and downstream of theregulator. Be careful not to install gauges that haveworking pressures lower than the pressure you arechecking. Make sure the gauge installed downstreamof the regulator will handle the upstream pressure ifthe regulator fails while being inspected and will notlockup.

4. Service and operate all valves.

Some of our regulator stations have two separateregulator runs with only one regulator in each run. Oneregulator has the pressure set at the required pressurefor the system it is feeding; we call it the primary run.The other regulator has the pressure set lower; we call itthe standby run. The purpose of the standby regulatorrun is to assist the primary regulator if there is a heavyfeed on the station or if the primary regulator fails. Thestandby regulator run should be inspected first. Whiledoing the inspection continue to monitor the downstreampressure of the primary regulator.

A few of our regulator stations only have one regulatorrun. In this case we have a bypass run with a valve thatwe can operate to maintain the required pressure for thesystem. The person operating the bypass valve mustcontinually monitor the gauge downstream and maintainthe required pressure.

We have some regulator stations with multiple regulatorsin each regulator run. There are numerous reasons touse multiple regulators:1. When the upstream pressure is very high and the

pressure needed downstream is lower.2. Keep from having to install a large relief valve.

3. To keep the regulator from freezing up due to a largepressure cut.

We use three Types of regulators:1. Spring loaded2. Pilot loaded3. Pressure loaded from a controller

RAILROAD COMMISSION OF TEXAS-PIPELINESAFETY RULES

192.195 Protection Against Accidental Over-Pressuring

A. General Requirement:Each pipeline that is connected to a gas source so thatthe maximum allowable operation pressure could beexceeded as a result of pressure control failure or of someother type of failure, must have pressure relieving orpressure-limiting devices.

B. Additional Requirements for Distribution System:Each distribution system that is supplied from a sourceof gas that is at a higher pressure than the maximumallowable operating pressure for the system must:(1) Have pressure regulation devices capable of meeting

the pressure, load, and other service conditions thatwill be experienced in normal operation go thesystem, and and that could be activated in the eventof failure of some portion of the system and

(2) Be designed so as to prevent accidental over-pressuring.

192.199 Requirements for Design of Pressure Relief &Limiting Devices

Except for rupture disc, each pressure relief or pressurelimiting device must:

(a) Be constructed of material such that theoperation of a device will not be impaired bycorrosion.

(b) Have valves and valve seats that are designednot to stick in a position that will make the deviceinoperative.

(c) Be designed and installed so that it can bereadily operated to determine if the valve is free,can be tested to determine the pressure at whichit will operate, and can be tested for leakagewhen in the closed position.

(d) Have support made of non-combustiblematerial.

(e) Have discharged stacks, vents, or outlet portsdesigned to prevent accumulation of water, ice,


or snow, located where gas can be dischargedinto atmosphere without undue hazard.

(f) Be designed and installed so that the size of theopenings, pipe, and fittings located between thesystem to be protected and the pressurerelieving device, and the size of the vent line,are adequate to prevent hammering of the valveand to prevent impairment of relief capacity.

(g) When installed at a district regulator station toprotect a pipeline system from over-pressuring,be designed and installed to prevent and singleincident such as an explosion in a vault ordamage by a vehicle from affection the operationof both the over-pressuring protective deviceand the district regulator; and

(h) Except for a valve that will isolate the systemunder protection from its source of pressure; bedesigned to prevent unauthorized operation ofany stop valve that will make the pressure reliefvalve or pressure limiting device inoperative.

John Johnson


REMOTE METER READINGMETHODS OF RETRIEVING DATA BY USE OF REMOTE DEVICES

Arun SehgalItron Inc.

2818 North Sullivan Road, Spokane, Washington 99216

INTRODUCTION

Remote meter reading is a fast growing trend in NaturalGas industry. As per industry estimates, close to 14million gas meters in the United States are read remotely.This paper presents the advantages of remote meterreading and explains the various technologies in use.

WHAT IS REMOTE METER READING?

Remote meter reading implies the use of acommunication device to automate gas meters, thatsends out the meter readings and other crucial meterinformation to a data collection system, thus eliminatingthe need to physically visit the gas meters.

Commercial and Industrial customers

Small commercial customers’ gas meters are usually readmonthly for consumption. However, large commercialand industrial customers have advanced meteringsystem- like volume correctors, which are read — usuallydaily at the gas day time. These remote-monitoringdevices attached to gas meters and correctors storehourly gas consumption profiles.

COMMUNICATION TECHNOLOGIES FOR RETRIEVINGDATA

RF (Radio Frequency) Communications

RF is the most commonly used method ofcommunication between a remote device and datacollection system. A gas meter is mounted with a smallmodule, usually under the index, where it converts themovement of wriggler into pulse counts and stores them,taking into account the count rate and other factors.These modules are essentially encoders and RFtransmitters, put together as a single unit.

These RF devices are activated by a “wake up” signalfrom the data collection system and the device sendsback the latest meter read and other information, liketamper status, to the data collection system. As thesedevices are battery powered, using this wake techniquehelps conserve battery life.

In some cases, RF devices are regularly transmitting thereadings and the data collection system does not “wakeup” the device.

RF-based remote meter reading devices are easilydeployable and offer a very reliable means of datacollection. These devices are used mostly for residentialand some commercial remote meter reading.

DATA COLLECTION SYSTEM FOR RF DEVICES

There are three most popular ways to “read” these RFdevices.

Radio Equipped Handheld Computer

A meter reader carrying a rugged handheld computerequipped with a radio receiver walks by homes, withoutactually entering the premises. The devices radio theirreads to the handheld computer, the meter readeraccepts the read and keeps moving on its route.

DATA RETRIEVAL

Residential Customers

Most Natural Gas customers are billed monthly and, thus,a monthly consumption read is sufficient for thesecustomers. However, some residential Natural Gascustomers in fully deregulated markets may need to beread more frequently, and some even daily. This may bea growing trend, especially when the potential benefitsof deregulation are made available to more residentialcustomers. Utilities also need to monitor whether thegas meter has been tampered with in any way — resultingin a potential loss of revenue.

Datacollectionsystem

CIS /Billingsystem

CustomerBill


The meter reader thus does not enter the readingsmanually, hence eliminating any manual entry errors. Thissystem is normally used to read those accounts withinthe utility service territory that have high-cost orhazardous-to-read meters. These meters may besituated in a basement, in a back yard with a dangerousdog or locked gate, or with an angry customer whodoesn’t want the meter reader on the property. The meterreaders can collect anywhere from 600 to 1000 meterreads on a typical day using a Radio equipped handheldcomputer.

“Drive By” or Mobile Data collection system

“Drive by” or mobile data collection is a very popularmethod of data collection from radio equipped remotedevices.

Mobile data collection uses vehicles equipped with radiounits to read RF module-equipped gas meters via radiowithout the need to access the meter. This readingsystem dramatically improves meter reading efficiency.

A radio transceiver is installed in a utility vehicle. Routeinformation is downloaded from the utility billing systemand loaded into this radio transceiver. While driving alonga meter-reading route, the transceiver broadcasts a radio“wake-up” signal to all RF meter modules within rangeand receives the meter readings when they respond.Completed reads are uploaded to the billing system forbill generation.

Mobile data collection system is used in saturated areaswhere there may be large populations, difficult-to-access,or hazardous-to-read meters. As a result of this level ofsaturation, meter reading efficiency is dramaticallyimproved. A single transceiver can read an average of10-12,000 meters in an 8-hour shift, and can read up to24,000 meters per day, depending on meter density andsystem use.

Fixed Network Data collection system

Some utilities, especially those who read both gas andelectric meters, are considering deploying a fixed networkfor remote meter reading.

Fixed network deployment is usually done as a migrationfrom the mobile data collection system. The FixedNetwork is usually installed over saturated areas whereadvanced metering data, variable reads; unscheduled

reads or operational improvements are required. Thissaturated deployment spreads the cost of the networkcomponents over multiple meters.

Other considerations for RF remote monitoringdevices

While deciding on the choice of remote metering devices,utilities often take into account the following:

• Compatibility with gas meters: Utilities often lookfor those remote-metering devices, which canbe directly mounted on most of their installedbase of gas meters. This minimizes the need tochange the gas meters just to suit the design ofremote metering device and helps by reducingthe cost of implementing a remote meteringsystem.

• Battery Life: Utilities like long battery life —ideally that lasts as long as the useful life of theremote monitoring device itself. This eliminatesthe need to do battery change outs. Currentdesigns of some remote metering devices havebattery life of 17-20 years.

• Safety: Safety is an important attribute and thesedevices should be intrinsically safe and certifiedto be used in natural gas environments.

• One of the factors that utilities have oftenconsidered is the availability of a migration pathfor reading technologies. Many utilities havestarted with the handheld walk-by solution andas their deployment base grew, they havemigrated to a mobile data collection solution,without any impact on the meter modules orneed to revisit them.

RF-BASED SYSTEMS: IMPLEMENTATION SUCCESSES

Radio based Remote Meter Reading systems have beeninstalled by gas utilities for nearly a decade. The earliestinstallations at Minnegasco, Keyspan energy,Philadelphia Gas Works, Atlanta Gas and at severalothers are more than 15 years old, and these utilitiesmoved towards saturation several years back. This is atestimony to the success of RF technology in the field.Several hundred utilities around North America haveimplemented RF based systems and are reaping thebenefits.

COMMUNICATIONS USING TELEPHONE

Large commercial and industrial gas customers consumea large amount of gas that needs to be measured andreported — often at daily or even hourly basis. Thesereadings are usually required at the Gas Day time. Insuch cases, using a telephone line provides an easy wayto communicate with these remote metering devices.

Telephone based remote metering devices are of twotypes based upon their mode of communication, inboundor outbound.


Inbound systems: In this type of remote meter readingsystem, the modules call a Central Master Data CollectionComputer at prescheduled times, usually at Gas day timedaily, and provide the hourly gas consumption data.Inbound phone systems have been very successful inefficiently sending consumption and other data in a giventime window.

Outbound systems: In this type of communication, theMaster Station calls the remote meter module andcollects the data. This is useful where on-demand readsmay be required as well.

The telephone-based remote metering modules mayeither be direct mount or remote mount. The directmounted modules encode and store the hourlyconsumption and communicate using telephone line.

The remote mounted modules take multiple pulse inputsfrom gas meters, electronic volume correctors, count andstore this data and use a telephone line to send this datato the collection software.

The utilities, especially those who run in a deregulatedenvironment, need to report the consumption profiles oftheir large commercial and industrial consumers toseveral third party systems which communicateinformation to customers, gas marketers, systemplanners etc. These utilities are now looking at ways toimprove their data collection methodologies and howthey can improve the dissemination of the metering data,using enterprise-wide deployment of newer datamanagement and analysis software.

OTHER COMMUNICATION METHODS

In places where it is difficult to make a regular phoneline available, cellular transreceivers are currently beingevaluated at some sites around North America as acommunication medium.

While cellular offers a reduction in installation costs,experience regarding the network availability, cost ofservice and battery life, etc. is being evaluated. Currentlyonly few hundred cellular-based devices are in field.

Besides the above, various other communicationmedium like satellite, powerline carrier (in conjunctionwith electric meters) etc have been experimented in smallnumbers.

CONCLUSION

Benefits of Remote Meter Reading

Traditionally, Remote Meter Reading has been thoughtof as a means to improve meter-reading operations andto reduce costs. But that’s just the beginning. Bydeploying advanced data collection and managementsolutions, utilities are already achieving a variety ofbenefits throughout their companies every day. Inaddition to reducing costs, other benefits include:

• Revenue cycle improvements and costreductions

• Increased billing accuracy and resulting revenue• Revenue assurance and protection• Reductions in theft of service

Today, gas utilities are beginning to realize that RemoteMeter Reading technologies deliver operationalimprovements and cost savings throughout a utility’s coreservice processes. In addition to meter reading services,the benefits of advanced data collection extendthroughout a utility’s operations to deliver quantifiableimprovements and savings in revenue assurance,customer service, field service operations, distributionsystem reliability and efficiency, marketing and businessdevelopment, and regulatory compliance. The economicvalue of remote meter reading technology doesn’t endat the meter shop. That’s where it begins.

Remote Meter Reading is a growing at a fast pace andmore and more utilities will be employing its benefits inthe future.

Arun Sehgal


PRINCIPLES OF ODORIZATIONThomas E. Tucker

Odor-Tech7591 Elser Field Road, Pineville, LA

INTRODUCTION

Modern industry conducts itself much differently than itdid at the turn of the century. Public safety and care forthe environment has gone from the bottom of the list ofgoals, right to the very top.

This has come about for a variety of reasons, includingincreased knowledge of the products, processes andservices offered together with the hazards associatedwith them, greater awareness of the effects on thepopulation and environment, and the introduction oflegislation to ensure compliance with the standardpractices necessary to ensure these goals are met.

For the gas industry, one essential way of providing thisservice to their customers is by odorization of natural gas.

Odorization of natural gas has evolved from afragmented, unregulated practice, into the current highlyregulated and monitored practice we see today. Theprimary focus of odorization is safety, and this must bekept in mind as we develop, maintain and improve ourodorization techniques and processes in a changingregulatory environment.

This paper will focus on the chemistry and characteristicsof gas odorants and will discuss related topics includingpipeline pickling and odor-fade.

HISTORY

Odorization of gas was first proposed in Germany in the1880s’ by Von Quaglios’ use of ethyl mercaptan as ameans of leak detecting the escape of blue water gas.

The manufactured gas used at the turn of the centurycontained by-products, which to some extent imparteda gassy odor to the gas. As high quality natural gasdisplaced lower quality manufactured gas, the by-products that caused the gassy odor in the lower qualitygas were no longer present.

Without these by-products, natural gas had little if anydetectable smell to warn of leaks or accumulation. Thisundetectable gas caused the disaster at the New LondonElementary school in 1937 that leveled the school, killingmany children.

The gassy odor of manufactured gas was originallyduplicated in natural gas by cheap refinery by-product

streams. However, these by-product streams wereunreliable and varied in quality. The growth of thechemical industry during World War II resulted in theavailability of high quality synthetic chemicals that provedwell suited for natural gas odorization. These chemicalsare the low molecular weight (C3-C4), branched chainedalkyl mercaptans, alkyl sulfides and a cyclic sulfide. By1960, virtually all natural gas odorization was done withblends of these synthetic chemicals.

ODOR FADE

Odor Fade can be a major problem. Gas may besatisfactorily odorized at source, but if it no longer hasthe necessary odor impact and intensity by the time itreaches the customer, escaping gas can go undetectedand result in a serious fire or explosion hazard.

To understand why it occurs and what can be done toovercome the problem, we have to consider thefollowing:

1. Odorant blend types and the chemistry of the variouscomponents

2. Pipeline conditions

3. The quality of the gas to be odorized

ODORANT BLENDS AND THEIR COMPONENTS

Odorant Characteristics

Odorant blends are extremely odorous, volatile,flammable liquids. Acceptable odorants must possesscertain physical and chemical characteristics. Theseinclude a gassy odor, low odor threshold, high odorimpact, resistance to pipeline oxidation and good soilpenetrability. Vapor pressure of blend components usedin vaporization type odorizers is also a very importantconsideration.

Odorant Components

The odorants used today are usually blends of two ormore components, which achieve the desirablecharacteristics.

Therefore, it is important to understand the characteristicsof the components. Basically there are three chemicalgroups from which odorants are blended:


1. Alkyl Mercaptans

2. Alkyl Sulfides

3. Cyclic Sulfide

Mercaptan Components

Tertiary Butyl Mercaptan (TBM)

CH3|

CH3 - C - SH|

CH3

TBM is the leading single component used in natural gasodorants. Its low odor threshold, gassy odor, good soilpenetration, and highest resistance to oxidation of themercaptans, make TBM very desirable. However, thehigh freezing point of TBM (34°F) results in the need forblending with other components to prevent freezing.Otherwise, TBM would be an excellent “stand alone”odorant.

Isopropyl Mercaptan (IPM)

CH3|

CH3 - C - SH|H

IPM has a strong, gassy odor and low freezing point(–202°F). Of the mercaptans it is the second mostresistant to oxidation. IPM is commonly blended withTBM to depress the freezing point while enhancing theodor impact. IPM is also a stand-alone odorant, butrarely, if ever, used as such.

Normal Propyl Mercaptan (NPM)

CH3 - CH2 - CH2 - SH

NPM is not a major component in odorant blends,typically 2-6%. It is more easily oxidized than othermercaptans. However, NPM has a low freezing point(–171°F) and a strong odor. NPM was originally a co-product in the IPM manufacturing process. It is not agood stand-alone odorant due to low oxidative stability.

Secondary Butyl Mercaptan (SBM)

CH3|

CH3 - C - CH2 - SH|H

SBM, originally an impurity in TBM manufacture, isprobably the least used component in odorant blends.On the rare occasions it is used, it is typically in the 2 –4% range. It is a branched chain mercaptan, whichresists oxidation. SBM has a strong odor, low freezingpoint, but high boiling point and low vapor pressure.Sometimes used at 100% in evaporative systems.

There is some evidence that SBM enhances the thresholddetection level of blended odorant.

Alkyl Sulfide Components

Alkyl sulfides are resistant to oxidation but they do nothave the odor impact of the mercaptans. They are notconsidered “stand alone” odorants. Their primaryfunction is to lower the freezing point of TBM.

Dimethyl Sulfide (DMS)

CH3 - S - CH3

DMS has been widely used as a blend component,particularly with TBM. DMS will not oxidize in the pipelineand has good soil penetrability. DMS has a much highervapor pressure than TBM; thus TBM/DMS blends arenot suitable for vaporization type odorizers.

Methyl Ethyl Sulfide (MES)

CH3 - S - CH2 - CH3

MES is the latest addition to odorant blends with TBM.MES will not oxidize in pipelines. MES has a vaporpressure similar to TBM; therefore TBM/MES blends aresuitable for injection or vaporization type odorizers.

Cyclic Sulfide

Tetrahydrothiophene (THT) or Thiophane

H2C - CH2_ _

H2C CH2\ /

S

THT is the most resistant to pipeline oxidation. It has agassy odor but low odor impact and poor soilpenetrability. The low odor impact makes it difficult toover-odorize with THT. THT may be used in pure form oras part of a blend with TBM. THT is a “stand alone”odorant.

Blend Composition

The odorant blends in use today fall into one of threemain categories, which are:

1. All mercaptan blends

2. Mercaptan/alkyl sulfide blends


3. Tetrahydrothiophene (THT) /mercaptan blends.

The following compositions (and minor variationsthereon) are the most common blend types in use today.Also listed is the type of odorizing equipment that canbe used.

All Mercaptan Blends

Component Blend 1(%)

TBM 79IPM 15NPM 6

Odorization Vaporization or LiquidMethod Injection

Mercaptan/Sulfide Blends

B L E N D SComponent 1(%) 2(%) 3(%)

TBM 75 80 10DMS 25 0 10MES 0 20 0IPM 0 0 70NPM 0 0 10

Odorization Liquid Vaporization LiquidMethod Injection or Liquid Injection

Injection

Tetrahydrothiophene (THT)/Mercaptan Blends

Component Blend 1(%) Blend 2(%)

THT 100 50TBM 0 50

Odorization Vaporization Liquid InjectionMethod or Liquid Injection

Pipeline Conditions and Gas Quality

Odorization, especially by an injection system, is anaccurate way of verifying that you have added therequired ratio of odorant to gas. However, there arecircumstances that occur within distribution systems thatcan mask the odorant level in the gas stream or causethe odorant to fade. There are basically three causes forthis phenomenon and they are the following:

Oxidation — formation of di-sulfides in the presence ofiron oxide (rust) and air (oxygen)

New pipe —

• Adsorption/absorption of odorant onto/into thesurface of synthetic (plastic) pipe

• Formation of patina layer inside steel pipe

Gas Quality —• Absorption, masking, or reaction of odorant

components with impurities in the gas stream.

Some of the causes of odorant fade are chemicalreactions whereas the others are physical phenomena.Let us explore the possible causes of odor fadementioned above.

Oxidation

The presence of rust and air within a pipeline will act asa catalyst on mercaptans causing them to oxidize intocompounds that have virtually no detectable odor. Ofthe common mercaptan odorants, the following listrepresents how they will react in the presence of a rust/oxygen environment:

TBM Most resistant to OxidationIPM —NPM Least resistant to Oxidation

All of the sulfide components (DMS, MES, & THT) usedin odorant blends are resistant to oxidation.

The solution: temporary increase in the odorantdosage rate

New Pipe

Plastic — the other potential cause for odor fade is aphysical reaction caused in the presence of new plasticpipe. In this case, the odorant is being adsorbed and/orabsorbed onto and into the plastic pipe. However, onceequilibrium is achieved, the amount of odorant going ontoand into the surface of the pipe wall equals the amountcoming back out. When this point is finally attained, odordetection with normal dosage levels should resume.


Steel - this same principle exists, to some extent, in thepresence of new steel pipe although a chemical reaction,not physical. However, introducing larger than normalquantities of odorant into the pipe at the start can picklenew steel pipe. Eventually an iron sulfide layer forms(patina) on the inside surface of the pipe and theconditions that would cause odor fade will diminish.


Gas Quality

The gas quality must also be considered wheninvestigating causes of odor fade. Is your gas supply?1. Dry - Not Naturally Odorized?2. Wet - Not Naturally Odorized?3. Dry - Naturally Odorized?4. Wet - Naturally Odorized?5. Peak Shaved Gas?


Dry Gas — Not Naturally Odorized

Dry Gas, not naturally odorized is the easiest to odorizeand does not cause odor fade. Any of the definedcommonly used odorant blends will perform satisfactorilyprovided that continuous odorization is practiced. Lowflow absorption may be an issue

Wet Gas — Not Naturally Odorized

Condensed liquids in the pipeline absorb odorantcomponents. Some odor masking may also occur dueto the odor imparted by the impurities in the gas. Bothgive rise to odor fade.

Odorants with the highest vapor pressure and lowestthreshold values work best. Blends high in IPM (with itshigh vapor pressure) are considered best in this situation.TBM blends work well in overcoming masking but arenot recommended where liquid levels are high.

Do not use THT or THT blends. Their low vapor pressureand low Kd values results in a higher degree of absorptionin pools of condensate resulting in more rapid odor fade.Also, if drier gas is later introduced, condensates with ahigh level of dissolved odorant can rapidly evaporate,resulting in overodorization of the gas stream.

Dry Gas — Naturally Odorized

Dry, naturally odorized gas can cause odor fade becauseit contains among others; methyl and ethyl mercaptan,which can cause oxidation of TBM to disulfides, whichhave low vapor pressure and low odor impact. As thelevels of natural mercaptans increase, it is best to usesulfide blends, which are oxidation resistant. THT blendsare best. DMS is oxidatively stable but lacks odor impact,and is not considered a “stand alone” odorant.

Wet Gas — Naturally Odorized

It is almost impossible to satisfactorily odorize this typeof gas. IPM based blends may work if liquids are the

main problem. THT blends may work if naturalmercaptans are the major contaminants. The bestsolution is not to purchase this quality of gas.

Peak Shaving

This practice which involves addition of propane dilutedwith air to natural gas results in a similar situation to thatof naturally odorized gas in which the ethyl mercaptanused to odorize the propane promotes oxidation ofmercaptan based odorants resulting in odor fade. Alsothe addition of oxygen and moisture increases thepossibility of mercaptan oxidation. So even if unodorizedpropane is purchased conditions for oxidation, albeitreduced, still exist. THT blends are considered best ifconditions are severe.

NOTE:Sulfides Oxidative Resistance — As previouslymentioned, alkyl sulfides (DMS, MES) and the cyclicsulfide (THT) are resistant to oxidation. However, THT isthe only sulfide that will act as an effective stand-alonegas odorant. Both DMS and MES do not possess therequired gassy odor and are therefore ineffective asstand-alone gas odorants. Additionally, both DMS andMES are typically used in minor concentrations (20%-30%) further reducing the chances of odor detectionshould 100% of the mercaptan in the blend be oxidized.

CONCLUSION

Hopefully, the information in this paper will increase yourknowledge of odorant behavior.

Always remember that safety of the public is the primeconcern. Proper odorization allows your customers tosafely use natural gas by providing an adequate warninglevel allowing them to recognize a leak, should one exist,prior to the gas reaching an explosive level.


Many gas pipeline companies struggle with lost-and-unaccounted-for-gas (L&U) and it can be a significantcost to their bottom as shown below.

As shown in Figure 1, by reducing L&U from 6/10 percentto 1/4 percent, a typical company with a 2 Bcf dailythroughput could achieve $10 million annually in bottomline benefits based on $4.00 gas prices.

NOTE: For simplicity, the formula assumed a Btu factor of 1,000.

FIGURE 1.

KEY CONSIDERATIONS

Now that you see the carrot, you need to consider thekey issues to achieving the goal, and making itsustainable.

• First, you will need management’s “buy in” and thecompany must be willing to make changes. Thosechanges involve implementing industry “bestpractices” for measurement and gas loss control,i.e.,– A really good, state-of-the-art, measurement

collection, balancing, and reporting system,– Pipeline segmentation for loss control,– Well trained people in both the field and

measurement support,– Regional Specialists,– The proper equipment for testing and calibration,– Good standards and procedures, etc.

HOW TO PERFORM A LOST ANDUNACCOUNTED-FOR GAS PROGRAM

Rick FeldmannQuorum Business Solutions, Inc.

13070 Richmond Ave. #1100, Houston, TX 77042

730,000,000 MMBtu Annual Throughput0.6% Average Percentage of L&U

4,380,000 MMBtu Lost Annually

730,000,000 MMtu Annual Throughput0.25% Average Percentage of L&U

1,825,000 MMBtu Lost Annually

4,380,000 MMBtu Lost Annually at .6 Percent

(1,825,000) MMBtu Lost Annually at .25 Percent

2,555,000 MMBtu Potential from Loss Reduction$4.00 Price per MMBtu

$10,220,000 Value Potential from Loss Reduction

• You will need a full-time project management teamthat fully understands the measurement process and“best practices.”

• You will need programs and checklists to ensure fulland adequate coverage of work.

• You will need someone with strong data miningexpertise, along with a thorough knowledge of gasmeasurement systems.

• Determine early on if you have the experts you willneed, or contract that experience.

TWO VERY IMPORTANT RULES TO FOLLOW

The first important rule to follow for a successful lostgas turnaround project is that you must have dedicatedexperts who make no assumptions.

You cannot have field technicians audit meter stationsunless you are sure of their expertise on all of the devicesthey will encounter, and that they will do a complete auditof the stations. Complete means to review everything,from technician skills, equipment used, certifications, andall equipment present at the site.

And you cannot have measurement systems supportpeople review their own procedures, processes, andsystems use. To not use outside expertise in these areaswill doom the project from the start.

One of the main benefits from performing an L&UTurnaround project with outside contract assistance isthat the outside contractors bring with them extensiveindustry experience and “best practices” that guide theprogram, train the field and office technicians, and outlinea reinvestment program that provides value in return forthe dollars spent.

The other very important rule to follow is that theprogram must be organized and a thorough “A to Z”program.

ONE ASSUMPTION YOU SHOULD MAKE

The one assumption you should make at the beginningis that you will find problems across all processes,systems, designs, operational procedures, people skills,test equipment, etc. This is typical and that is why youneed to perform a complete, thorough review with peopleyou are sure are experts.


FIELD AUDIT PREPARATION

We focus on determining which meter stations to auditas soon as possible because it takes significant leadtime to schedule the field audits with third partywitnesses, and to match the normal operating schedule.

There are two major considerations in selecting meterstations for field audits and you should accomplish both:

1. Getting the biggest bang for the buck.

2. Getting an oversight on all types of meters, EFMs,people, divisions, etc. across the entire system.

Factors that should be considered in selecting meterstations for field audits that give the quickest reductionsin L&U are:

• High volumes stations should be audited firstbecause of the 80/20 rule where 80% of the volumesmay be flowing through 20% of the stations.

• Audit delivery meters before receipt meters because,99 times out of 100, measurement error caused byphysical devices will record less gas than actuallyflowing through the meters. Exceptions shouldhowever be made if large volume receipt stationsare suspected of pulsation problems.

• Consider performing full gas plant audits to includeall gas and liquid meters if gas plants are not isolatedfrom the pipeline system balancing with inlet andtailgate meters.

• Include large volume receipt meters near recipcompressors because of potential square root error.

• Include stations using V-Balls for flow control.

• Include stations reflecting problems found in metertest/inspection reports, especially very dirty platesand liquids in the runs.

Other factors for selecting meter stations for field auditswill depend on what is found through analysis of thephysical balancing and reporting system, review ofsystem balancing reports, and other data analyses.

Once you make the initial selection, make sure you havecovered:

• All types of meters (orifice, turbine, rotary,ultrasonic, etc.)

• All types of station designs (chromatographs,accumulated samplers, flow control, run control,filter separators, bidirectional meters, etc.)

• All districts or regions because of differentprocedures, assignments, skill sets, etc., and

• Some non-custody meters used for pipelinesegmentation.

You should ensure a full coverage as shown abovebecause of differences across the system. You don’t needto visit every site that is identified with the same problem.After seeing a small number of sites with the sameproblem, you can most likely determine what may becausing the problem(s) and therefore give direction tofield technicians on how the problem should be fixed.

PREPARING FIELD AUDIT PACKAGES

The following documents should be printed, evaluated,and sent out to the field with the field audit teams. Aninitial evaluation of these also serves to fine-tune thestations selected for field audits.

• Current configuration log (EFMs)• 3 most recent months of event logs & error logs

(EFMs)• 3 most recent months of hourly flow data (EFMs)• 12 most recent months of charts (Dry Flows).• 12 most recent months of Meter Volume

Statements.• 3 most recent Meter Test/Inspection/Proving

Reports• Meter Change Reports associated with the

period beginning with the first Test/Inspection/Proving Report.

• Copies of the station design schematics, iflocated.

• A copy of the original meter mic form.

A review of these documents may provide clues tosignificant problems that would benefit prioritizing thefield audit performance.

SCHEDULING LOGISTICS

It is extremely important to notify gas control of plannedfield audits because a station can only be fully auditedwhen gas is flowing. It is also a good practice to havegas control operate station actuators and flow gas underboth normal and unusual conditions that may be foundin the station historic data.

Important factors in scheduling the audits are:• You want to witness the technicians normally

assigned the stations so you can evaluate theirknowledge and skills, and their equipment usedfor testing and provings.

• You want to audit the station at the normal timesfor testing and provings to both reduce theamount of field time required, and to see thestations in there normal condition (You do notwant to see plates just cleaned a couple of daysbefore the audit).

• You want to follow an organized travel scheduleto reduce the amount of travel between stationsbeing audited.


ANALYZING SYSTEM DATA

Another data source for selecting stations for field auditswill be analysis of the measurement system data. Thisshould begin on the first day of the engagement. UseSQL, or another database query tool, to analyze thesystem data for anomalies. Consider including thefollowing in the data mining analysis:

• Default factors such as 1000 Btu, 60 degreestemperature, etc.

• Meters showing a change in plate size withoutan appropriate change in differential.

• Meters showing differential or pressureexceeded the respective spans.

• Meters showing differential or pressure < 20%or > 80% of their spans.

• Meters showing flow > 1440 minutes for any dayexcept for the change back from daylightsavings.

• Meters showing flow time evenly divisible by 60minute increments.

• Meters where the calculated flow extension((DP*AP) ^.5) <> 2% of reported flow.

• Meters where the flow extension * C Prime variesfrom the final volume by > 2%.

• Meters where the adjusted energy factors variesby > 2%.

• Meters showing DP, SP, flow time, and flowextension but reporting zero flow.

• Meters showing zero DP, SP, flow time, and flowextension but reporting flow.

• Btu values remaining unchanged from one testto another, or where chromatograph readingsappear unchanged.

• Meters where original and adjusted Btus vary >2%

• Gas qualities where Btu changed > 2% or gravitychanged >.005 compared to the previoussample.

• Meters reporting flow when DP is below the lowpoint cutoff.

• Fuel meters with expected flow showing zeroflow.

• Meters with beta >.6 without flow conditioners.• Meters with hard edits.• Meters where original and final volumes changed

>1%.• Meters flagged to recalculate and overlay EGM

data.• Meter ids and orifice plate sizes that are exact

standard or default sizes.

ANALYSIS OF THE MEASUREMENT SYSTEMBALANCE REPORTING

Another source of information for both selecting stationsfor audits and for determining overall accuracy ofreporting is the system balance report.

The report should follow the format shown below:

FIGURE 2

Key items for consideration:

• Consider the reasonableness of the fuel beingreported. Is all fuel metered and the meters reflectedin the balance report, in the right segments? If somecompressors operate without fuel meters, is the fuelcalculated and reported under dummy meters in thebalance report?

• Does the report include known “gas to atmosphere”for blowdowns, blowouts, flared gas, etc.? If not, itshould be. Estimates can be included for all juniorruns tested each month to cover the gas released toatmosphere. This should be done by segment.

• If liquids (hydrocarbons and water) are beingcollected and hauled away, they should be reflected.Use best estimates to equate Mcf and MMBtu tothe amounts of liquids believed removed eachcalendar month, and periodically verify with runreports.

• Does the report show changes in linepack from onemonth end to another? Is it reasonable and is itreflected within each appropriate segment? If not,you will want to begin doing this. You will want toreview the formulas used along with the points,pressures and temperatures are taken along pipewith major swing potential.

• Compare the Mcf and MMBtu gas loss percentages.Note that if they track (fairly close from month-to-month) both in the segments and for the entirepipeline, then you should suspect no seriousproblems with gas quality determination. If they varysignificantly however, you will need to consider a

Mcf MMBtuCustody Receipts

LessCustody Deliveries

Fuel Used

Known Gas to Atmosphere

Hydrocarbons & Water Removed

Change in Linepack (+/–)

Lost & Unaccounted For Gas

Percentage (L&U / Receipts)


very thorough review of how gas quality isdetermined and applied.

Another key piece of information from the monthlysystem balance report is the gas loss percentages. Whencomparing these percentages, you should suspectindependent problems with “quality errors” if the twopercentages do not track.

These problems will be identified through office dataanalysis and field audits, with one exception. Thatexception is the assignment of chromatographs tometering stations.

The proper time to analyze chromatographassignments will be on the conclusion of the MeterConfirmation Process, when you know the locationof all stations and what is immediately upstream anddownstream of each station.

THE METER CONFIRMATION PROCESS

When beginning a Lost Gas Turnaround Program youshould ask yourself the following questions:

• Are all meters that should be included in thebalance report in the report?

• Are they in the correct segments?• Are they reflected correctly as receipts and

deliveries?

To determine the answers, you will need to confirm theaccuracy of the system data with the field technicians.

All key meter data should be extracted from the systemdatabase, analyzed and sent to the field for confirmation.Include data on meter type, characteristics (EFM?,chart?, size run?, size plate?, etc.), purpose (custody?,check?, zone balancing?, receipt?, delivery?, etc.),location, etc.

Strongly consider using a “positive confirmation” processthat requires each technician to review, confirm, orcorrect information, and return the form(s).

Also ask the field technicians to:• Identify the meters immediately upstream and

downstream of each meter.• Identify meters they know of that are not included

in the system list.• Identify any known problems at each meter

station (i.e. equipment not working, pulsation,liquids, not designed correctly, etc.)

During the normal course of a Gas Loss TurnaroundProject there is an ongoing office review team and a fieldaudit team. This confirmation process should begin assoon as possible and will continue until completed. Theoffice team will perform the confirmation process butwill coordinate with the field audit team for assistance,

to identify reported problems, and to seek additionalinformation as necessary.

SYSTEM EDITOR REVIEW

In addition to supporting the field audit team throughoutthe project, the office review team will need to reviewand evaluate the use of the system editors.

Measurement systems sold by third party vendorscontain edit capabilities to both identify and reportproblem data, missing data, etc. These third partysystems also have the capabilities of setting ranges onmeters to identify potential problems with Btu factors,differentials, static pressures, temperatures, etc. Verifythat the editors are being used appropriately, and actedon timely.

If the measurement system was developed in-house,verify that it contains full edit capabilities and the abilityto monitor anomalies as discussed above.

Data editors should address whether data received isaccurate, complete, etc. Some are included in the listpresented in the section titled “Analyzing System Data”on the previous page.

PROCESS REVIEW — MEASUREMENT SUPPORTGROUP

The office team should also perform a detailed review ofall processes performed by the Measurement SupportGroup from a monthly close through a subsequentmonthly close.

Key questions you will be seeking answers to are:• Is the group adequately staffed?• Do they have the appropriate skills?• Are they receiving adequate training in new

systems and equipment?• Do they know the pipeline system, segments,

stations, meters, etc.?• Are they fully capable of using the system

balancing system and its full functions?• Are they working correctly, efficiently, and timely

to identify potential errors, control gas losses,and close data to the allocation or sales systems/

• Do they work appropriately with field techniciansand measurement specialists?

• Are they auditing third party measurement data?

REVIEW — OUTSIDE CONTRACT SUPPORT

If support services, such as chart integration, has beenoutsourced, consider having a selection of charts re-integrated by other service providers to test the accuracyof the current service.


FIELD AUDIT PERFORMANCE

GUIDELINES FOR THOROUGH FIELD AUDITS

The number one guideline to ensure that a station auditis complete and accurate is “never assume” anything.Don’t assume the technician’s equipment is accurate.Don’t assume the technician knows how to perform hisor her duties correctly. Don’t assume the EFMconfiguration data is correct, etc.

Some key guidelines for performing Lost Gas TurnaroundProjects include:

• Determine early if you will be developing an ongoingMeasurement Specialist Group (see the later sectionon Measurement Best Practices). If so, involve themearly in the process of station selection, logistics,etc.

• If you are unsure of the knowledge and capabilitiesof your field auditors (potentially MeasurementSpecialists in training), contract expert assistancefor a minimum of one week to assist and train eachfield auditor. A mistake here can doom the entireprogram.

• The audit team should carry certified meter testequipment as a backup should they find thattechnicians don’t have proper equipment, or that itisn’t certified. The equipments should be as accurateas the transducers being used.

• Carry a Square Root Error (SRE) indicator, borescopes, and different size micrometers.

• Carry extra orifice plates and seal rings of standardrun sizes should you need to make quick fixes andfind the technicians don’t have the items.

• Carry plates with very small holes, below minimumbeta, to determine if gas is flowing below cutoffpoints.

• Carry a couple of gas sample bottles should the needarise to take a spot sample and to use when atechnician doesn’t have a sample bottle on hand fordemonstrating his or her skills in taking samples.

• Carry sample bottles to capture liquids for lateranalysis, should it be necessary.

• Carry a previously determined certified gas sampleto be run against chromatographs as an “unknown”should you discover that certified gas being used ata site is not appropriate or old and doesn’t calibratecorrectly.

• Carry a digital camera and laptop to capture picturesof each station and problems identified. The pictureswill aid later discussions, reporting, and station fixes.

• Assign a minimum of two “trained” auditors to eachaudit team. If using “Regional” MeasurementSpecialists, one should be from the region underaudit and the other should be from an outside region.This helps eliminate mistakes that may be createdby technician or station familiarity.

• Witness the technician first in all tests, then correctand teach later. It is very important to observe eachtechnician to evaluate his or her knowledge andskills. You may want to consider having a memberof the training staff assigned to assist on the auditso they can make first hand judgments also on thetraining the technicians need and are getting.

• If a member of the training department accompaniesthe field audit team, he or she should be assignedthe additional duties of 1) logistics for the week, 2)getting lunch for the team so they can eat at thestation and not lose valuable time, 3) taking digitalpictures of the station and its problems, 4) overseeingthe end-of-day write-up of each station, 5) makingaudit checklists available, and 6) assisting in takingmeasurements of each meter run.

• Consider performing field audits on a four-day week,Monday through Thursday. This allows the fieldauditors time on Fridays to conduct other necessarybusiness, to finish all write-ups from the weekfinished, and time to review plans for the next week.

• Consider having “Sunday Night Supper Meetings”where the field audit team meets the area managersand technicians being audited during the followingweek. Be open about the process, the extensiveamount of time required at each station, and thatthe audit is not a “witch hunt” to identify anyindividual poor performers.

• Begin each morning by meeting the assignedmeasurement technician early to 1) review stationschematics, 2) review original micing sheets if keptin the field, 3) to review his or her equipmentcertifications, and 4) to resolve questions the officeteam may have on the meter confirmation process.

• On completing the write-up at the end of each day,review the audit package for the next day.

Considering an audit team consists of two MeasurementSpecialists (one possibly an outside contractor to trainthe Specialist) and a member of the training department,predetermine responsibilities.

The first decision is to determine who is “in charge.”Normally that person would be the one with the mostexperience. If both Specialists are equally experienced,assign that responsibility to the Specialist from outsidethe region and most unfamiliar with the technicians orstations.


The Specialist with the least amount of familiarity to thestation technician should be responsible for reviewingthe technician’s testing (meters, chromatograph, spotsamples, etc.). The unfamiliarity makes it easier to bethorough and not overlook steps.

THE USE OF CHECKLISTS

Based on experience, it is extremely valuable to usechecklists throughout the conduct of the field audits.They will serve to make documentation easier, to list AGAstandards, and to ensure nothing is overlooked.

The standard checklists used by Quorum and a numberof major transport companies include:

Technician’s Test EquipmentStation DesignMeter Tube - OrificeMeter Tube - UltrasonicMeter Tube - TurbineMeter Tube - RotaryOrifice PlatesGauge LinesGas SamplersChromatographsUltrasonic Meter TestOrifice Meter TestTurbine Meter Test/ProvingRotary Meter Test/ProvingSRE and GLE

The checklists should contain the industry standards andguidelines, along with the procedural steps for testing.They should also contain sufficient space fordocumenting information and problems.

INITIAL ASSIGNMENTS

Basic assignments for a field audit would include:

Lead Auditor (Specialist)

The Lead Auditor should make all decisions and beresponsible for witnessing the meter tests and provings,chromatograph tests, and technician’s qualifications. Ifthe Lead Auditor determines that the technician’sequipment may not be accurate, he should have thetechnician perform the first test with the normal testequipment. It should then be re-performed with the auditteam’s backup equipment, and that backup equipmentshould be used for the balance of the tests.

The Lead Auditor should have the technician performthe first test on his or her own without interruption. If notdone correctly, correct the technician and work togetheron the remaining runs. This similarly applies to testingthe chromatograph.

Assistant Auditor

The Assistant Auditor should walk the station and reviewit against the schematics. He is responsible for lookingfor leaks, checking valves, and reviewing all peripheralequipment such as filter separators, flow control devices,run control devices, dampening bottles, etc.

With the assistance of the Trainer (or someone else), theAssistant Auditor performs measurements andcompletes the appropriate checklists for the stationdesign, meter runs, orifice plates, gauge lines, gassamplers, etc.

Trainer (or other assistant)

The Trainer, in addition to the duties discussed earlier inthe “Guidelines for Thorough Field Audits” will take digitalpictures of the station, runs, and problems noted by thespecialists.

He or she will also assist in taking measurement and willcall Gas Control to change flows as necessary to properlytest all equipment under normal and extreme operatingconditions.

SPECIAL NOTE ON THIS PAPER

It is not the author’s intent in this paper to provide adetailed step-by-step list of procedures for performingmeter tests and provings, nor to list all of the AGA andindustry standards. These are understood in the industryand should be documented in the checklists.

From this point forward the discussion on field auditswill address key points and special review work.

KEY NOTES — STATION DESIGNS

It is important to take pictures of the station design, andespecially the upstream and downstream headerconfigurations.

A review of the station schematics will identifyunderground header configurations which should bedrawn on the Station Design Checklist if you can’t makea photocopy.

KEY NOTES — METER RUN DESIGNS

The meter runs need to be measured to ensurecompliance with AGA specifications. The measurementscan be handwritten on the checklist designs, or writtenon “stickman” drawings (should your checklists notcontain run pictures with specifications).

In addition to taking measurements, ensure:

• Straightening vanes, or other flow conditioners, arein place, and in the right location (verification ofcondition will be accomplished by bore scoping or


removing “end caps”).

• The temperature probe and test well are in the rightlocations, set to the right depths of the pipe, andcarry a proper medium.

• The static pressure tap is in the proper location anduses the right size tubing.

• The differential taps are in the proper location onorifice runs and are not configured with takeoff tees.

• Gauge lines on orifice meters, if used, slope correctlyand are the right size tubing.

• The gas analysis tap for chromatograph oraccumulated sampler, and spot sample tap, are inthe right locations and are also of the right depth.

• The sample line for the chromatograph or samplershould be proper size tubing and the shortestdistance possible from the run to the chromatographor sampler. There should be no liquid traps, and theline should be “heat traced” to a chromatograph.

• All valves should be “full opening.”

There may be other devices installed on meter run piping,such as flow regulators, valves, controllers, etc. Theseshould be properly located on piping, off the run, so notto interfere with a conditioned gas flow.

Other notes:

Grandfathered Runs

Just because the AGA has grandfathered some rundesigns doesn’t mean that they are accurate. Determinethe purpose and accuracy required. If justified, replaceold grandfathered runs being used for custodymeasurement. Those runs can be used for zonebalancing meters.

Oversized Orifice Runs

There are a lot of twenty inch runs being used and theseare not accurate for custody measurement and shouldbe considered only for zone balancing.

Pipe Welds

A number of companies have made their own meter runsor made changes through welding. Cutting and weldingon a meter run will create measurement problems andcan cause significant measurement error.

Fittings

While taking measurements and checking runs forcompliance, leak test all fittings.

KEY NOTES — ORIFICE PLATES

Be prepared to take a close up digital picture of both theupstream and downstream side of the orifice plate as itis removed.

Pictures of dirty plates provide evidence of problems inthe gas stream, support estimating potentialmeasurement error, and show evidence on how the gasflows through the meter through striations.

Consider using the special plates with small holes totest for flowing gas below cutoff points.

Consider changing plates as necessary to maintaindifferentials between 20 and 80 percent of the transducerrange. Alternative procedures would include realigningrun control, if used, and changing transducer ranges.

In addition to testing the plate don’t forget to mic theseal rings.

ON CONCLUDING TESTS AND PROVINGS

On concluding all meter tests and provings, the LeadAuditor should have the technician verify the accuracyof the characteristics loaded in the EFMs.

The Run ID should be verified to both the mic sheet andthe flange stamp. They should all agree.

Another step to perform with orifice runs, not part of aroutine test, is to do a “field pulsation test” by locking ina true zero, closing both the high and low dp sides andobserving to see if there is any differential shift in theEFM.

BORE SCOPING RUNS

Be prepared to blowdown and bore scope runs:• If you need to determine if straightening vanes

are in place.• If you are unsure when the tubes were last

cleaned and you need to determine the conditionof the tubes.

• If you note dirty plates.

Note that it is not necessary to bore scope all runs offthe same common header. The decision on which runsto look at, if looking only for dirt and/or liquids, shouldbe based on 1) the amount of flow, and 2) the run at thefarthest end of a header.

When gas flows equally through a series of runs off acommon header, dirt and liquids tend to accumulatemore at the last run off a common header.


TESTING FOR SQUARE ROOT AND GAUGE LINEERROR

Earlier we discussed performing the field test for potentialsquare root error. If you detect any pulsation with a visibleshifting of differential during the field test, if the stationis located near any reciprocal compressors, or if yoususpect any noise in the lines due to misaligned flowregulators, perform a square root error test.

Take three separate readings and average them for thesquare root error at the flange taps. When transmittersare not close mounts, the steps should be repeated atthe ends of the gauge lines. Gauge line error is thendetermined by subtracting the second averagecalculation from the first calculation of square root error.Note that the acceptable level for square root error isset at 1/10 percent.

MISCELLANEOUS DEVICES

Don’t overlook miscellaneous devices around themetering station. These may include the following:

Scrubbers and Filter Separators

These would normally have been installed to remove dirtand liquids. If you note dirty gas or liquids on plates,check these devices to see that they are workingcorrectly. They may be clogged and not dumpingautomatically or filters may be sized incorrectly to beworking effectively.

Run Actuators

Test these to see that they are set correctly to performrun control based on correct differential ranges. Alsoverify that they open valves slowly enough to even outpressures and not dish plates.

Flow Control Devices

Ensure these are properly located to not affect the gasflow upstream of measurement. Also, if you find morethan one used side-by-side check to see that theyoperate in tandem so not to cause noise in the gasstream.

Pressure Regulators

Ensure these are properly located to not affect the gasflow upstream of measurement. Also check the pressuredrops to ensure they aren’t significant to cause water orhydrocarbon liquids to drop out of the gas stream.

FOLLOW-UP STEPS FOR DIRTY GAS AND LIQUIDS

When observing dirty plates and liquids in the gas stream,it becomes necessary to isolate the source of the problemand correct it.

This may require determining the makeup of liquids: Is itwater, hydrocarbons, compressor oil, or something else?You may have to send a sample to a lab to get afingerprint of the manufacturer as a way of determiningwhere it comes from.

QUANTIFYING THE RESULTS AND MAKINGRECOMMENDATIONS

Quantifying the results is as simple as multiplying themeter’s throughput by the potential error factor that canbe extrapolated from various studies that have beenpublished over the years.

Most primary element problems will cause the meters torecord lower volume deliveries than actual. An exceptionis square root error that will have an opposite effect inthe lower ranges.

Errors in secondary recording devices can be more easilycalculated by running laptop flow calculations with boththe wrong and correct data.

The errors can be looked at in two ways: Either as anamount lost that has a direct relationship to the amountof gas being retained for L&U on a percentage basis, orby multiplying the losses by a conservative price perMMBtu.

Once these calculations are determined, you should buildvalue propositions by developing the offset costs forimplementing the recommendations. Those with the bestvalue propositions should be implemented first.

KEY BEST PRACTICES

Below are two of the more significant industry bestpractices to reduce and control L&U.

Accountability

Leading pipeline companies have made operating teamsaccountable for controlling the L&U in their responsibleareas. Some of the key components to this concept are:

• Teams are awarded bonuses when they achievepreset results,

• They are given equipment and training necessaryto achieve and control accurate measurement,

• The pipelines have been segmented and theyare aligned with specific pipeline segments andmeters,

• They are given access to the physical measurementsystem data,


• They have Regional Measurement Specialists toassist them in both the field and office.

Measurement Specialists

A large number of companies have created positions ofregional or division measurement specialists whoseresponsibilities include:

• Meeting routinely with facility planners andengineering to review new station designs andexisting design changes,

• Meeting routinely with and gas control to identifypotential measurement problems on thesystems,

• Assisting the training function through fieldhands-on training and the certification of skillsets,

• Leading routine field audits on large volumestations, new stations, and stations that haveundergone change,

• Working with field technicians and gasmeasurement services to design systembalancing controls, and to monitor gas lossesacross the systems.

• Meeting regularly with gas measurementservices to monitor system changes andreporting, and

• Participating in (and guiding) emergencyresponse audits if significant errors occur.

People assigned these responsibilities need to beknowledgeable of:

• AGA specifications and industry measurementstandards,

• measurement equipment and techniques,• how to access and use the physical

measurement balancing and SCADA systems,and

• how to perform analytical analyses.

BACKGROUND ON THE AUTHOR

The Pipeline and Gas Journal recognizes Rick Feldmannas “one of the industry’s leading experts on lost-and-unaccounted-for gas. He has over ten years experienceperforming “Lost Gas Turnaround Services” on contractfor most of the major pipeline transportation companies,in addition to gathering and processing companies, anda number of LDCs.

His credentials include:

“Lost & Unaccounted-For Gas: Chasing the Silver Bullet,”July 1998, Pipeline & Gas Journal

“Controlling Lost and Unaccounted For Gas inDistribution Systems,” July 2000, Pipeline & Gas Journal

Accounting & Operational Issues with Lost &Unaccounted-For Gas, a 60 minute live televisedbroadcast, March 9, 2001, Southern Gas AssociationCorporate Telelink Network

“How to Implement a Successful Lost Gas TurnaroundProject,” July through October 2002, (4 articles in a 4part series), Pipeline & Gas Journal

How to Implement a Successful Lost Gas TurnaroundProject, November 14, 2002, a 90 minute live televisedbroadcast, Southern Gas Association Corporate TelelinkNetwork

Rick is also the author of an industry white paper thatidentifies the industry’s “Best Practices for MeasurementControl,” a number he helped develop.

Rick leads the Process Solutions Group at QuorumBusiness Solutions, Inc., a company with over 150professionals supporting the natural gas industry withoffices in Houston, Dallas, and Calgary. Quorum is theleading provider of integrated computer softwaresystems for the natural gas industry, and is best knownin the industry for its “TIPS System” for gas plants.

Rick Feldmann


AUTOMATION SYSTEMS FOR GAS TRANSMISSION ANDDISTRIBUTION PIPELINES

Doug OsburnAutomation Solutions Inc.930 Gemini, Houston, TX 77058

The automation systems that control and measure naturalgas flow in transmission and distribution pipelines ofteninvolve two systems. The system that controls the gasflow in a pipeline is called a SCADA (Supervisory Controland Data Acquisition) system and an AMR (AutomatedMeter Reading) system measures the amount of gasflowing into, and out of, the pipeline. These automationsystems which can be completely separate or combinedare widely distributed throughout the service area of apipeline and must rely heavily on long-distancecommunication technologies to telemeter the datanecessary to coordinate and monitor pipeline activities.This is the most challenging aspect of the automationproblem as the communication system is the mostexposed component to service interruptions anddegradation. Strategies should be implemented in thesystem design that will maintain continued safeoperations of the pipeline during periods ofcommunication failure.

SCADA SYSTEMS

Traditionally, SCADA systems have been distinguishedfrom plant automation systems as they are designed toautomate and monitor assets that are widely distributedgeographically. SCADA systems are commonly deployedto automate remote assets that are distributedthroughout cities, states, continents, or on a global basis.The control and monitoring activities necessary tomaintain appropriate operation of these distributedassets are usually performed by field devices that arelocated at each site. The performance of the field devicesand the assets they control are monitored and supervisedfrom host computers at a central location such as acontrol room.

Typically, host computers provide visual performancemeasures to allow operational personnel to evaluateasset conditions and status. The data representing theseconditions and status is received by the host computerfrom the field devices over a telemetry system. Data entrymethods are also provided to enable adjustments tocontrol targets and thresholds in the field devices tomaintain appropriate operation of the assets. In thisdiscussion of SCADA systems there are three major areasof focus.

• Field Devices — Field devices are equipment that islocated at a remote asset to perform on-site control,data acquisition and monitoring.

• Telemetry System — The telemetry system connectsthe field devices and the host computer.

• Host Computer(s) — Host computers providecollection, presentation and analysis of data fromthe field devices. In addition, the host computerssupport data entry methods to enable supervisoryadjustments to the control that is performed by thefield devices.

AMR SYSTEMS

AMR systems gather metering data that documents thepurchase and sales of natural gas. This metering data iscrucial to the transmission or distribution company as itdocuments how much gas has passed through a point,a custody transfer point, on the pipeline and in doing sohas changed ownership. The metering data issubsequently passed on to accounting departments forvalidation and billing purposes. AMR systems havearchitectures similar to SCADA with field devices thatmeter gas flow at distributed assets, host computers thatcollect the meter data, and telemetry systems thatconnect the field devices and host computer.

FIELD DEVICES

Field devices that were used in early SCADA systemswere primary data acquisition devices with very littlecontrol capabilities. As microprocessors and memorycosts became more affordable more control capabilitieswere incorporated into the field devices. Thesecapabilities are modifiable through programming orconfiguration to address different automationapplications. This autonomy decentralized control andimproved the overall reliability of the system by reducingcommunication loads on the telemetry system and bylocating control at the remote site where it cannot beinterrupted by communication failure or degradation. Incontrast to SCADA, field devices used in AMR systemsusually only perform data acquisition and the calculationsnecessary for metering natural gas.

There can be two fundamentally different kinds of datain SCADA and AMR field devices. Users typically thinkof data that is updated in the field device on a scheduledbasis as control and data acquisition functions areprocessed. When this data is communicated to the hostcomputer the most current value of the data is sent. Thisdata is often referred to as real-time or current data.

In some applications however, the field device saves datavalues along with the date and time at which the datavalue was sampled, in its local memory. This data maybe retained in the memory of the field device for the


purposes such as auditing gas meter calculations. Oftenthere is also a relationship between the values and timestamps that is significant to interpreting the events thatgenerated the data. The block of data is referred to as a“data set”. An example of a data set could be a snapshotof a spectrum of vibration data associated with a pieceof rotating machinery. Another example could be thecycle of an oil field rod pump which is called a “Card”. Inall of these cases the loss of a piece of data can reducethe integrity of the whole block of data. As a result, thecommunication of data sets such as gas metering auditdata from field devices to host computers requiresspecial consideration.

RTUsThe field devices that are used in SCADA systems havetraditionally been called RTUs, Remote Terminal Units.Today RTUs are generally hardened, compact;computers that can be installed in environmentallyexposed locations. Other industrial devices such as PLCs(Programmable Logic Controllers) which were originallydeveloped for automotive and discrete partsmanufacturing have also been used in SCADAapplications where less stringent environmentalrequirements allow. SCADA specifications such as thosebelow can have an impact of the selection of RTUs versusPLCs.

• Operating temperature range: -40 to 85 DegreesCentigrade

• Low power consumption: Less than 100 mV• Class 1 Division 2 hazardous area certification

Additionally, RTUs have other internal data managementdifferences from PLCs such as the capability to collectand communicate data sets, which supports moreefficient usage of a particular telemetry method.

RTUs generally have a flexible mix of I/O to allow a RTUdesign to be applied to different automation applications.

• Analog Inputs: 0 to 5 VDC, or 4 to 20 ma• Analog Outputs: 0 to 5 VDC, or 4 to 20 ma• RTD inputs interface platinum resistance

temperature measurement devices which arecommonly used in EFM applications.

• Pulse inputs for which the SCADA standard is 0 to10 kHz

• Digital Inputs and Outputs can support DC or ACsignals however, AC inputs/outputs can requireinterfacing relays which can be problematic inmeeting hazardous area classifications.

The I/O supports data acquisition to monitor theconditions and status of an asset, and also supportsadjustment of control elements such as heaters, motorsand valves to affect changes in the measured conditionsor status. RTUs usually do not initially have theintelligence to perform these data acquisition and controltasks. A user that has knowledge about how to operatea facility or piece of equipment must first enter thisknowledge into the RTU by programming or configuring

the device for the targeted application. This programmingcan include complicated control algorithms, alarmdetection and notification strategies, and communicationlogic. The overall performance and success of theSCADA system can be greatly affected by the quality ofthis programming.

Some RTUs have the capability to perform control anddata acquisition algorithms and gas metering calculationssimultaneously. Combining these functions into onedevice can significantly reduce the installation andcommunication costs.

EFMS AND CORRECTORS

EFMs and gas correctors are field devices that narrowlytarget gas metering. The automation logic and I/Ocapabilities of these devices are specifically tailored toperforming gas metering calculations in compliance withstandards that are sanctioned by AGA, American GasAssociation. The AGA calculations compensate for thechanges in density that occurs when a gas changestemperature and pressure. The result of the calculationis a mass flow rate with units of mass per time such asmillion pounds per hour. In addition to the AGAcalculations federal and state regulatory agencies requirethat gas metering devices store the raw data that is usedin the density calculation for 35 days. This archived datais made available to host computers through thetelemetry system for the purpose of auditing thecalculations performed by the metering device.

EFMs and correctors usually have environmental andpower requirements similar to those establish for RTUs.They are often installed in remote locations off the gridwhere solar cells and batteries are the only poweralternatives. The power requirement in some gascorrectors is small enough to allow them to operate overextended periods on D-size batteries with no rechargingsystem.

TELEMETRY SYSTEM

Telemetry systems in SCADA provide a long-distancedata connection between the field devices and the hostcomputer. Traditional telemetry methods have includedcommunication over leased telephone lines which aresimply called “Leased-lines”, radios of all types,microwave systems and satellite. The advent of long-distance networking standards for the Internet however,has made a big impact on SCADA communications.Routers, bridges, firewalls, and access servers haveextended corporate intranets further into the field. Thesefar flung networks based on IP technologies have broughtgreater performance and reliability than the previousSCADA technologies. Emerging standards in the cellphone industry such as CDMA and GPRS providewireless connections between a field devices and theInternet. Once the field device is connected to theInternet a host computer is able to communicate withthe device from any location.


The key consideration in the design of a SCADA systemis that communication interruptions and performancedegradation will occur regardless of the technology. Thisis in part due to the many points of exposure toenvironmental and human factors in long-distancecommunication technologies.

HOST COMPUTERS

The host computer system can involve one computer ormany networked computers which are connected to thefield devices by the telemetry system. The purpose ofthe host computer depends on the mission of the system.In SCADA systems the host computer provides a meansto monitor and evaluate the performance of the controlsbeing performed at the remote assets. In addition dataentry methods are provided to enable adjustments tocontrol targets to correct for deviations from the desiredoperating points. Generally, this activity is performed froma central control room that is typically manned byoperations personnel on a 24/7 basis.

In simpler SCADA systems the host computer may runautonomously in an unmanned facility but is designedto notify appropriate personnel through annunciation,paging, telephone call-out and email when performancemeasures exceed alarm tolerance. The system shouldalso aid service personnel responding to the alarm withdata to isolate and diagnose the problem.

These SCADA host systems are often based on HMI(Human Machine Interface) software products thatprovide tools for the development of interactive graphicsrepresenting conditions and status received from the fielddevices. Additionally, HMI products provide alarmindication and annunciation capabilities, data entrymethods to change control targets and alarm thresholds,and data logging capabilities.

AMR host computers have requirements similar toSCADA host computers for alarming, presentation andfunctions. However, the primary mission of the AMR hostcomputer is to collect the substantial amount of auditdata that is retained in the EFMs and correctors. Thisdata is received by the host computer from the fielddevices over the telemetry system and is subsequentlypassed on to gas accounting for data validation andbilling purposes.

SCADA COMMUNICATION STRATEGIES

The measure of the amount of information that can beconveyed over a given telemetry method is call the“bandwidth”. The bandwidth of a long-distance telemetrysystem is a commodity that has significant value and thatcan be purchased and sold. One can think of a telemetrysystem as a data transportation system or an informationpipeline through which information can be shipped. Thebandwidth of this data transportation system is a measureof the greatest amount of information that can flow throughthat pipeline at a given time.

Generally communication loads which consumebandwidth on a given telemetry system are not constantor deterministic. There can be momentary bursts of trafficthat can consume significant bandwidth for a very shortperiod of time. There can also be extended periods ofhigh communication traffic during working hours andperiods of low volume traffic during evening and earlymorning hours. If at a time of high volume, an extra loadoccurs that causes the entire traffic level to exceed thebandwidth of the telemetry system, data delivery can bedelayed or interrupted. In the worst cases data can belost. In order to achieve reliable and economical SCADAor AMR communications it is not advisable to utilize allof the bandwidth of a given telemetry method. In addition,scheduling communication transactions to distributeloading over periods of time, to take advantage of periodsof low traffic, and to leave overhead for spontaneouscommunication events that can result during upsets, iskey to the overall performance and success of a SCADAor AMR system.

In both SCADA and AMR systems the communicationtransactions to convey data from field devices to thehost computer can be automatically triggered on ascheduled basis. The rate at which these communicationevents are scheduled is one of the first considerationsin the design and implementation of a SCADA or AMRsystem. Users often attempt to schedule communicationtransactions at rates that are not reasonable for a givetelemetry method. For example, the rates used tocommunicate to field devices over PSTN (PublicSwitched Telephone Network) would not be the sameas those used in a 100 MB TCP/IP Ethernet network.Additionally, a high rate of communication could bescheduled for one field device that would prohibitcommunication with any other devices in the application.The judgment as to how often it is necessary tocommunicate with a given field device is a critical stepin the implementation of a SCADA or AMR system.

In larger SCADA systems there can be thousands ofremote sites and field devices. Due to the bandwidthconstraints it is generally not possible or economical tocommunicate with each of these field devices fastenough to perform routine control functions from the hostcomputer. If an individual control loop has an executioninterval of one second, it would be necessary for thehost computer to send a control output to the appropriatefield device at least once per second. As a general ruleone second communication rates with even one devicewould use an unacceptable amount of bandwidth. Inaddition, long-distance communication technologieshave many points of exposure to service interruptionsor degradations. If communication between the hostcomputer and the field device is interrupted or delayedthe control located in the host computer would no longerbe affective. A better strategy that can survive telemetrydegradation and failures is to perform control, dataacquisition and alarm or event monitoring functions inthe field device which is located at the remote asset. Inthis architecture the host computer performs a“supervisory” function under which adjustments to


control targets and alarm thresholds in the field deviceare made on a relatively infrequent basis.

In SCADA and AMR systems communicationtransactions are typically scheduled to occur over aperiod of time or at a particular time of day in order touse available telemetry bandwidth effectively. Thescheduling of communication events is controlled by thehost computer. As a result, Master-Slave protocols asopposed to Peer-to-Peer protocols are used. Thecommunication event which is initiated by the hostcomputer is called a “Poll”. In a simple Polling scenariothe host computer will send a request for data messageover the telemetry system to a field device. Subsequentlythe field device responds with a message that containsthe requested data. In more complicated protocols theremaybe other steps such as acknowledgements however,the key is that the host computer initiates thecommunication transaction and the field deviceresponds. This is a Poll.

There are several other types of communication eventsin addition to Polls and there are several types of Pollsdepending on how the transaction is initiated.

Interval Polling

In Interval Polling, the host computer initiates acommunication transaction by making a request forcurrent or real-time data from a field device periodicallybased on a pre-established interval. This interval isgenerally defined for each field device in an application.The polling rate for a particular device may be faster orslower than others depending on the criticality of theasset or application to which the field device is applied.Generally, the polling interval is a compromise betweenthe timeliness of the data and the need to establish abaseline communication loading on a particular telemetrymethod. The polling interval can also be synchronizedwith the host computer clock to synchronize polls withperiods of low traffic and to further controlcommunication loading.

In RTU, EFM and Corrector field devices performing gasmetering calculations Interval Polls can be initiated torequest the EFM data that is archived in the field device.These communication transactions are initiated with arequest for data from host computer however, the fielddevice can respond with a massive block of data thatmust be managed as a group or a “data set”.

Demand Polling

In Demand Polling the communication transaction, therequest for current or real-time data, is initiated from thehost computer on an event such as a user input. As withInterval Polling the host computer initiates a request fordata message that is sent to the field device. The maindifference is that a Demand Poll is not scheduled andcan happen at any time.

As with Interval Polls, Demand Polls can also initiaterequests for natural gas metering data that is archivedin RTU, EFM and Corrector field devices for auditpurposes.

Unsolicited Messages

Unsolicited messages are initiated by a field device onan event or alarm circumstance and are referred to asException Reports or Cryouts. In these situations thefield device initiates the communication transaction bysending a message without a prior request message fromthe host computer. As with Demand Polling an unsolicitedmessage can be received by the host computer at anytime.

Exception Reports are unsolicited messages that aregenerated by a field device to convey specific dataregarding an event or alarm detected by the device.

Cryouts are unsolicited messages that are generated bya field device to request that the host computer Poll thedevice in response to an event or alarm detected in thedevice.

It is usually a good practice to use Interval Pollingsparingly as a compromise between receiving data on atimely basis and establishing a baseline load on atelemetry system for reliable service. For the specialcircumstances such as alarm notification or to servicepersonnel with critical data Demand Polling andUnsolicited Messaging can have a low impact with regardto added bandwidth consumption but can provide timelydata at critical junctures.

INTEGRATION FRUSTRATION

The management of communication protocols andtelemetry methods associated with large geographicallydistributed data gathering and control systems hasbecome a complex application that is not adequatelysolved in leading SCADA and AMR systems. In the past,software manufacturers have focused on communicationsoftware only as a necessary add-on to their coreproduct. These “drivers” were generally softwaremodules called DLLs (Dynamic Link Libraries) that couldnot run in a standalone mode and that usually supportonly one protocol and telemetry method. Additionally,the drivers were written to the proprietary API (ApplicationProgramming Interface) of a particular core productinsuring that they will be incompatible with othermanufacturer’s products. Finally, the current businessclimate further emphasizes the need to integrate differentprotocols and telemetry systems into SCADA and AMRsystems.

• Today’s corporate strategy of growth throughacquisitions often necessitates the consolidation ofequipment and systems communicating in variousprotocols over different telemetry methods, bothcurrent and legacy. Consolidation is difficult however,


Doug Osburn

in an atmosphere of reduced budgets and restrictedcapital expenditures which is typical following mostcorporate acquisitions.

• As businesses expand, contract or redirect theirresources to meet changes, the automation systemmust also react to these changes in a timely manner.Lost opportunity, increased operating costs, andmounting profit losses, are incurred if changes mustbe accumulated for a window of opportunity whenthe automation system can be taken out of servicefor updates. Even safety is impacted when facilitiesand equipment are operated under temporarymethods while awaiting integration into themainstream automation system.

• Historically, parallel data acquisition systems havebeen installed and maintained to accomplish theautomation missions of different business groups.For example, the gas measurement and the gascontrol business groups within pipeline companiesand utilities often have completely separate systems.Consolidation to reduce operating and support costshas been difficult due to incompatiblecommunication methodologies.

• Due to the unacceptable costs of replacing installedequipment and the competing standards movementsthe need for SCADA and AMR systems to supportdifferent protocols will continue for the foreseeablefuture.

CONCLUSION

Automation systems that support both SCADA and AMRhave three major component areas.

• Field Devices• Telemetry Systems• Host Computers

The issues of communication affect each one of theseareas and often have the largest impact on the successof the system.

• It is rare that a SCADA or AMR application can befound that relies on devices from a singlemanufacturer speaking a single protocol. While mostdevice manufacturers provide there own softwareor firmware that will allow the host computer toreceive data from the device, it is unusual that itwould also support and communicate with acompetitor’s device. Additionally, SCADA and AMRapplications typically utilize different class of devicesthat are not intended to communicate within thesame platform.

• Different telemetry systems may be necessary toprovide coverage to all of the areas to which thesystem is applied.

• Host computer communication software shouldintegrate the protocols and telemetry into a cohesivesystem


METER SELECTION FOR VARIOUS LOAD REQUIREMENTSMike Haydell

Centerpoint Energy2500 Hwy 14, New Iberia, LA 70560

INTRODUCTION:

Gas meters have become known as the “CASHREGISTER” of the natural gas industry. With today’scompetitive energy markets and the environment of FERCorder 636, natural gas measurement has become anincreasingly important issue. It is therefore the duty ofmeasurement departments, to select equipment anddesign installations that are both efficient and economical.

HISTORY OF GAS MEASUREMENT:

The diaphragm meter, as we know it today, was firstdeveloped in 1847. Rotary measurement was introducedin 1923. Development of the turbine meter began afterWorld War II and was introduced in 1963.

POSITIVE DISPLACEMENT METERS:

Positive displacement (diaphragm and rotary) meters,measure gas quantities by the successive filling andemptying of chambers of known quantity. This can becompared to a cook adding quantities of ingredients to arecipe with a measuring cup. This motion is transferred, bya mechanical linkage, to an “index” or read-out device whichis graduated in the appropriate units (usually cubic feet).

Diaphragm meters accomplish this with a set of bellowswhich are filled and then emptied through a set of slidingvalves. As one compartment of the bellows fills, the otheris emptying. (See Figure 1.)

Diaphragm meters are available with capacities rangingfrom domestic loads (175 cubic feet per hour) tocommercial and industrial loads (10,000 cubic feet perhour) per hour, at base conditions.

Because of their construction, these meters are limitedto working pressures of 100# or less. Some models,though are available with 500# cases.

Rotary meters have a set of rotating vanes or impellersthat counter rotate inside a chamber. Each rotation ofthe vanes carry a known quantity of gas though the meter.(See Figure 2.)

Capacities of rotary meters range from 800 to 102,000cubic feet per hour, at base conditions.

These meters typically have working pressures of 175#,with some sizes ranging to 400#, 600#, 900# and 1440#.

INFERENTIAL METERS:

Inferential meters measure gas volumes based on physicalproperties of the gas being measured. Turbine and orificemeters are two types of inferential measurement.

Similar to a child’s pinwheel, Turbine meters use gasvelocity to spin a turbine wheel. This spinning motion islinked to an index by means of a gear train. The higher therate of flow, the faster the spin, therefore registering highervolume on the index over a given period of time. Figure 3shows the construction of a typical Turbine meter.

FIGURE 1. Operation of Positive Meters

FIGURE 2. Operation of Rotary Meters

ROOTS Rotary Positive Displacement Operating Principle

Position 1 Position 2 Position 3 Position 4

FIGURE 3. Construction of a Turbine Meter


Turbine meters have the advantage of high capacity, fora given size but limited low flow characteristics.Orificemeters consist of “PRIMARY” devices (orifices) and“SECONDARY” devices (chart recorders or flowcomputers). Measurement is achieved by recordingpressure and the differential created by the flow acrossthe orifice. Calculations are then made using variousparameters (such as meter tube size, orifice diameter,etc.) to determine gas volumes. Rules governing orificemeasurement can be found in the A.G.A. Report No. 3.

DESIGN PARAMETERS:

Several parameters have to be considered when selectingmeasurement equipment for a particular load. Thesewould include: load characteristics, equipment pressurerequirements, system capabilities, site restrictions, etc.

The Louisiana Division of CENTERPOINT ENERGY usesthe form shown in fig. 4 to provide this information tothe various departments involved in meter selection.

This form must be submitted when application for serviceis made for connected loads of 1500 cfh or more or forpressure requirements above 4 oz.

The form is completed by Marketing personnel and sentto the Engineering Department for a determination ofsystem capabilities.

Engineering “plugs” the proposed load into it’s pressurestudies and determines if the system can support the load.

Once it is determined that the system can handle the load (orwhat improvements or changes have to be made) the form issent to the Measurement Department for meter selection.

Measurement studies the load and decides on the metersize based on the type of load.The Division Chief Engineer makes the final approval andthe form is then sent to the field for installation.

SIZING FOR VARIOUS TYPES OF LOADS:

Various types of loads require different sizing criteria.For example; consider the load in Figure 4. It consists ofseveral types of gas burning equipment and may besized, based on a “DIVERSITY FACTOR.” Experienceshows that this type of load may be sized using a factorof 60% to 75% of total connected load.

As a contrast, the load shown in Figure 5 is a DEMANDtype load and has to be sized for full capacity of theconnected equipment.

If a load is in question, sometimes it is practicable toresearch actual usage of similar operations in other areas.This technique is possible, for example, in the case ofthe new “Super Stores” where the connected load mayexceed the actual usage.

METER CAPACITIES:

Once a load for a customer has been determined, it istime to select the equipment. Meter capacity tables are

FIGURE 4. FIGURE 5.


pressure of the regulator upstream of the large meterthus allowing flow through both meters.

In locations where seasonal loads vary, parallel meterscould be the used. During the “off season” one of themeters can be shut-in. This would allow the single meterto operate towards the upper limit of its range rather thanboth meters operating in the lower ends of their capacities.

With measurement personnel now using “Transfer Provers,”accessibility to field test meters is an issue. Normally theprovers are installed in a vehicle or trailer necessitatingample “maneuvering space” near the meter installation. Alltoo often, location becomes a problem. In situations suchas this, meters should be selected that are small, compactand easily removed from the line. This allows the techniciansto bring the meter to the prover for test.

MEASUREMENT AT ELEVATED PRESSURES:

When it is necessary to measure natural gas at pressuresother than “base pressure”, the Measurement Supervisorhas several options. These would include: pressurecompensating indexes, fixed factor measurement,mechanical or electronic integrators and flow computers.

Basic economics will dictate the best choice for selectingfrom these options. Pressure compensating indexes orfixed factors should be used only on smaller loads (suchas steam cleaners or small boilers) where a regulator, ofadequate capacity, is installed upstream of the meter.This will assure a constant pressure for measurement.Mechanical or electronic integrators can be used onlarger commercial loads. These instruments willcompensate for any fluctuating pressure (or temperature)on the meter. Full blown flow computers should be usedon large industrial loads, where there may be multiplemeter runs. These devices will store information, suchas pressure, temperature, uncorrected and correctedvolumes, etc. for retrieval by laptop or “on-line”.

CONCLUSION:

Because meter sizing is not an “exact science”, priorexperience and common sense should be used whenselecting measurement equipment for various loadrequirements. Load profiles and operating characteristicsshould be carefully studied and equipment selectedaccordingly. Using over sizedmeters will result in poormeasurement during periods oflow flow; under sizing will causeundue wear on meters andpossibility customer equipmentproblems.

Even though the ±1/4% accuracyMeasurement Personnel strivefor, is not always possible, it ispossible to select measurementequipment that is economical,efficient and safe.

necessary for this procedure.

Manufacturers rate the capacities of their meters at a “BASE”pressure and at pressures up to the working pressure of themeter. Figure 6 is a typical meter capacity table.

SPECIAL CONSIDERATIONS:

When selecting meters, for various applications, specialconsiderations need to be taken into account. Forinstance: rotary and turbine meters should be used onlywhere gas is relatively free of foreign materials (such asdust) that may cause undue wear on close tolerance parts.Strainers or filters should be used is such applications.

In situations where the minimum flow is only a smallpercentage of the maximum flow, (such as a pilot loadon a large boiler) low flow measurement must be takeninto account. This can be accomplished by using acombination of large and small capacity meters inconjunction with regulators, set at different pressuresand an orifice plate (choke). Figure 7 illustrates the layout.

The regulator upstream of the small meter is set at 10#,for instance. The regulator upstream of the large meteris then set at 5#.

Under low flow conditions, the regulator upstream of thelarger meter (T-60) is “locked-up” because of the higherpressure of the regulator ahead of the smaller meter (3M).This allows flow through the small meter and preventsflow through the large meter. As the flow increases, thechoke downstream of the small meter reaches criticalflow causing the pressure downstream to fall to the set

Mike Haydell

FIGURE 6.

FIGURE 7.

GAUGE MTR. SIZE 750 1600 1000 3000 5000 10000PRESSURE DIFF.

4 OZ. 2" W.C. 1600 1600 2200 3000 5000 10000MAX. ALLOW. 1600 1600 2200 3000 5000 10000

5# 2" W.C. 1840 1840 2530 3450 5750 11500MAX. ALLOW. 2070 2070 2840 3880 3450 12900

10# 2" W.C. 2080 2080 2860 3900 6500 13000MAX. ALLOW. 2480 2480 3420 4660 7760 15550

15# 2" W.C. 2320 2320 3190 4350 7250 14500MAX. ALLOW. 2840 2840 3960 5400 9000 18000

25# 2" W.C. 2640 3630 4950 8250 16500MAX. ALLOW. 3750 4900 6700 11100 22300

50# 2" W.C. 3360 6300 10500 21000MAX. ALLOW. 5000 9370 15600 31200

100# 2" W.C. 4480 8400 14000 28000MAX. ALLOW. 7170 13400 22400 44800


DESIGN AND IMPLEMENTATION OF MARKET BASED:SAFE, RELIABLE & COST EFFECTIVE METERING AND

REGULATING FACILITIESTom Quine

Northstar Industries126 Merrimack St., Methuen, MA 01844

PATENT AND COPYRIGHT STATEMENT

Northstar Industries, Inc. has prepared this educationalmaterial.

Certain designs and concepts are presented which areprotected by US Patent #6,176,046: Portable Pre-manufactured Natural Gas Delivery Systems.

This material is copyright protected and cannot bereproduced without the written permission of NorthstarIndustries, Inc.

AGENDA

• FIRST WAVE LESSONS LEARNED• GAS AND ELECTRIC REQUIREMENTS FOR

SITING• DESIGN CONSIDERATIONS• CONCLUSIONS

OVERVIEW

The first wave of natural gas fired and market basedpower generation assets have been installed. The resultshave been mixed from a fuel supply, electric market andproject finance basis.

Load growth is still occurring and new assets will berequired.

The next wave of power projects will benefit from thelessons learned over the past three years.

The successful players in the next wave will adapt theirdevelopment process and address these core issues.

LESSONS LEARNED

Natural Gas interconnects for Power• Interconnect Process Not Well Defined Up Front• Too many cooks: Pipeline, LDC, EPC Contractor

at Site.• Project Financing and Business Case Did Not

Consider This.• Interconnect Agreements Had Gaps• Fuel Management SCADA System Not Clearly

Defined• Market Base Electric Sales Versus Fixed

Contract

GAS INTERCONNECT REQUIREMENTS• Adequate Natural Gas Fuel Supply: Pressure and

Capacity• Understanding of Pipeline Tariff to Close the Deal• Project Financing and Business Case in Place• Workable Interconnect Agreements• Pipeline Hot Tap and Side Valve Site• Pipeline Metering, FCV, SCADA, and Power• Lateral Alignment to Plant Site• Lateral Facilities: Scrappers, MLV, CP, Odor,

Power• Fuel Management SCADA System• Pre-heating for Pressure Cut and Dew Point• Gas Regulation and Sub-meters for Ct and

Domestic Needs• Fit for Duty and OQ Project Team• O&M, Emergency and Field Services Plan

PROJECT IMPLEMENTATION

• Feasibility Analysis• Preliminary Engineering• Permits• Contract Supply and Transport• Procurement• Final Design• Qualified, Materials, People, and Procedures• Pre-manufacturing• Installation• Project Documentation• Commission and Train


PRELIMINARY ENGINEERING

• Explore the Project Background and Need• Determine Authority Having Jurisdiction• Define Design Criteria by Discipline• Select Major Equipment• Determine Permits and Approval• Establish Siting Criteria• Establish Site Selection and Site Layout• Create Process Flow Diagrams• Establish Cost and Schedule• Present Data to Authority

DESIRED FEATURES OF M&R FACILITIES

Safety, Reliability and Cost Effectiveness

The following section illustrates traditional and Alternatedesigns for facilities as well as highlighting featuresfavorable to all metering and regulating facilities.

• Ease of Maintenance• Ease of Access• Ergonomic Design• Simplicity• Appropriate Levels of Redundancy• Cost Effectiveness

TRADITIONAL CONFIGURATION

Traditional sites were comprised of multiple buildings withseparate ownership from the pipeline and enduser. Thesefacilities were stick built on site. Schedules were longand economics were rate based.

TRADITIONAL SITE

ALTERNATE CONFIGURATION

Deregulation:

• Has resulted in staff reductions• Has resulted in new players in the marketplace• Has resulted in new functionality requirements• Has resulted in shorter schedules• Has resulted in market base projects• Has resulted in new new power projects• Has resulted in alternate new configurations:

Gas Quality, Remote Monitor and Control,Functionality, M&R CP, SCADA, SharedMeasurement.

ALTERNATE

PRE-MANUFACTURED MODULAR GAS FACILITIES:US PATENT #6,176,046


Each Module Independent Unit or Combined CompleteFacility: Filtering, Metering, Preheating, Odorization,SCADA and Regulatory Facility.

DESIRED ATRIBUTES OF ALL M&R FACILITIES

Safety, Reliability and Cost Effectiveness

ERGONOMIC DESIGNS

HIGH PRESSURE GAS

METERING AND FCV

REGULATORY AND OPP

GAS HEATING


REMOTE HEAT SOURCE

ODORIZATION

SCADA SECURITY AND POWER

ELECTRICAL DESIGN

ESTABLISH HAZARDOUS AREA PLAN IN PHASE 1:• Design, Procure and Install Equipment

Accordingly

CONDUCT AC - MITIGATION WHERE NECESSARY:• Personnel and Equipment Protection

HAZARDOUS AREA PLAN

SYSTEM ARCHITECTURE PLAN

VOLTAGE MITIGATION GOALS

SCADA DESIGN

• Develop P & ID• Develop I/O List• Develop System Architecture

Select EquipmentDetermine PMISelect Communication Links

PeopleProtection

EquipmentProtection


CONTROL ROOM SCADA, COMMUNICATION,SECURITY DOCUMENTATION AND PMI

CONTROL ROOM HARDWARE

SOFTWARE

CONCLUSION

The implementation of a successful project dependsupon a capable team performing preliminary engineering,design, procurement, installation, training, anddocumentation.

Deregulation has added new pipeline interconnects,shortened the schedules, created the need for new typesof facilities and created market based projects.

The successful players in the next century will accomplishtheir objectives in concert with customers and suppliersby sharing facilities, sharing information, finding win-winsituations and UNDERSTANDING THEIR ROLES IN THEMARKETPLACE

Tom Quine


GTI METERING RESEARCH FACILITY UPDATEEdgar B. Bowles, Jr. and Marybeth G. Nored

GTI Metering Research Facility Program, Southwest Research Institute6220 Culebra Road, San Antonio, TX 78238-5166

INTRODUCTION

The Gas Technology Institute (formerly the Gas ResearchInstitute) sponsors a comprehensive flow measurementresearch, development, and commercialization (RD&C)program aimed at improving natural gas meteringperformance in the field. This paper summarizes someof the recent accomplishments of the research programat the Gas Technology Institute (GTI) Metering ResearchFacility (MRF), a high-accuracy natural gas flowcalibration laboratory capable of simulating a wide rangeof operating conditions for the industry’s research,calibration, and testing needs. The MRF, located atSouthwest Research Institute (SwRI) in San Antonio,Texas, supports a variety of GTI-sponsored research andthird-party test and calibration activities. Major researchinitiatives currently being funded by GTI (formerly knownas the Gas Research Institute or GRI) include ultrasonicand turbine flow meter research and gas samplingmethods research. Over the past year, GTI has alsofunded Coriolis flow meter research and the developmentof a new energy flow rate meter concept. Through itsportfolio of projects addressing priority research needs,the GTI natural gas measurement program providessignificant benefits to the natural gas industry.

FIGURE 1. GTI Metering Research Facility

METERING FACILITY DEVELOPMENT

The Metering Research Facility program was initiatedby GTI in the late 1980s in direct response to a growingneed within the natural gas industry for improvement inthe state-of-the-art of gas flow measurement. Theconcept of a natural gas industry metering research andcalibration facility was first proposed in the Gas Industry

Measurement Plan prepared by the Operating Sectionof the American Gas Association (AGA). This planrecommended the development of an independent,qualified flow test facility that would be operated underthe sponsorship of the gas industry. This facility had tobe capable of providing performance data on a broadspectrum of meter types and sizes over a wide range offlow conditions. In response to this recommendation,GTI initiated the MRF program with Southwest ResearchInstitute in 1987.

To cover the wide range of flow conditions necessaryfor the research and testing needs of the natural gasindustry, development of the MRF1 included three primarycomponents: a High Pressure Loop (HPL), a LowPressure Loop (LPL), and a Distribution Meter Test Stand(DTS). Table 1 lists the operational ranges for all threesystems. The HPL and LPL are re-circulating flow loops,while the DTS is a “blow-down” type system. Eachsystem can flow either natural gas or nitrogen.

An on-line, laboratory-grade gas chromatograph is usedfor detailed analysis of the test gas.

Flow measurement accuracy on the order of 0.1 to 0.25%is achieved in the HPL, LPL, and DTS through the use ofindividual gravimetric primary calibration systems. Sonicnozzles, gas turbine meters, and laminar flow elementsare also used as secondary transfer standards. Thegravimetric calibration system for the High Pressure Loopis shown on Figure 2.

An example of the performance of the MRF is illustratedby Figure 3, which plots orifice meter dischargecoefficients (Cd) from the HPL against similar data fromother international flow calibration laboratories.

In addition to serving as a test bed for GTI-sponsoredresearch, the MRF is available to any other interestedparties for test and calibration services. The MRF

Parameter HPL LPL DTSMax. Rate (MSCFH) 7,084 610 2.5Max. Rate (ACFH) 84,000 43,600 2,500Pressure (psig) 150-1,200 20-210 0-40Gas Temp. (°F) 40-120 40-120 AmbientPipe Size (inches) 2-20 1-8 up to 2Specific Gravity 0.55-0.97 0.55-0.97 0.55-1.0

TABLE 1. Metering Research FacilityOperational Ranges


technical staff also provides assistance with meter stationdesign and helps troubleshoot metering problems in thefield. In addition, the MRF offers training courses on meterstation design and operation.

ULTRASONIC GAS FLOW METER RESEARCH

In June 1998, the AGA Transmission MeasurementCommittee (TMC) published its Report No. 9, entitledMeasurement of Gas by Multipath Ultrasonic Meters.2

This document represents the first industry guidelineson the use of ultrasonic flow meters for natural gasapplications. Report No. 9 is fundamentally different frommost other gas flow meter standards in that it isperformance-based and does not include dimensionaland other mechanical specifications associated with themeter installation. Instead, the report states that the flowmeter must perform within specified measurement errorlimits when installed per the meter manufacturer’srecommended installation configuration.

Since the publication of Report No. 9, the AGATransmission Measurement Committee has proposedthat a flow performance verification test be included inthe next revision of the report to help ensure that metersperform within the specified measurement error limits.This verification test would allow a manufacturer tovalidate recommended installation configurations andallow meter users to compare meter performance andinstallation requirements under a common set of pipingconfigurations. It is also anticipated that the next revisionof Report No. 9 will include a small number ofrecommended installation configurations.

During the past year, the MRF research program hasassisted the AGA TMC with the revision of Report No. 9by producing test data on various meter installation andoperational effects. Commercially-available multi-pathultrasonic gas flow meters from ABB Totalflow, DanielFlow Products, FMC/Kongsberg, and InstrometUltrasonic were tested during the course of this test work.

Figure 4 shows a flow meter skid used for part of theMRF test program in 2002 and 2003. This skid includedthree parallel meter runs containing an 8-inch diameter4-path Daniel ultrasonic flow meter, a 6-inch diameter4-path Daniel ultrasonic flow meter, and a 3-inchdiameter Instromet positive displacement flow meter. Thetwo ultrasonic flow meter runs were fitted with aperforated-plate type) flow conditioners upstream of themeters.

FIGURE 3. GTI MRF High Pressure Loop 10-inchDiameter Orifice Discharge Coefficient Data

FIGURE 2. GTI MRF High Pressure Loop Weigh Tank

Pipe Reynolds Number

105 106 107

Orif

ice

Cd

0.590

0.592

0.594

0.596

0.598

0.600

0.602

0.604

0.606

0.608

0.610

β = 0.75250 mm Orifice Meter Tube Calibration

NISTCEAT

NELDHL

Ruhr GasBritish GasGasunie

AGA - 3 (RG) equation for flange tapped orifice coefficient

GRI MRF HPL

95% confidence interval for RG equationWater data Natural gas data

FIGURE 4. Ultrasonic Flow Meter Skid Package


The flow meter skid shown in Figure 4 was used to testfor various meter installation effects, including the effectsof an upstream and downstream header. In addition, thelow flow rate performance of the ultrasonic meters wascompared to the performance of the positivedisplacement meter to help determine if the low-endrange of the ultrasonic meters can achieve comparablemeasurement accuracy levels to that of a typical positivedisplacement meter. Other operational effects, such asline pressure variation, on meter performance were alsoinvestigated. At the time this paper was being written,the results of the MRF test work had not yet beencompiled in a GTI Topical Report. However, as of thedate of the 2003 American School of Gas MeasurementTechnology, a GTI report documenting the findings ofthis MRF research work should be available to the public.

Previous ultrasonic meter installation configuration testsperformed at the MRF[3],[4],[5],[6] demonstrated that eachflow meter/flow conditioner combination producesunique operational characteristics due to differences inthe flow rate calculation algorithms and acoustic pathconfigurations. The ability of a meter to compensate fora velocity profile distortion determines the amount of biasthere will be in the meter error.

In addition to the recent test work, the MRF researchprogram has also quantified meter performance effectsdue to a diameter mismatch between the meter bodyand the adjacent meter tube. Diameter mismatches, ineither the upstream or downstream direction, within the+1.0% of pipe diameter limit specified in AGA ReportNo.9, did not result in any additional error for theultrasonic flow meters tested.

TURBINE GAS FLOW METER RESEARCH

Due to advances in recent years in gas turbine flow metertechnology, the AGA Transmission MeasurementCommittee is in the process of revising AGA Report No.7 - Measurement of Gas by Turbine Meters.[7] In supportof this effort, a research program at the MRF has beenevaluating the measurement performance of bothconventional and new, high-capacity turbine meterdesigns. Both single- and dual-rotor meters have beenincluded in the study.

A number of turbine meter installation and operationaleffects are being investigated in detail as part of the MRFresearch program. Recent testing has focused onnominal 6-inch diameter meters having eitherconventional 45º rotor blade angles or ‘extended-capacity’ 30º rotor blade angles. These meters have beentested in the meter installation configurations referencedin AGA Report No. 7 (i.e., the ‘recommended,’ ‘short-coupled,’ and ‘close-coupled’ configurations (see Figure5). The meters have also been evaluated with theInternational Standards Organization (ISO) 9951 ‘high-perturbation’ piping configuration installed upstream. Theeffect of line pressure variation on meter accuracy hasalso been studied.

FIGURE 5. Turbine Meter Under Test in a ‘Close-Coupled’Configuration at the MRF

Test results have found that the single-rotor and dual-rotor meter configurations do not exhibit significantlydifferent measurement performance and both metertypes demonstrate improved measurement accuracywith effective flow field conditioning at the meter inlet.

Additionally, the MRF test results have demonstrated thatchanges in line pressure (between atmospheric and 160psig) can produce significant shifts in meter calibration.The magnitude and direction of the calibration shifts aregenerally a function of meter design. The current MRFresearch is expected to result in an updated specificationfor meter performance tolerances.

A summary of recent MRF turbine meter research resultswas presented in a technical paper[8] at the 2002 AGAOperations Conference and was also published in a GTITopical Report.[9]

NATURAL GAS SAMPLING RESEARCH

Proper natural gas sampling methodologies are criticalto the accurate determination of natural gas heatingvalue. Improper sampling technique can distort thecomposition of the natural gas sample, which will directlyaffect the accuracy of the heating value and indirectlyaffect the accuracy of the volumetric flow rate (througherrors in the gas properties, such as gas density).Because of the importance of accurate sampling of thenatural gas flowing through a pipeline, GTI, the AmericanPetroleum Institute (API), and the U.S. MineralsManagement Service (MMS) have been funding aconsortium research project to document the causes ofgas sample distortion and to implement procedures thatproduce accurate sample analysis. Much of thisconsortium research work has been carried out at theMRF. The research results have lead to a recent revisionof the industry standard for gas sampling methods, i.e.,the API Manual of Petroleum Measurement Standards(MPMS), Chapter 14.1 - Collecting and Handling ofNatural Gas Samples for Custody Transfer.[10]


This comprehensive research program has looked at allaspects of natural gas sampling methodology. Varioustypes of spot- and composite-sampling methods, as wellas on-line analysis methods, have been studied. Causesof gas sample distortion that have been identified includethermodynamic phase changes, molecular adsorption1(on the surfaces of solids or liquids), sample probelocation, filtering of the sample gas, and cleanliness ofthe gas sampling equipment. Once a gas sample hasbeen distorted, its composition has been altered and,thus, its heating value will be erroneously determinedbased on the altered composition.

Recent MRF gas sampling methods research has beendirected toward improving the accuracy of thehydrocarbon dew point temperature calculated from thecommon equation of state (EOS) models used by theU.S. natural gas industry. Inaccurate prediction of thehydrocarbon dew point temperature by analytical meanscan be a contributing factor in sample gas distortion.Accurate estimation of the hydrocarbon dew pointtemperature is highly dependent on the percentages (orcharacterization) of the heavier hydrocarbons in a gasmixture (i.e., hexane (C6) and heavier). Experiments arebeing run on a wide range of natural gas blends tomeasure the hydrocarbon dew point temperature forcomparison with calculated values (see Figure 6).

Department of Energy. The device is designed to measurea small number of process variables (e.g., gas soundspeed, diluent concentrations, pressure, temperature,and volumetric or mass flow rate) to determine energyflow rate in real time, and at normal pipeline operatingconditions. This new device is expected to be a morecost effective approach to determining energy flow ratethan the conventional approach of combiningmeasurements from a volumetric flow meter and gaschromatograph. Further sensor development, laboratorytesting, and field-testing at gas transmission pipelinesites are currently being completed. A photographshowing one of the prototype measurement sensorenclosures is shown in Figure 7 below. Technical reportson this device are available from GTI.[14],[15]

FIGURE 6. Natural Gas Phase Diagram ShowingHydrocarbon Dew Point Temperature

Results of recent gas sampling research at the MRF weresummarized in a technical paper[11] presented at the 2002AGA Operations Conference. In addition, two GTI TopicalReports[12],[13] on the subject have been published.

OTHER MRF RESEARCH

In addition to the ongoing work described above, theMRF program is currently involved in other natural gasmeasurement research. For example, a prototype energyflow rate meter is under development at the MRF. Asnoted above, this work is receiving funding from the U.S.

FIGURE 7. Energy Flow Rate Meter PrototypeMeasurement Sensor Enclosure

Also in 2003, orifice flow meter research, funded by GTI,is planned for the MRF. At the time this paper was written,this work had not yet begun. There are two componentsto the 2003 orifice flow meter research program at theMRF. Test work will be performed to acquire additionaldata on the orifice meter expansion factor. TheInternational Standards Organization recently adopteda new equation for the calculation of the orifice meterexpansion factor coefficient. The new ISO equationproduces values that vary slightly from those producedby the equation referenced in the AGA orifice meterstandard (i.e., AGA Report No. 3 - Orifice Metering ofNatural Gas[16]). The new MRF data will provide betterinsight as to the differences between the two expansionfactor equations.

The second component of the MRF orifice flow meterresearch program in 2003 is an investigation of the effectof bent orifice plates on meter accuracy. Some data areavailable on this subject, but additional data are needed.An extensive test matrix is planned to assess the effectof bent orifice plates. A GTI Topical Report summarizingthe findings of the MRF orifice flow meter research in2003 should be available to the public by the end of thecalendar year.


CONCLUSIONS

GTI’s applied flow measurement research programcontinues to address the priority needs of the naturalgas industry. This paper summarizes some of the recentmeasurement research activities at the MRF. A completelisting of all MRF research reports and technical papersis available from GTI (www.gri.org) or the MRF website(www.grimrf.org).

REFERENCES

1. Johnson, J. E., et al., “Metering Research FacilityDesign,” GRI Topical Report No. GRI-91/0251, GRI,Chicago, IL, March 1992.

2. Measurement of Gas by Multipath Ultrasonic Meters,American Gas Association TransmissionMeasurement Committee Report No. 9, AmericanGas Association, Arlington, VA, June 1998.

3. Grimley, T. A., “Performance Testing of UltrasonicFlow Meters,” 15th North Sea Flow MeasurementWorkshop, Kristiansand, Norway, October 1997.

4. Grimley, T. A., “The Influence of Velocity Profile onUltrasonic Flow Meter Performance,” Proceedingsof the American Gas Association OperationsConference, Seattle, WA, May 1998.

5. Grimley, T. A., “Performance Testing of 12-InchUltrasonic Flow Meters and Flow Conditioners inShort Run Installations,” GRI Topical Report No. GRI-01/0129, GRI, Chicago, IL, January 2002.

6. Grimley, T. A., “Effects of Diameter Mismatch andLine Pressure Variations on Ultrasonic Gas FlowMeter Performance,” GRI Topical Report No. GRI-02/0031, GRI, Chicago, IL, April 2002.

7. Measurement of Gas by Turbine Meters, AmericanGas Association Transmission MeasurementCommittee Report No. 7, American Gas Association,Arlington, VA, June 1996.

8. George, D. L., “Turbine Meter Test Results -Installation Configuration Effects,” Proceedings ofthe American Gas Association OperationsConference, Chicago, IL, May 2002.

9. George, D. L., “Progress Report: Effects of LinePressure and Gas Density on Turbine MeterMeasurement Accuracy,” GRI Topical Report No.GRI-03/0050, GRI, Chicago, IL, May 2003.

10. Collecting and Handling of Natural Gas Samples forCustody Transfer, Manual of Petroleum MeasurementStandards, Chapter 14.1, 4th Edition, AmericanPetroleum Institute, Washington, D.C., August 1993.

11. Kelner, E., D. L. George and M. G. Nored, “NaturalGas Sampling Techniques - Recent ResearchResults,” Proceedings of the American GasAssociation Operations Conference, Chicago, IL,May 2002.

12. Behring II, K. A. and E. Kelner, “Metering ResearchProgram: Natural Gas Sample Collection andHandling-Phase I,” GRI Topical Report No. GRI-99/0194, GRI, Chicago, IL, August 1999.

13. Kelner, E., C. L. Britton (Colorado EngineeringExperiment Station, Inc.), K. A. Behring, II, and C. R.Sparks, “Natural Gas Sample Collection andHandling - Phase II: Experimental Testing UnderSimulated Field Conditions,” GRI Topical Report No.GRI-01/0069, GRI, Chicago, IL, January 2003.

14 Behring II, K. A., E. Kelner, A. Minachi, C. R. Sparks,T. B. Morrow, and S. J. Svedeman, “A TechnologyAssessment and Feasibility Evaluation of Natural GasEnergy Flow Measurement Alternatives,” FinalReport, Tasks A and B, Southwest Research Institutefor the U.S. Department of Energy, MorgantownEnergy Technology Center, Morgantown, WV, May1999.

15. Morrow, Thomas B., E. Kelner, and A. Minachi,“Development of a Low Cost Inferential Natural GasEnergy Flow Rate Prototype Retrofit Module,” TopicalReport to GRI and the U.S. Department of Energy,GRI Contract No 5097-270-3937, DOE CooperativeAgreement No. DE-FC21-96M33033, October 2000.

16. Orifice Metering of Natural Gas, American GasAssociation Transmission Measurement CommitteeReport No. 3, American Gas Association, Arlington,VA, June 2000.

Edgar B. Bowles, Jr.


Cd Y π d 2 2 gcρ

f∆P

1–β4 4Q

V = ρ

b

AGA CALCULATIONS — OLD VS NEWBrent E. BerryABB-Totalflow

Pawhusk Road, Bartlesville, OK 74005

SECTION 1 — BACKGROUND

This paper is intended to help bridge the gap betweenthe Old AGA-3 equation (hereafter referred to as AGA-3-1985) and the New AGA-3 equation (hereafter referredto as AGA-3-1992). As such the paper begins with abackground section aimed at assisting those who aremostly familiar with the factored form of the orificemetering equation.

Factored VS Fundamental Flow Rate Equation Form

Of the following two equations, which are published inthe AGA-3-1985 standard?

(a) equation 1(b) equation 2(c) both equation 1 and 2(d) don’t know

(eq 1)

where,

(eq 2)

The correct answer is (c). Both equations are actuallypublished in the 1985 standard and they are bothequivalent within their scope of applicability. Equation 2is often referred to as the factored form of the AGA-3equation. It can be found on page 38 of the 1985standard as equations (59) and (60). Equation 1 is thefundamental orifice meter equation. It can be found onpage 25 of the 1985 standard and is actually acombination of equations (3) and (7) on that page.

Equation 1 describes the theoretical basis, the physicaland practical realities of an orifice flow meter. Equation2, the factored equation, is based on or derived fromequation 1. Why bring all this up now? If you are like thisauthor was at one time, you might only be familiar withthe factored equation. If so, I recommended you becomemore familiar with the fundamental form of the equation.Firstly, because this form more readily facilitatescomparing AGA-3-1985 and AGA-3-1992. Secondly, youwill be more comfortable with the new AGA-3-1992standard. Thirdly, because it more clearly describesorifice meter dynamics.

Why are there different forms of the equation anyway?The first factored form of the equation was introducedin 1935 with the publication of Gas Measurement Com-mittee Report No. 2. Before that time the factors werenot in use. Factors are a convention that allow variousterms of the fundamental equation to be calculatedindividually. This allows tables to be generated for eachfactor which can then be used to estimate volumes.These tables were especially useful before the availabilityof computers and programmable calculators. They arestill used and will continue to be used, but their usage isdiminishing with the advent of electronic instrumentation.

Cd, Coefficient of Discharge

If the old factored equation is all you have been using, youmay have never really dealt with Cd. Basically, Cd isimbedded in the old factored equation as part of (Fb * Fr).Reviewing equation 1 of this document you will notice Cd

is included as part of AGA-3-1985’s fundamental equa-tion. It has always been there, simply hidden by the factors.Most of the research and development undertaken overthe last several years was for the purpose of deriving amore accurate, technically defensible correlation betweena published Cd equation and actual laboratory data. Thisis at the heart of the changes in AGA-3-1992.

What is the purpose of Cd in the fundamental equation?

Cd = true flow ratesppi

theoretical flow rate

The true flow rate is determined in a laboratory byweighing or by volumetric collection of the fluid over ameasured time interval and the theoretical flow rate iscalculated. Then a discharge coefficient (Cd) is computedas a correction factor to the theoretical flow rate. Thisdata is all generated over varying flow rates, fluid types(Reynolds number conditions) and various geometries(diameters). Once all the data is taken then an empiricalequation is derived which allows us to compute Cd overmany combinations of conditions.

That is what Cd, the coefficient of discharge is. It is ahigh falutin’ fudge factor. The developers of the newequation have taken advantage of newer technology inmore numerous testing labs to gather more real worlddata over a wider set of operating conditions. They havealso postulated on a new form of the Cd equation thatthey believe more closely correlates to the fluid dynamicsassociated with the physics of an orifice meter. Thismeans the new Cd is based more on first principles thanthe older one. You might say it has a higher falutin’ indexthan the older coefficient of discharge.

C ’ = Fb Fr Y Fpb Ftb Ftf Fgr Fpv

Qv = C’ hw Pf


Why does the theoretical equation not match the realworld exactly?

In trying to keep orifice metering practical, simplifyingassumptions are sometimes made. It is simply not alwayspossible, practical or necessary to perfectly model thereal world. Some of the things influencing the theoreticalequation, causing it not to model the real world exactlyare:

(1) It is assumed there is no energy loss between thetaps.

(2) The velocity profile (Reynolds number) influences arenot fully treated by the equation. It is assumed thatsome installation effects and causes of flow pertur-bations (changes) are insignificant.

(3) Different tap locations affect the flow rate. Taplocation is assumed for a given Cd.

Through rigorous testing, you could develop a uniqueCd for each of your orifice meters. This technique, referredto as in-situ calibration, is something like proving a linearmeter. However it is somewhat bothersome since youneed a unique Cd for each expected flow rate. Economicsusually make in-situ procedures unfeasible.

Therefore, the goal is to develop a universal Cd thateveryone can use. To accomplish this, one must controltheir orifice meter installation well enough so that itreplicates the same orifice meters used in the laboratoryfrom which the universal Cd equation was derived. Thisis referred to as the law of similarity. If your orifice metersystem is acceptably similar to the laboratory’s then yourCd will be acceptably similar to the laboratory derivedCd. That is why edge sharpness, wall roughness,eccentricity and flow conditioning, etc. are so important.Ideally your flow measurement system would be exactlythe same as was used in the laboratory.

Density

If the old factored equation is all you have been using,you may have never really dealt with density. Lookingback at Equation 1 of this document you will notice twosymbols, ρf and ρb. The symbol ρ (pronounced rho) isused to represent density.ρf = density at flowing conditionsρb = density at base conditions

Most measurement systems do not have density as alive input, so density is computed from other data that isavailable. When the fluid being measured is a gas, densityis computed from other data as follows:

Density at Flowing Conditions

(eq. 3)

Density at Base Conditions

(eq. 4)

Notice Pf, Tf, Zf, Gi, Pb, Tb, and Zb in eq 3 and eq 4. Theserepresent temperatures, pressures, specific gravities andcompressibilities. It is these variables that eventuallymake there way into the old factors Ftb, Fpb, Ftf, Fgr, andFpv (see Section 3 of this document for more information).Leaving densities in the fundamental equation, ratherthan hiding them in a plethora (abundance) of factors,seems less confusing and more instructive.

Real VS Ideal Gas Specific Gravity

One other item of note regarding the density equationsis that they are based on Gi, ideal gas specific gravity.Most systems have historically provided Gr, real gasspecific gravity, which is different. Additionally AGA-8requires Gr as an input, not Gi. Strictly speaking,Gi isrelated to Gr with the following equation.

Practically speaking the measurement system does notusually have enough information available to solve theabove equation. Part 3 of the new standard makes thefollowing statements regarding this issue.

“the pressures and temperatures are defined to be atthe same designated base conditions....”. And again in afollowing paragraph, “The fact that the temperature and/or pressure are not always at base conditions results insmall variations in determinations of relative density(specific gravity). Another source of variation is the useof atmospheric air. The composition of atmospheric air —and its molecular weight and density — varies with timeand geographical location.” 3

Based on this, and several of the examples in the stan-dard, the following simplifying assumptions are made:

(eq. 5)

This equation for Gi is exactly like the one shown asequation 3-48 on page 19 of Part 3 of the new standard.


Conclusion

These are the major new equation concepts you mightneed to learn if the older factored equation is all you arefamiliar with. A more detailed comparison between thefundamental equation and the factored equation ispresented in Section 3. The following section summarizeschanges to the new standard.

SECTION 2 — SUMMARY OF CHANGES TO THESTANDARD

Change 1, Document Organization

Rather than one document, the standard is nowpublished in 4 parts, each of which is a unique document.

Part 1 – General Equations and Uncertainty Guidelines

The mass and volumetric flow rate equations and theirtheory are discussed. The empirical equations for Cd andY are presented. Uncertainty guidelines are presentedfor determining possible errors associated with using thestandard.

Part 2 – Specification and Installation Requirements

Mechanical specifications are presented for the orificemeter. In particular orifice plates, plate holders, sensingtaps, meter tubes and flow conditioners are discussed.

Part 3 – Natural Gas Applications

The fundamental equation, as presented in Part 1, is notrestricted to a specific fluid or system of units. Part 3provides a guide for forcing the equation to computevolumes assuming the fluid is natural gas and the inch-pound system of units. It is in appendix B of this partthat a factored form of the equation is developed.

Part 4 – Background, Development, ImplementationProcedures and Subroutine Documentation

The history and evolution of the equation is included inthis part beginning with Report No. 1 from the 1920s. Adescription of the research that was undertaken to derivethe new equation is presented. Implementation proce-dures, guidelines, and subroutine recommendations arealso documented to assist programmers with implement-ing the new equation on computers. Intermediate resultsare also available to assist with program verification.Inclusion of computer related documentation of this sortis completely new for the AGA-3 standard and recognizesboth the need for a computer to solve the new equationand the availability of computers to accomplish this.

Change 2, The nomenclature of the fundamental equationwas modified slightly.

It is important to note that the fundamental equation didnot actually change. Since it is based on the actual

physics of an orifice meter you would not expect it tochange. However, nomenclature was slightly modified.

eq. 1, AGA-3-1985 Fundamental Equation

eq. 7, AGA-3-1992 Fundamental Equation

Change 3, Cd — New coefficient of discharge solutionrequires an iterative solution

As was stated earlier, Cd is at the heart of the changesfor AGA-3-1992. Many people spent substantial time andeffort in various countries conducting tests to providenew data that could be used to empirically derive a better,more technically defensible coefficient of discharge.

Once the data was gathered and accepted, talentedpeople, using computers, derived a Cd equation that hasa high degree of correlation with all the new data. Thismeans that, within the stated uncertainties of Part I ofthe new standard, you can feel confident that whenapplied as specified by the standard, the new Cd equationwill produce dependable answers.

The tricky part about Cd is that one needs to know theflow rate to compute it. But one also needs the Cd tocompute the flow rate. This is a sort of catch 22. In asituation like this we say the equation is not in closedform. This is why you will hear people say an iterativesolution is required to compute the new Cd.

What does iterative solution mean? It means you, (or morelikely a computer) begin with an estimate for Cd. Basedon that Cd, an estimated Reynolds Number, Flow Rate and

TABLE 2-1Fundamental Equation Nomenclature Changes

AGA-3-1992Equation

AGA-3-1985Equation

NumericConstant*Dimension

ConversionConstant

velocity ofapproachequation

Velocity ofApproachSymbol

*Nc is combined with other constants before solving the equation.These other constants are a function of the system of units chosen.


a subsequent new Cd are computed. The two Cd values(Cd_old and Cd_new ) are compared and if they differ bymore than an acceptable threshold, the process isrepeated. Each time the process is repeated the mostrecent Cd is retained for comparison with the next onebeing computed. Eventually, the difference betweenCd_old and Cd_new become so small it is safe toassume the proper Cd value has been obtained. APIdesigners estimate that, most of the time, no more thanthree iterations will be required. I believe no more than10 iterations were ever required on the test cases. Anexact procedure is outlined in Part 4 of the new standardunder Procedure 4.3.2.9. In that procedure the thresholdfor determining acceptable convergence is six significantdigits (0.000005).

Change 4, Thermal effect corrections on Pipe and Orificediameters are required

In the AGA-3-1985 standard an optional orifice thermalexpansion factor, Fa, was specified to correct for the errorresulting from thermal effects on the orifice plate diameter.

In the AGA-3-1992 standard this type of correction isnot optional. It is required. Additionally you must alsomake corrections for thermal effects on the pipe diameter.

Another new requirement is that these corrections cannotbe tacked onto the end of the equation as a factor. Theyare to be applied on the front end as adjustments to thediameters themselves. Therefore, the end user shouldbe supplying diameters at a reference temperature(68 DegF), and the device solving the equation shouldbe adjusting the diameters based on the differencebetween the reference temperature and the actual fluidtemperature.

This means that virtually none of the equation can be pre-computed and re-used. Even though the new equationdoes not have Fb, there are portions of the equation thatdepend only on the diameters. In the past, we would com-pute those portions of the equation only when thediameters manually changed. Now, since the diametersare a function of temperature they, and everything basedon them, must be computed on a continual basis.

Assume the measurement system is supplied dr, orificediameter at reference temperature and Dr, pipe diameterand reference temperature. Before these diameters canbe used anywhere in the flow rate calculation they mustbe corrected for thermal effects with the followingequations.

Corrected orifice diameter

Corrected pipe diameter

Change 5, Downstream expansion factor, requiresadditional compressibility

The equations for upstream expansion factor have notchanged. However to compute the downstreamexpansion factor, real gas effects must now be accountedfor. This means an additional Z, compressibilitycalculation is required when computing the downstreamexpansion factor.

If your system measures static pressure downstream,but you do not want to incur the additional processingto compute another Z for the expansion factor, there issomething you can do.

You can compute the upstream pressure as follows anduse it to compute the upstream expansion factor.

where N is a conversion constant from differentialpressure to static pressure units.

If you employ this technique, you must be careful to usePf1 for all occurrences of static pressure in the flow rateequation. You cannot use upstream pressure in someplaces and downstream pressure in others.

Change 6, Fpv, supercompressibility is computed usingAGA-8

Many people have been using NX-19 to compute Fpv fornatural gas. The new standard specifies AGA-8.

A new AGA-8 standard was published in late 1992. Thatstandard documents two possible ways to compute Fpv.One method is referred to as gross method, the other isreferred to as detailed method. The gross method issupposed to be simpler to implement and require lesscomputing power than the detailed method. Havingworked with both, I can tell you that compared to eitherof these methods NX-19 processing requirements arerelatively minuscule (small).

As a user, there are two major distinctions between thegross and detailed methods you should consider.

1. The gross method accepts the same compositiondata you are used to supplying for NX-19 (specificgravity, percent CO2 and N2). The detailed methodrequires a total analysis. What constitutes a totalanalysis depends on each measurement site.Generally, composition through C6s is considered atotal analysis. Sometimes C7s or C8s or C9s mightneed to be broken out. The detailed method of theequation will support this if needed.

2. The gross method is applicable over a narrowerrange of operating conditions than the detailedmethod. The gross method was designed to beapplicable for pipeline quality natural gas at normal


pipeline pressures and temperatures. For example,the gross method supports up to 0.02% HydrogenSulfide, while the detailed method supports up to100% Hydrogen Sulfide.

The following table summarizes the range of applicabilityfor the two methods. The Normal Range column appliesto the gross method. The Expanded Range columnapplies to the detailed method.

Change 7, Implementation guidelines for computers areprovided

As mentioned earlier, Part IV provides these guidelinesand test cases to check out a program. Inclusion ofcomputer related documentation of this sort iscompletely new for the AGA-3 standard and recognizesboth the need for a computer to solve the new equationand the availability of computers to accomplish this.

Change 8, Older factored form of equation not asprominent

This has already been discussed. The factored approachis relegated to an appendix in Part 3 of the new standard.Strictly speaking, appendices are not considered abinding part of the standard. They exist for informationalpurposes. The implementation guidelines in Part IV donot even mention factors as such. See Section 3 of thisdocument for more detail.

Change 9, 50 ppm tolerance on computer solutionsexpected

Part 4 of the standard states:

“The implementation procedures in this documentprovide consistent computed flow rates for orifice meterinstallations which comply with other parts of thisstandard. A particular implementation may deviate fromthe supplied procedures only to the extent that the finalcalculated flow rate does not differ from that calculatedusing the presented implementation procedure usingIEEE Standard 754 double precision arithmetic by morethan 50 parts per million in any case covered by thestandard.” 4

Change 10, Pipe Taps not supported by new standard

The coefficient of discharge research did not include pipetaps. Since Cd is a function of tap location, the new Cd

equation does not support pipe taps. The standard directsyou back to the AGA-3-1985 standard to handle pipe taps.

Change 11, Zb for air changed

Air’s compressibility at base conditions was changedfrom 0.99949 to 0.99959

Change 12, Uncertainty statement was revised

Optimistically, the AGA-3-85 uncertainty statement wasapproximately 0.5%. The new statement is approximately

TABLE 2-2 - AGA-8 Ranges of Applicability

QUANTITY NORMAL RANGE EXPANDED RANGERelative Density (Gr) 0.56 to 0.87 0.07 to 1.52Gross Heating Value 477 to 1150 Btu/scf 0.0 to 1800 Btu/scfMol Percent Methane 45.0 to 100.0 0.0 to 100.0Mol Percent Nitrogen 0.0 to 50.0 0.0 to 100.0Mol Percent Carbon Dioxide 0.0 to 30.0 0.0 to 100.0Mol Percent Ethane 0.0 to 10.0 0.0 to 100.0Mol Percent Propane 0.0 to 4.0 0.0 to 12.0Mol Percent Butanes 0.0 to 1.0 0.0 to 6.0Mol Percent Pentanes 0.0 to 0.3 0.0 to 4.0Mol Percent Hexanes Plus 0.0 to 0.2 0.00 to Dew PointMol Percent Helium 0.0 to 0.2 0.0 to 3.0Mol Percent Hydrogen Assumed 0.0 0.0 to 100.0Mol Percent Carbon Monoxide Assumed 0.0 0.0 to 3.0Mol Percent Argon Assumed 0.0 0.0 to 1.0Mol Percent Oxygen Assumed 0.0 0.0 to 21.0Mol Percent Water 0.0 to 0.05 0.0 to Dew PointMol Percent Hydrogen Sulfide 0.0 to 0.02 0.0 to 100.0Flowing Pressure 1200 psia (8.3 MPa) 20,000 psia (140 MPa)Flowing Temperature 32 to 130 Deg F (0 to 55 DegC) -200 to 400 DegF (-130 to 200 DegC)

Note: This table taken from Table 1, page 3 of AGA-8 Standard 5


0.5% for Cd plus uncertainty in other measured variables.Typical is probably between 0.6% and 0.7%.

This may sound as if the new equation has as muchuncertainty as the old. However, it appears the AGA-3-1985 uncertainty statement was very optimist and, strictlyspeaking, was not technically defensible over all theoperating conditions for which it was being used.

The new standard is expected to improve the uncertaintyby 0.1% - 0.5%. 6

Regarding this issue, a summary of statements takenfrom Part 4 of the new standard follows:

The orifice equation in use through AGA-3-1985 wasbased on data collected in 1932/33 under the directionof Professor S.R. Beitler at Ohio State University (OSU).The results of these experiments were used by Dr. EdgarBuckingham and Mr. Howard Bean to develop thecoefficient of discharge equation.

In the 1970’s, researchers reevaluated the OSU data andfound a number of reasons to question some of the datapoints. This analysis identified 303 technically defensibledata points from the OSU experiments. Unfortunately itis not known which points were used by Buckingham/Bean to generate the discharge coefficient equation.

Statistical analysis of the Regression Data Set (the newdata set) showed that in several regions, the Buckingham/Bean equations did not accurately represent that data.4

This means that the uncertainty statement in the AGA-3-1985 standard cannot be substantiated in all cases.

Changes 13-18, Changes to Part 2 – Specification andInstallation Guidelines

Since this paper mostly deals with the equation, detailsabout changes to the installation requirements are onlymentioned in brief here.

• Diameters’ reference temperature is 68 DegreesFahrenheit.

• Minimum orifice bore thickness is specifiedold no statementnew bore must be larger of

(e >= 0.01d) or (e > 0.005) inch

• Orifice plate thickness specification was changedTable has same values, but statement restrictingrange of applicability to (hw < 200 in. H2O) and (Tf <150DegF)

• Meter tube roughness specification was changedold 300 microinches in all cases

new 300 microinches if Beta < 0.6 and250 microinches if Beta >= 0.6

• Meter tube diameter tolerances were changed

For Any Diameterold range of 0.1 to 0.75 %

depending on Beta

new 0.25% regardless of Beta

For Max-Min Diameter

old range of 0.1 to 0.75 % dependingon Beta

new 0.5% regardless of Beta

• Eccentricity requirement was changed

old

new

• Perpendicularity requirement addedNew statement that orifice plate plane must be keptat an angle of 90 degrees to the meter tube axis.

This concludes the overview of changes in the new orificemetering standard.

SECTION 3 - MORE ON FACTORS

In this section a table is developed to more clearly showthe relationships between the fundamental and factoredequation forms (both AGA-3-1985 and AGA-3-1992). Acomplete derivation of factors will not be shown here. Boththe 1985 and 1992 standards already document thosederivations. To be more instructive, the density terms ofequations 1 and 2 are shown being calculated using densityequations 3 and 4. Additionally, within the densityequations, Gi is computed based on equation 5. Theseequation numbers refer to equations in this document.

1985-AGA-3 fundamental equation shown with densityequations included

Substituting density equations (eq. 3 and eq. 4) and idealgas gravity equation (eq. 5) into the AGA-3-1985fundamental equation (eq. 1) results in equation A-1 asshown on page 180.


TABLE 3-1 Fundamental Equation Terms included in Factored Equation Terms


Substituting equations as shown above results inequation A-1 below.

Equation A-1

AGA-3-1992 fundamental equation shown with densityequations included

Substituting density equations (eq. 3 and eq. 4) and idealgas gravity equation (eq. 5) into the AGA-3-1992fundamental equation (eq. 7) results in equation A-2 asshown below.

Substituting equations as shown above results inequation A-2 below.

Equation A-2

SECTION 4 — PART 4’S NEW EQUATIONPROCEDURES

To date, most publications in the public domain haveincluded summaries of the new equations as presentedin Parts 1 and 3 of the new standard. Since Part 4 of thenew standard is intended as a guide for those taking onthe task of implementing the equation, it seems appro-priate to include a summary of it here. This does notprovide all the necessary information to completelyimplement the new equation, but it should give you afeel for the scope of work involved.

General Outline of Solution Procedures

For this example it is assumed the fluid being measuredis natural gas. The general outline of the solutionprocedures for flange-tapped orifice meters is as follows:

1. At Tf, calculate terms that depend only upon orificegeometry: d, D, b, EV and Cd correlation terms.

2. Calculate the upstream flowing pressure, Pf fromeither Pf2 or Pf1 and Dp.

3. Calculate required fluid properties (Gi, Rhof, andRhob) at Tf, Pf and other specified fluid conditions.

4. Calculate the upstream expansion factor.

5. Determine the converged value of Cd.

6. Calculate the final value of Qb.

Detailed Outline of Solution Procedures

1. At Tf, calculate terms that depend only upon orificegeometry: d, D, b, EV and orifice coefficientcorrelation terms.


Calculate corrected orifice diameter

Calculate corrected pipe diameter

Calculate Beta

Calculate velocity of approach term

Note: In the following equations A0 through A6 and S1through S8 are references to constants that aredocumented in the standard.

Calculate orifice coefficient of discharge constants

Additional Tap Term for small diameter pipe

2. Calculate the upstream flowing pressure, Pf fromeither Pf2 or Pf1 and Dp

3. Calculate required fluid properties (Gi, Rhof, andRhob) at Tf, Pf and other specified fluid conditions.

Using AGA-8 Compute Zb Gas at (Tb and Pb) and Zf Gasat (Tf and Pf), Then, compute Gi, Rhof and Rhob using thefollowing formulas


Compute orifice differential to flowing pressure ratio, x

Compute expansion factor pressure constant Yp

Compute expansion factor

5. Determine the converged value of Cd.

5.0 Calculate the iteration flow factor, Fi, and itscomponent parts, Flc and Flp, used in the Cd

convergence scheme.

Compute Cd’s Iteration flow factor, FI

5.1 Initialize Cd to value at infinite Reynolds numberCCdd0

Ideal Gas Gravity

Flowing Density

Base Density


5.2 Compute X, the ratio of 4000 to the assumedReynolds number

5.3 Compute the correlation value Fc and its derivativeDc, of Cd at the assumed flow, X

5.4 Calculate the amount of change to guess for Cd

5.5 Update the guess for Cd

5.6 Repeat steps 5.2, 5.3, 5.4 and 5.5 until the absolutevalue of is less than 0.000005.

6. Calculate the final value of Qv, the flow rate at baseconditions.

SECTION 5 — PART 3’S NEW EQUATION FACTORS

As stated earlier, a factored form of the new equation isdeveloped in appendix B of Part 3 of AGA-3-1992. To date,most publications in the public domain, have includedsummaries of the new equation as presented in Parts 1and 3 of the new standard. Since these presentations havenot covered the factored form of the equation, apresentation of new equation procedures based on thefactored equation form is included in this section.

For reasons stated earlier, the factored form is notrecommended for most implementations. However, inthe context of comparing the old and new equations,the factored equation is presented in this section forinstructional purposes. As derived in Appendix B of Part3, the factored equation form is as follows:

General Outline of Solution Procedures

For this example it is assumed the fluid being measuredis natural gas and that the inch-pound units of measureare used. The general outline of the solution proceduresfor flange-tapped orifice meters is as follows:

1. At Tf, calculate terms that depend only upon orificegeometry: d, D, b, Ev and Fn.

2. Calculate the upstream flowing pressure, Pf fromeither Pf2 or Pf1 and Dp.

3. Calculate factors associated with densities at Tf, Pf

and other specified fluid conditions. These factorsinclude Fpb, Ftb, Ftf, Fgr, and Fpv.


5. Determine the converged value of Cd e.g. (Fc + Fsl).

6. Calculate the final value of Qb.

Detailed Outline of Solution Procedures

1. At Tf, calculate terms that depend only uponorifice geometry: d, D, b, Ev and Fn.

Calculate corrected orifice diameter

Calculate corrected pipe diameter

Calculate Beta


Calculate Fn

2. Calculate the upstream flowing pressure, Pf fromeither Pf2 or Pf1 and Dp


3. Compute Factors associated with densities (Rhob

and Rhof)





5. Determine the converged value of Cd , e.g. (Fc +Fsl).

Step 5.1 Assume a value for Cd

Assume Cd = (Fc + Fsl) = 0.6

Step 5.2 estimate a value for Reynolds Number by firstestimating Qv

Step 5.3 Calculate the orifice Calculation Factor Fc

Step 5.4 Calculate the orifice Slope Factor Fsl

Step 5.5 Repeat steps 5.2 through 5.4 until Cd, e.g. (Fc +Fsl) changes are acceptably small

7. Calculate the final value of Qv, the flow rate atbase conditions.

SECTION 6 — TOTALFLOW’S IMPLEMENTATION OFNEW ORIFICE EQUATION FOR GAS

This section describes Totalflow’s realtime implementa-tion of the new orifice metering equations. As previouslyshown in Section 2, equation 7 of this document, thefundamental equation for volumetric flow rate is statedas follows.

eq. 7 (restated), AGA-3-1992 Fundamental Equation

Form of the Equation

Part 4 of the new standard exists for the purpose ofproviding implementation procedures that, whenfollowed, produce consistent results for most allcomputer systems. Additionally, Part 1 of the newstandard recommends Part 4 procedures be followed.

The recommended implementation proceduresprovided in Chapter 14.3, Part 4, allows differententities using various computer languages ondifferent computing hardware to arrive at nearlyidentical results using the same standardized inputdata.1

Compute Zbgas and Zfgas usingappropriate AGA-8 method.


Additionally, since Part 4’s implementation uses theequation’s fundamental form it is more easily adaptedto a mass flow equation and can also be handily adaptedto other sets of engineering units.

For these reasons this implementation is based on Part 4of the new standard. This means that factors, as such,are not part of this implementation. However, theequation is still solved as a collection of various terms.These terms are themselves factors of the equation, butthey are not the classic collection of factors historicallyassociated with the AGA-3 equation.

The new standard has clearly relegated the older factoredform of the equation to a less prominent position byputting it in an appendix. It is clear the authors of thenew standard are moving toward the more fundamentalform of the equation.

Integration and Time Related Issues

Equation 7 is a rate equation which must be integratedover time to produce a quantity (volume or mass). Sincethe orifice metering standard does not specify integrationrequirements, these techniques are left to each systemdesigner. Much of this section is devoted to describingtechniques for integrating the fundamental flow rateequation to produce volume.

As illustrated below, portions of the equation arecomputed at different times. The possible times are:

Table 6-1 Names of Calculation Time Periods

Time Period DescriptionCONST (constant) Computed once, never changeSEC (second) Computed once per second

(sample period)VOLP (Vol Period) Computed once per volume

calculation period (useradjustable)

NEW_VOL_CONST Computed when static valuesare manually changed

NEW_COMP Computed when new gasanalysis data is received

Fip

To begin describing these time domain issues, thefundamental equation is rewritten such that the portionof equation 7 under the radical (e.g. ) is set apart as aseparate entity. Part 4 of the standard refers to thisportion of the equation as Fip. For consistency we referto it likewise here.

This results in equations:

eq. 8

where, eq. 9

Equation 9 above contains a flowing density term( )ρ f

that, as discussed in Section 1 of this document, is com-puted using the following gas density equation.

eq. 3 (restated) Density at Flowing conditions

Substituting equation 3’s density solution into equation 9,results in the following equation for Fip.

eq. 10 Fip with gas density equation included

Equation 10 above contains a ideal gas gravity term Gi

that, as also discussed in Section 1 of this document, iscomputed using the following equation.

eq. 5 (restated) Gi computed from Gr

Substituting equation 5’s ideal gravity solution intoequation 10, results in the following equation for Fip.

eq. 11 Fip with Gr used instead of Gi

Equation 11 shows the form of Fip used in this imple-mentation to compute gas volumes. However, portions ofFip are computed on different time periods. To illustratethose portions of Fip, the following equations are provided.

eq. 12 Constants within Fip equation

eq. 13 Supercompressibility within Fip equation

eq. 14 Extension within Fip equation


Restating the Fip equation in terms of the variables solvedfor in equations 12, 13 and 14 results in an Fip equationof the following nomenclature.

eq. 15 Fip with time dependent factors shown

With this final representation of Fip, we can now constructa table showing each portion of the flowrate equation(equation 8) and their respective computation timeperiods. See Table 6-2 above.

Static Pressure and Expansion Factor

As mentioned in Section 2 of this document, if down-stream expansion factor is used then an additional Z(compressibility) calculation must be performed. To avertthe need for this additional processing, this implemen-tation always uses the upstream static pressure therebyallowing computation of the upstream expansion factor.

The user is allowed to specify either up or down streamfor location of the static pressure sensing element. If theupstream location is specified, that pressure measure-ment is used without modification. However, if the down-stream location is specified then the upstream pressureis computed as:

This logic and math execute each second thereby alwaysproviding the upstream static pressure for use throughoutthe whole equation.

Averaging Techniques

Type 1 AveragesAverages constructed from one second samples takenonly during times of flow are maintained for the real timemeasured variables of differential pressure, staticpressure, and flowing temperature.

Type 2 AveragesAverages constructed from all one second samples(regardless of flow) are also maintained for the samevariables.

Type 1 averages are stored in the historical record forperiods in which some quantity (volume or mass)accrued. Type 2 averages are stored for periods in whichzero quantity accrued. This technique provides adequatevolume adjustment averages for downstream processingbut also supports site operations with averages forpressure and temperature even when there is no flowrate.

In older Totalflow devices Type 1and 2 averages werealways based on linear values. In newer Totalflow deviceseither linear or square root averages can be specified.

Other New Implementation Features

• Different Z (compressibility) calculation methods areavailable. These include the latest AGA-8 methods andNX-19. Additionally Fpv can be turned off if desired.

TABLE 6-2 Summary of Calculation Time Periods

SECThe extension is computed andintegrated each second untilVOLP, when it is used in thevolume calculation.

VOLP

CONSTComputed once, neverchanges.

VOLPBut portions are computed ondifferent time periods as shownin following three table entries for

NEW_VOL_CONST andNEW_COMP

VOLPSee equations in Part 4 of standard orSection 4 of this document.

Time Period for ComputationEquation Being ComputedVariable Name


• VOLP, Volume calculation period defaults to one hour,but is user selectable. Selections offered are 1, 2, 5,10, 30, and 60 minutes.

• Up to 23 composition variables for supporting AGA-8detailed method are supported.

• Selectable static pressure tap location is supported.

• Selectable differential pressure tap type is supported.

• Higher static pressure transducers are supported. Upto 3500 psi is currently in use.

Algorithmic Detail of Realtime Implementation of NewEquation for Gas

The following is a more detailed summary of periodiccomputations performed by this implementation forsolving the new orifice equations (AGA-3-1992). Theperiods referred to in this section are those same periodssummarized in Table 6-1. Please note that the followingequations are based on using linear averages, if squareroot averages are selected, then square roots areperformed before they one second summations takeplace.

CONST PERIOD

NEW COMP PERIOD & NEW VOL CONSTS PERIOD

Currently the same calculations are being performed foreach of these two periods. Future optimizations couldresult in different calculations being performed for eachof these two periods.

Perform Fpv Pre-Calculations

IF (Fpv Method = AGA-8 gross)Compute AGA-8 gross method precalcs (e.g. AGA-8 terms that are function of composition) UsingAGA-8 gross method Compute Zbgas

ELSE IF (Fpv Method = AGA-8detail)Compute AGA-8 detail method precalcs (e.g. AGA-8 terms that are function of composition) UsingAGA-8 detail method Compute Zbgas

ELSE IF (Fpv Method = NX19_FIXEDFTFP)Accept user supplied Ft and Fp values

ELSE IF (Fpv Method = NX19)

IF ((Gr < 0.75) AND (CO2 < 15%) AND (N2 < 15%))Compute Ft and Fp using NX19 Gravity Method

ELSECompute Ft and Fp using Methane Gravity MethodENDIF

Fip

Fipconst

ELSE IF (Fpv Method = NX19_GRAVITY)Compute Ft and Fp using NX19 Gravity Method

ELSE IF (Fpv Method = NX19_METHANE-GRAVITY)Compute Ft and Fp using NX19 Methane GravityMethod

ENDIF

Calculate Base Density

SEC PERIOD

IF (Pressure Tap Downstream)Calculated Upstream Static Pressure, Pf

IF (DP > DP_ZERO_CUTOFF) (If Flow Exists)

ENDIF

VOLP (VOL PERIOD)(THIS ONLY EXECUTES IF THERE WAS FLOWDURING THE VOLP)

ELSE

ENDIF


Construct averages from one second accumulators

At Tf, calculate terms that depend only upon orificegeometry: d, D, b, Ev and orifice coefficient correlationterms.

Calculate corrected diameters and Beta


Calculate orifice coefficient of discharge constantsNote: In the following equations A0 through A6 and S1through S8 are references to constants that are docu-mented in the standard.

Additional Tap Term for small diameter pipe

Calculate Fpv at Tf, Pf and other specified fluidconditions (using NEW_COMP precalcs).

IF (Fpv Method = OFF)Fpv = 1.0

ELSE IF (Fpv Method = AGA-8gross)Calculate Zfgas using AGA-8gross method thencalculate Fpv

ELSE IF (Fpv Method = AGA-8detail)Calculated Zfgas using AGA-8detail method thencalculate Fpv

ELSE IF (Fpv Method = NX19_FIXEDFTFP ORNX19_GRAVITY OR NX19_METHANE-GRAVITY)Calculate Fpv using NX19 method and previouslysupplied ft and fp

END IF

Calculate the upstream expansion factor.





Calculate Fip

Determine the converged value of Cd.

Cd_step.0 Calculate the iteration flow factor, Fi, andits component part, Fip. Re-use the Fip com-puted earlier. Then use these in the Cd con-vergence scheme.

Compute Cd’s Iteration flow factor, FI

If Fic < 1000 Fip

Cd_step.1 Initialize Cd to value at infinite Reynoldsnumber

Cd_step.2 Compute X, the ratio of 4000 to the

assumed Reynolds number

Cd_step.3 Compute the correlation value Fc and itsderivative Dc, of Cd at the assumed flow, X

Cd_step.4 Calculate the amount of change to guessfor Cd

Cd_step.5 Update the guess for Cd

Cd_step.6 Repeat steps 2,3,4 and 5 until the absolutevalue of is less than 0.000005.

Calculate the final value of qm , the mass flow rate atline conditions.

Calculate the final value of Qv , the volumetric flowrate at base conditions.

Calculate the final value of Volb , the volume at baseconditions for the Volume Period

Section 7 - NOMENCLATURE

a1 Linear coefficient of thermal expansion of theorifice plate material

a2 Linear coefficient of thermal expansion of themeter tube material.

b Beta. Ratio of orifice plate bore diameter to metertube internal diameter (d/D) at flowing tempera-ture,Tf.

Cd Orifice plate coefficient of discharge.

Cd0 First flange-tapped orifice plate coefficient ofdischarge constant within iteration scheme.

Cd1 Second flange-tapped orifice plate coefficientof discharge constant within iteration scheme.

Cd2 Third flange-tapped orifice plate coefficient ofdischarge constant within iteration scheme.

Cd3 Forth flange-tapped orifice plate coefficient ofdischarge constant within iteration scheme.

Cd4 Fifth flange-tapped orifice plate coefficient ofdischarge constant within iteration scheme.

Cd_f Orifice plate coefficient of discharge bounds flagwithin iteration scheme.

d Orifice plate bore diameter calculated at flowingtemperature Tt.

D Meter tube internal diameter calculated atflowing temperature Tf.

dr Orifice plate bore diameter calculated atreference temperature Tr.

Dr Meter tube internal diameter calculated atreference temperature Tr.

V


Dc Orifice plate coefficient of dischargeconvergence function derivative.

DP Orifice differential pressure.

e Napierian constant, 2.71828.

Ev Velocity of approach factor.

Fc Orifice calculation factor for Cd (Used differentlyin Parts 3 and 4)

F s l Orifice Slope Factor for Cd

F l Iteration flow factor

F l c Iteration flow factor — independent factor.

F lp Iteration flow factor — dependent factor.

Fmass Mass flow factor.

Fb Basic orifice factor.

F r Reynolds number factor.

Fpb Pressure base factor.

Ftb Temperature base factor.

F t f Flowing temperature factor.

Fgr Real gas gravity factor.

Fpv Supercompressibility factor.

Fa Orifice thermal expansion factor.

gc Dimensionless conversion constant.

Gi Ideal gas relative density (specific gravity).

Gr Real gas relative density (specific gravity).

k Isentropic Exponent.

m Mass.

Mrair Molar mass (molecular weight) of dry air.

Nc Unit conversion factor (orifice flow).

N1 Unit conversion factor (Reynolds number).

N3 Unit conversion factor (expansion factor).

N4 Unit conversion factor (discharge coefficient).

N5 Unit conversion factor (absolute temperature).

Nvtime Time Interval Constant used in flowrateintegration algorithm to produce quantity volume

Pb Base pressure.

Pf Static pressure of fluid at the pressure tap.

Pf1 Absolute static pressure at the orifice upstreamdifferential pressure tap.

Pf2 Absolute static pressure at the orifice down-stream differential pressure tap.

Pmair Measured air pressure.

Pmgas Measure gas pressure.

π Pi, 3.14159

qm Mass flow rate at actual line conditions

qv Volume flow rate at actual line conditions.

Qv Volume flow rate per hour at base conditions.

R Universal gas constant.

ReD Pipe reynolds number.

ρb Density of the fluid at base conditions, (Pb, Tb).

ρbair Air density at base conditions, (Pb, Tb).

ρbgas Gas density at base conditions, (Pb, Tb).

ρf Density of the fluid at flowing conditions, (Pf, Tf).

Tb Base temperature.

Tmair Measured temperature of air.

Tmgas Measured temperature of gas.

Tf Flowing temperature.

Tr Reference temperature of orifice plate borediameter and/or meter tube internal diameter.

Td Downstream tap correction factor.

Ts Small meter tube correction factor.

Tu Upstream tap correction factor.

Volb Quantity Volume at base conditions

X Reduced reciprocal Reynolds number (4000/ReD).

Xc Value of X where change in orifice plate co-efficient of discharge correlation occurs.

Y Expansion factor.

Yp Expansion factor pressure constant.

Zb Compressibility at base conditions (Pb, Tb).

Zbair Air compressibility at air base conditions (Pb, Tb).

Zbgas Gas compressibility at gas base conditions (Pb,Tb).

Zf Compressibility at flowing conditions (Pf, Tf).

Zmair Air compressibility at air measurement con-ditions, (assumed Pb, Tb).

Zmgas Gas compressibility at gas measurement con-ditions, (assumed Pb, Tb).


Section 8 — CITED PUBLICATIONS

1. American Petroleum Institute Measurement onPetroleum Measurement Standards (API MPMS)Chapter 14.3, Part 1; Also recognized as AGA ReportNo. 3 Part 1; Also recognized as GPA 8185-92, Part3; Also recognized as ANSI/API 2530-1991, Part 1




5. American Gas Association (AGA) TransmissionMeasurement Committee Report No. 8; Alsorecognized as API MPMS Chapter 14.2.

6. Teyssandier, Raymond G.; Beaty, Ronald: New orificemeter standards improve gas calculations, Oil & GasJournal, Jan. 11, 1993

7. ANSI/API 2530: Second Edition, 1985, OrificeMetering Of Natural Gas and Other RelatedHydrocarbon Fluids; Also recognized as AGA ReportNo. 3; Also recognized as GPA 8185-85; Alsorecognized as API MPMS Chapter 14.3, API 2530.

Brent E. Berry


A REVIEW OF THE REVISIONS TO API 14.3/AGA 3–PART 2Tom Cathey

JW Measurement Company515 N. Sam Houston Parkway, Houston, TX 77060

INTRODUCTION

In April of 2003, revisions to the specification andinstallation requirements for orifice meters was publishedby the American Gas Association in the form of the AGAReport No. 3–Part 2, Fourth Edition. The revisions orchanges in the following categories are significant whencompared to the 1991 Third Edition publication of AGAReport No. 3 and will be discussed in greater detailthroughout this paper:

• Flow Conditioners• Required Meter Tube Lengths• Meter Tube Surface Roughness• Orifice Plates• Thermometer Well Location• Pulsation Environment

FLOW CONDITIONERS

Flow conditioners are divided into two categories:flow straighteners and isolating flow conditioners. Flowstraighteners are defined as “Devices that remove or havelimited ability to accurately replicate the orifice platecoefficient of discharge database values.” Isolating flowconditioners are defined as “Devices that effectivelyremove the swirl component from the flowing streamwhile redistributing the stream to produce flow conditionsthat accurately replicate the orifice plate coefficient ofdischarge database values.”

The 1998 Uniform Concentric 19-Tube bundle isconsidered a flow straightener and the physicalspecifications are considerably different from therequirements listed in the 1991 publication forstraightening vanes. The individual tubes must be ofuniform smoothness, outer diameter and wall thickness.Commercially available seamless carbon steel tubing isreadily available and most commonly used. The individualwall thickness of the tubes shall be less than or equal to2.5% of the published internal diameter. For example:The individual wall tube thickness for a 3” schedule 40meter tube (3.068”) must be .025 X 3.068 or 0.0767” orless. The individual tubes must be chamfered on bothends not less than 50% of the wall thickness by 45degrees. The tubes must be arranged in a cylindricalpattern and the individual tube outer walls must come indirect contact with each other. The outside diameter ofthe tube bundle must be a minimum of 95% of thepublished internal diameter of the meter tube and canobviously be no greater than the published internaldiameter. In order to ensure that the bundle outer tube

walls come in direct contact with each other and achievean outside diameter greater than or equal to 95% of thepublished inside diameter of the meter tube, the correctO.D. tubing must be used as illustrated in the table belowfor schedule 40 piping.

Meter Vane Minimum Vane Tube ID ” Tube O.D. ” Bundle O.D. ”

3.068 19/32 2.9146 4.026 13/16 3.8247 6.065 1-3/16 5.7618 7.981 1-5/8 7.5820 10.020 2 9.5190

The required length of the tube bundles must beas follows:

• 3 X NPS for 2”• 2.5 X NPS for 3” & 4”• 2 X NPS for 6” and above

Flow conditioners not meeting the requirements of the1998 Uniform Concentric Tube Bundle are consideredas Other Flow Conditioners. While the 2000 publicationdoes not recommend any particular type of flowconditioner, specific criterion for evaluation of installationand/or flow conditioner testing is provided. These testsdefine the meter tube lengths and flow conditionerlocations for acceptable performance. Significantresearch and numerous flow studies have beenconducted since 1991 to test various flow conditionerdesigns with repeatable and acceptable results.

REQUIRED METER TUBE LENGTHS

A beta ratio of.75 should be used as the design criteriafor new orifice meter installations. The 2000 publicationprovides required minimum installation lengths for metertubes with no flow conditioners in Table 2-7 and minimuminstallation lengths for meter tubes with the 1998Concentric 19-Tube Flow Straighteners in Tables 2-8aand 2-8b. Two configurations to be noted for tubeswithout flow conditioners in Table 2-7 are:

• Two 90°elbows in perpendicular planes whereS<5Di

• Any other configuration (catch all category)

For the two 90° elbows in perpendicular planes, therecommendation is 95 published inside diameters


beginning at beta ratios of .50 through .75. Forconfigurations not specifically addressed in Table 2-7,145 published inside diameters are recommended forbeta ratios of .40 through .75.

For tubes with The Uniform Concentric 1998 19-TubeBundle, there are multiple configurations that are notallowed at the higher beta ratios. This means that it isnot possible to find an acceptable location for the 1998Concentric 19-Tube Bundle downstream of the fittingfor all values of upstream length. A few examples follow:

• Single 90° tee used as an elbow but not as aheader element for beta ratios of .67 and higher.

• Partially closed valves (at least 50% open) forbeta ratios of .60 and higher.

Required minimum lengths for other types of flowconditioners are not specified although flow testing insitu and at flow-testing laboratories is addressed. Manycompanies have now changed their engineeringstandards to accommodate other types of one and twopiece flow conditioners such as those produced byGallagher, Canadian Pipeline Accessories and Daniel.

METER TUBE SURFACE ROUGHNESS

In the 1991 and 2000 publications, sections 2.5.1.1through 2.5.1.1.3 address the inside surface of metertubes. There are several changes for tubes greater than12 inches in diameter and a minimum surface roughnessis specified for the first time. To illustrate the changesmore clearly, the 1991 requirements are listed followedby the 2000 requirements.

1991

• 300 micro inches for beta ratios less than 0.6

• 250 micro inches for beta ratios greater than orequal to 0.6

2000

For meter runs with nominal diameters of 12 inches orsmaller:

• 300 micro inches for diameter ratios equal to orless than 0.6

• 250 micro inches for diameter ratios greater thanor equal to .60

• The minimum roughness shall not be less than34 micro inches for all diameter ratios

• For meter runs with nominal diameters largerthan 12 inches:

• 600 micro inches for diameter ratios equal to orless than 0.6

• 500 micro inches for diameter ratios greater thanor equal to 0.6

• The minimum roughness shall no be less than34 micro inches for all diameter ratios.

ORIFICE PLATES

There are revisions to 8 inch and 24 inch plates in the2000 publication as well as maximum allowabledifferential pressures for all meter tube diameters.Changes in recommended thickness for 8 inch and 24inch orifice plates are as follows:

Recommended Orifice Plate Thickness

1991 8” .125”2000 .250”1991 24” .375”2000 .500”

The maximum recommended differential pressure was200” W.C. for all tube diameters and plate bores in the1991 publication. The exception was the 8 inch platewith a .125” thickness which was limited at 150” W.C. toprevent plate deflection. In the 2000 publication,maximum recommended differential pressures are listedin Table 2-3. These differential pressures apply tostainless steel orifice plates at a maximum operatingtemperature of 150°F. For all orifice flange fittings orunions, the maximum differential pressure is listed at1000 inches of water column. For other orifice fittings(single and dual chamber) differential pressuremaximums vary from a minimum of 180” to a maximumof 1000”.

THERMOMETER WELL LOCATION

The required lengths downstream of the orifice platerelated to thermometer well location remain unchangedfrom the 1991 to the 2000 publication. In the 1991publication, the thermometer well could be locatedbetween 12 –36” upstream of the straightening vane. Inthe latest revision, the thermometer well may be locatedno closer than 36” upstream of the flow conditioner inlet

PULSATION ENVIRONMENT

Pulsation is addressed in section 2.6.4 of the 2000publication. The allowable pulsation environment isstated as follows:

D P rms / D P avg £ 0.10

or:

“Accurate measurement of flow with an orifice meteroperating under pulsating flow conditions can be ensuredonly when the root mean square of the fluctuatingdifferential pressure amplitude normalized overdifferential pressure time mean does not exceed 10%.”


This applies to single frequency pulsations caused byreciprocating compressors or releif/blowdown valves andto broadband flow pulsation or noise caused by throttlingvalves. There is currently no reliable means to adjust for

indicated pulsation and it is stated that attempting to doso may actually introduce more error.

Every effort should be made to eliminate the source ofthe pulsation which can usually be accomplished throughproper design of the facility.


A NEW PERSPECTIVE ON MEASUREMENTTHE IMPACT OF MEASUREMENT IN ACHANGING BUSINESS ENVIRONMENT

David WoffordShell Gas Transmission, LLC

1301 McKinney, Suite 700, Houston, TX 77010

The measurement of hydrocarbons has evolvedsignificantly through the years, from both a technical andbusiness application perspective. Developments andadvances in technology have made the measurement ofhydrocarbons more precise, efficient and available.Changes in the energy business environment haveplaced the measurement of hydrocarbons into a moresignificant role within organizational and industrybusiness processes.

A HISTORICAL OVERVIEW OF THE NEED TOMEASURE

The Roman’s discovered the value in measuring andcontrolling the flow of water throughout the aqueductsof their cities in order to better manage resources andserve the needs of the populace. The Chinese firstdeveloped pipeline systems made of bamboo and woodto transport hydrocarbons and water from supply regionsto consumption areas. These early concepts andapplications of natural resource acquisition, delivery andmanagement were provisional to the needs of primarilyAgrarian societies. The resources were consumed withinthe context of meeting the basic needs of people withinthe scope of their existence.

The onset of the Industrial Age changed the value andapplications of the available natural resources tosocieties. No longer were these resources only applicableto agriculture, cooking, and lighting (via torches andlamps). These resources now were used as energysources to ultimately create products and power. Oil andgas could be burned to create steam from water, whichin turn could be applied to drive machinery that couldproduce more work output than humans alone.

As the properties of hydrocarbons became betterunderstood, such could be processed and refined intosub products that had even greater uses. These productscould be used as raw materials to create other products,such as textiles, fertilizers, chemicals and additives forother products, or as end products such as fuels. As theuses for hydrocarbons grew, so did the value. Themeasurement of hydrocarbons and the subsequentrefined products began to take on greater importancebecause of the increased value of such.

One of the greatest technological innovations thatdramatically affected the hydrocarbon industry was theinvention of the internal combustion engine. Thisinvention enabled the direct use of hydrocarbons as afuel to power machinery. Since the energy contained

within the hydrocarbons could be directly utilized to drivethe machine, rather than indirectly as with the generationof steam, machines became more compact and mobile.Machines became available to virtually everyone as anaffordable means of transportation, such havingrevolutionized the means by which the entire worldfunctions.

But the key to it all was still the hydrocarbon. Thehydrocarbon based fuel must be available to operatethe machinery around which the world functions. Thishas given even greater value to the hydrocarbon andthe ultimate measurement and quantification of such.As technology has continued to advance, not only arethe quantities of hydrocarbons impor tant, but thecomposition and quality of the hydrocarbons as well.

MEASUREMENT IN THE “OLD” ENERGY BUSINESS

The “Energy” business was born per the Industrial Age.Fuels were needed to power machines. An entire industryevolved from the need to discover, gather, transport,refine and deliver fuels. Because these fuels were ofvalue, such had to be measured.

The measurement of fluids was accomplished byphysically determining the amount of the fluid that passeda given point. The principals of these measurements werefounded in applying the physical properties of the fluidsto the principals of geometry and physics of the mediumin which the fluid traveled. These measurements werepart of operating the transport systems to move the fluidsfrom one point to another.

The measurement of fluids was accomplished by strictlymechanical means. The most common means tomeasure was by passing the fluids though a restrictionto create a pressure drop. The drop in pressure wouldbe measured and related to the conditions under whichthe fluid was contained and a rate of flow determined.Devices were developed to record the measurement ofpressure drop (differential pressure), static pressure, andtemperature of the fluid. These recordings were theninterpreted in relation to time in order to determinequantities.

The most common means of measuring hydrocarbonfluids became the orifice plate. A thin plate was placedin the pipe. The plate had a bore (usually concentric withinthe plate and pipe) through which the flowing fluid wouldpass. The pressure on each side of the plate would bemeasured to determine the pressure drop. This means


of measuring the flowing fluid proved very reliable in thatthe systems withstood very harsh conditions andrequired little maintenance.

Other types of meters were also developed to measurehydrocarbon fluids. Turbine meters, wedge meters,displacement meters and others began to becomecommonly employed to measure hydrocarbons. All weremechanical systems and were employed per theconditions that existed relative to the needs andquantities of the fluids to be measured. The systemsemployed to record the measurements of the fluids werealso of strictly mechanical means, and requiredsubsequent interpretation to determine quantities.

Measured quantities of hydrocarbons were then reportedto administrative groups within the energy organization.The accountants would then apply the price per unit offluid to the total quantity for the time period and astatement would be forwarded to the party with whichthe transaction took place. The process was no differentthan any other business transaction, such as dry goodsor groceries. The product was priced, the quantitiesmeasured or counted, the total price determined andthe transaction finalized. The sequence of events increating and finalizing the transaction were fixed and verysimple. A to B, B to C, C to D, and so on. The need tomeasure or count units was just a necessary step in theprocess and was a physical function left to those whooperated the pipeline systems.

These were the means by which the energy businessand the measurement of hydrocarbons existed for many,many years. Subtle changes would occur per thedevelopment of some new mechanical measuring orrecording systems, but for the most part, no significantimpacts were realized as far as the role of measuringwithin the organization.

AN EVOLUTIONARY STEP

Along came the microprocessor, a small silicone basedchip that could perform a multitude of mathematicalfunctions that normally required the efforts of manypeople operating many machines to reach the sameresult. Computing had been around for a while, but suchwas not available to everyone. Large mainframesdominated the information management world and theuse of such was limited in scope to the rigid structuralenvironment of the system. Mainframe time was limitedand expensive. The PC microprocessor introducedaffordability and accessibility into the equation.

Until this time, hydrocarbons were traded virtually interms of only quantity. “How much” was the only realissue. Content of the quantity and how much total energywas available were not so important. Gas was still cheapand the determination of quantities was very laborintensive. Detailed determinations of compositions andenergy values were even more complex and had marginalmonetary value to the transaction.

This all changed in 1978 per the passage of the NaturalGas Pricing Act. This legislation was the cornerstone forthe means by which natural gas is traded today, in termsof energy rather than just quantity. Sudden and dramaticincreases in the value of natural gas, as well as otherhydrocarbon based fuels, brought to light the need toconsider the quality of the product, not just the quantity.It became the difference in buying a bicycle or a sportscar. Both are transportation vehicles to get you from pointA to point B, but one has significantly more features andperformance characteristics than the other, and thus, thedifference in value to the consumer.

Now natural gas was exchanged in terms of BTU’s orTherms. The total amount of energy delivered was thebasis for the transaction. Computing equipment hadbeen implemented into the business processes, butmeasurement still was very labor intensive.

MEASUREMENT AND ENERGY GO “HIGH TECH”

The introduction of the microprocessor meant a new erafor measurement. Computers could now be employedto record measurements, calculate volumes, storeinformation and communicate with other informationsystems. Microprocessors also allowed the developmentof gas quality systems that could be implemented at themeasurement facility. Such instruments were previouslylimited to the central laboratory. Fluid composition andquality could now be determined “On Site”.

Various devices could be interfaced together in order tocombine the recordings of measured physical variables,compositions and thermal values to render a totaldelivered energy quantity. This information could bemade available to business units within organizationsmuch faster in order that such could be applied tocommercial transactions on a more timely and precisebasis.

Energy could now be transacted upon in virtual “RealTime”. The historic waiting period to gain access to vitalinformation because of energy measurement, dataprocessing and information provision to business unitshad been reduced tremendously by the development andimplementation of computing and informationtechnologies.

The microprocessor also enabled new research anddevelopment to occur with existing primary measurementdevices, as well as the development of new primarymeasurement systems. Greater precision could beobtained by employing newly developed measurementand calculation mechanisms per the acquisition andimplementation of better research data. New primarymeasuring elements, such as ultrasonic meters, weredeveloped for use in measuring fluid flows. Computingpower enabled these meters to be implemented on aproduction basis because the huge amounts of dataacquisition and processing required to precisely andsuccessfully utilize these systems could now beaccomplished.


These technologies that spawned the creation of newmeasurement, recording and communication systemscreated huge amounts of data to be managed. Client-Server based information systems were developed toefficiently receive, process, validate and transmit thisinformation to business systems. More and betterinformation could now be derived, processed and utilizedwith much less required work and human intervention.

ENERGY GETS “REAL TIME”

Readily available and precise information changed thescope of the energy business. Energy could now betraded in “Real Time”. Precise quantities of energy couldbe transacted upon on a daily or even hourly basis, atdifferent prices and under different conditions. Theenergy market place now resembled the stock market.

A key to this market concept being successful, though,was the integration of the measurement of products intothe business cycle mainstream. No longer couldmeasurement be maintained as an upstream processwhose product slowly trickled into the business cyclefor “end of period” processing. The function had to bean integral part of the dynamic process that enabledbusiness to be conducted and finalized “Now”. Hugemonetary benefits to all interested parties were at stake.Not only was measurement an engineering andoperations process, but now a vitally integrated businessnecessity.

Because of this new integration into the businessrelationship, measurement personnel had to be moreknowledgeable of the energy business process. Also,personnel had to have greater cognizant understandingof other business processes, such as accounting,finance, contract administration and commodities tradingpractices and implications. Measurement was now a truly“Business” process, not just another step that wasnecessary in the core operation of the pipeline system.

This also created a need for measurement personnel withnew skills and education. Not only were engineering andoperations principals important, but also proficienciesin business disciplines and processes. The new“Measurement Man” was now required to wear morehats, have broader understandings of the entireengineering, operations and business processes, andcontinually develop new proficiencies as technologiesand businesses practices progressed.

Discussions among today’s industry measurementmanagers not only focus upon the engineering andoperations aspects of the process, but as much uponthe business related implications of the measurementprocesses as well. Highlighted topics at today’s industryconferences are often the impact of LAUF (Lost andUnaccounted For) product, minimizing the “after the fact”processes of having to “scrub” measurement data inorder that such may be immediately applicable to currentbusiness transactions, and the continuing developmentand implementation of measurement informationtechnologies that are integrated directly into the primarybusiness system platforms of the organization.

ENERGY MEASUREMENT AND THE FUTURE

So where does measurement go from here? What arethe important aspects of the entire measurement processin relation to the dynamic environment of today andtomorrow’s energy business? These are issues that shallbe explored, defined and implemented by energymeasurement professionals.

All of the answers are not currently known, but there isone significant aspect of the energy businessenvironment that will be at the top of the agenda. Industrymeasurement professionals must work in concerttogether in the development and implementation ofmeasurement technologies, applications and practices.No longer is any energy organization an “island”. Industrywide participation and acceptance must be achieved inorder to ensure that energy measurement is, andcontinues to be, performed with technical, business andethical integrity.

This concept will not be limited to energy in America,but will have global implications. Energy today, all energy,whether raw or produced, is a global commodity. Industryorganizations worldwide are working to developstandards and practices that are globally applicable andacceptable. Today and tomorrow’s measurementprofessionals must actively participate within thisendeavor in order to ensure that our industry andorganizational interests, concerns and ideas aresatisfactorily addressed and considered in order that thefuture of measurement in the energy operations andbusiness environments is an integral part of the futureenergy industry.

David Wofford


PROBLEMS UNIQUE TO OFFSHORE MEASUREMENTWayne T. Lake

Independent Measurement Consultant

INTRODUCTION

As the worldwide demand for oil and gas forces offshoreexploration into waters off the continental shelves intodepths of over a mile deep, capital expense spending(CapEx) and production operation expense (OpEx)budgets are slashed and the Environmental Health andSafety (EH&S) requirements as well as some companies’goals for a ‘greener image’ raises the standards ofoperations even higher, the demands placed on accuratehydrocarbon measurement with minimal maintenance,space and weight requirements becomes increasinglygreater. These financial, governmental and technicalchallenges coupled with normally high flow rates andtherefore wide flow range requirements have enhancedthe development and application of new technology suchas ultrasonic gas and liquid meters, multiphase flowmeters, microwave and near infared (NIR) water cutanalyzers, coriolis flow meters for oil and gas andcompact orifice meter tubes utilizing isolating flowconditioners and liquid meter provers. This paper willattempt to provide guidelines in selecting, installing andoperating this equipment to insure cost effective designsand reliable operation with a high degree of accuracy.Since the author’s background is primarily in projectdesign, emphasis will be placed on the decision processof selecting, installing and commissioning meteringequipment.

DESIGN DILEMMAS

Nearly every E&P project group responsible for thedesign, engineering and fabrication of an offshoreproduction platform goes through a decision processwhereby conventional measurement equipment (orificemeters for gas sales and pipe provers for liquid turbine/displacement meter calibration) are compared toalternative equipment based on space and weightrequirements. The project groups’ responsibility toreduce CapEx by installing compact metering must bebalanced by operational factors such as reliability ormean time between Failure (MTBF) which translates toOpEx, safety, governmental and contractualrequirements and approval of interested parties (partners,purchasers and pipeline operators).

Recently, regardless of country location, depth of wateror fluid application (oil or gas field), there is nearly alwaysa very strong consideration for the use of MultiphaseFlow Meter (MPFM) for well testing and/or allocationmainly due to the estimated reduction of topsides weightand space when compared to conventional separators

and associated metering equipment. The vendors havewhat appears to be an infinite supply of information tosupport the installation of MPFM in various applications.

Engineering studies have shown that ‘alternative meteringconcepts (multipath ultrasonic meters for gas and compactprovers for liquids) with the same accuracy of today’sconventional concepts, might reduce space and weightby more than 50% compared to present layouts. The totalcost savings might be twice the actual procurement costof the metering skid.’1 Although this does not addressreliability, it is inferred that these alternative devices areas reliable as their conventional counterparts. Anotherengineering study conducted on behalf of North Seaoperators estimated (based primarily on vendor’s input)the MTBF for multi-path ultrasonic meters to be 2 hoursof downtime over a period of 78 years or an estimated‘uptime’ of 99.9997%! However, this may not be the caseas an offshore platform in the North Sea that recentlyinstalled 13 ultrasonic meters have reported 11 failureswithin the first two years of operation. The lack of reliabilityin emerging technology is not always the case, as withcompact provers that have proven to be a real workhorsein offshore applications, some with well over 100,000cycles or ‘piston strokes’ between failures in offshore crudeoil applications.

Lately, a very common equipment selection discussionrevolves around the application of conventional orificeversus multi-path ultrasonic measurement equipment forthe custody transfer of natural gas. Arguments supportingthe use of the ultrasonic meter over orifice include spaceand weight saving, increased flow rangeability, reducedpressure drop, inherent diagnostics, reducedmaintenance (calibration), tolerance to entrained liquids(wet gas) and improved accuracy to name a few.

DEEP WATER EXPLORATION CHALLENGES

The incremental cost ($/lbm) to support topsides facilitieson a deep water floater such as a Tension Leg Platform(TLP) is estimated to be 5.5 $/lbm (excluding deck, drillingfacilities and hull) which translates into $180,000 for 100’of 20” Sch 120 pipe with two pair of 600# RF flanges.This estimated incremental cost does not reflect anyassociated cost savings for possibly decreasing the decksize and weight by reducing the size of facility equipmentsuch as orifice meters or liquid provers. Another deepwater challenge is the potential for hydrate formation inflow lines as the seabed temperature at 3000+’ of wateris 34°F which is accentuated by the unbelievableapproximate cost to work over a subsea completed well


at this depth of $12,000,000. These significant costfactors force every deep water project team to investigatenew ways to reduce weight and space or ‘footprint’ oftopside measurement equipment and to work withvendors to develop new equipment to measure relativelysmall amounts of free water and/or water vapor forhydrate control.

ENVIRONMENTAL CONCERNS

There are various international, regional and nationalconventions, agreements and laws that defineoperational standards for oil-contaminated effluents anddischarge water from offshore platforms and facilities.Typically, the average discharge limits of oil in water is29-40 mg/l (36-50 ppmv for 0.8 SG oil) over a period of30 days with maximum discharge levels 42-100 mg/l.Some major oil companies have endorsed self imposed‘greener’ guidelines to further reduce emissions;BPAmoco plans to maintain total current emissions levels(that meet or exceed regional guidelines) regardless ofnew field development or production rates. Theseconservative discharge limits place increased demandson separation facilities and associated measurementequipment in mature oil fields where water cuts aregreater than 60-80%.

FLUID PROPERTY AND OPERATIONAL ISSUES

Today’s typical offshore platform location of sales orallocation measurement equipment is downstream offinal phase separation facilities (no dehydration of gas,if not compressed) and immediately before the fluidleaves the platform in a subsea pipeline. The measuredfluids at this point, although separated as well aseconomically possible, are typically at hydrocarbon andwater dewpoint for gas and at bubble point for liquids.Neither of these above described fluids conditions areconsidered ideal for custody transfer measurement andsampling. In the event of any upset in the productionseparators, liquids may carry over the top thus allowingliquids in the gas line or gas may carry under allowingfree vapors in the liquid line. Even without an occasionaloperational upset, any normal cooling of gas due toambient temperatures or inevitable pressure drop dueto frictional piping losses will cause liquids to condenseand likewise any pressure drop in the liquid line will allowfree gas to evolve. These potential separator and resultingmultiphase fluid problems will cause numerous meteringproblems such as liquid accumulation near the orificeplate, cavitation in liquid meters, problems in obtaininga representative sampling from gas streams, repeatabilityin proving of liquid meters and if liquids are introducedto an on-line gas chromatograph, catastrophic failure ofthe analyzer.

EQUIPMENT SELECTION FOR OPTIMAL DESIGNSOLUTIONS

For offshore fiscal gas metering applications whereeconomic space constraints require a compact designwith a high degree of reliability, life of field design criteria

requires extreme rangeablity in flow rate and of courseaccuracy is considered to be essential, the author’spreference is to use two or more conventional, concentricorifice meters installed in parallel as follows:

• low loss, isolating flow conditioner with aminimum of thirteen (13) pipe diametersupstream meter tube

• maximum thickness allowed orifice plates• single, 0-400 IWC differential pressure range

smart type transmitter• orifice flange taps oriented above the pipe

centerline (12 o’clock preferred) with transmittersinstalled on direct mount, full bore manifolds

The above design when using 0.2-0.6 orifice to pipe ratio(ß) and 30-150 inches of water column (IWC) differentialpressure for normal operations and a maximum ß of 0.66and 300 IWC differential pressure for emergency capacityoperations will provide a flow range of 80 to 1 with anestimated random uncertainty in volume of less than±0.75%. This uncertainty may be validated from thefollowing sources:

• offshore, wet gas pipeline accumulated systemenergy and volume balance of <0.2%

• flow conditioner tests results from SouthwestResearch Institute (SWRi)2

• orifice discharge coefficient data from API 14.3Part 1

• mass error due to plate bending by Jepson andChipchase3

• differential pressure transmitter field calibrations

With all the vendor ‘information’ available and theemphasis to reduce deck space to save CapEx andreduce maintenance in order to save OpEx, it would bevery easy to conclude that the multi-path ultrasonic is abetter choice over the orifice meter for offshore, wet gasapplications. After all, the ultrasonic meter is reported tobe more accurate than the orifice (when wet calibrated),more tolerant of the effects of wet gas, requiresignificantly less deck space and maintenance and havegreater flow range capability.

However, let’s take an objective look at each of thesecomparison claims starting with the accuracy claim.Regardless of the vendors’ statements on meteraccuracy, keep in mind that the reliability or mean timebetween failure (MTBF) is also extremely important whendepending on the meter’s output for the monthlyaccounting statement so that a loss of data for any reasonwill always produce negatively biased errors (losses tothe seller) such that a downtime of one (1) hour in acontract month will cause a –0.14% error and an eight(8) hour downtime will cause a –1.1% error.

The implementation of an isolating flow conditionerinstalled at a proper distance upstream (13-17 pipediameters overall from last piping disturbance to theplate) will not only reduce the upstream lengthrequirements historically required for an orifice meter but


lab tests have shown a near perfect correlation andexcellent precision with the API 14.3 Reader-Harris/Gallagher (RG) empirical coefficient of discharge equation(data base using 45-80 diameters of straight pipeupstream) over a wide range of ß ratios. Thisimprovement in measurement is due to the isolating flowconditioner’s capability to eliminate any effects fromupstream piping and create an ideal flow pattern oraxisymmetric velocity profile, free of swirl for virtually allworst case disturbances.

There is an inherent overall uncertainty advantage of theorifice over the ultrasonic in that the orifice is an inferentialhead type device with flow computed as a function ofthe square root of differential pressure and fluid densityas opposed to the ultrasonic meter, a linear device sothat any error in density will have roughly twice theadditive effect on the ultrasonic.

As stated above, the use of dual orifice meter runs withisolating flow conditioners upstream, taps rotated abovecenterline, smart transmitters mounted directly on fittingsby means of full bore manifolds, higher differentialpressure ranges and thicker plates with ( ratios up to0.66, allows for a flow range of 80 to 1, prevent dishingof orifice plates from inadvertent blow downs and providefor an accurate, reliable wet gas system balance(<±0.2%) with minimal maintenance requirements.

Although, the issue of improved wet gas tolerance havenot been fully evaluated at this date (data is currentlybeing compiled as part of the GRI sponsored Wet GasMetering JIP conducted at CEESI), the use of selfdraining, full bore direct mount manifolds and tap rotationabove pipe centerline minimizes any detrimental effectsin the impulse lines. Regarding reduced maintenance,the smart type transmitters appear to be very stablerequiring less frequent calibrations, making this a mootpoint as most companies prefer to have qualifiedtechnicians carefully check all metering components ona monthly basis, especially if the gas volume is significant.

Offshore gas volume measurement facilities may becomplimented with reliable, accurate on-line gaschromatographs (GC) to providing real time energymeasurement provided the GC is installed properly asfollows:

• sample probe installed with the tip in the centerthird of the pipe

• use heat traced, 1/8” SS tubing to insure noliquid drop out and minimal lag time

• heated regulator (located near the probe) toinsure no condensation due to J-T cooling

• 1/8” SS heat traced tubing from the regulator tothe GC sample inlet

• emergency shut off solenoid valve in the sampleline — fail upon high-high level alarm from theproduction separator

• inlet sample filter types and sizes to minimizepossibility of liquid contamination withoutremoving any heavy end hydrocarbons

• protect sample exhaust manifold from windvelocity effects

• appropriately blended, tested and heatedcalibration gas

For offshore fiscal liquid metering applications (custodytransfer and allocation), where deck space, costeffectiveness, pressure drop, fluid stability (bubble point)and accuracy are critical issues, meters may be installedas follows:

• dual (parallel) metering is preferred• locate meter/prover at least one deck below

separator• use oversized, low loss piping to minimize

pressure drop• operate separators at highest liquid level,

especially during proving• install small volume prover upstream of meter(s)• install separator control valve(s) downstream of

metering• locate sample probe downstream of meter(s) in

a vertical pipe section

The above design does not require a pump to increasepressure above the bubble point, but simply uses thefluid hydraulic head pressure and meter componentlocation to maintain sufficient pressure for metering andproving. This design has been validated to providerepeatable3 results (repeatability <0.05% for fiveconsecutive runs) where repeatability is defined asfollows:

Repeatability (%) = (Vhigh –Vlow) * 100/Vavg

The type of meter selected should be based on theparticular application depending on gravity, flow rate,viscosity, sand production and water cut (if operatingseparator in two phase mode). Regardless of meterprincipal of operation or type, low pressure drop sizesand models are required.

CORROSION SOLUTIONS AND PREVENTION

The issue of external corrosion due to high humidity, seaspray, salt water washdowns and deluge systems iscommon to all offshore facilities. Solutions to corrosionproblems include the use 316 SS over 304 due toincreased resistance from chloride pitting due to 3-4%molybdenum content. Care should be given to insure allcomponents are resistant to corrosion as if two stainlesssteel components are fastened with a mild steel, even ifcadium plated, the result will be evident in a matter ofweeks or even days. The use of Denso ‘Petrolatum tapesystems’ in highly corrosive environment such as offshorefacilities can significantly reduce the effects of a salt ladenatmosphere. All electrical conduit should be PVC coated


using SS fasteners and bulkhead connectors with SS orfiberglass enclosures.

PREVENTING UNNECESSARY PROBLEMS DURINGCOMMISSIONING

Several common, some preventable and some inevitableproblems are encountered during the commissioningphase of the construction project. These problems arecaused from the use of sea water for hydrostatic pipelinetesting, careless deposits of foreign materials and debrisfrom drill bit cuttings, welding slag and sand blastparticles, acids and produced sand during the wellcompletion process and the application of extremephysical force to overcome unexpected resistance. Manyof these problems are preventable and with a littleplanning and control may be completely avoided byfollowing a few simple guidelines:

• allow measurement technicians to commissionnew equipment — this will allow technicians tobecome familiar with equipment before actualoperation begins as well as protectingequipment from destruction by the constructiongorillas

• remove turbine and displacement meters andorifice plates from the line and bypass the proveruntil final commissioning is complete

• clean taps and orifice fitting slot of rust anddebris

• provide for the supply of air free water for proverwaterdraw

• do not operate the GC during the first week tomonth of production operations — use a fixedcomposition in the flow computer and edit thedata as required

NET OIL MEASUREMENT ON HIGH WATER CUTPLATFORMS, EARLY WATER DETECTION FORHYDRATE CONTROL IN DEEPWATER SUBSEA FLOWLINES AND THE APPLICATION OF MULTI-PHASEFLOW METERS (MPFM)

Due to the gradual watering of wells in mature oil fieldsand the eventual use of water flood techniques toenhance production, a well’s water cut (fraction of waterof produced in total liquids) may increase to 90% andabove. This increase in water cut will significantlyincrease the total produced fluid resulting in problems inadequate phase separation and water handlingcapabilities of an offshore platform. When trying toaccurately measure net oil for allocation and reservoirmanagement purposes with real time reporting, meetdesired production expectations at minimal OpEx budgetof management and adhere to increasingly stringenteffluent requirements, the implementation of emergingtechnology measurement devices is essential. Thisequipment ranges from the use multiphase flow meters,coriolis meters for volume and water cut, microwave andnear infrared or NIR principle devices for water cut andfor some applications, the combination of these devices

to complement and work in concert. Extreme diligenceis required when selecting the types of equipment to beemployed to insure the user’s objectives are met. Coriolisand microwave techniques may be used successfully, ifinstalled and applied appropriately. However, both ofthese methods are subject to increased errors in net oilat very high water cuts (i.e., ±10% error in net oil at 90%water cut). NIR devices, relatively new on the market,may be a better fit for very high water cut applications ormonitoring interstage rejection water processing.

The investigation into measurement equipment for earlywater detection for hydrate control in deepwater, subseaflow lines has caused project groups to consider a widerange of equipment and methods. These includedownhole devices using a combination of venturis inseries and annular capacitance techniques, sandmonitoring (acoustical) devices to listen for the sound ofice crystals bouncing along the pipe, system pressuredrop to predict pipeline clogging due to reducedhydraulic area from ice and the use of modified MPFM.Some of these methods might work but none have beenproven in the field.

Multiphase flow meters for well test and allocation arebeing considered for several reasons both onshore andoffshore. The potential for economic benefits from usingMPFM for well testing offshore range from increasedproduction by use of test lines as flow lines, reducedsize and weight as compared to a test separator, reducedwell test time and possibly, improved measurement. Eachapplication must be carefully evaluated consideringrange of types of wells to be tested, gas void fractions,effects of salinity, viscosity, accuracy of data and usuallygovernment or royalty owner approval.

When considering the MPFM for allocation keep in mindthat although this is not sales, it is fiscal measurementand a 10% error could be very costly to your company’sbottom line. However, MPFM may be the best fit forservice method when marginal fields are introduced intoexisting facilities and the only other alternative isadditional processing facilities or isolated phaseseparation for measurement purposes only.

SUMMARY AND CONCLUSIONS

In summary, the fiscal measurement of hydrocarbons onoffshore facilities, although sometimes more expensivethan onshore counter parts, can be very accurate, reliableand cost effective if common sense is employed:

• work around the problems you cannot control• apply the KISS principle (Keep it Simple Stupid)

and apply emerging technology carefully• respect Mother Nature and protect the

equipment• use the pipeline balance to monitor results

Working around the problems you cannot control requiresthat you first recognize the problem such as wet gas,


bubble point crude or liquid carry over from separatorsand then finding tools, equipment, orientation andlocation to prevent failure thus insuring reliable, accuratemeasurement.

The application of the KISS principle could not be moreimportant than when selecting high volume meteringequipment for the fiscal measurement of natural gasoffshore in today’s project management economy.Emerging technology equipment should be carefully,realistically and objectively evaluated before beinginstalled offshore.

Respecting Mother Nature means protecting theequipment by using corrosive resistant materials andadequately protecting equipment from the forces ofnature.

A Gas Pipeline Energy Balance is defined as the %difference between the total re-delivered energy from thepipeline and the total delivered energy into the pipelineas follows:

Energy Balance= ∑Re-Delivered-∑Delivered (MMBTU)∑Re-Delivered

A well designed and operated system with a tight balancemay be used to monitor the performance ofmeasurement equipment and identify problems early.

REFERENCES:

1. ‘Metering Study to Reduce Topsides Weight’ by Nil-RikHannisdal Aker Engineering presented at North SeaWorkshop, October 1991.

2. ‘Effects of Swirl and Velocity Profile Asymmetry onFlow Conditioner Performance for Orifice Meters’ by K.A.Behring and T.B. Morrow presented at 1998FLOWMEKO.

3. ‘Field Experience with Hod Metering ’ by Sidsel E.Corneliussen Amoco Norway Oil Co. presented at theNorth Sea Workshop, October 1991.

4. ‘Effect of Buckling on Orifice Meter Accuracy ’ by P.Jepson and R. Chipchase J. Mechanical Eng. Sc. Vol 17No 6 1975.


OVERALL MEASUREMENT ACCURACY —DETERMINATION AND INFLUENCE

Paul J. La NasaCPL & Associates

P.O. Box 801304, Houston, TX 77280-1304

ABSTRACT

This paper presents methods for determining theuncertainty of both differential and positive meteringstations. It takes into account the type of meter, numberof meters in parallel, type of secondary instruments, andthe determination of physical properties. The paper thenrelates this information to potential influence on systembalance.

INTRODUCTION

Gas measurement uncertainty is a function of thefollowing items:

• Selection of the appropriate metering device• Correct Installation of the metering device• Proper operation and processing of the metering

information• Proper maintenance of the metering device

Understanding how measurement uncertainty applies tometering requires a basic knowledge of the terminologyand assumptions used in the calculation of measurementuncertainty.

Measurement uncertainties can be categorized as thefollowing:

Pseudo Pseudo uncertainties are potential humanerrors or those associated with themalfunction of an instrument. Onceidentified, these errors can usually becorrected and are not included in thecalculation of measurement uncertainty.

Random Random uncertainties are potentialmeasurement errors that have an equalchance of being higher or lower than thetrue value of the measured variable. If alarge number of measurements are made,the random error in the individualmeasurements cancel and the mean of themeasurements will be approximate to thecorrect value.

Systematic Systematic uncertainties are measurementerrors that are directional or contain a bias.Because these errors are directional, theydo not cancel as additional measurementsare made.

Random uncertainty reduces as the number ofmeasurements increases. However, additionalmeasurements will not reduce the systematic uncertainty.

Because the random and systematic uncertainties arecharacteristically different, the calculation of each mustbe performed independently. The combination of the twoindependently performed calculations then forms thetotal measurement uncertainty.

The elements of the random and systematic uncertaintiesare classified as either independent or dependent andmust be determined before the total measurementuncertainty can be obtained. The determination of gasmeasurement uncertainty has been and is addressed innumerous industry articles and standards publications.Three such articles and publications that were referencedin the preparation of this paper are: (1) Norman andJepson, (2) Tiemstra, Rans, and Backus, (3) AGA ReportNo.3 Part 1 — 1990. However, for the purposes ofevaluation, this paper will not concern itself with theinfluence of the interdependence of variables but willutilize the calculation procedure given in A.G.A. ReportNo. 3 (API MPMS 14.3, ANSI 2530, GPA 8185-90) Part I— 1990 to determine the orifice meter measurementuncertainty and will apply the same metrology to thepositive meter (turbine, rotary, or diaphragm meter)measurement uncertainty.

The uncertainty for a single meter run is evaluated fromthe random and systematic uncertainty of the primaryelement (orifice, turbine, rotary, or diaphragm meter) andits instrumentation. The uncertainty of the primaryelement includes the uncertainty associated with the flowcoefficient, expansion factor, diameter of the meter run,diameter of the orifice plate bore, and calibration of thepositive meters.

For an Individual meter run:

UTM = URM + USM

WhereUTM = Total meter run uncertaintyURM = Meter run random uncertainty

URM = √Σ(URi )

USM = Meter run systematic uncertainty —

USM = √Σ(USi )


The percent random uncertainty contributed by eachvariable, URi , is defined as follows:

URi = ( Xi × Ai )2

The sensitivity coefficient of each variable at the point ofevaluation, Xi, can be determined by calculating theresults for the conditions of evaluation, R, and the changein the result, DR, produced independently by theaccuracy variation of each variable at the conditions ofevaluation and substituting into the following equation:

Xi = ∆R( R )The percent accuracy of each variable at the point ofevaluation, e.g., average differential pressure, isdetermined as follows:

For variables whose accuracy is stated as function of itsfull-scale value, the percent accuracy is the value of theaccuracy at full scale, AF, divided by the value of thevariable at the conditions of evaluation, VCE.

Ai = AF × 100 (VCE)

As an example, assume that one of the variables is adifferential pressure value whose accuracy is stated as0.1% of full scale, its full scale is 100 and the point ofevaluation is 50. The percent accuracy of the variable atthe point of evaluation, Ai, would be:

Ai = 0.001 × 100

× 100 = 0.2% ( 50 )The sensitivity coefficient of the variable at the point ofevaluation, Xi, could be determined by calculating theresults, R, using the point of evaluation value of 50, thencalculating the change in results, ∆R, using the point ofevaluation value, 50, changed by 0.2%. These twonumbers would be inserted into the equation for Xi todetermine the sensitivity coefficient of the variable at thepoint of evaluation.

For variables whose accuracy is stated as a function ofreading, the percent accuracy is the same though outits range.

The systematic uncertainty of each variable is determinedusing the accuracy of the calibration standards as an estimateof its contribution to the total systematic uncertainty (USM ).

USi = ( Xi × Ai )2

The accuracy of the calibration standards (Ai) areexpressed as a percent of reading so they can besubstituted directly into the USi equation along with thesensitivity coefficients (Xi ) calculated for the appropriateelement to determine the systematic uncertaintycontribution by each variable.

The total systematic uncertainty is determined using theUSM equation.

USM = √Σ(USi )

However, since the criteria applied for the determinationof the sensitivity coefficient, can vary and is specific toan application, the sensitivity coefficients used for theorifice meter uncertainty were chosen from A.G.A. ReportNo.3 (API MPMS 14.3, ANSI 2530, GPA 8185-90), Part1 — 1990 and similar sensitivity coefficient weredeveloped for the positive meters. The use of theseparticular sensitivity coefficients can result in a smallunderstatement of the uncertainty estimates resultingfrom not accounting for the interdependence of some ofthe elements.

DIFFERENTIAL METER UNCERTAINTY

The variable elements of a gas orifice meter measurementuncertainty calculation are as follows:

Differential Pressure, dpStatic Pressure, PfFlowing Temperature, TfGas Relative Density, GrGas Compressibility Factor, Zf & Zb (Fpv)Orifice Meter Coefficient of Discharge, CdOrifice Bore Diameter, dMeter Tube inside Diameter, DExpansion Factor, YDifferential Pressure Calibrator, dpcStatic Pressure Calibrator, PfcFlowing Temperature Calibrator, TfcGas Relative Density Calibrator, Grc

To calculate the measurement uncertainty for a multiplemeter run station, the variables that are independent ona per run basis are differential pressure, static pressure,temperature, and meter run tolerances. The variablescommon to all runs in the station are the relative density(specific gravity), gas composition, and calibrationstandards.

The total percent measurement uncertainty for a meterstation is as follows:

UTS = URS + USS

Where

UTS = Total orifice meter station uncertaintyURS = Total orifice meter station random uncertaintyUSS = Total orifice meter systematic uncertainty

The total orifice meter station random uncertainty is given as:

URS = Σ URi

2

per Run + Σ(URi )2 per Station √ ( √n )


Where n is the number of meter runs. And the total orificemeter station systematic uncertainty, USS, as:

USS = √Σ(URi )2 per Run

Since there are numerous combinations of equipment,operating conditions, and calculation methods existingfor orifice metering, it is impossible to establish a singlebase line uncertainty relationship. The most practicalapproach is to provide uncertainty ranges for the mosttypical orifice metering combinations.

POSITIVE METER UNCERTAINTY (ULTRASONIC,TURBINE, ROTARY, AND DIAPHRAGM)

The variable elements of a gas positive metermeasurement uncertainty calculation are as follows:

Static Pressure, PfFlowing Temperature, TfGas Relative Density, GrGas Compressibility Factor, Zf & ZbPositive Meter Linearity, PML

Positive Meter Calibrator, PMpc

Static Pressure Calibrator, Pfc

Flowing Temperature Calibrator, Tfc

Gas Relative Density Calibrator, Grc

To calculate the measurement uncertainty for a multiplemeter run station, the variables that are independent ona per run basis are positive meter calibration or proof,static pressure, and temperature. The variables commonto all runs in the station are the relative density (specificgravity), gas composition, and calibration standards.


UTS = URS + USS

Where


The total positive meter station random uncertainty isgiven as:

URS = Σ URi

2



USS = √Σ(USM )2 per Run

Since there are numerous combinations of equipment,operating conditions, and calculation methods existingfor positive metering, it is impossible to establish a single

uncertainty relationship. The most practical approach isto provide uncertainty ranges for the most typical positivemetering combinations.

ENERGY DETERMINATION UNCERTAINTY

The measurement of total energy received or deliveredis customarily the product of the measured volume andthe heating value (Hv) per unit volume. The heating valueper unit volume is typically an inferred measurementresulting from a chromatographic analysis of arepresentative sample of the gas being received ordelivered. In addition to heating value per unit volume,relative density (specific gravity) used in the determinationof volume is also obtained from the chromatographicanalysis. Industry standards, which address theperformance of chromatographic analysis, thecalculation of heating value per unit volume, and relativedensity of a gas sample, are:

• (4)ASTM D 1945-96 (GPA 2261-95) — StandardTest Method Analysis of Natural Gas by GasChromatography

• (5)ASTM D 3588-98 (GPA 2172-96) — StandardPractice for Calculating Heat Value,Compressibility Factor, and Relative Density(Specific Gravity) of Gaseous Fuels

The industry standards, ASTM D 1945-96 (GPA 2261-95) and ASTM D 3588-98 (GPA 2172-96) provide aprecision statement for repeatability and reproducibilityas a function of the mole fraction of each component inthe gas mixture. The repeatability is the expectedprecision within a laboratory using the same equipmentand the same analyst. The reproducibility is the expectedprecision when different laboratories using differentequipment and different analysts use the same method.Tables 1 and 2 provide the given repeatability andreproducibility tolerances.

Component RepeatabilityMole % %

0 to 0.1 0.010.1 to 1.0 0.041.0 to 5.0 0.075.0 to 10 0.08Over 10 0.10

TABLE 1.ASTM D 1945-96 Precision Repeatability Criteria

Component ReproducibilityMole % %

0 to 0.1 0.020.1 to 1.0 0.071.0 to 5.0 0.105.0 to 10 0.12Over 10 0.15

TABLE 2.ASTM D 1945-96 Precision Reproducibility Criteria


The individual component reproducibility tolerances werecombined using the square root of the sum of the squaresmethod as shown in ASTM D 3588-98 (GPA 2172-96) toobtain a precision statement. This is a commonmethodology employed when determining the toleranceof calculated values containing random individualelemental tolerances. Chromatographic analysis and thecalculations of Hv and relative density performed usingindustry standards, ASTM D 1945-96 (GPA 2261-95) andASTM D 3588-98 (GPA 2172-96), will produce heatingvalue results to within ± 0.25% and relative density resultsto within 0.003 relative density units for a typical pipelinenatural gas having the following composition:

Mixture MoleComponent %Methane 96.5222Ethane 1.8186Propane 0.4596Isobutane 0.0977n-Butane 0.1007Isopentane 0.0473n-Pentane 0.0324n-Hexane 0.0664Nitrogen 0.2595Carbon Dioxide 0.5956

BTU/Ft3 1036.06 Ideal Gross Hv per Real Ft3

@14.73 & 60°F0.582 Real Relative Density

@14.73 & 60°F

Since the ASTM D 1945-96 repeatability andreproducibility criteria originated from a statisticalexamination of interlaboratory test results, it includes theinfluences of properly prepared calibration gas standards.The Hv and relative density precision values assume thatthe sampling methods and sampling systems utilizedprovide a representative sample of the flowing gas streamfor analysis.

SYSTEM BALANCE INFLUENCE

Engineering departments can use metering stationuncertainty information in selecting the type of equipmentto be use in a meter station. Equipment can be selectedto meet a system balance expectation or uncertainty. Itcan be used by gas control departments to estimatewhen the uncertainty of a meter station’s measurementis increasing. It can be used to help manage lost andunaccounted-for numbers. If all one type of equipmentis installed on the inlet and all of another type on theoutlet, the metering system may not produce the desiredsystem balance results. It can be used by maintenanceto understand on which pieces of equipment toconcentrate their efforts.

REFERENCES

1. Calculation defines uncertainty of unaccounted-forgas, Norman, R and Jepson, P., Oil & Gas Journal ReportApril 6, 1987

2. Comparison of Orifice and Turbine Meter Accuracy,Tiemsyra, P., Rans, R., and Bacus, H., American GasAssociation Distribution/Transmission Conference April,1991, Nashville, Tennessee

3. Orifice Metering of Natural Gas and Other RelatedHydrocarbon Fluids, Part 1 — 1990, General equationsand uncertainty guidelines. American Gas AssociationReport No. 3, Third Edition, Arlington, VA, October 1990.

4. Standard Test Method Analysis of Natural Gas by GasChromatography — ASTM D 1945-96 (GPA 2261-95)

5. Standard Practice for Calculating Heat Value,Compressibility Factor, and Relative Density (SpecificGravity) of Gaseous Fuels — ASTM D 3588-98 (GPA2172-96)

Paul J. LaNasa


can be judged to be acceptable or unacceptable foraccurate flow measurement. When the meter siteconditions are found to be unacceptable for accuratemeasurement, pulsation control techniques are available,as part of the SGA/GMRC program, to eliminate thepulsation.

PRIMARY ELEMENT ERRORS DUE TO PULSATION

Pressure and flow variations in the flow stream cause achange in the differential pressure across an orifice, whichis a basic error in indicated flow.

Square Root Error

The most fundamental error at an orifice is the squareroot error (SRE). SRE results from averaging thedifferential pressure from a square law device before thetaking the square root. The differential pressure acrossan orifice is shown in the basic orifice relationship,Equation 1, which confirms that ∆P is proportional toflow rate squared. This square law relationship is alsoshown in Figure 1.

∆P = KQ 2 [1]

FIGURE 1. Orifice Relationship thatResults In Square Root Error

Let us consider a flow that is modulating or pulsating,the differential pressure will follow the square law curveso that the average value of ∆P will not occur at the samepoint at which the average value of velocity occurs.

PULSATION EFFECTS ON ORIFICE METERINGCONSIDERING PRIMARY AND SECONDARY ELEMENTS

Robert J. McKeeSouthwest Research Institute

6220 Culebra Road, San Antonio, TX 78238

INTRODUCTION

The use of orifices for commercial flow measurementhas a long history dating back more than 50 years.Orifices are extensively used in the United States naturalgas, petroleum and petro-chemical industries and areimportant as one of the most practical ways to meterlarge volumes of gas flow. These meters are very reliableand cost effective; and if properly used, can be reliedupon to give accurate results.

Proper use normally requires the steady flows for whichorifices were intended and for which the orifice coefficientswere developed. In actual field installations, flow is oftennot steady but subject to the periodic changes in pressureand velocity that are referred to as pulsation. Pulsationcan be caused by compressors, pressure regulators,control valves, fluctuating loads, or by flow-inducedphenomena within the piping. It is known and wellrecognized that pulsation causes errors in orifice meterresults. In fact, A.G.A. Report No. 3 on Orifice Metering ofNatural Gas, which is also API 14.3 and ANSI/API 2530,clearly states that: “Reliable measurements of gas flowwith an orifice cannot be obtained when appreciablepulsation . . . are present at the measurement point.”

Although the potentially adverse effects of pulsation arewell recognized, the nature of resulting error mechanismsand what should be done to correct such problems havenot been well publicized. Discrepancies in orifice flowmeasurement are often reported by operating companiesand in the open literature, but the amplitude of theassociated pulsation and the resultant errors are seldommeasured or even considered. This paper will describethe types of pulsation-induced errors and measurementsthat should be made to identify pulsation effects on orificemetering. Pulsation affects both the orifice itself and thesecondary system causing different errors in each partof the measurement system.

The first step towards eliminating pulsation-induced errorslies in identifying and understanding the error producingmechanisms. Through a continuing research effort,sponsored by the Gas Machinery Research Council(GMRC) of the Southern Gas Association (SGA), progresshas been made in understanding the effects of pulsationon orifice metering. The dominant error mechanisms cannow be identified and quantified. Effects on thesecondary measurement equipment can also beidentified and their relative importance considered. Withproper pulsation measurements, conditions at an orifice


Because of the square law relationship, the average valueof differential pressure is always higher than the ∆P, whichcorresponds to average flow. The difference betweenthe average differential pressure and the differentialpressure that corresponds to average flow is referred toas SRE. This is a real condition or increase in ∆P thatexists at the orifice plate. The SRE is independent of thesecondary system used and is a direct result of averagingthe differential pressure before determining the flow rate.

A mathematical definition of SRE can be developed fromthe fact that the average flow is proportional to theaverage of the instantaneous square root of ∆P asopposed to the square root of the average ∆P. That is tosay, if the square root of the differential pressure is takenat every instant in time, and these square root valuesare averaged, the result will represent the correct flow inthe absence of other errors. The incorrectly indicatedflow is represented by the square root of the averagedifferential pressure. Comparing the incorrect to thecorrect averaging can be expressed in Equation 2, wherethe percentage error in flow is related to the differencebetween the square of average ∆P and the average ofsquare roots of ∆P:

[2]

There are many approximate equations for SRE in theliterature, but most of those that predate 1983 areincomplete. Because of the details that were omitted from

earlier analysis, little progress was made in quantifyingSRE until recent developments under the GMRC’sprogram at SwRI, which developed Equation 2. As aresult of these advancements, a device called the squareroot error indicator (SREI), which uses Equation 2 with amicroprocessor to measure SRE, has been developedand patented. The SREI digitizes the differential pressuresignal and determines the amount of SRE at an orifice.Using a fast response differential pressure transducerand an SREI, any company can determine if they havepulsation problems at their orifices.

To illustrate how SRE functions to measure or representthe error at an orifice, a sequence of pulsating flow testshave been performed. In these tests, two orifices in seriesmeasured the same flow stream. One orifice was isolatedfrom pulsation by upstream and downstream pulsationisolation filters, so that it experienced steady flow at thesame time the second orifice was exposed to pulsationof various frequencies. The difference between the flowinferred at the orifice subjected to pulsation, and at thesteady flow orifice is the total pulsation induced error.The heavy line in Figure 2 shows the amplitude of thistotal error over a range of frequencies used in the testing.The SRE was also measured during these tests and isplotted as the light line in Figure 2. The results of thesetests clearly indicated that SRE accounts for most ofthe total error at an orifice in pulsating flow. Thus, SREcan be measured and used to indicate the presence ofpulsation-induced error at an orifice. It is evident in thegraph of Figure 2 that there are some other smaller errorsat an orifice in pulsating flow.

FIGURE 2. Total Pulsation-Induced Error and SRE at Various Frequencies

* 100 –

∆P ∆P

∆PEf% =


Inertial Error

Inertial error is one of the other primary element errorsthat can be clearly identified. Equation 1 does notdescribe the complete pulsation response of an orificebecause it was developed with the assumption of steadyflow. Developing an orifice equation from time dependentunsteady flow considerations, including the unsteadycontinuity and momentum equations results in Equation 3:

[3]

In this equation, L is a coefficient of the rate of velocitychange, which is dependent on certain dynamicproperties of the orifice. Just as with a solid object, whengas is in motion through an orifice and the differentialpressure changes (decreases), the gas tends to remainin motion due to its inertia. As a result of this inertia, flowvelocity changes lag behind differential pressure changesand the simple relationship between ∆P and flow isdistorted. On the orifice relationship plot as flow changes,the differential pressure is not always on the square lawcurve, which can be seen in Figure 3.

Inertial effects do distort the differential pressure, but areunlike SRE in that inertia does not increase the average∆P. Therefore, if we correctly average ∆P, the inertial erroris eliminated but SRE is present. If the instantaneous ∆P’sare square rooted to eliminate SRE, the inertial errorappears because it involves a change in the time varying

∆P(t) = KV(t)2 + LdV(t)

dt

FIGURE 3. Inertial Effect on OrificeCoefficient for Rapidly Varying Flow

differential pressure. Fortunately, the inertial error is notas large as the SRE. In fact, inertial effects are insignificantunless the pulsations are fairly large in amplitude and at arelatively high frequency. It is, however, partially becauseof inertial errors that SRE measurements should not beused to correct measured flows, but primarily to determineif pulsations are adversely affecting an orifice installation.

Results from a large number of tests are plotted inFigure 4. The plot shows the residual error versus thedifferential pressure modulation. Residual error is definedas actual measured error in flow minus the SRE. If SRE

FIGURE 4. Orifice Response to Pulsations


were the only error in orifice measurement, then theresidual error would be zero.

In Figure 4, the residual error is near zero in the regionwhere the differential pressure modulation is below 40percent. Differential pressure modulation is the zero topeak variation in differential pressure divided by theaverage differential pressure. At larger modulations, above40 percent, Figure 4 clearly demonstrates that residualerror becomes negative as a result of the inertial effect.Figure 4 also shows that for pulsation frequencies below100 Hz, the inertial effect is generally small. Inertial effectsare important if the pulsations are above 40 percent ∆Pmodulation, and the frequency is above approximately100 Hz. When inertia of the flow through an orifice doesbecome significant, the differential pressure for a givenflow is reduced. Thus, a negative error with respect tothe SRE is produced. The most important featuredemonstrated in Figure 4 is that up to a level of 40 percentdifferential pressure modulation, which corresponds toapproximately 1 percent SRE for simple pulsationpatterns, the response of an orifice to pulsation ispredictable. At most metering installations, inertial effectswill not cause significant errors.

Coefficient Shifts

The final error that affects the primary element, as a resultof pulsation, is coefficient shift. In all of the discussionto this point, it has been assumed that the orifice flowcoefficient, K, is constant. However, it is known that undersome conditions a change in the orifice coefficient ispossible. Figure 5 shows the results of a coefficient shift.

FIGURE 5. Effect of a Shift in Orifice Coefficient

Swirl and flow profile distortions are known to causecoefficient variations up to several percent. Furthermore,pulsation are known to cause distortion of average velocityprofiles so there is reason to believe that pulsation wasthe cause of coefficient shifts. Because of other errors, itis difficult to specifically isolate pulsation-induced shiftsin orifice flow coefficients. Throughout most of the orificeflow testing sponsored by SGA/GMRC and conductedat SwRI, with pulsation sources both upstream and

downstream of the orifice, there has been no verifiableor consistent indication of coefficient shifts. However,many of these tests were performed at relatively lowpressures and without the highest accuracy obtainable.This area of identifying possible coefficient shifts is thesubject of continuing research efforts.

There is one exception that is significant for a fewapplications. The testing described above has beenconducted on orifices with flange taps. SRE and inertialeffects are applicable to other pressure tap locationssuch as vena contracta or pipe taps. The orificecoefficient for these other tap locations, and particularlyfor pipe taps, is not constant and is adversely affectedby pulsation. For pipe tap orifices, pulsation definitelycauses coefficient shifts. The dynamic response of a pipetap orifice installation, with respect to changingcoefficients, is such that it is not recommended for usewhere pulsation might be present.

SECONDARY ELEMENT ERRORS DUE TOPULSATION

Secondary recording systems always include pressuresensing lead lines, which will be referred to as gage lines,and chart recorders or electronic transmitters. Pulsationcauses several types of secondary systems errors orincorrect responses, which can be as important or evenmore-significant than the primary element errors.

With flow, especially pulsating flow, through an orificerun, acoustic resonances can be excited in the attachedgage lines. Typically, a quarter-wave pressure pulsationwill develop in a gage line with the main piping or orificeend as a node, and the closed end or pressure transmitterend as a pulsation maximum. Such acoustic resonancesin gage lines can have two very significant effects on therecorded differential pressure.

Gage Line Amplification

The first effect is an amplification of the pulsation. Anexample of this is shown in Figure 6, which illustrates anactual case history and exemplifies the potentiallyadverse effects of high gage line pulsation. In thisinstallation, both a chart recorder and an electronictransmitter are connected to the orifice. If pulsationreadings at the transmitter had been used to determineSRE instead of pulsation at the flange taps, greater errorswould have been predicted than actually existed at theorifice. SRE measured at the orifice was 0.17 percent,while SRE measured in the gage line at the electronictransducer was 11.8 percent. The SRE indicated at thechart end of the gage line was low at approximately 0.3percent. If the flow rate indicated by the chart had beencorrected for SRE measured at the transmitter, then theinferred flow would have been more than 11 percent low.

In these tests, an SREI was used to measure the SRE atthe orifice and also at other points along the gage lines.The SREI has been licensed for commercialization; andthese devices make monitoring the pulsation and


determining the SRE simple and direct. When using anSREI, it is essential to recognize that SRE is afundamental error that occurs at the orifice, and can becorrectly measured only at the orifice taps. SRE for orificeflow condition acceptability should not be measured atthe recorder or at the distant ends of pressure sensinglines.

FIGURE 6. Results of Gage Line Amplification of Pulsations

Gage Line Attenuation

Gage lines can also attenuate, distort or hide the presenceof pulsation as shown in Figure 7. In this situation, whichhas been observed at a number of field installations, thereare pulsations in the orifice run. The average ∆P, as seenacross the orifice taps and on the chart, is high andincorrect.

FIGURE 7. Attenuation of Pulsation in a Gage Line


Because of gage line attenuation, if an attempt were made(using proper high frequency response instrumentation)to determine the differential pressure modulation or theSRE at the chart location, no indication of the error wouldbe present. The chart recording is in error, but there is noway for the measurement engineer, technician or operatorto be aware of the pulsation. Proper instrumentationneeds to be placed across the orifice taps without thenormal lengthy gage lines attached.

Gage Line Shifts

The second type of problem that pulsation in a gage linecan cause is a shift in the indicated pressure. Whilepressure shifts due to gage lines have long beensuspected, they have only recently been measured andverified. Figure 8 is an indication of what has beenobserved when careful measurements are made at points“a” and “b” on a gage line. Not only is the pulsationamplitude different between points “a” and “b,” but theaverage static pressure is different. This distortion of thedifferential pressure is totally a result of the gage line,and not the orifice, but it can substantially affect the flowmeasurement.

FIGURE 8. Typical Pressure Signalsin a Gage Line with Pulsations

As a practical matter, gage line shift can sometimes beobserved when metering valves are pinched. If themanifold valves leading to the recorder are pinched untilthe chart paint reduced, a change in average indicateddifferential pressure could often be observed. Thisprocess merely indicates that gage line errors areprobable, but neither reading is necessarily correct.Instead of pinching valves, excessive pulsation at theorifice should be eliminated.

Research into gage line shifts is relatively new and theprecise causes have not yet been completely quantified.Gage line pressure shifts have been measured underlaboratory and field conditions. Figure 9 shows the resultsof some careful laboratory measurements where a gradientin static pressure along the gage line was verified.

FIGURE 9. Static PressureMeasurements Along a Gage Line

The causes of pressure distortions in gage lines includerectification and kinetic energy effects. These occur mostpredominantly in gage lines that do not have a uniforminside diameter. Rectification results from the fact thatlosses for flow into the gage line are not the same aslosses for flow out of the gage line. Research is still inprogress to define these pressure distortions, whilemethods for predicting the amount of gage line shift arebeing developed. Typically, gage line shifts are one to afew inches of water. In measuring a single static pressure,gage line shift makes little difference; however, indetermining a differential pressure, the error can besignificant. It has been found that if the pulsation in themain pipe and the gage lines are small, gage line shiftsare less likely. If pulsations are kept to a minimum, thechances of avoiding any significant error due to gage lineshift are greatly improved.

Pressure Transmitter and Recorder Response

Secondary system errors can result from the responseof the differential pressure transmitter or chart recorderto pulsation. Most of the differential pressure transducersused in gas metering, including mercury and bellowsrecorders, and electronic ∆P cells and transmitters, havea very poor frequency response. Typically, these deviceswill not track changes in the differential pressure thatare more rapid than two cycles per second. Therefore,they do not follow the actual ∆P across the meter. Theseinstruments cannot be used to determine the amplitudeor even the presence of pulsation. Because such devicestend to average the differential pressure, they contain


some SRE; but because of gage line amplifications orattenuation, it is not necessarily the correct SRE. Thesignal from a pressure transmitter, which has a very lowfrequency response, does not contain the informationnecessary to determine the amount of SRE. It would beof no value to connect a computer that samples andcalculates flow 200 times each second to a transmitterwith a 2 Hz frequency response.

Row computers have an important place in flowmeasurement, but they are not a panacea for pulsation-induced errors. The electronic monitoring and calculationof flow can also cause errors or inconsistency in results.An example of this, in Figure 10, illustrates a situationwhere a differential pressure transmitter with an adequatefrequency response was being used on a process gasorifice meter. The plant’s computer was used to computethe gas flow rate, but as indicated, the sample rate wasevery 6 seconds. Because of a dump valve cycle in theplant process, there was a 3 second transient in the flowevery 12 to 13 seconds. The computer did not calculatethe correct flow for several reasons, including the factthat the apparent average differential pressure was notthe true time average, and the amplitude of modulationthat might be associated with SRE was also not a correctindication. No matter what flow computer is used, if itssample rate is not adequate for the pulsation in differentialpressure, it will not yield reliable results.

FIGURE 10. Computer Sampling inthe Presence of Flow Transients

Chart Paint

Chart recorders are commonly used to monitor thedifferential pressure at an orifice. It is frequently thoughtthat chart paint is a result of pulsation at the orifice.However, chart paint is almost never a direct reproductionof pulsation by the recorder. Pulsations in gas pipingsystems are typically found above 5 Hz and below 100Hz, with the 15 to 45 Hz range being very common. Thefrequency response of a chart recorder is usually lessthan 2 Hz. Chart recorders are not able to accuratelyrespond to most pulsation.

At best, chart paint is a distorted, damped, and incompleterepresentation of pulsations which are greatly suppressedin amplitude. In these cases, use of a fast clock will notproduce an accurate ∆P record because the pen of the

chart recorder is unable to follow the pulsation waveform.Typically, charts exposed to pulsation will not respondor paint at all and will appear smooth. If, on a 100 inchchart, the average flow corresponds to 72 inches andpulsation in the pipe are such that a 10 percent SRE ispresent, then the chart should show pulsation from justover 100 to below 45 inches. In other words, the upperhalf of the chart should be solid paint, but this is notwhat is actually observed.

Chart paint is often caused by phenomena other thanorifice run pulsation. Sometimes chart paint is causedby actual periodic changes in the average flow rate suchas engine governor hunting, regulator valve adjustments,or small adjustments in flow balance between parallelorifice runs. Even with steady flow and no pulsation inthe orifice run, if there are pulsations that develop in thegage lines, a change in ∆P on the chart can result whichwill return to normal, causing paint, when the gage linepulsation decay. Figure 11 is a chart record with paintfrom an installation where no pulsations were detectablein the main line piping. It should be clear that the presenceof chart paint does not mean that pulsations are present.

The last example in Figure 12 was taken from aninstallation at which a pulsation filter vessel could bevalved in or out of service. With flow passing throughthe filter and no pulsation presence, the chart showssome paint that due to flow variation from the variablespeed compressor. With no change in average flow orlocal conditions, the filter bottle was bypassed such thatpulsation developed at the orifice. Although the natureof the paint change was little, the important change wasthe increase in ∆P and the corresponding error in flowthat resulted from the pulsation.

SUMMARY AND CONCLUSIONS

In view of the different types of pulsation-induced errorsthat can occur at the primary element and in secondarysystems, it is useful to review what should be done tominimize pulsation effects on orifice measurement.Because the less definable influences of pulsation onorifice measurements, such as the inertial effect, averageto zero over time, and because most of the available ∆Ptransmitter or chart recorders have a slow response, thefirst step is to obtain the average differential pressure.This average should be taken often enough to accountfor normal variation in the flow rate. The second importantstep is to then place a fast response pressure transduceracross the orifice taps and using an SREI, or at least anoscilloscope, measure the SRE, or determine if there areany significant differential pressure modulation at theorifice. If the measured SRE is small, such as less than0.2 percent, or if there are no noticeable pulsation, theorifice installation will have a high probability of yieldingacceptable readings. If the measured SRE is at or over1.0 percent, or a very significant modulation in differentialpressure exists, then the pulsation are large enough tocause other errors and will probably cause gage lineproblems. Such pulsation should be corrected or


eliminated through the use of acoustic filtering andpulsation control technology. If the measured SRE is inthe middle range between 0.25 and 1.0 percent, thenthere are some additional checks that should be madeand decisions that must be addressed by the operatingcompany. For example, at a measured SRE of 0.5percent, there will be approximately 1/2 percent positivebias in the flow and the possibility of other errors which,

FIGURE 11. Chart Record from an Orifice without Pulsations

FIGURE 12. Orifice Chart Record without and with Pulsations

depending on the service, may or may not be acceptable.The third step, in any event, should be to use the fastresponse pressure transducer at the chart recorder ortransmitter locations. This will not indicate the correctSRE value, but it will indicate if the pulsation have beenamplified or significantly distorted along the gage lines.If not all of these dynamic checks of the orifice installationindicate an error, the installation will most likely be


acceptable. If, however, there are pulsation errorspresent, they should be corrected.

In conclusion, the following statements are derived fromthe work accomplished to date, and can hopefully guidethe application of this information and futureinvestigations:

• Pulsations cause errors at the orifice primary elements,including the most significant square root error, andother errors, such as inertial effects and coefficient shift.

• The amount of SRE can be measured with aninstrument or the proper techniques and transducers,and can be used as an indicator of the severity ofpulsation at an orifice.

• Pressure sensing gage lines can amplify, attenuateand distort pulsation such that recorders or

transmitters are not exposed to the actual pulsationenvironment, resulting in incorrect indications of flow.

• When pulsation are present in the gage line, shifts inaverage pressure can cause direct errors in measuredflow.

• Chart recorders and the normal metering transmitterscannot follow the frequency of pulsation and cannotbe used to detect, measure, identify or compensatefor pulsation.

• Where pulsations are eliminated or properly controlled,orifices can provide highly reliable and accurate gasflow measurements.

Robert J. McKee


PULSATION REDUCTION BY ACOUSTIC FILTERSFOR METERING APPLICATIONS

Robert J. McKeeSouthwest Research Institute

6220 Culebra Road, San Antonio, TX 78238

INTRODUCTION

Because of the adverse effects of pulsations on orificeand other types of flow meters there is for manyinstallations, a need to eliminate or decrease theamplitude of pulsations in the piping. This task has beenthe primary domain of acoustical piping designers whohave had both theoretical and practical field experiencein such areas. The most common and effective treatmentfor pulsation control is the design and installation ofacoustic filters. However, most filters designed by novicesare not effective and are costly to operate because ofpressure drop losses. This paper discusses the basicprinciples and considerations in acoustic filter design.

There are many small compressors such as well-headgathering compressors that cannot justify the cost of athorough acoustic analysis in order to protect the nearbyorifice meter from excessive pulsations andaccompanying square root error. This paper will make aneffort to demonstrate design procedures related to aspecific type of acoustic filter to be used to reducepulsations in most simple metering applications. Thespecific filter is a symmetrical in-line low pass filter. Theimportant elements of this filter can be summarized inthe following points:

1. The inlet line is located at the acoustic center of thefirst chamber of the filter.

2. The first chamber of the filter, the choke tube, andthe last chamber are all the same acoustical length.

3. The choke tube connects the acoustical centers ofeach filter chamber.

4. The outlet line is located at the acoustical quarterpoint of the final chamber.

It is fully realized that this is a specialized filter designand other types of designs could be recommended butthis particular filter design was chosen because of itsconservative and foolproof aspects. Drawings of thephysical aspects of the acoustic filter piping are shown inthe Appendix.

WHERE DO PULSATIONS COME FROM?

Pulsations in metering applications can be generated byreciprocating compressors, centrifugal compressors, flowinduced phenomena, or turbulent sources. The primaryfocus of this paper is directed toward filtering the mostpersistent pulsations that are produced by reciprocatingmachinery.

A reciprocating compressor produces pulsation atcompressor crankshaft RPM and its multiples. Pulsationfrequencies are generally expressed in cycles per second(Hertz). A 300 RPM compressor produces pulsation at 5hertz, 10 hertz, 15 hertz, and higher multiples. Mostcompressors (for natural gas services) are double-actingand compress gas on the head and crank end of thecylinder. Double-acting cylinders tend to produce morepulsation at the even multiples of RPM and less at theodd multiples. Therefore, a 300 RPM double-actingcylinder will produce its strongest pulsation at 10 hertz.When compressors have more than one cylinder, thecrankshaft phasing of the cylinders will also cause certainmultiples to be higher than others. For example, if twodouble-acting cylinders are phased 90’ apart they willproduce a significantly higher pulsation level at four timesRPM. In the case of a 300 RPM machine and two double-acting cylinders phased 90 degrees apart there will be ahigh fourth harmonic or 20 hertz. Generally, the highermultiples (sometimes called compressor harmonics orcompressor orders) will contain less energy than thelower orders.

WHAT ARE THE IMPORTANT CHARACTERISTICS OFLOW PASS FILTERS?

The single most important factor of a low pass acousticfilter is its natural frequency. The volume-choke-volumeconfiguration exhibits a relatively low natural frequencyin comparison with the acoustic halfwave naturalfrequencies of the chamber and choke tube lengths. Atfrequencies below the filter natural frequency, there willbe no attenuation of pulsations passing through the filter.There will be a sharp attenuation starting at about 20percent above the natural and extending out to severalhundred hertz. It is very important to understand thatpulsations at the natural frequency of the filter will actuallybe magnified by as much a 10 to 40 times. Therefore, itis very important that the natural frequency of the filtersystem not be frequency-coincident with pulsations. Thiscan be accomplished in two ways. The filter naturalfrequency can be placed 20 percent below the RPM ofthe compressor or it can be placed halfway between thefirst order of the maximum RPM and second order ofthe minimum RPM.

HOW TO CALCULATE THE BEST FREQUENCY FORTHE FILTER NATURAL FREQUENCY

This procedure will assist in a correct design for placingthe filters natural frequency below the fundamental RPMor between the first and second RPM orders of the


compressor. The following two cases illustrate thecalculations and limitations:

Case 1: Placing the filter natural frequency F0 below thefundamental order.

F0 = RPMmin

(60)(1.2)

Where:F0 = filter natural frequency of filter system

RPMmin = the minimum actual operating RPM of thecompressor

Case 2: Placing the filter natural frequency between thefirst and second orders.

F0 = RPMmin * 2 + RPMmax

(60) (2)

This frequency F0 must be at least 20 percent separatedfrom RPMmax and RPMmin * 2 . This 20 percent detuning willensure that the filter natural frequency is not excited bythe compressor.

CALCULATING THE INSIDE DIAMETER OF THECHOKE TUBE

The inside diameter (ID) of the choke tube should be assmall as pressure drop limits will allow because largerchoke tubes require large volumes to create effectivefilters. A standard procedure is to limit the gas flow velocityto 100 feet per second as implemented in the followingequations.

q = 0.327 MMSCFD(T+460)Z ft3

P sec

IDC = 1.354 q(in.)

P

Where:MMSCFD = million standard feet3 per day flow

T = gas temperature (°F)P = gas pressure (psia)Z = real gas correction factorq = flow rate (feet3 per second)

IDC = inside diameter of choke tube (inch)

( )

60 60

CALCULATING THE FILTER CHAMBER INSIDEDIAMETER

To ensure sufficient attenuation of pulsation above thefilter natural frequency, the bottle or filter chamber IDBshould be at least 2.5 times the choke tube diameter IDC.

IDB = 2.5* IDC

CALCULATING THE FILTERELEMENT ACOUSTIC LENGTH

To complete the acoustic filter design, the acoustic lengthof each bottle and the acoustic length of the choke tubeare all equal. This length is calculated based on thefollowing equation.

Lf = c IDC = 0.225 c IDC (ft) √2 F 0 IDB F0 IDB

Where:c = gas velocity of sound (feet per second)

An equation and method for calculating gas velocity ofsound is shown in the Appendix.

Because of acoustic end effects, the physical length chokeof the tube should be slightly shorter than the calculatedacoustic length. The acoustic length must be reduced atotal of 1.2 times the ID of the choke so that the physicallength has the proper acoustic effect.

EXAMPLE CALCULATIONS (CASE 1)

General Information

A set of reciprocating compressors operating over anRPM range of 300 to 400 RPM is causing pulsation at ametering site approximately 500 feet from thecompressors. The compressors single-act on occasion.The maximum flow rate through the meter site is approxi-mately 31.5 MMSCFD. The discharge temperature is 84°Fand the pressure is 925 psia. The ratio of specific heatsis 1.32. The gas specific gravity is 0.62 and thecompressibility is estimated to be 0.947. The velocity ofsound is calculated to be 1371.9 feet per second.

Solution

The natural frequency of the filter system should beplaced below the fundamental compressor frequency,since single-acting is possible. The highest possibleplacement of the filter natural frequency is 4.17 Hertz.

√


The choke tube volume flow rate is 5.735 feet3 per second.The choke inside diameter of 3.548 inches is elected.The bottle inside diameter is calculated to be at least8.87 inches but this would produce a very longcumbersome installation; therefore, a 22.626-inch insidediameter is selected instead. This will produce a bottlechamber acoustic length of 11.62 feet (single chamber).To account for the acoustic end effects, 1.2 times thechoke diameter is subtracted from the characteristic filterlength to give a choke tube length of 11.26 feet.

EXAMPLE CALCULATION (CASE 2)

General Information

A reciprocating compressor operating over an RPM rangeof 300 to 360 RPM is causing pulsation at a meteringsite approximately 200 feet from the compressor. Thecompressor has no mechanism to single-act. The maxi-mum flow rate through the meter site is approximately10.5 MMSCFD. The discharge temperature is 120°F andthe pressure is 925 psia. The gas specific gravity is 0.62and the compressibility is estimated to be 0.947. The ratioof specific heats is 1.32. The velocity of sound is calculatedto be 1416.5 feet per second.

Solution

The natural frequency of the filter system can be placedbetween the fundamental and second compressorfrequency since single-acting is not possible. The bestplacement of the filter natural frequency is 8.0 Hertz. Thisis 25 percent below the minimum second order and 33percent above the fundamental compressor order. Thechoke tube flow rate is 2.038 feet3 per second and thechoke tube is calculated to be 1.933-inch inside diameterso a 1.939-inch pipe is selected. The calculated bottlediameter is 4.848, which would again produce anawkward layout, so a 13.124-inch ID is selected. Thefilter characteristic length of 5.888 feet is calculated andthe physical length of the choke tube is calculated to be5.694 feet.

MECHANICAL CONSIDERATIONS

If a single-bottle design is selected the internal choketube should be restrained at both ends to ensure thatthe cantilever lengths on each side of the baffle are notresonant mechanically. The baffle should be dished headwith thickness approximately the same thickness as thebottle wall. If a two bottle design is selected, the branchconnections should be saddles or pads. Weld-o-letsshould be avoided because of the high stress riser theyinflect on the bottle wall. Elbows in the choke tube areacoustically irrelevant as long as the center line length isused for the length. Elbows do serve as acoustical tomechanical coupling points and should be avoided ifpossible. The filter bottle should be securely restrainedto ensure mechanical stability and an external chokeshould be restrained to ensure that the mechanical natural

frequency is not coincident with the half wave resonantfrequency of the acoustic length.

OTHER PULSATION REDUCTION APPROACHES

In some cases, careful design of meter installations canbe effective in reducing or eliminating the chances ofpulsations. Meter installations should be designed so thatthe piping is not resonant and so that the orifices or othermeter is located at a velocity minimum. This involves themeter tube length being unrelated to acoustic wave lengthand being symmetrically about the orifice or other meter.

Restrictions or additional orifice plates installed at variouslocations in a pipe can dissipate some of the energy inpulsating waves. To be effective these pulsationrestrictions are sized to create pressure drops and therebyrequire an increase in compression horsepower.Pulsation plate performance is dependent on adjacentpipe configuration and frequency content of thepulsations. In some cases, pulsation plates cause higherfrequency resonances and increased amplitudes ofpulsations.

Surge volumes or single-bottle filters such as scrubbersor separators generally do reduce pulsation levels,however, they are not always effective and their specificfrequency responses vary with the installation.

CONCLUSIONS

It is theoretically possible to reduce pulsations at meteringsites with acoustic filters so that little or no pulsation willinfluence flow measurement. The two examples illustratethat the bottle-to-choke diameter ratio of 2.5 is usuallyincreased to ensure practical physical construction onthe system. Reduction of choke tuber pressure dropcould be reduced even further by the use of a bell mouthentrance fitting. Bell mouths are common when a single-bottle design is selected. It is advantageous to have anexperienced acoustical engineer evaluate your filterdesign practically if the installation or the meteringaccuracy is critical. If the filter natural frequency is placedbelow the compressor fundamental, it is possible to beconservative on the low side. However, if the naturalfrequency is to be placed between the first and second,it is important to be accurate in both calculations andphysical construction. Other methods for reducingpulsation are not as effective and foolproof as acousticfilters.

APPENDIX

Speed of Sound Calculation

The velocity of sound can be approximated by thefollowing expression:

c = 1.354 √ kTZ

G


where:c = velocity of sound (feet per second)k = ratio of specific heatsT = gas temperature (degree R)G = gas specific gravityZ = real gas correction factor at flowing conditions

Robert J. McKee

FIGURE 2.

FIGURE 1.


LESSONS LEARNED FROM THEAPI CHAPTER 14.1 GAS SAMPLING PROJECT:

AN OVERVIEW OF COMMON CAUSES OF GAS SAMPLE DISTORTIONAND INFORMATION NEEDED FOR PROPER GAS SAMPLING

Eric Kelner and Darin L. GeorgeSouthwest Research Institute

6220 Culbera Road, San Antonio, Texas

INTRODUCTION

Over the past seven years, the Gas Technology Institute(GTI), the American Petroleum Institute (API) and theUnited States Minerals Management Service (MMS),have co-sponsored an extensive natural gas samplingmethods research program at the GTI Metering ResearchFacility (MRF), located at Southwest Research Institute(SwRI). The results of this research provided a basis forthe revision of Chapter 14.1 (i.e., Collecting and Handlingof Natural Gas Samples for Custody Transfer) of the APIManual of Petroleum Measurement Standards (MPMS).The revision is complete and was published in 2001.

The API Chapter 14.1 Working Group, a research steeringcommittee consisting of natural gas sampling expertsfrom major oil and gas companies, provided input thathelped focus the project on improving current fieldpractices. The research identified several causes ofnatural gas sample distortion, as well as techniques foravoiding gas sample distortion. The research dataillustrated how errors in calculated gas properties, suchas heating value and density, can occur as a result ofpoor sampling technique.

THE 2001 REVISION OF API CHAPTER 14.1

The 2001 revision of Chapter 14.1 of the API MPMSprovides guidance for obtaining representative samplesof natural gas through spot, composite, and continuoussampling methods. It focuses on the practical applicationof thermodynamic principles that, if ignored, can causea gas sample to become distorted, resulting in a biasedgas analysis. If a biased analysis is used to calculate theheating value or other properties of the sampled gas,errors in excess of 10 percent may occur.

The revised Chapter 14.1 builds on the knowledgecontained in the previous version by identifying severalspecific causes of natural gas sample distortion. Chapter14.1 discusses gas sample distortion in detail andprovides recommendations for avoiding it. The standard

is suitable as an instructional tool and as a guide tosampling system design and sampling techniques. Thestandard is written primarily for field personnel. It providesthe technical background necessary to understand andapply basic hydrocarbon mixture thermodynamics tonatural gas sampling in order to avoid gas sampledistortion.

This paper draws on the information in the revisedChapter 14.1 and presents an overview of three commoncauses of gas sample distortion: (1) sample distortiondue to equipment and processes that cause the samplegas temperature to drop below the hydrocarbon dewpoint temperature, (2) sample distortion caused by dirtyor contaminated sampling systems and (3) sampledistortion caused by sampling system componentsfabricated from materials known to affect the integrity ofa natural gas sample. Recommendations for avoidinggas sample distortion according to the revised Chapter14.1 are presented below. This paper begins with adiscussion of the importance of the hydrocarbon dewpoint in gas sampling methodology and sampling systemdesign, followed by an introduction to the natural gasphase diagram.

HYDROCARBON DEW POINT

The hydrocarbon dew point is defined as the pressureand temperature at which hydrocarbon constituents ina natural gas mixture begin to change phase. Forinstance, if the temperature of a natural gas mixture isreduced while the pressure remains constant,1 thetemperature at which hydrocarbon condensation beginsto occur is the hydrocarbon dew point temperature. Ifthe pressure of a natural gas is increased while thetemperature remains constant,2 the pressure at whichhydrocarbon condensation begins is the hydrocarbondew point pressure.

The hydrocarbon dew point of a natural gas differs fromthe water dew point in that the latter describes thepressure and temperature at which water vaporcontained in the gas mixture begins to condense. Somegas mixtures will reach the water dew point temperaturebefore reaching the hydrocarbon dew point temperatureduring an isobaric temperature reduction. This paperfocuses on the hydrocarbon dew point because of itsinfluence on heating value. This distinction should be

1This process is known as an isobaric (or constant pressure)temperature reduction. It is the process that occurs when using a“chilled mirror device” to determine dew point in the field.2This process is also known as an isothermal (or constant temperature)pressure increase and is similar to processes used for determiningthe dew point in a laboratory.


kept in mind during any discussion of natural gasthermodynamics.

Retrograde condensation is a phenomenon that occursin many common natural gas mixtures. It is characterizedby the presence of two hydrocarbon dew points at a givenpressure or temperature. Retrograde condensation canoccur during isobaric temperature increases, or duringisothermal pressure reductions. Retrograde behavior ischaracteristic of natural gas and should be consideredwhen sampling a natural gas stream and when designinggas sampling systems.

THE PHASE DIAGRAM

A phase diagram or phase envelope describes the phasechange behavior of a natural gas mixture. It can be usedto illustrate the effect of natural gas sampling processeson natural gas.

Figure 1 shows a typical phase diagram for natural gas.The line A-B is the bubble point curve. The bubble pointis reached when an infinitesimal amount of gas appearsduring an isothermal pressure reduction of a liquidhydrocarbon mixture.

The line B-E is the dew point curve. It represents therange of pressures and temperatures at which gas/liquidphase changes occur with a natural gas mixture.

The points along line B-D represent the pressures andtemperatures at which retrograde condensation occurs.The points on this line represent retrograde phasebehavior. Retrograde condensation can occur duringcommon natural gas sampling processes.

The line D-E is the lower or normal dew point curve.Condensation associated with the conditions defined bythis curve may occur during a pressure increase, suchas when compressing a gas sample from a vacuumgathering system into a sample cylinder.

HYDROCARBON DEW POINT AND NATURAL GASSAMPLING

The hydrocarbon dew point is perhaps the single mostimportant property in natural gas sampling. If the sampletemperature is allowed to drop below the hydrocarbondew point temperature, a significant loss in hydrocarboncontent can occur, resulting in errors in volumetric flowrate, heating value and other gas property calculations.

Tests conducted at SwRI on spot sampling methodsshowed that the impact of dropping below thehydrocarbon dew point temperature contributes toincreased random and bias error in the calculated heatingvalue and density. The phase diagram shown on Figure2 illustrates how different processes common in naturalgas sampling can cause the temperature of the sampledgas to fall below the hydrocarbon dew point.

Path 1-2 represents the process that occurs when naturalgas flows through a regulator or partially closed valve.The cooling associated with the pressure reduction isknown as the Joule-Thompson effect. Condensation andsample distortion can occur during this “throttling”process.

The cooling can be offset through the application of heatto the sampling system. Path 1-3 shows the potentialimpact of adding sufficient heat to the system to offsetthe cooling effect.

Path 4-5 shows how condensation of a sample can occurif the sample container is exposed to an ambienttemperature below the hydrocarbon dew pointtemperature.

THE EFFECT OF A PHASE CHANGE ON HEATINGVALUE

The potential impact of condensation on the heatingvalue of natural gas can be illustrated graphically. Figure3 shows the potential effect of 41OF gas samplingequipment on a 1,500 BTU natural gas with 0.85 mole

FIGURE 1.A typical natural gas phase (P-T) diagram.

FIGURE 2. A typical natural gas phase (P-T) diagramshowing several processes common in natural gas

sampling. These processes can cause condensation — afundamental cause of gas sample distortion.

Temperature (F)-200-180-160-140-120-100-80 -60 -40 -20 0 2 0 4 0 6 0 8 0 1 00120 14 0

Pres

sure

(psi

a)

010 020 030 040 050 060 070 080 090 0

100 0110 0120 0130 0140 0150 0160 0170 0180 0190 0200 0

GGGGaaaassss

2222 ---- PPPPhhhhaaaasssseeeeRRRReeeeggggiiiioooonnnn

CompositionN2 = 2.05533CO2 = 0.5132C1 = 82.6882C2 = 6.9665C3 = 4.5441i-C4 = 1.1559n-C4 = 1.2856i-C5 = 0.3983 nC5 = 0.2624C6 = 0.0836C7 = 0.0273C8 = 0.0167C9 = 0.0009

LLLLiiiiqqqquuuuiiiidddd

CCCCrrrriiiitttt iiiiccccaaaallllRRRReeeeggggiiiioooonnnn

1111

2222

Path 1 -2: Retr ograde conden sation during throttling to a lower pressure.

No, or not enough he at tracing on sample line.Path 1-3: Heat tracing can offset the coolingwhich occurs during the throttling process,

thereby avoiding condensation.

4444

5555 Condensat ionth at occurs when a sampleor calibration standard is

exposed to ambient temperaturesbelow th e hydrocarbon dew point

temperature.

3333


percent n-hexane through n-decane (labeled as C6+)components and a hydrocarbon dew point of 91OF. TableI shows the gas composition associated with the figure.

The vertical axis on the left shows the vapor (gas) fraction,on a molar basis. The liquid fraction is simply 1 minusthe vapor fraction. The vertical axis on the right showsthe change in vapor fraction BTU.

As the temperature is reduced below the hydrocarbondew point temperature, hydrocarbon constituentscondense in order of decreasing molecular weight. Thiscondensation causes the vapor fraction of the mixtureto decrease and a corresponding decrease in the heatingvalue of the vapor fraction.

As the gas temperature is reduced below thehydrocarbon dew point, there is a large decrease inheating value associated with a small amount of liquidcondensation. At a temperature of 41 OF, the loss inheating value amounts to over 70 BTU/SCF.

The mass and weight of liquid produced by condensationof the natural gas in Figure 3 can be estimated. Assumethat a representative sample of natural gas is containedin a standard 300 cc constant volume sample cylinder.When the temperature is reduced to 41OF, condensationoccurs and the vapor fraction decreases byapproximately 2.9 mole percent. The percent decreasein the mass of the vapor fraction for this particular gasmixture is 7.5 percent. The total mass of gas (vapor) at75 psia and 91OF is approximately 0.0045 lbm. If 7.5percent of the mass condenses, then the liquid containedin the cylinder will weigh approximately 0.00034 lbf. Thisis only a fraction of the weight of a dime. This smallamount of liquid can account for significant losses ingas sample heating value, yet it is difficult to detectwithout sensitive laboratory instruments.

The process described is similar to the process shownby Path 4-5 in Figure 2 and illustrates how a phasechange can cause a decrease in heating value. Now

consider a very small amount of condensedhydrocarbons (weighing less than a dime) located withina component of a sampling system, such as a samplevalve. This could occur during the pressure reductionprocess described by Path 1-2 in Figure 2. If the smallamount of condensed hydrocarbons is swept into thesample cylinder during sampling, a significant increasein heating value can result.

This discussion illustrates the magnitude of the impactof phase changes on natural gas samples. In practice,the effect of a distorted gas sample on calculated gasproperties is very difficult to predict. The effects of poorsampling technique on gas samples taken under actuallaboratory and field conditions are far more complicatedand cannot be accurately predicted using currenttechnology.

THE NEED TO CORRECTLY CHARACTERIZE THEGAS WHEN COMPUTING HYDROCARBON DEWPOINTS

Clearly, it is crucial for a natural gas stream to be aboveits hydrocarbon dew point during sampling to avoidsample distortion. If enough is known about the gascomposition before the sample is taken, the dew pointcan be estimated using any of a number of commercialsoftware packages. Small amounts of heavyhydrocarbons, n-hexane and heavier, can strongly affectthe dew point of a gas. Unfortunately, many process gaschromatographs (GCs) cannot identify these heaviercomponents separately, and only report a “lumped C6+fraction” in their results. Using this lumped fractionincorrectly can lead to a significant underestimate of thedew point, and to sample distortion.

Ongoing research at SwRI has sought a useful, accuratemethod for predicting hydrocarbon dew point

FIGURE 3. The vapor fraction and change in vaporfraction BTU associated with condensation of a 1,500

BTU/SCF natural gas mixture.

MRF 1,500 BTU MixComponent Mole Percent

Methane 64.107Ethane 10.33

Propane 7.128Iso-butane 2.174

Normal-butane 6.386Iso-pentane 1.874

Normal-pentane 2.307Normal-hexane 0.538Normal-heptane 0.187Normal-octane 0.086Normal-nonane 0.023Normal-decane 0.016

Nitrogen 3.939Carbon Dioxide 0.906

Total 100.001

TABLE 1. The 1,500 BTU/SCF natural gas mixture used tocalculate the values in Figure 3.

Vapor Fraction and Change in Vapor Fraction BTUfor MRF 1500 BTU Mix with 0.834% C6 + at 75 PSIA

Temperature [F]30 35 40 45 50 55 60 65 70 75 80 85 90 95 10 0

Vap

or F

ract

ion

0 .95

0 .96

0 .97

0 .98

0 .99

1 .00

1 .01

Cha

nge

in V

apor

Fra

ctio

n [B

TU/s

cf]

-125-120-115-110-105-100-95-90-85-80-75-70-65-60-55-50-45-40-35-30-25-20-15-10-505

Vapor/Feed RatioChange in Vapor Fraction BTU

Dew Point91OF

~ 97.1% vaporby mole

~ 71BTU/SC Floss due to

condensation


temperatures using only process GC data through C6.Phase diagrams have been computed using severalsoftware programs, different equations of state(equations that predict the behavior of a gas mixture withtemperature and pressure changes), and different“characterizations” of the heavy hydrocarbons in the gasmixture. These characterizations make assumptionsabout the relative amounts of hexanes and heaviercomponents in the lumped C6+ fraction. Not surprisingly,the characterization method had the largest influenceon the accuracy of the predicted phase diagram.

Figure 4 shows an example of several phase diagramsfor the same gas, all computed by assuming differentcompositions for the hexanes and heavier hydrocarbons.The results were compared to the phase diagramcomputed from the certified gas composition, todetermine the potential errors due to poorcharacterizations. The worst case was obtained bytreating the lumped C6+ fraction as 100% normal hexane.Using this characterization gave computed dew pointsas much as 35°F below the curve computed from thecertified gas composition. The best characterization inthis example used the actual composition of the gasthrough nonane (C9), and treated the heavier componentsas a “lumped C9+ fraction.”

In general, the SwRI research indicates that a naturalgas composition must be known through nonane for itsdew point to be computed with reasonable accuracy.This information can be obtained by analyzing the streamusing a GC capable of detecting hexanes and heavierhydrocarbons separately. This may require a sample tobe carefully obtained and sent offsite for analysis.Treating a lumped C6+ fraction as pure hexane willconsistently cause the dew point to be underpredicted,and can lead to sampling errors.

GAS SAMPLE DISTORTION DUE TO THE SAMPLEGAS TEMPERATURE DROPPING BELOW THEHYDROCARBON DEW POINT TEMPERATURE –LABORATORY AND FIELD TEST RESULTS

The Gas Processors Association (GPA) spot samplingmethods and three composite samplers were tested with

several gravimetrically prepared natural gas mixtures instatic (non-flowing) conditions. The deviations incalculated heating value and density for the gas samplesobtained during the tests are shown in Figure 5.

FIGURE 4. Potential errors in phase diagrams computedusing incorrect characterizations of heavy hydrocarbons.

FIGURE 5. The deviation in calculated heating value anddensity using analyses obtained from samples taken

with equipment at temperature below the hydrocarbondew point temperature.

When the temperature of the sampling hardware wasbelow the hydrocarbon dew point, all methods produceddistorted gas samples. Some sampling methodsproduced samples that were enriched, causing anincrease in the sample heating value and density. Othermethods produced samples that were depleted, causinga decrease in the sample heating value and density. Allof the methods produced highly variable results,suggesting that a phase change occurred duringsampling.

The results from the composite sampler tests in Figure 5(Methods 10, 11 and 12) led to subsequent field tests tobetter understand the operational limits of compositesamplers. In-situ tests were conducted at a gas pipelinesite in northwest Colorado. Four composite samplerswere installed on a pipeline that was flowing 1,100 BTU/SCF natural gas, and exposed to ambient conditions 60-70OF below the hydrocarbon dew point. The compositesamples were analyzed and the calculated heating valueand density of each sample were compared to theaverage values acquired from on-line gas chromatograph(GC) analyses. After approximately six months of testing,heated enclosures were installed on the compositesampling systems to stabilize the ambient air temperaturearound the samplers at approximately 100OF.

Test results from one composite sampling system areshown in Figure 6. Results after the heated enclosureswere installed showed a significant improvement, witheach system providing samples with heating values anddensities that agreed with the reference to within +/-0.75%, nominally.

The 2001 revision of Chapter 14.1 of the API MPMSstates that the sample gas temperature must remain 20-50OF above the hydrocarbon dew point temperature atall times during sampling. If the sampling processinvolves a pressure reduction, additional heat must be

Pre

ssur

e (p

sia)

1 2 3 4 5 6 7 8 9 10 11 12

Diff

eren

ce [%

]

-10 .0 0

-5.0 0

0.00

5.00

10 .0 01,500 Btu/scf MRF Mixes (4), 75 psia, 110°F

OOOOnnnn----LLLLiiiinnnneeee1: Probe Reg.SSSSppppooottt2: Cont r. Rate3: Fill/Empty4. Evac. Cont.5: Reduced P6: Helium Pop7: Water Displ .8: Glyc. Displ.9: Pis ton Displ .CCCCoooommmmppppoooossssiiiitttteeee10: P-Throttle11: Pumped12: Vanish Chmb.

r [lbm/ft3] at 60° F, 14.73 psiaHv [Btu/real cf] at 60° F, 14.73 psia


added at or upstream of the point of pressure reductionto offset the Joule-Thomson effect (approximately 7OFper 100 psi of pressure decrease).

Chapter 14.1 recommends the use of steam, hot water,or electrical heat tracing, catalytic heaters and insulationto provide heat to the sample gas. If the ambienttemperature will keep the sample gas at 20-50OF abovethe hydrocarbon dew point temperature, heating is notrequired. Previous research suggested that the fill-and-empty spot sampling method might heat the samplecylinder enough to avoid hydrocarbon phase changeproblems in cold ambient conditions. Experiments wereconducted in 2001 and 2002 to monitor the samplecylinder temperature during the fill-and-empty process,and determine the method’s self-heating ability atdifferent line conditions and ambient temperatures. It wasfound that the method does heat the sample cylinder,

FIGURE 6. Composite sampler field test results showingthe effect of placing a heated enclosure around the

sampling system.

especially in cold weather, but does not generate enoughheat to be used on saturated gas when the initial samplecylinder temperature is lower than the pipeline gastemperature.

If sample cylinders are exposed to temperatures belowthe hydrocarbon dew point temperature after sampling(Path 4-5 in Figure 2), the sample can be recovered byheating it to 140OF or 20-50OF above the hydrocarbondew point temperature, whichever is lower, for at leasttwo hours prior to analysis.

GAS SAMPLE DISTORTION CAUSED BY DIRTY ORCONTAMINATED SAMPLING SYSTEMS

The previous discussion showed that the presence ofliquid hydrocarbons affects the integrity of a natural gassample. Liquid hydrocarbon contaminants are not alwaysnatural gas constituents. Occasionally, heavier

hydrocarbons contained in compressor or machine oilcontaminate sampling equipment. Figure 7 shows theimpact of several residues left in a sample cylinder. Thehydrocarbon residue is a 50/50 mixture of SAE-30compressor oil and n-hexane through n-decane. Theliquid hydrocarbon residue caused a reduction in theheating value and density of the gas sample.

FIGURE 7. The impact of several residues left in asample cylinder. The hydrocarbon residue is a 50/50

mixture of SAE-30 compressor oil and n-hexane throughn-decane.

Composite Sampler #3 — All CylindersInstalled at Powder Wash, CO

Heated Enclosures Installed on 1/26/00

9/1/1999

9/15/1999

9/29/1999

10/13/1999

10/27/1999

11/10/1999

11/24/1999

12/8/1999

12/22/1999

1/5/2000

1/19/2000

1/26/2000

2/9/2000

2/23/2000

3/8/2000

3/22/2000

4/5/2000

4/19/2000

5/3/2000

5/31/2000

Diff

eren

ce fr

om O

n-lin

e G

C [%

]

-4

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

10ρ, [lbm/f t3] at 60°F, 14.73 psiaHV, [BTU/real cf] at 60° F, 14.73 psia

64.5% of ReferenceGC data used for

calculation.

12.9 % of ReferenceGC data used for

calculation.

NNNNooootttt HHHHeeeeaaaatttteeeedddd HHHHeeeeaaaatttteeeedddd

- 10

-8 0

-6 0

-4 0

-2 0

0.0

2.0

4.0

6.0

GlycCV

WaterDispl

CV

Displ

CPw/HRes

CVw/HCRes

CVClean

ρ [ lbm/ft3 ] at 60°F, 14.73 psia

Hv [BTU/real cf] at 60 ° F, 14.73 psia

* 1500 BTU/SCF, 100 psia, MRF Mix (8-11-97)

* 300cc SS Cylinders Stored at 110 ° F for 2-3 Days

If there is reason to believe that any part of the samplingsystem has been contaminated, the system must bethoroughly cleaned to obtain a representative natural gassample. Several cleaning methods were tested oncontaminated sample cylinders during the researchproject. Figure 8 shows the results of one series ofcleaning tests. The results indicate that most methodsleave some residual. Steam cleaning was the most robustmethod evaluated during the research.

If a gas sample is taken using a cylinder that has notbeen sufficiently purged, or if air or other contaminantshave leaked into the sampling system, then the integrityof the sample will be compromised, even if the samplingmethod is performed correctly. Figure 9 shows the effectof a nitrogen leak on the heating value and density of agas sample obtained during sample cylinder cleaningtests at SwRI. The nitrogen increases the density andreduces the heating value of the sample. The errors canbe significant and might not have been discovered if onlythe heating value of the sample was considered. Theeffect is similar if air is introduced into the system duringsampling.

The 2001 revision of Chapter 14.1 recommends thatsample systems be designed so that they can bethoroughly cleaned and that a procedure for cleaningsampling systems and sample containers be established.Chapter 14.1 recommends that sample cylinders becleaned prior to each sample collection.


Steam is identified as the most effective sample systemcleaning agent. Water containing corrosion inhibitors orother chemicals that may contaminate the samplingsystem should be avoided. Solvents, such as acetoneand liquid propane, that do not leave a residue afterdrying are considered acceptable. Decon Contrad 70®,or equivalent, is also acceptable. Other cleaning methodsmay be used, if testing can prove their effectiveness.

tested that did not adversely impact the gas heating valueand density determination (Figure 10).

The 2001 revision of Chapter 14.1 recommends the useof inert and non-porous materials in gas samplingsystems. 304 or 316 stainless steel is generallyrecommended as a suitable material for samplingsystems. Carbon steel is not recommended because ofthe potential for chemical reactions with the componentsin the gas, which can result in sample distortion.Furthermore, carbon steel is susceptible to high corrosionrates, particularly when used in wet or sour gas sampling.The use of dissimilar materials is discouraged due tothe potential for high corrosion rates and gas sampledistortion.

Valve seats, o-rings, and other types of seals should bemade of elastomers that will not degrade when thesample gas makes contact. With the exception of Nylon11, or its equivalent, Chapter 14.1 does not recommendthe use of plastic tubing in natural gas sampling systems.

For sour gas applications, Chapter 14.1 recommendsthe use of special linings or coatings, such as epoxy or

FIGURE 8. Cylinder cleaning tests conducted with 300 ccconstant volume sample cylinders.

Sample containers and sampling systems must be driedand purged, or evacuated, after cleaning. Nitrogen,helium, and dry instrument-quality air are acceptable fordrying sample containers and sampling systems. Blanketgases may be used to pre-charge sample containers.The blanket gas must be selected so that the analyticaldevice will not interpret it as part of the sample.

GAS SAMPLE DISTORTION CAUSED BY SAMPLINGSYSTEM COMPONENTS FABRICATED FROMMATERIALS KNOWN TO AFFECT THE INTEGRITY OFA NATURAL GAS SAMPLE.

Many materials commonly used to fabricate samplingsystem components can distort gas sample integrity.Highly porous materials and components with largesurface areas are likely to cause adsorption or ‘sticking’of hydrocarbon molecules to the surface of the material,that produce a corresponding reduction in heating value.If the temperature and/or pressure of the systemchanges, adsorbed molecules can be released from thematerial surface and can re-enter the sample stream,causing an increase in the heating value. Thisphenomenon cannot be eliminated by cleaning thesampling system.

Several common tubing materials of different diametersand lengths were tested to determine the impact on gassample integrity. Clean stainless steel tubing was foundto have little or no impact. Most plastic tubing materialshad an impact on sample integrity, with polyethylenecausing a reduction in heating value and density of over6 percent. This is believed to be caused by solid-surfaceadsorption. Nylon 11 was the only plastic tubing material

FIGURE 9. The effect of nitrogen that has leaked acrossthe seals of a floating piston cylinder during a particular

set of tests.

Ave

rage

Gas

eous

Res

idua

l H

[Btu

/scf

] (5

Rep

eats

)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0* 300cc SS CCCooonnnssstttaaannnttt VVVoollluuummmeee CCCyyyllliiinnndddeeerrrssss* UHP Meth ane Solvent at 45 psia, 75° F, 3-4 days* An alysis at 115° F (4 5 min. heat ing)* C8-C10 Calibration Corrected

Contrad70 (15%)

We tSteam

Acetone N2 CH4 Evac. Liqu idPropane

CCCClllleeeeaaaannnniiiinnnngggg TTTTeeeecccchhhhnnnniiiiqqqquuuueeee

+0.2% for a 1,000 Btu/scf

SCCO2

Diff

eren

ce fr

om C

ontro

l [%

]

-8

-6

-4

-2

0

2

4 Density Difference from ControlHeating Value Difference from Control

Control Repeat Cleaning Method 2Cleaning Method 1

IIIInnnnccccrrrreeeeaaaasssseeee iiiinnnn ddddeeeennnnssssiiiittttyyyy aaaannnndddd ddddeeeeccccrrrreeeeaaaasssseeeeiiiinnnn hhhheeeeaaaatttt iiiinnnngggg vvvvaaaalllluuuueeee aaaassssssssoooocccciiiiaaaatttteeeedddd wwwwiiiitttthhhhnnnniiiitttt rrrrooooggggeeeennnn ccccoooonnnnttttaaaammmmiiiinnnnaaaatttt iiiioooonnnn....

FIGURE 10. The effect of several types of plastic tubingon a 1,250 BTU/SCF natural gas.

Diff

eren

ce [%

]

-8.00

-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

8.00

* 1/8 inch X 50 ft Tube Lengths* 1,250 Btu/scf, #C3 6388 Mix* 10 0 psia, 110° F* 65 psia, 110° F (Tygon Tube)

ρ [lbm/f t3] at 60 °F, 14.73 ps iaHv [ Btu/real cf ] at 60°F, 14.73 psia

Teflon( PTFE)

Polyethy lene (Linear Low Density)

Nylon 11 Tygon(R-3603)

Control(No Plastic Tube)


other suitable coatings. Soft metals, such as brass,copper, and aluminum are not recommended.

CONCLUSIONS

Unrepresentative (distorted) natural gas samples can becollected through improper sampling methods and poorlydesigned sampling systems. Errors derived from thecalculation of gas properties, such as heating value anddensity, based on analyses of distorted gas samples mayexceed 10 percent. These errors will propagate throughthe energy flow rate equations, resulting in an error inthe calculated energy flow rate.

Gas sampling research conducted at SwRI over the lastseven years has identified several causes of gas sampledistortion. Three causes of gas sample distortion are (1)sample distortion due to equipment and processes thatcause the sample gas temperature to drop below thehydrocarbon dew point temperature, (2) sampledistortion caused by dirty or contaminated samplingsystems and (3) sample distortion caused by samplingsystem components fabricated from materials known toaffect the integrity of a natural gas sample.

The 2001 revision of Chapter 14.1 (i.e., Collecting andHandling of Natural Gas Samples for Custody Transfer)of the API Manual of Petroleum Measurement Standardsincludes discussions of these and other causes of gassample distortion. Chapter 14.1 providesrecommendations for obtaining representative gassamples by avoiding gas sample distortion. Itemphasizes the importance of avoiding the hydrocarbondew point temperature and recommends the use of the

hydrocarbon phase diagram as a design tool to keepsampling equipment temperatures above thehydrocarbon dew point temperature. The revision alsoemphasizes the critical need for sampling equipmentcleanliness and recommends that sampling systemdesigns include features that allow them to be thoroughlycleaned in a timely manner. Chapter 14.1 also addressesthe importance of avoiding components fabricated frommaterials known to cause gas sample distortion, suchas many types of plastic tubing.

Technicians and engineers responsible for obtainingnatural gas samples should be aware of the factors thatcause sample distortion. The causes of sample distortion,such as the hydrocarbon dew point temperature,sampling equipment material selection, and equipmentcleanliness should be considered when obtaining gassamples and when designing gas sampling systems.

REFERENCES

GPA Standard 2166, Obtaining Natural Gas Samples forAnalysis by Gas Chromatography, Gas ProcessorsAssociation, Tulsa, 1998.

Manual of Petroleum Measurement Standards, Chapter14-Collecting and Handling of Natural Gas Samples forCustody Transfer, American Petroleum Institute,Washington D.C., 2001 Revision.

Metering Research Program: Natural Gas SampleCollection and Handling –Phase I, Behring, K.A. III andKelner, E. GRI Topical Report No. GRI-99/0194.

Eric KelnerDarin George


ULTRASONIC METER FLOW CALIBRATIONSCONSIDERATIONS AND BENEFITS

Joel ClancyColorado Engineering Experiment Station, Inc. of Iowa

2365 240th Street, Garner, IA 50438

INTRODUCTION

The primary method for custody transfer measurementhas traditionally been orifice metering. While this methodhas been a good form of measurement, technology hasdriven the demand for a new, more effective form of fiscalmeasurement. Ultrasonic flowmeters have gainedpopularity in recent years and have become the standardfor large volume custody transfer applications for avariety of reasons. Most users require flow calibrationsto improve meter performance and overall measurementuncertainty. Although AGA Report No. 9, Measurementof Gas by Multipath Ultrasonic Meters [Ref 1], currentlyonly recommends flow calibration for ultrasonicflowmeters, the next revision will likely require flowcalibration for all ultrasonic custody transfer applications.

What considerations then, should be taken whenchoosing to flow calibrate an ultrasonic flowmeter? Whatare the benefits to the user? What should a user expectfrom a flow calibration? What kind of performance shouldthe customer expect or accept from an ultrasonic meter?These areas, as well as others will be explored andconsidered.

PRE-CALIBRATION INSPECTION AND METERINSTALLATION

Upon receiving the ultrasonic flowmeter at the calibrationfacility, a thorough inspection is started. Ultrasonicmeters are often very large with attached electronicinstruments so the inspection of the ultra-sonic flowmeterbegins before it comes out of the box. A damagedshipping container indicates that the meter may havevisible damage or damage to electronic components thatwill be harder to find. Open the crate and inspect theelectronics. Ensure that all electronic boards are securelyfastened to the junction box that houses the electronics.Look for any signs of damage or any loose parts orfittings. Inspect the meter body. Ensure that thetransducers are not damaged. Ensure all cables aresecurely fastened.

When installing the meter in the piping system, inspectthe holes where the pressure taps penetrate the meterbody on the inside surface. Any burrs or protrusions onthe pressure taps can create pressure reading errors andmust be removed prior to calibration. Ultrasonicflowmeters are often sold with upstream and downstreamspool pieces. There may be identification stamps on themeter and accompanying spool pieces, make sure theidentification numbers match. The meter and spool

pieces may have alignment pins. Check the alignmentas it is not unusual to find the pins do not provide goodalignment. Spool pieces may come separately from adifferent supplier. In this case ensure that the internaldiameter of the spool pieces matches the internaldiameter of the meter. Drawings typically accompanythe meter and spool pieces. Assemble the meter partsas shown in the drawings. Flow conditioners sometimesfail when first used. Inspect the flow conditioner to ensurethe manufacturing and assembly is complete. Some flowconditioners need to be pinned as they can move aroundinside the pipe when installed. It is important that allupstream components that can affect the flow conditionsat the meter remain exactly the same in use as they wereduring the calibration.

Meters that have been in use in the field are oftenrecalibrated. These meters do not have original shippingcontainers and are often partially disassembled forshipping. Inspect the cables carefully ensuring all cablesare with the meter and that no damage has occurred.Inspect the inside surfaces of the meters. There is oftena build-up of contaminants. Ensure the pressure tapsare clear. The customer may want the meter calibratedin the condition it arrives in, referred to as an “As Found”calibration, then cleaned and recalibrated clean, referredto as an “As Left” calibration. In the past few years, moredata has become available on how well ultrasonicflowmeters perform with thick layers of contaminants onthe transducers and pipe walls. Any difference inperformance of the meter between the “As Found” and“As Left” calibration may be very useful to the customer.

Once the meter and spool pieces have been installed inthe test section the instrumentation can be installed andthe meter can be powered up. Pressure and temperaturetransmitters are now installed. Dual instrumentation ispreferred. When dual instrumentation is used anydifferences in readings can be identified quickly allowingthe calibration to proceed smoothly. The communicationlines are then connected to the meter. The most commoncommunication options are RS-485, RS-232, or Ethernetcommunication. One communication output is used tocommunicate flow and meter status information to acomputer running software provided by themanufacturer. Another output is also connected to themeter. This output is a second flow signal from the meter.The second output may be an RS-485 output or themeter may produce a frequency output, which isproportional to flow passing through the meter. If an oldermeter is received from the field for recalibration it mayrequire some communication switch changes to allow


communication with the calibration facility. Thesechanges are well documented and are returned to theirinitial settings once the calibration is complete.

The location of the thermal wells should be noted. AGAReport No. 9 discusses the appropriate placement of athermal well stating 2 to 5 pipe diameters downstreamof the ultra-sonic meter. In the case of a bi-directionalmeter, AGA 9 calls for a thermal well placement of 3diameters from either ultrasonic meter flange face.Although thermal well placement is defined in AGA 9,some users elect to choose a different location for theirthermal well(s). Caution should be taken here. Oftenthermal wells are placed upstream of a flow conditioner.The pressure drop created by the flow conditioner alsocreates a corresponding temperature drop known as theJT (Joules Thompson) effect. You have, therefore, adifferent temperature at the meter than is being recordedby the temperature transmitter. Proponents of this typeof temperature measurement design look to the designand expense that has gone into the conditioning of theflow. Flow conditioners can help produce a good,symmetrical flow profile; given the upstream flowconditions are not extremely severe. A great deal ofthought, design and cost goes into ensuring a good,uniform flow profile, before the gas reaches the meter. Ifthe thermal well is placed at a position downstream ofthe flow conditioner, the profile becomes slightly skewed.However, research has shown that placing the thermalwell directly upstream of the meter does not adverselyaffect meter performance. This technique also ensuresaccurate temperature measurement at the meter. Manyusers, therefore, place the thermal well at 3 or 5 diametersupstream of the meter.

METER CALIBRATION

Test section pressurization, leak check, and pre-flow arenow performed. As the meter body pressurizes, dualpressure instrumentation is checked for good agreement.Pre-flow is generally conducted at 60 to 80% of the metercapacity. Pre-flow typically lasts for 15 to 30 minutes.The pre-flow allows the meter and test section piping tocome to the flowing temperature of the gas. Dualtemperature instrumentation is checked for goodagreement. During pre-flow, several piping andinstrumentation conditions are checked. Flowconditioners are often a source of flow noise. The amountof noise being generated by the flow conditioner ismonitored during pre-flow. Any unusual mechanicalnoises may be an indication that the flow conditioner iscoming apart or vibrating violently. Installing a thermalwell too close to a flow conditioner can cause thermalwell vibration. This installation can produce severalproblems. This vibration can cause problems for theultrasonic meter. The introduction of noise inhibits themeter’s ability to function properly. The thermal wellvibration also creates a heating effect that will producea temperature measurement error at the ultrasonic. Whenperforming pre-flow, the performance of all the flowtransducers is monitored to ensure there are no chord

failures. Unusual signals can be produced from a varietyof problems to include a bad set of transducers, incorrectwiring, etc.

Once pre-flow is finished, the calibration begins. The flowis taken up to the highest flow rate requested by thecustomer. If no flow-rates have been specified by thecustomer, the flow-rate is taken to the maximum flow-rate suggested by the manufacturer. At the high flow-rate there may be enough flow noise to cause chordfailure. That is, the flow noise is of a sufficient level toweaken the signal received by the meter. It is importantto monitor the system carefully when increasing flow tothe highest flow-rate. If any components like flowconditioners are going to fail then this is the time whenfailure is most likely to happen. Any unusual noises orlarge changes in noise may indicate that a systemcomponent is experiencing failure.

When flow at the highest flow-rate has been established,the calibration system is allowed to stabilize. Ultrasonicmeter calibration systems may be composed of largepiping systems with a considerable amount of volumebetween the standards used to accurately measure flowduring the calibration and the ultrasonic meter beingcalibrated. It is important that any pressure fluctuationsthat may be present in the system due to changes inflow-rate are allowed to dissipate. When stable flowconditions have been observed for an adequate lengthof time, calibration data can be taken from the ultrasonicmeter being calibrated and the calibration system.Several data points may be taken at a single flow-rate.The number of data points may be specified by thecustomer, or it may be left to the judgment of thecalibration system operator.

Data may be acquired using two separate computersystems. One system will be running software suppliedby the manufacturer that will interrogate the meter whilea data point is taken and another system will acquiredata from the calibration system. Typically, these twosystems acquire data during the same time period.

Obtaining the calibration log file from the meter’s softwaredata logs can prove to be an important tool once themeter is put into service. This initial log collected at thetime of the calibration can provide information such asspeed of sound, or gain level to limit ratios on a chordby chord basis to name a few. This initial log file collectedat the time of the flow calibration is often referred to asthe meter’s baseline or fingerprint. When collecting logsthroughout the life of the meter, the baseline logs can beused as a reference. Any deviations from the ratiosobserved at the calibration can be used as a way totroubleshoot potential problems with meter performance.

GENERAL DESCRIPTION OF A CALIBRATION ANDCALIBRATION SYSTEMS

Calibrations are performed by placing a flow standard inline with the ultrasonic flowmeter being calibrated. The


flow standard is used to accurately measure flow andhas been calibrated using standards that are traceableto NIST or some other national standard. As long as thereare no leaks in the system between the standard andthe meter being calibrated it can be assumed that thetwo meters are passing the same amount of flow. Theremay only be one standard or there may be many thatcan be placed in parallel in the flow stream to produce awide flow-rate range.

There are two basic types of calibration systems. Figure5 shows a calibration system on an existing natural gaspipeline. When the large valve on the pipeline is closedslightly a differential pressure across that valve isproduced. The differential pressure across that valveprovides a motive force to push flow through thecalibration system. As the main pipeline valve is closedfurther, more flow is pushed through the calibrationsystem. In this manner a wide flow-rate range can bepassed through the calibration system allowing thecalibration of a wide range of meter sizes. At very lowflow-rates, fine flow control can be accomplished bythrottling with a smaller valve inline with the meter beingcalibrated.

There are advantages to this type of system. Becausethe pipeline is passing flow constantly, very long datapoints or many data points at a single flow-rate can betaken. The meter being calibrated is flowing under actualpipeline conditions. There are potentially many gaschromatograph outputs available to monitor gascomposition so a stable gas composition can be assured.Because there are gas chromatographs placed atmetering stations along the pipeline any variations in gascomposition can be seen well in advance as the gasproceeds through the line. This type of system can hitvery high flow-rates allowing calibration of the largestultrasonic meter sizes.

The disadvantages of this type of system vary. Thepressure drop at a given location will have limitations.This does not affect ultrasonic calibrations but can be aconsideration when calibrating meters like orifice metersthat need to create a pressure drop. The calibrationsystem has to operate at the pressure in the pipeline.This means that a meter to be used in a low-pressuresystem may have to have a flange with a higher pressurerating installed temporarily for the calibration. The qualityof the data may be affected by the pipeline stability. Ifthe line pressure is rising or falling, it may not be possibleto acquire data. Typically however, pressure, temperatureand gas composition are quite stable.

The second type of system typically used is shown inFigure 6. This type of system is referred to as apressurized loop. Prior to flowing, the loop is pressurizedwith gas to the desired flowing pressure. Flow throughthe system is created by a compressor that must runcontinuously while calibrating. Flow control valves canbe placed in the system for flow control. Heat exchangersin the system allow some temperature control.

A pressurized loop system also has certain advantages.The flow in pressurized loop systems can be preciselycontrolled. The temperature in a pressurized loop canbe varied over a limited temperature range allowing theeffects of different flowing temperatures to beinvestigated. The composition of the flowing gas can bevaried by injecting different components into the loop.

The disadvantages of a pressurized loop calibrationsystem include high operating expenses. This is becausethe compressor(s) must be operating continuously. Thesuction pressure on the compressor must be maintainedabove some minimum value, which places a limitationon the differential pressure across the loop. Thislimitation, as well as limitations to compressor capacitiesand line size is typically the primary contributing factorsto flow limitations on a pressurized loop based calibrationsystem. [Ref. 2]

AS FOUND, AS LEFT AND AGA 9

Although not a standard, AGA Report No. 9 is a usefultool for the end user to use as a basic guideline forultrasonic meters. While most users use this as aguideline for acceptance criteria, not all use this as hardand fast pass-fail criteria. Many users look at the “as-left” performance of the meter.

Observe the following example of two twelve inchultrasonic meters: Meter “A” (Table 1, Figure 1) meetsAGA 9 criteria both in linearity and in offset (% error).Meter “B” (Table 2, Figure 2) does not meet AGA 9 criteriain offset; however it does meet the linearity criteria. Notethe “as-left” results of meter A. This example uses theAGA 9 Flow Weighted Mean Error (FWME) adjustment.Applying this adjustment to the meter does not correctfor the non-linearity of the meter and therefore makesthe meter read slow at the high end and fast at the lowend. Conversely, Meter “B” is quite linear. The as-leftadjustment allows this meter error to be reduced to lessthan 0.04%, after the FWME adjustment is applied,throughout the entire range. While Meter “B” does notmeet AGA 9 criteria, this meter exhibits betterperformance after adjustment.

AGA REPORT NO. 9 CALIBRATION FACTORCALCULATION EXAMPLE

A 12-inch ultra-sonic flowmeter is to be calibrated witha maximum flow-rate of 100 ft/sec. The customer hasrequested that calibration data be taken in compliancewith AGA Report No. 9 with a minimum flow-rate at ameter velocity of 1 ft/sec. The calibration facility operatorsets up the calibration plan shown in Table 1.


TABLE 1. 12-INCH METER “A” RESULTS SUMMARY Velocity % Error % Error % Full % F.S x ft/sec As Found As Left Scale Error 100 -0.69 -0.19 100% -0.69 70 -0.5 0 70% -0.35 50 -0.4 0.1 50% -0.2 30 -0.32 0.18 30% -0.096 20 -0.3 0.2 20% -0.06 10 -0.26 0.24 10% -0.026 4 -0.2 0.3 4% -0.008 1 -0.01 0.49 1% -0.0001

Sum = 2.85 -1.4301

From the above results, a FWME can be calculated.Utilizing the data from the above table, the Percent FullScale must be determined. This value is calculated asfollows.

Now the percent full-scale values are multiplied by thepercent error (as found) values. This leaves the resultingvalue in the percent full-scale times error column in table1. The percent full-scale and percent full-scale times errorcolumns are then summed. With these summed values,a FWME can be calculated. This is done by dividing thesummed percent full scale, times percent error value, bythe summed percent error value. From table 1, wecalculate the following.

Once the FWME is derived, a calibration factor can becomputed. The calibration factor (or calibrationcorrection) is then entered into the meter software andthe meter is adjusted electronically. The calibration factoris calculated as follows.

TABLE 2. 12-INCH METER “B” RESULTS SUMMARY Velocity % Error % Error % Full % F.S x ft/sec As Found As Left Scale Error 100 0.75 0.03 100% 0.75 70 0.71 -0.01 70% 0.497 50 0.72 0 50% 0.36 30 0.74 0.02 30% 0.222 20 0.71 -0.01 20% 0.142 10 0.72 0 10% 0.072 4 0.74 0.02 4% 0.0296 1 0.77 0.05 1% 0.0077

Sum = 2.85 2.0803

FIGURE 2.Meter “B” Results with Correction Applied.

AGA 9 acceptance criteria are a good tool for the userto utilize. Many custody partners will agree to use thesecriteria as a basis for “pass or fail”, in order to have anaccepted agreement prior to flow calibrating theultrasonic meter. However, the above example showswhy some users may choose to accept a meter that doesnot meet AGA 9 criteria.

ALTERNATIVE METHODS FOR METERADJUSTMENT

AGA 9 also allows for alternative methods for adjustingthe meter. One popular method that is being used byseveral ultrasonic meter manufacturers is a second orderpolynomial curve fit. This method would better fit oradjust meter A in the previous example. Note the belowresults (Table 3, Figure 3) showing the same meter andit’s new as-left condition utilizing a second orderpolynomial curve fit. This linearizes the meter quite well,leaving the meter error at less than 0.04% through theentire range.

PercentFullScaleIndicatedFlowrate

MaximumFlowratex= 100

FWME = − = −1 432 85

0 50175.

..

CalibrationFactorFWME

=+

=100100

1 0050.

FIGURE 1.Meter “A” Results with Correction Applied.

Velocity (Ft/sec)

Meter “A” Results

% E

rro

r

Velocity (Ft/sec)

Meter “B” Results

% E

rro

r


TABLE 3. METER “A” WITH 2ND ORDERPOLYNOMIAL ADJUSTMENT APPLIED Velocity % Error % Error % Full % F.S x ft/sec As Found As Left Scale Error 100 -0.69 -0.03 100% -0.69 70 -0.5 0.03 70% -0.35 50 -0.4 0.03 50% -0.2 30 -0.32 0.02 30% -0.096 20 -0.3 -0.01 20% -0.06 10 -0.26 -0.03 10% -0.026 4 -0.2 -0.03 4% -0.008 1 -0.01 -0.03 1% -0.0001

FIGURE 3.Meter “A” with 2nd Order Polynomial Adjustment Applied

Another method being used by one meter manufactureris a Piece-Wise Linearization (PWL) technique. Thismethod allows the user to adjust the meter at the timeof the flow calibration, such that all points theoreticallyfall directly on the 0.0% error line throughout the entirerange (Table 4, Figure 4). This same technique can alsobe incorporated in the user’s flow computer, however,this typically does not allow for a verification check aftersuch adjustments have been made; where the PWLcoefficients installed directly in the meter allow for averification checkpoint to be run at the time of the flowcalibration. Note Figure 4 below showing a PWLadjustment.

TABLE 4. METER “A” RESULTS WITH PWLADJUSTMENT APPLIED. Velocity % Error % Error % Full % F.S x ft/sec As Found As Left Scale Error 100 -0.69 0 100% -0.69 70 -0.5 0 70% -0.35 50 -0.4 0 50% -0.2 30 -0.32 0 30% -0.096 20 -0.3 0 20% -0.06 10 -0.26 0 10% -0.026 4 -0.2 0 4% -0.008 1 -0.01 0 1% -0.0001

FIGURE 4.Meter “A” Results with PWL Adjustment Applied

METER RANGEABILITY AND CALIBRATION POINTSELECTION

Depending upon the application and station throughput,users may choose to install one large ultrasonic meter,two mid-sized meters, or several, small ultrasonic meters.This is true for both new meter stations, and stationsbeing retro-fitted with ultrasonic flowmeters.

Note one example where nine, 12 inch orifice meter runswere replaced with one 30 inch ultrasonic flow meter.This decreased the user’s maintenance costsconsiderably. For a new meter station, with the samethroughput, this would decrease costs for pressure,temperature, and differential pressure transmitters.Additionally, this station requires less flow computers,valves, and piping to name a few. Proponents for suchan installation point to these areas as the benefits ofinstalling one, large volume meter.

The above installation does have disadvantages,however. Having one, large volume meter can makemeter re-calibration difficult. Additionally, many users feelthat installing a single meter would be “putting all youreggs in one basket”. If the meter becomes inoperable,the pipeline’s measurement would be adversely impactedif a second meter was not available for measurement.Because of this philosophy, some users would, forexample, choose to install two 16 inch meters for theirfiscal measurement. This would enable the user to havea second meter in the event that one might fail. It alsoallows for ease in maintenance, such as pulling the meterand meter run for cleaning or sending off for re-calibration.

The above examples can and do drive where the meterwill operate within its range. Once again, depending uponstation design, users should flow calibrate the meterwhere they expect to operate this meter. In the case ofthe 30 inch meter above, the user, knowing the meterwould not be used below 10 ft/sec, chose to calibratedown to 5ft/sec. Many users are pushing the lower limitsof ultrasonic meters in an effort to reduce the costs of

Velocity (Ft/sec)

Meter “A” Results – Polynomial Adjustment

% E

rro

r

Velocity (Ft/sec)

Meter “A” Results with PWL Adjustment

% E

rro

r


installing secondary low-flow custody meters (such as aturbine or PD meter). Because of this, and improvedcurve fitting methods, many users are flow calibratingdown to 1 or 2 ft/sec with expectations of operating atthat low end for short periods of time. Due to potentialthermal gradients and meter repeatability at this lowoperating range, the meter’s uncertainty is reduced, butfor many users the tradeoff is beneficial, as the costs ofinstalling a second low flow meter are not incurred.

Data is typically taken at a minimum of six flow-rates.AGA Report No. 9 specifies flow-rates of qmin, 0.10qmax,0.25qmax, 0.40qmax, 0.70qmax, and qmax. Additional datapoints may be requested at specific flow-rates by thecustomer if the meter is to be used in a specific flow-rate range. Once the initial calibration is complete, theappropriate adjustment(s) should be appliedelectronically to the meter. After the adjustment has beenmade, a verification calibration is typically conducted.The user typically chooses between one and threeverification points to be run in order to ensure theadjustment was applied correctly and worksappropriately.

POST CALIBRATION CONSIDERATIONS

Once the calibration is complete, data should bereviewed by the user. Once the data has been accepted,a security jumper is set (if available) to set the meter intoa “read-only” mode. This security feature inhibits anyaccidental (or intentional) changes that might be madeto the meter’s metrology configurations.

Shipping instructions should be provided to thecalibration facility in order to ensure proper and timelydelivery to the end location. The meter and its associatedpiping should be end capped and sealed such that nodebris or dirt can enter the meter and/or meter tubeinternal section.

Once the meter is installed in the user’s pipeline, acommunications check should be performed. Ensureproper responses to the flow computer. A log file shouldbe collected for your records. This data should be

reviewed paying specific attention to individualtransducer path data. If the meter is under flowingconditions, this log file should be compared to the metersbaseline logs taken at the time of the flow calibration.Some users choose to take daily logs for the first weekto add to the meter’s baseline data. This allows for severalsets of baseline data to be collected while the meter isstill typically in a new and clean condition. Most userswill then take monthly logs. These logs can then betrended with previous logs to ensure path ratios areremaining constant and that no anomalies have occurredwith other log file data. During the course of monitoringthese data, differences may be noted. For example, aspeed of sound difference may be observed. Often, thisdeviation from the trend may not be due to problemswith the ultrasonic meter. Frequently, the user mayinstead discover problems with pressure or temperaturemeasurement, or a problem with the gas chromatograph.If these instruments are checked and known to be inproper working condition, further investigation shouldbe conducted by reviewing the meter’s log files. Specificattention should be given to individual path speed ofsound, gains, performance, and signal to noise ratios toname a few. If any of these differ from recent trends inthe meter’s performance, this may point to potentialproblems with the ultrasonic meter. Your company’sultrasonic measurement specialist and/or the metermanufacturer should be consulted if an abnormality isdiscovered.

SUMMARY

Today’s users continue to demand improved performanceon measurement. Improved technology, and the wide-spread use of ultrasonic metering has allowed enhancedperformance and improved overall measurementuncertainty. A key part of this equation is the flow calibrationof ultrasonic meters being used for fiscal measurement. Aknowledgeable user should have a good understanding ofareas to look at during the calibration process, as well asoptions available with respect to operating range, numberof data points, and meter rangeability. Once operational,the user should use the diagnostic tools available to monitorthe health of the meter.

FIGURE 5.Pipeline Based Calibration Facility.


Joel Clancy

FIGURE 6.Pressurized Loop Based Calibration Facility.

REFERENCES

1. AGA Report No. 9, Measurement of Gas by MultipathUltrasonic Meters, June 1998

2. William Johansen and Joel Clancy, Calibration ofUltrasonic Flowmeters, International School ofHydrocarbon Measurement, May 2003


UNDERSTANDING THE DIFFERENT STANDARDSTHAT GOVERN MEASUREMENT

An Overview of the Different Standards That Govern MeasurementRonald E. Beaty PE

Premier Measurement Services, Inc.19302 Allview Lane, Houston, TX 77094

A number of questions should be answered regardingthe origin of Measurement Standards used in the UnitedStates.

Why do we need measurement standards?Who decides when a standard will be written?What group undertakes the preparation of a

document?How much time is spent writing a document?Who decides the document is adequate and

approves it use?Are there any checks and balances?

No national standards for measurement were accepteduntil the early 1930’s. The use of installation andvolumetric calculation procedures required an individualcontract agreement for each custody transfer delivery.The founders of the American Gas Association (AGA)commissioned research by the United States NationalBureau Standards (NBS) to development data that wouldlead to their first orifice meter standard, AGA Report No.1. The AGA issued the first revision of the orifice meterstandard Report No. 2 later in the 1930’s. In 1963, theinitial version of AGA Report No. 3 governing orificemeasurement was published. AGA voted to retain thetitle, AGA Report No. 3 for all subsequent revision.

The American Petroleum Institute (API) assumed the leadin getting the Orifice Meter Standard approved as anAmerican National Standard by the American NationalStandards Institute (ANSI) in 1978. The Gas ProcessorsAssociation (GPA) joined with the AGA and the API todevelop the first truly national standard for Natural GasFluids Measurement for Concentric Square-EdgedOrifice Meters that was published in 1990 thru 1992. Part2 of the standard, Specification and InstallationRequirements, was revised in 2000.

A significant question regards the personnel that staffthe API, AGA and GPA Measurement Committees. TheAPI Committee On Gas Fluids Measurement is staffedwith representatives of the member companies, expertmeasurement consultants, academics, and governmentemployees of the Bureau of Land Management (BLM)and Mineral Management Service (MMS). The AGA andGPA Committees are similarly staffed. The committee isdriven by new technical advanaces or new data aboutthe meters. The parent committees of the variousMeasurement Committees must approve thedevelopment of a new standard. The API allows aspecified period of time to complete the development ofthe standard. If the standard is not completed asscheduled, it will be rescheduled and or allowed to be

sunseted. Meetings are characterized by open technicaldiscussion. Care is taken during these discussions torestrict the conversation only to the technical issues.Standards were developed for the positive displacementmeter, the turbine meter, the ultrasonic meter, etc. in thesame way.

API STANDARDS

The API publishes standards covering both oil and gasmeasurement. All of these standards are availableindividually or in a compilation known as the Manual ofPetroleum Measurement Standards. Chapter 14 isdedicated to natural gas fluids and contains eightsections.

Section 1–Collecting and Handling of Natural GasesSamples for Custody Transfer was published in 2001 andgenerally was not widely used in 2002.

Tests showed that several incorrect statements and theresults from limited data should be corrected. The testsshowed how important a properly cleaned samplecylinder is for assuring a precise analysis. Steam cleanedcylinders produced the most accurate samples foranalysis. This was accomplished by advocating the useof heated sample lines and heated sample cylinders toensure single-phase measurement. Tests of the multiplemethods contained in the old standard revealed thatsome of the methods differ greatly from the resultsexpected. The two best methods identified for samplingnatural gas based on all of the tests are the purge and filland the helium pop procedures. Pigtails are needed toassure temperature stability with any of the multiple fillingprocedures.

Section 2–Compressibility Factors of Natural Gas Otherand Related Hydrocarbon or AGA Report 8 wasreaffirmed in 1999.

This document contains the formulas to correct thevolume measured for compressibility (formerly supercompressibility). The method was developed by AGA incooperation with the Gas Research Institute (GRI). Theresearch work was extensive requiring a number ofresearch man-hours. The result is a document that maybe used for a partial analysis through a defined range oftemperatures and pressures. An extended or full analysisis required beyond this range. The uncertainty of theformulas increase as the pressure and temperaturedeviate more from the first defined range that waspreviously discussed.


API STANDARDS

Section 1–Collecting and Handling of Natural GasSamples For Custody TransferThe standard was published in 2001. It is intended toassist the user in correctly sampling of single phasenatural gas production. Proper preporation of the samplecontainer is discussed. The proper sampling proceduresare outlined. The need for heated regulator and heattraced tubing is explained.

Section 3–Concentric, Square-Edged Orifice Meters isdivided into 4 Parts.

Part 1–General Equations and UncertaintyGuidelines was reaffirmed in 1999.Part 2–Specification and InstallationRequirements was revised in April 2000.Part 3–Natural Gas Applications was reaffirmedin 1998.Part 4–Background, Development, Imple-mentation Procedures and SubroutineDocumentation was reaffirmed in 1999.

The orifice meter standard is comprehensive and lengthy.The committed decided to separate the document intofour parts to make revisions only in the areas where theywere needed. The function of each section is evidentfrom the titles. Part 1 provides the derivation of thegeneral calculation equations and uncertainty guidelines.Part 2 provides the specifications of the upstream anddownstream meter tubes and fittings. The specificationsfor the orifice plates are redefined. The meter tube lengthsare totally revised based on extensive testing with eachdata set verified by two or more laboratories withworldwide creditability. Part 2 has been rewritten andpublished in 2000. Part 3 uses the formulas developedin Part 1 and applied them to natural gas measurement.Part 4 provides implementations procedures for thecalculation of the measured volume and neededsubroutine documentation to allow individualprogrammers to be assured the program is accurate.Parts 1 and 3 are being rewritten to include new dat thatis currently developped or is under development.

Section 4–Converting Mass of Natural Gas Liquids andVapors to Equivalent Liquid Volumes was reaffirmed in1999.

The standard was developed by the GPA to standardizethe calculation procedure to convert the mass measuredto gallons of a given component in the case of liquids orMCF of a given component in the case of natural gas.

Section 5–Calculation of Gross Heating Value, SpecificGravity, and Compressibility of Natural Gas Mixtures FromCompositional Analysis–was revised in 1996 and is GPA2172-96. The document contains the approved methodfor calculating relative density and saturated, partlysaturated and dry heating value at any selected pressurebase.

Section 6–Continuous Density Measurement was revisein 1998. An erratum was also published in 1998.

This standard addresses the installation, operation andcalibration of densitometers. A densitometer is calibratedusing a pycnometer. A pycnometer is simply a spheroidwith an inlet valve attached to n internal filling tube withan exit valve located in the perpendicular plain. Thevolume and mass of the pycnometer is known at apressure and temperature defined in the standard.Correction factor formulas for other temperatures andpressures are supplied. A concern has arisen concerningthe sealing of the shut off vaules that may require anerrata to be prepared. The standard is also undergoinga revision which may make an errata unnecessary.

Section 7–Mass Measurement of Natural Gas Liquids(GPA 8182-95) was reaffirmed in 1999.

This standard brings together the best designs for massmeasurement metering. A standard mass meter stationconsists of a standardized meter (turbine meter, PD meteror orifice meter), a densitometer, and a continuoussampler operated proportional to flow. Excellentaccuracy has been obtained when the measurementdevices are properly installed, operated and tested. Thestandard requires the temperature and pressure of themeter and the densitometer to be the same or vary onlywithin a range prescribed in the standard. The range oftemperature and pressure differences allowed iscontrolled by the product being measured.

Section 8–Liquefied Petroleum Gas Measurement waspublished in 1997.

The major thrust of this standard is to provide directionregarding how to calculate liquid volumes using an orificemeter.

API publishes two additional gas measurement standardsof major interest to the community:Chapter 20.1–Allocation MeasurementChapter 21.1–Electronic Gas Measurement.

Allocation Measurement addresses how gas and liquidproduction will be divided between the variousproducers. The producers and the pipeline operators arefaced with a significant task as environmental conditionsand pressure losses in the system normally result incondensation of hydrocarbons. The normal result isgreater amounts of liquid being delivered than weredelivered to the system. Greater problems will occur withdeeper water production in the Gulf of Mexico. Thestandard guides the operator to an equitable allocationof the product. It is being rewriten to address theconcerns serious that have been documented.

Electronic Gas Measurement addresses flow computerinstallations. It begins with pressure transducers and thetemperature transducer. The standard defines theamount of data to be reported and stored. Testing


frequencies are discussed. It is being rewriten to addressthe serious concerns that have developped since thedocument was published.

AGA STANDARDS

The AGA publishes several reports that should beextremely useful to the total user community

AGA Report No. 4A: Natural Gas Contract Measurementand Quality ClausesThe revision of this report was completed in June 2001to reflect the new gas quality requirements.

AGA Report 5: Fuel Gas Energy MeteringDevelop of new techniques for gas analysishasnecessitated the revision of the document and the TMChas begun the revision.

AGA Report No. 6: Transfer and Critical Flow ProversA new task group has been formed to revise thisdiscontinued report for proving pulse output meter.

AGA Report No. 7: Measurement of Gas by TurbineMetersThe report is being jointly revised by the DMC and theTMC. The expected completion date is 2004.

AGA Report No. 9: Measurement of Gas by MultipathUltrasonic MetersThe report is being revised to incorporate the knowledgegathered from the field and laboratory tests. The revisionis expected to be available by mid june 2004

AGA Report No. 10: Speed of Sound in Natural GasThe report has recently been completed. It is expectedto be released in early June 2003. The report contains

computer codes for C+ language programming. This willprovide a means to verify the ultrasonic meter calculation.

AGA Report No. 11: Measurement of Gas by CoriolisMetersThe report is expected to be released by August 2003.

GPA STANDARDS

The GPA publishes a number of standards used for gasmeasurement and gas quality determinations. Mostnatural gas contracts rreference GPA analyticalprocedures.

The GPA has published four revised standards since2000. They are;

2100-00; Tentative Method for the QualutatuveDetermination of COS in Propane.

2186-02: Extended Analysis of Liquids by GC

2261-00: Analysis of Natural Gas by GC

2145-03: Physical Constants for Hydrocarbons

A new version of GPA Engineering Data Book is beingperpared. The tentative release date 2004.

CONCLUSIONS

Standards bring the best, most knowledgeable personstogether to develop industries.

Standards are sound technically grounded documents.

Standards eliminate confusion in the industry and assurefair dealing.

Ronald E. Beaty


REPORT ON API 21.1 EGM STANDARDBrent E. BerryABB-Totalflow

Pawhuska Road, Bartlesville, OK 74003

A QUICK WORD ABOUT NOMENCLATURE

Since this report references both itself and the 21.1standard, the following nomenclature has been adoptedto make it clear which document is being referenced.

report - references this document, the one you arenow reading.

standard references the 21.1 standard, unlessotherwise noted.

section and subsection both refer to portions ofthe API 21.1 standard.

document is a generic term that could bereferencing either document. Hopefully the contextwill make it clear which document is beingreferenced.

INTRODUCTION

In September of 1993 API published a new section ofthe Manual of Petroleum Measurement Standards titledChapter 21 Flow Measurement Using ElectronicMetering Systems, Section 1 Electronic GasMeasurement. This report provides an overview of theAPI 21.1 document with the intent of serving as a primerand something of an introduction to the publication.

The 21.1 standard was developed by representativesfrom the American Petroleum Institute (API), AmericanGas Association (AGA) and Gas Processors Association(GPA) member companies with input from equipmentmanufacturers and others. The 21.1 standard representsthe first API publication in the field of electronic meteringsystems.

Work on the standard began much earlier than 1993 withinitial meetings taking place as early as mid 1989. Earlyon, ground rules were established that served as missionstatements guiding the efforts of all involved in thepublication of the standard. Those ground rules aresummarized as follows:

• The intent was to define such things as algorithmsand audit trail requirements for using electronic flowcomputers and associated equipment for custodytransfer.

• The intent was not to produce a comparative studybetween the accuracy of chart and electronic basedsystems.

• The intent was to address minimum requirementsbased on current technology, yet not circumvent theuse of future, more capable technology or morestringently specified systems.

• The scope included both differential and linearmeters. The scope originally included various typesof hydrocarbon fluids, but in the interest ofexpediency, it was decided to treat different fluidsindividually, starting with gas and moving to liquidslater. As of this writing, work is currently underwayon API 21.2 Electronic Liquid Measurement.

API 21.1 CONTENTS

API 21.1 consists of nine major subsections, three appen-dices and 10 figures. The subsection titles are:

1. Introduction and Scope2. Description of an Electronic Gas Measurement

System3. Referenced Publications4. Electronic Gas Measurement System Algorithms5. Data Availability6. Audit and Reporting Requirements7. Equipment Installation8. Equipment Calibration and Verification9. Security

The appendices are:

A. Rans Methodology for System Flow MeasurementAlgorithms

B. Averaging TechniquesC. Calibration and Verification Equipment

Subsection 1.1 – INTRODUCTION AND SCOPE

The major paragraph of the scope statement is as follows,“This standard describes the minimum specificationsfor electronic gas measurement systems used in themeasurement and recording of flow parameters ofgaseous phase hydrocarbon and other related fluids forproduction and transmission custody transferapplications utilizing industry-recognized primarymeasurement devices. For the purpose of this standard,electronic correctors of the type used on linear meterswere not considered to constitute an electronic gasmeasurement system.”


Some notable elements of the scope are,

• Minimum specifications• Gaseous phase hydrocarbon• Production and Transmission applications• Linear electronic correctors not considered

Subsection 1.2 — ELEMENTS OF A GASMEASUREMENT SYSTEM

Main system elements presented in subsection 1.1include:

• Primary Device is the basic meter run including theorifice plate, turbine, rotary or diaphragm meter.

• Secondary Devices are used for sensing suchprocess variables as static pressure, differentialpressure, temperature, and density.

• Tertiary Device is an electronic computer that isdesigned to correctly calculate flow and that canreceive information from the primary and/orsecondary devices.

Also introduced in subsection 1.2 are the concepts oftransducers, transmitters and signal processing. Inpractice, transducer and transmitter are often used assynonymous terms. Strictly speaking, this is not correctand the standard attempts to provide instruction in thisregard.

“Transducers respond to changes in the measuredparameters with a corresponding change in electricalvalues. These devices are referred to as transmitterswhen they have been specifically designed to convertthe transducer’s electrical output to a signal suitable fortransmission over distances greater than could otherwisebe achieved.”

Hybrid EGM systems currently exist which render thedefinitions of these traditional terms (secondary devices,tertiary devices, transducers and transmitters) inadequateas descriptors of the systems’ components. These hybridsystems have evolved to lower cost and decreaseuncertainties.

Several different configurations are currently in use. Forexample, some flow computers interface to standard1-5Vdc or 4-20ma external transmitters. Others interfacedirectly to transducer low level analog signals, while yetother flow computers connect directly to newer digitaltransducers using either serial or parallel techniques.Sometimes the transducers share electronics with theflow computer and both reside in the same enclosure.

Add to this the concept of smart transducers/transmittersand yet another set of scenarios exists. Smart trans-ducers/transmitters generally provide lowermeasurement uncertainties by digitally compensating fortemperature and pressure effects on the transducer’ssensing elements and associated components.

Usually, smart transducers/transmitters performcompensation algorithms internally and providecompensated values to the flow computer, but somesystem designs rely on the flow computer to performtransducer compensation algorithms.

Therefore, in addition to the traditionally defined terms,the 21.1 standard attempts to allow all theseconfigurations by including statements such as “Thetertiary and secondary devices, as well as the primarydevice, may be contained in one or more enclosures, orpackaged separately.” And in another statement, “Theelectronic flow computer has no effect on the accuracyof either the primary or the secondary device, exceptwhere characterization may be performed.”

Subsections 1.2.2 and 1.2.3 define terms and symbolsused throughout the document. Subsection 1.3references publications cited within the 21.1 standard.These sections are not reviewed in this paper.

Subsection 1.4 – ELECTRONIC GAS MEASUREMENTSYSTEM ALGORITHMS

Subsection 1.4 defines algorithms for both differentialand linear measurement systems. Only differentialsystems are presented in this report.

The standard defines component algorithms that, whenapplied as recommended, collectively result in acomposite algorithm suitable for computing a desiredquantity such as mass, energy or volume. Thesealgorithms are not intended to supplant already publishedwork. Instead the standard references other equationstandards when possible. Only generalized equationsgermane to the topic at hand were included in the 21.1standard.

The primary goal of subsection 1.4 is to define theminimum acceptable frequencies for solving equationsand measuring inputs to those equations.

1.4.2 Differential Meter Measurement

In differential metering, a total quantity (volume, forexample) is determined by integration of a rate equation(AGA3 / API 14.3, for example) over a specified timeinterval. The equation form of this integration operationis presented in the 21.1 standard as:

(eq 1)

where,

t0 ∫ t

= integration operation from time t0 to time t

Qt = quantity accumulated between time t0 and time t

qt = rate equation for quantity per unit of time

dt = delta time between integration samples

For illustration purposes, this report assumes the rateequation of eq 1 is based on AGA3/API 14.3. The 21.1

Qt = ∫ tqt * dt t0


standard allows this rate equation to be factored intotwo component algorithms that can be solved on twodifferent time domains. This is loosely related to the olderAGA3 rate equation that often was described by twocomponent algorithms as:

(eq 2)

where,

Qh = Flow Rate (cubic feet / hour)

C ’ = Fb * Fr * Y ..... (one component)

Extension = Hw * Pf (second component)

Since the new AGA3/API 14.3 standard no longer usesthe older [C prime * Extension] paradigm and since the21.1 standard’s algorithms were to generically discusstime domain issues rather than re-state equations in otherstandards, new terms were introduced to discuss thetime domain issues. Those terms are:

imp = Integral Multiplier Period, a unit of time forspecifying the frequency for performing acomplete volume calculation.

Qimp = Quantity (volume) accumulated for the integralmultiplier period.

IVimp = Integral Value, one of the two components ofthe flow rate equation (like the older Extension).

IMVimp = Integral Multiplier Value, the second componentof the flow rate equation (like the older C prime).IMVimp is the value resulting from the calculationof all other factors of the flow rate equationnot included in IVimp.

Using these new terms, the volume algorithm ispresented in the 21.1 standard as:

(eq 3)

Where IVimp is the portion of the flow rate equation thatmust be integrated on relatively fast time periods, andIMVimp is the portion of the flow rate equation that canbe computed on a slower time period, the IntegralMultiplier Period.

Four minimum acceptable criteria are established insubsection 1.4 of the 21.1 Standard. They are:

• The minimum sampling frequency for any dynamicinput variable shall be once every second. Thisincludes such variables as static pressure, differentialpressure and temperature.

• The IVimp component of the flow rate equation mustinclude at least the Square Root of Static Pressure *Differential Pressure.

• The IVimp component of the flow rate equation shallbe calculated and summation performed at leastonce per second. It is further recommended that thesampling frequency and the integral value calculationfrequency be performed at the same time interval.

• The Integral Multiplier Period, the period at whichIMVimp and Qimp are computed, shall not exceed onehour. An Integral Multiplier Period of less than onehour shall be such that an integral (whole) numberof multiplier periods occurs during one hour.

Putting all this together into a composite algorithm resultsin:

(eq 4)

where,

imp = Integral Multiplier Period of time period not toexceed one hour.

dt = integration period not to exceed one second

IVimp = hw * Pf , at a minimum

IMVimp = the remainder of the flow rate equation notincluded in IVimp.

Rans Methodology

For all timing requirements mentioned in section 1.4,slower periods can be used if they can be qualified asacceptable. The 21.1 standard provides a tool to helpconduct this qualification process. This tool is includedin appendix A and is called the Rans Methodology.

The Rans Methodology is a statistical evaluation method,developed by Mr. Rick Rans, to estimate the maximumamount of measurement uncertainty that exists for anygiven flow pattern across an orifice plate or linear typemeter.

First, The Rans Methodology evaluates the uncertaintyas a function of both flow pattern and calculationfrequency. Second an evaluation of the additionaluncertainty resulting from doing portions of the flow ratecalculation using averages is conducted.

A detailed analysis based on the Rans Methodology isbeyond the scope of this report.

Averages

Portions of the IMVimp equation, such as expansion factorand compressibility, are dependent on the dynamic inputvariables.

Since IMVimp can be computed on a time period slowerthan the dynamic input variables are sampled, thedynamic input variables must be averaged over the

Qh = C’ * Extension

Qimp = IMVimp * IVimp

Qimp = IMVimp * timp IVimp dt

t0


longer time period. The 21.1 standard allows four differenttypes of averages.

These averages are described algorithmically in appendixB of the standard. They are:

• Flow-dependent time-weighted linear• Flow-dependent time-weighted formulaic• Flow-weighted linear• Flow-weighted formulaic

Flow dependent means that each sample is used toupdate the average’s accumulator only during times offlow. Therefore, each sample’s contribution to theaverage is turned on and off as a function of the presenceof flow resulting in an average during times of flow.

Flow weighted means that each sample in the average’saccumulator is multiplied by the flow rate at the time thesample is taken. Therefore, each sample’s contributionto the average is weighted by the flow rate. This is aninteresting concept, since the flow rate isn’t really knownuntil some time later (the integral multiplier period).Therefore, the flow rate used in this case is someapproximation of the actual flow rate.

Linear means that each sample in the average’saccumulator is simply applied in engineering units. (e.g.,PSI for pressure, Degrees F for temperature, in. H2O fordifferential pressure).

Formulaic means that each sample in the average’saccumulator is applied as a function of the form of theprimary element’s fundamental equation. Thus, fordifferential meters the square root of each sample is takenbefore addition takes place. For linear meters, eachsample is simply added in engineering units.

Low Flow Detection

As indicated by the term flow dependent above, decidingwhen there is and when there is not flow is somethingthe EGM is expected to do.

If everything was perfect and stayed that way, therewould be no need for special low flow detection logic. Adifferential pressure of zero would mean no flow and thatwould be the end of it.

The reality is that things, such as transducer calibration,drift and therefore a need to define an artificial zero (lowflow cutoff point) exists.

Therefore, the 21.1 standard mentions this requirementin subsection 1.4.2.3 Low Flow Detection by simplystating, “A low flow cutoff point for differential metersshould be determined by the contractually concernedparties based upon realistic assessment of site con-ditions.”

1.5 DATA AVAILABILITY

Subsection 1.5 defines data availability requirements forboth differential and linear measurement systems. Onlydifferential systems are presented in this report.

Although many people think the 21.1 standard appliesonly to electronic, remote, battery operated devices thatcompute rates and quantities, the 21.1 standard’sauthors did not intend to limit the scope that much. Theauthors intended to write a standard that applied tosystems of various acceptable configurations and, assuch, they intended to allow rates and quantities to becomputed in office systems too. They also allowed forportions of the audit trail to be constructed of eitherelectronic or hard copy records. This subsection and thenext (1.6 Audit and Reporting Requirements) reflectthis intention more than anywhere else in the standard.These two subsections (1.5 and 1.6) are closely relatedand frequently restate the same requirements fromdifferent perspectives and in different ways.

1.5.1.1 Differential Meter On-site Calculations

This subsection describes information that must beavailable on-site, or be collectable on-site with a portabledata collection device. For the most part, subsection 1.5is a succinct itemization of required data items, most ofwhich are as follows:

• 1.5.1.1.1, Historical data spanning the time since thelast completed data collection period including, butnot limited to the following:

• At least hourly average values for temperature,pressure and differential pressure. Also, relativedensity, energy content, composition anddensity, if they are live inputs.

• At least hourly quantity totals.

• Dates and times for all averages and totals.

• Total quantity accumulated during each contrac-tually specified measurement period.

• 1.5.1.1.2, Input variable values affecting measure-ment such as meter run reference diameter (Dr),orifice bore reference diameter (dr), the calibratedspan of the pressure, differential pressure andtemperature transducers.

• 1.5.1.1.3, Instantaneous readings or displays for thelive values of pressure, differential pressure,temperature, flow rate, accumulated quantity andalarm conditions. Other live inputs (such as density)shall be available if used.

• 1.5.1.1.4, An electronic or hard copy recordincluding, but not limited to:


• “As found” and “As left” equipment calibrationvalues.

• Old and new values for changes to any inputvariable that will affect the calculated quantities.

• A summary of alarm or error conditions affectingmeasurement.

• A daily summary indicating the hours or per-centage of time for flow or no flow.

• The date and time of all events in the record shallbe identified chronologically.

• 1.5.1.1.5, A quantity statement including, but notlimited to, daily custody transfer totals and averagepressure, differential pressure and temperature andother live inputs (such as density) if they are used.

• 1.5.1.1.6, The unique identification number of themetering system.

• 1.5.1.1.7, All original data, plus all subsequent localedits to that data.

1.5.1.1 Differential Meter Off-site Calculations

• Data required on-site includes, but is not limited to:

• 1.5.1.2.1, Instantaneous readings or displays for thelive values of pressure, differential pressure, tem-perature, flow rate, accumulated quantity and alarmconditions. Other live inputs (such as density) shallbe available if used.

• 1.5.1.2.2, The unique identification number of themetering system.

• 1.5.1.2.3, Data required off-site includes, but is notlimited to the same data as is required in 1.5.1.1 (on-site calculations) with the exception of 1.5.1.1.3.

• 1.5.1.2.4, Indications of alarm or error conditionsshall be available off-site.

1.6 AUDIT AND REPORTING REQUIREMENTS

Subsection 1.6 defines audit and reporting requirementsfor both differential and linear measurement systems.Only differential systems are presented in this report.

Some key elements of the introduction to subsection1.6 include:

• The audit trail shall include, but is not limited to,transaction records, configuration logs, event logscorrected transaction records and field test reports.

• The records and reports in this section may becreated on-site, off-site or a combination of both.

• The primary reason for retaining historical data is toprovide support for the current and prior quantitiesreported.

• The data specified in this subsection will providesufficient information to apply reasonable adjust-ments when the electronic gas measurement systemhas stopped functioning; is determined to be out ofaccuracy guidelines; or measurement parameters areincorrectly recorded.

Within the introduction of subsection 1.6 some newterms, such as quantity transaction record and event logare used. Most of the succeeding paragraphs insubsection 1.6 are devoted to defining these terms andthe minimum acceptable data that composes the datarecords to which this term refers.

1.6.2 Quantity Transaction Record

“The quantity transaction record is the set of historicaldata and information supporting the quantity orquantities of volume, mass, or energy. The quantitytransaction record is to be identified by a uniquealphanumeric identifier denoting a specific electronicmetering device and primary device.”

1.6.2.1 Daily Quantity Transaction Record forDifferential Meters

This subsection itemizes the data elements that, atminimum, must be included in the daily record. They are:

• date period• time• quantity• flow time• differential pressure average• flow temperature average• static pressure average• relative density average (if live)

Although flow integral is not required, it is stated that, incertain situations, it can provide valuable information.

When do Day’s begin and end?

The daily quantity transaction record is the average orsummation of data collected during a contract day. Onedaily quantity transaction record ends and a new onebegins:

• Once, at the end of each contract day• Any time one or more constant flow parameters are

changed

1.6.2.3 Hourly Quantity Transaction Record forDifferential Meters

This section itemizes the data elements that, at minimum,


must be included in the hourly record. They are:• date period• time• quantity• differential pressure average• flow temperature average• static pressure average• relative density average (if live)

Although flow integral is not required, it is stated that, incertain situations, it can provide valuable information.

When do hours begin and end?

The hourly record is the average or summation of datacollected and calculated during a maximum of 60consecutive minutes. One hourly quantity transactionrecord ends and a new one begins,

• Once, at the end of each hour• Any time one or more constant flow parameters are

changed

There are 24 hourly quantity transaction records for eachcontract day plus additional quantity transaction recordsfor each time one or more constant parameters arechanged.

1.6.3 Algorithm Identification

An algorithm identifier is to be provided to identify thecalculations performed by the system. This can beprovided in various ways, such as a software ormanufacturer’s version code.

1.6.4 Configuration Log

This log is to contain and identify all constant flowparameters used by the system to generate quantitytransaction records. The 21.1 standard provides a tableof elements that must be in the configuration log. Asalways, this is a minimum acceptable list that can beexpanded on.

For differential meters, this table contains such thingsas,

• Meter Identifier• Date and Time• Contract Hour• Atmospheric Pressure (if appropriate)• Pressure Base• Temperature Base• Meter Tube Reference Inside Diameter• Orifice Plate Reference Bore Diameter• Etc.

1.6.5 Event Log

Each time a constant flow parameter (see 1.6.4 above)that can affect the quantity transaction record is changed,the old and new value, along with the date and time ofthe change, shall be logged.

The date and time of all events in the log shall beidentified chronologically.

1.6.6 Corrected Quantity Transaction Record

With this subsection, the 21.1 standard recognizes thatadjustments to original quantity transaction records aresometimes necessary, and specifies the data to be keptwhen this occurs. The need to make these changesresults from:

• Constant flow parameters were not available at thetime of calculation

• Constant flow parameters were found to be in errorat a later date

• Dynamic flow parameters were found to be in errorat a later date, usually due to• calibration error• transducer failure• adverse operating conditions

The corrected quantity transaction record reflectschanges to the original constant and/or dynamic flowparameters used in the calculation of the final quantitytransaction record. The purpose of the record is to:

• identify reasons for all corrections

• provide the original and corrected constant anddynamic parameters used

• to clarify the adjusted quantities to be applied to themeter and quantity accounting statements

The original quantity transaction record is to remain intactas a permanent record. This original record, in combi-nation with the most recent corrected quantity transac-tion record, provides a detailed tracking of the custodytransfer quantities.

The next two subsections pertain, for the most part, toinstallation, verification and calibration of the secondarydevices. Since many people are already familiar withthese devices, a detailed description is not presentedhere. In lieu of this, an outline of the next two subsectionsof the 21.1 follows,

1.7 EQUIPMENT INSTALLATION

1.7.1 Transducer/Transmitters1.7.2 Gauge/Impulse Lines1.7.3 EGM Devices and Assoc. Communications1.7.4 Peripherals1.7.5 Cabling1.7.6 Commissioning

1.8 EQUIPMENT CALIBRATION AND VERIFICATION


1.8.1 Scope1.8.2 Devices Requiring Calibration/Verification1.8.3 Calibration and Verification Procedures

1.8.3.1 Pressure and Temperature Devices1.8.3.2 Pulse Counters1.8.3.3 Analyzers1.8.3.4 Densitometers and Gravitometers

1.8.4 Frequency of Verification1.8.5 Ambient Temperature and Pressure Effects1.8.6 Calibration and Verification Equipment

1.9 SECURITY

1.9.1 Access — This subsection restricts access to themetering system to the owner or the owner’s contractuallydesignated representative for the purpose of calibratingor altering the function of the metering system.

1.9.2 Restricting Access — States that the systemshould deny unauthorized access for the purpose ofaltering any input variables that may affect measurement.A unique security code of at least four characters is tobe provided in support of this requirement.

Instead of a security code, other measures, includingmechanical, may be used to restrict access.

A security code may also be used to grant access forthe purpose of collecting data.

1.9.3 Integrity of Logged Data — Restates the require-ment for an event log and calibration reports.

1.9.4 Algorithm Protection — Changing the algorithmsused to calculate quantities is to be protected even morerigidly than the restriction method described in paragraph1.9.2. Field operations and accounting office personnelare not to be authorized to change these algorithms.

1.9.5 Original Data — Simply states that there shall beno changes to the original data.

1.9.6 Memory Protection — This subsection requires abackup power supply, or nonvolatile memory, capableof retaining all data in the unit’s memory for a period notless than the normal data collection interval for the unit.

1.9.7 Error Checking — Simply states that an effectivesystem of error checking shall be utilized each time datais transferred from one data storage device to anotherand that detected errors shall prevent the use of incorrectdata.

CONCLUSION

Being the first of its kind, the 21.1 standard probablyhas room for improvement. However, considering the

FIGURE 1.0 - Elements Of A Gas Measurement System


issues being dealt with, it is a very good initial effort. Itshould provide valuable guidelines for evaluatingElectronic Gas Metering Systems for Custody Transfer.

Although not a complete treatment of the 21.1 standard,it is hoped this report will serve as an adequateintroduction to it. Time and space did not allow linearmeter presentation within this report. However, the linearmeter subsections within the 21.1 standard are presentedwith the same level of detail as are the differential metersubsections. Perhaps a sequel to this report, includingthe linear meter presentation will be forthcoming

If you are interested in receiving a copy the API 21.1standard it can be ordered from the American PetroleumInstitute. When ordering, refer to Chapter 21.1,Electronic Gas Measurement, First Edition, August1993, Order No. 852-30730.

FIGURE 2.0 Prominent Data Flows within an EGM System

Brent E. Berry


TRANSIENT LIGHTNING PROTECTION FOR ELECTRONICMEASUREMENT DEVICES

Patrick S. McCurdyPresented by: Dick McAdams

Phoenix Contact Inc.P.O. Box 4100, Harrisburg, PA 17111-0100

INTRODUCTION

Technology advances in the world of semiconductorsand microprocessors are increasing at a breathtakingpace. The density of transistor population on integratedcircuits has increased at a rate unimaginable just a fewyears ago. The advantages are many: faster dataacquisition, real time control, and fully automatedfactories, to name a few.

Semiconductor technology is also prevalent in fieldmounted instrumentation and electronic measurementdevices. Unfortunately, a tradeoff to the increasedperformance is the susceptibility of these semiconductordevices to voltage and current transient events. Theminimum results are unreliable instrumentation readingsand operation, with periodic failures. The worst caseresult is a completely destroyed measurement device.

Such power surges are often the work of mother nature.Lightning, which according to the National WeatherService strikes some 40 million times annually in the U.S.,is a leading cause of failure in electronic measurementdevices. When these devices are field mounted thevulnerability greatly increases due to their remote locationand outdoor installation.

It should be noted that although the most devastatingsource of transient voltage and current activity islightning, there are other sources. Some of these includestatic buildup, human error, inductive load switching, andutility capacitor switching.

This paper explores lightning effects on electronicmeasurement devices, and methodologies for preventingdamage including lightning arrester technology,shielding, grounding and surge protection devices (SPD).The discussion will also cover various coupling methodsfor transients, and national and international standardsthat can help in evaluation and application of the propersurge protection network.

LIGHTNING MAGNITUDE AND FREQUENCY

Most of the continental United States experiences at leasttwo cloud to ground flashes per square kilometer peryear. About one half of the United States will see threecloud to ground events per square kilometer per year.This is equivalent to about 10 discharges per square mileper year. An Isokeraunic map of thunderstorm daysdeveloped by ANSI/NFPA 780-1992 is shown in Figure 1.This shows the average number of thunderstorm days

per year by geographic region. A relationship can bemade between thunderstorm days and flash density (i.e.,the number of lightning-to-ground flashes in a specifiedperiod over a certain area, expressed in flashes/km2

[Table 1]).

As can be seen, depending on the region, locations suchas Tampa, Florida or Columbia, South Carolina can haveover 70 thunderstorm days per year. This corresponds toabout 25 flashes/km2/year. Although this is on the highside, roughly one half of the continental U.S. experiencesup to three cloud-to-ground events/km2/year, or equivalentof about ten discharges per square mile per year.

Lightning damage to electronic equipment might bethought of as a remote possibility. Perhaps somethingthat is as unlikely as winning the lottery. In fact thechances of winning the lottery in Florida are 1 in 14million; the chances of getting hit by lightning are 1 in600,000.

FIGURE 1.Isokeraunic Map Showing Mean Number of

Thunderstorm Days per Year.

Number of Thunderstorm Flash Density per Square Days per Year km per Year

10 1

25 4

40 10

80 30

100 50

TABLE 1.Relationship between thunderstorm days per year andnumber of lightning-to-ground strokes in a specified

period and geographic area.


COUPLING METHODS

Lightning effects can cause damage to sensitiveelectronic equipment in several ways, including: directcoupling, inductive coupling, and capacitive coupling.Direct coupling (Figure 2) is associated with lightningenergies on the incoming conductors. Ground pathsbecome saturated, and the transient energy seeks otherdirect paths to ground. Unfortunately, this is usuallythrough unprotected electronic equipment such asmeasurement instrumentation.

An analogy can be made between direct coupling oflightning transients and flooding that occurs because ofoverfilled storm drains. Just as ground circuits becomesaturated with electrons in the case of a lightning strike,storm drains become saturated with fast flowing waterin torrential downpours. The results are also similar inthat the water seeks another path to ground which couldbe across a busy highway or through someone’sbackyard.

Inductive coupling (Figure 3) is produced by the magneticflux lines generated during a lightning strike. As muchas 70V/meter of cable can be induced by lightning strikesthat are more than a three-dimensional mile away. In otherwords, lightning does not have to strike the ground toinductively couple transients. Electronic equipment andmeasurement devices can experience inductive couplingfrom a cloud to cloud lightning event.

Capacitive coupling (Figure 4) is derived from positiveor negative charge ions passing over conductors.Shielding is one method used to bleed off these chargesto negate any capacitive coupling effects. However, tominimize ground-loop problems caused by potentialdifferences along the cable shielding, it’s best to groundthe shielding on instrumentation signals only at one end.

FIGURE 3.Inductive coupling is produced by the magnetic

flux lines generated during a lightning strike.

FIGURE 2.Direct coupling results from lightning energies

on the incoming lines.

FIGURE 4.Capacitive coupling is derived from positive andnegative charge ions passing over conductors.

STANDARDS AND WAVEFORMS

In the days before microprocessor devices, engineersdidn’t have to worry about protecting their equipmentfrom voltage and current transients. Since theproliferation of microprocessor based equipment, theneed for protection against transients has becomenecessary. To define the areas of hazard, characteristicsof transients, and protection product performanceseveral national (ANSI/IEEE) and international (IEC)standards were created.

The ANSI/IEEE C62.41 is used primarily in North Americaas a guide for selection of surge protection devices (SPD).Table #2 shows the different equipment locations, testwaveforms, open circuit voltage, and short circuit currentspecifications as defined by ANSI/IEEE C62.41. Factoryfloor measurement devices would be covered underCategory B and field mounted measurement devicesunder Category C.

When the rate of current flow is greater thanthe earth’s ability to absorb the current flow,an alternate path is found.

Surge currentseeks an alternate

path Equipmen tDamage

High

Rate of

Current Flow

Rate of

Current FlowI I1 2

• Magnetic flux line’s (MFL) intensities are greatest near stroke.

• Greater (MFL) frequencies create higher voltages.

• As (MFL) crosses circuit area, damaging voltages are generated.

Surge currentseeks an alternate

path EquipmentDamage

High

Rate of

Current Flow

Rate of

Current FlowI I1 2

• A lightning stroke creates strong local voltage fields.

• Highly charged electrons are attracted or repelled to electricalcircuits.

• High inrush or outrush currents are formed which can damageelectrical circuits.

EquipmentDamage

High

Rate of

Current Flow

Rate of

Current FlowI I1 2


Location/ Waveform Open Short Category Circuit Circuit

Voltage Current

Category A: 0.5µs - 100kHz 6,000V 200A Long Branch Ring Waveform Circuits and Test Outlets, Data Wall Outlets.

Category B: 0.5µs - 100kHz 6,000V 500A Major Feeders Ring Waveform and Short Test Branch Circuits (Distribution 1.2 x 50µs High 6,000V n/a Panels). Impedance Test

8 x 20µs Current n/a 3,000AImpulse Test

Category C: 8 x 20µs Current n/a 10,000A Outdoor Impulse Test Overhead Lines, Service Entrance.

TABLE 2.ANSI/IEEE C62.41 Transient Environment Test Criteria

The International Electrotechnical Commission (IEC)specification IEC 1312 defines several specificationsrelated to lightning susceptibility. The categorizations,test waveforms, voltage and current peaks are similar tothe ANSI/IEEE C62.41. However, one uniqueclassification of the IEC 1312 is the definition of a truelightning test waveform, the IEC 1024 10 x 350µS (Figure#5) waveform. The IEC 10 x 350µS waveform describestransient characteristics which more closely representthe amount of energy associated with a real field lightningevent. The IEC 10 x 350µS waveform describes atransient current event with a 10µS rise time to 60kAand a 350µS decay to half energy or 30kA. This isapproximately 200 times greater energy than the ANSI/IEEE 8 x 20µS waveform, Figure 5). The IEC 10 x 350µSallows a true lightning survivability test to be performedon a surge protection device. This will be discussedfurther in the following pages.

GROUNDING

One of the key factors to protect any sensitive industrialequipment is a low impedance earth ground. TheNational Electrical Code (NEC) targets 25 Ohms.However, 25 Ohms is a target. A ground resistance of 5Ohms or less is preferred. Since the performance of anysurge arrester is enhanced by a highly conductive pathto earth ground, lightning currents are diverted fromseeking other sensitive industrial equipment paths. Theimpedance of the earth ground consists of both aresistive and a reactive component of the groundingconductor. The formula V-IR + L(di/dt) can be used toshow the voltage potential that a SPD must deal withwhen shunting transient energy to ground. The voltagedrop is a function of the resistance (R) and the inductance(L). Skin effect and the path taken by the conductor(straight, bends, etc.) have a direct effect on the groundpath inductance. A ground path of both low resistance

and low inductance ensures the maximum performanceof a SPD.

FIGURE 5.The IEC 1024 10 x 350µS waveform shown in comparison

to the 8 x 20µS waveform. The IEC 10 x 350µSallows a true lightning survivability test to be

performed on a SPD.

PROTECTION STRATEGIES

Surge protective device (SPD) is the common term usedto describe products that protect sensitive electronicequipment from damaging transient voltages. A surgeprotective device acts as a high impedance connectionbetween power or signal line and ground, under normaloperating conditions. Upon sensing a fast rising voltage,the SPD becomes a low impedance “short circuit” toground to divert the unwanted energy safely away fromthe sensitive electronic equipment. Since the transientcondition occurs for a maximum time duration ofapproximately 20 microseconds, this short lowimpedance state occurs very quickly. After diverting theenergy to ground the SPD then “resets” itself back tothe normal high impedance state.

Industrial equipment, control systems, and electronicmeasurement devices can be damaged from lightninginduced transients entering the control system throughthe power source. Also common is damage caused fromsuch “back door” entrances as data communicationlines, antenna connections, analog and digital I/O andphone modems. A typical control cabinet could containa variety of instrumentation systems. A commonapplication would consist of a programmable logiccontroller (PLC) with standard 24V dc relay logic, 4-20mAfield transmitter inputs and a radio modem for sendingcompressed data packets back to a central controlstation. The distributed I/O system utilizes an RS-422full duplex data protocol. By evaluating this standardsystem shown in figure #6 we find that there is a potentialfor transient damage at any of the points from the dataline to the power line.

µ

µµ


FIGURE 6.To be completely protected a typical remote terminal

unit (RTU) requires protection for all copperpathways leading into the device.

The ANSI/IEEE C62.41 specifications shown in table 2are a good guideline for selecting SPDs for protectingagainst the damage of indirect lightning effects. Thecomponents used by most SPD manufacturers includethe independent use or a cascade use of Metal OxideVaristors (MOVs), Silicon Avalanche Diodes (SADs) andGas Tube technology. In our typical field mounted controlsystem example indirect lightning effects are commonon the 24V dc discrete I/O, the 4-20mA analog signals,the RS-422 data communication and the telephonemodem applications.

For protecting the power system the effects of a directlightning strike need to be considered. AC power linesare directly linked to the outside world. The use of largergauge wire in power conductors allows these conductorsto carry higher direct lightning surges. Because of theseanticipated long duration lightning surges, a test standardusing the IEC 1024 10 x 350µS waveform should be usedwhen selecting the protection device. A “lightningarrester” utilizing an ARC Chopping Principle (Figure 7)can safely divert this energy from AC power lines. Alightning arrester specified at 60kA based on the 10 x350µS waveform will handle energy that would easilydestroy one rated at 100kA to 150kA based on the ANSI/IEEE 8 x 20µS. A comparison of energy handling,response time, and clamping voltage for differentprotection components is shown in Table 4.

NETWORKED APPROACH

To safely protect sensitive industrial control systems fromthe direct and indirect effects of lightning a networkedapproach is necessary. Unfortunately, you cannot simplyinstall one product and protect everything. Theinternationally recommended practice is to build anetworked surge suppression system. This includes alightning arrester technology on the AC power lines, andSPDs on the analog and digital I/O, communication lines,and data lines.

AC POWER LINES — THREE STEPS

Step One — Main Service Entrance. Facility AC powerlines are the direct link to the outside world. Since thewire gauge of power conductors is large, so is thepathway for lightning induced surges. The IEC 1312guidelines suggest that surge currents of up to 60,000amps based on the IEC 1024 10 x 350µS waveform(Figure 5) can be expected at the facility service entrance.To take this initial hit a surge protection device utilizingthe arc chopping principle must be employed. The ARCChopping device is a new technology spark gap. Theproblem with traditional spark gap technology has beenthe susceptibility to follow-on currents which can causenuisance tripping of upstream circuit breakers or fusing.The spark gap devices that feature the patented ARCChopping Principle avoid this by their unique follow-oncurrent extinguishing capabilities. The technology shownin Figure 7 consists of two electrodes positioned oppositeeach other, held in place by a barrier and separated by abaffle. This arrangement and spacing of electrodes iscalled “ARC Chopping” and provides reliable ignition ofthe arc, which is then chopped by the baffle into severalsmaller arcs. This effect diverts the lightning current whileself extinguishing the follow-on current before up streamcircuit protection devices have time to trip. A protectionsystem featuring this technology can be chosen in threephase enclosure systems with UL listing.

Pow er sup ply120 or 220 a c

D a tacom /te leco m p ort,RS-232, RS-422 , RS-485

D ig ita l I/O24 V dc

A na log I/O0-10 V & 4-20 m A

A ntennacoa x

µ

µ

FIGURE 7.The ARC Chopping Principle. A lightning arrester

technology which will handle a direct lightning strikecondition (IEC 10 x 350µS) on the AC main serviceentrance while minimizing follow through current.

FLASHTRAB Metal Gas Suppressor Spark Gap Oxide Discharge Diode

Varistor Tube 60kA 40kA 10kA 0.5kA 10 x 350µS 8 x 20µS 8 x 20µS 8 x 20µS 1,000V 5V - 600V 90V 5V-300V rise time nanosec. microsec. picosec. dependent

TABLE 4.Performance comparison of different SPD technologies.


Step Two — Distribution and Subdistribution Circuits.The lightning arrester just described will let through aresidual transient of approximately 5kA which takes theform of the ANSI/IEEE 8 x 20µS waveform. The secondstep for protection of the AC power system will addressthis transient energy. A metal oxide varistor (MOV) basedproduct with the capability to handle at least 40kA basedon the 8 x 20µS waveform should then be installed atyour critical power distribution circuits. Systems areavailable for all popular three phase and single phasepower configurations. It is important to select a productthat features pluggable and testable protection elementswith failure indication. This is important because of theinherent characteristics of MOVs to fail over time. Eachtime a surge is discharged, the metal oxide crystalsweaken and eventually form a direct short to ground. Therate at which MOVs fail is totally dependent on the numberand magnitude of transients. If a MOV fails at this stage,the direct short to ground poses several problems. Theleast critical of which would be a circuit interrupt from anupstream breaker or fuse. More critical is a potential firehazard. A MOV can actually leak enough current on anAC power line to heat up to the point of catching fire.

Step Three — Equipment Power Protection. Thetransient energy left after the step two protection shouldbe too small to cause destructive damage to the powersource of your sensitive equipment. However, it will stillcause dissipative damage over time. Dissipative damageoccurs when semiconductor materials are subjected tocontinuous small energy transient voltage. Thesemiconductor junctions are actually pitted away overtime until pathways open and the device eventually fails.Transient protection at this stage should feature a hybridprotection circuit with MOV, gas tube, and surge arrestingdiode technology. The three stage hybrid circuit givessufficient coarse protection with the gas tube and MOV,while providing fine protection with the surge arrestingdiode. Packaging is available in a convenient DIN-railmounting format and with the option of removableprotection elements. Also, features such as LED orremote warning contacts should be considered forindication of MOV failure.

ANALOG AND DIGITAL I/O PROTECTION OFCONTROLLERS AND INSTRUMENTATION

A typical programmable controller or industrial computerhas many areas of surge threat. We have just coveredthe power supply protection. However, there are several“back door” areas that also need to be considered. Signallines typically run between a control device and a fieldmounted measurement and / or transmitter device. Tohave total coverage a surge protection device (SPD)should be installed at both ends of the I/O.

A digital I/O circuit can utilize MOVs in the circuit from lineto ground. However, an analog circuit, which is a floatingground circuit, cannot be referenced to ground. As anMOV ages, it leaks to earth causing a ground referencethat can damage or cause signal error. Since most digitalor power supply I/O circuits are referenced to earth ground,

there is no problem with designing protection circuits withMOVs in common mode, line to ground. The most effectivedigital or analog protection circuits utilize a three-stageprotection scheme with MOV, suppresser diode and gasdischarge tube. This hybrid circuit is in series with theapplication. Packaging is available in DIN-rail mountingfor protecting the control side and also in a field mountedpipe nipple for protecting the field side.

DATA AND TELECOMMUNICATION CIRCUITS

Probably the most sensitive (and almost never protected)circuits are the data and telecommunication ports ofcontrol systems and electronic measurement devices.Any time a data or telecommunication line is routedthrough a facility or outdoors, it is susceptible to transientenergies that can damage network cards and spreadthe damage through the entire control system.

To protect these data systems again requires protectionat both ends of the transmission path. Twisted wire pairsand coaxial protection can be added to divert damagingtransients from getting into the systems. These systemsare usually high speed and cannot withstand high levelsof line impedance. MOVs should never be designed intotelecom or data circuits. They will lower and mis-shapedata signals that travel on cables or phone lines. In manycases the signal will just stop transmitting if the wrongtype of surge protection device is installed. Every datanetwork is different. Specific surge arresters are designedto function properly with networks like ethernet and tokenring. In industrial applications, networks that utilize RS-485 serial communications can be connected togetherby D-Sub style transient protection devices or hard wireddata protection devices. To protect telecommunicationlines, a multiple-pair protection system of single outletprotection devices can be utilized.

CONCLUSION

To properly protect industrial equipment, electronicmeasurement devices and control systems from thedirect and indirect effects of lightning a networkedapproach of lightning arrester and surge protective devicetechnologies is required. Coupling methods, grounding,and surge protection device performance need to beconsidered. By addressing both the lightning energiesand the I/O data structure, you can engineer a highlyreliable protection network tosafely protect your entire controlsystems from even a directlightning strike.

Patrick S. McCurdy


AN OVERVIEW AND UPDATE OF AGA 9Charles W. Derr

Presented by John LansingDaniel Division of Emerson Process Management

9720 Old Katy Road, Houston, TX 77055

ABSTRACT

The American Gas Association has published (June,1998) a recommended practice; Report No. 9;Measurement of Gas by Multipath Ultrasonic Meters.,This paper reviews some of the key contents of A.G.A.-9including recommended meter performancerequirements, design features, testing procedures, andinstallation criteria. An update for the committee work Inprogress for year 2001 and beyond is integrally included.The paper addresses some of the most commonly askedquestions by new users of the document.

A.G.A.-9 was drafted by the A.G.A. TransmissionMeasurement Committee (TMC) which incorporated manyof the recommendations in the GERG TechnicalMonograph 8 (1995) and certain related OIMLrecommendations. After two years of technical discussions,balloting, and revisions, the document represents theconsensus of several dozen metering experts in the USand Canada. They represent a cross-section of seniormeasurement personnel in the natural gas industry. TheISO/TC 30 standard currently being written for ultrasonicmeters has mostly adapted A.G.A.-9 information and isadding various additional pieces of operating practiceinformation, precautions and recommendations. A muchlarger data base now exists for performance andcalibration history of USMs, including additional test datafor piping, flow conditioners and valve interaction.

SELECTIVE REVIEW OF A.G.A. REPORT

Scope of Report

A.G.A.-9 was developed for multipath ultrasonic transit-time flow meters, used for the measurement of naturalgas. A multipath meter is defined as one with at leasttwo independent acoustic paths used to measure transittime difference of sound traveling upstream anddownstream.

Meter Requirements: Codes and Regulations

A.G.A.-9 makes statements which are design instructionsfor manufacturers of meters. One reason this approachwas taken was to insure the end user that an ultrasonicproduct would be safe and consistently manufactured.Unless otherwise stated, the meters are to be suitablefor use in an area which is subject to the requirements ofthe U.S. Department of Transportation’s regulations in49 C.F.R. Part 192, (Transportation of Natural and OtherGas by Pipeline: Minimum Federal Safety Standards).

Meter Requirements: Meter Body

Manufacturers are urged to publish the overall lengthsof their ultrasonic meter bodies. This is to help skid andother designers who may not be familiar with ultrasonicmetering to define metering section dimensions. Themeter length itself is Not considered part of the approachor discharge meter tube minimum length requirements.The inside diameter of the ultrasonic meter should bewithin 1% of the upstream tube’s diameter.

The value of 1% was based mainly on early Europeanstudies and also work performed at the SouthwestResearch Institute’s GRI/MRF (Gas Research Institute/Metering Research Facility) in San Antonio, Texas.

Other meter requirements in this document include anti-roll devices (feet), pressure tap location on the meter,and standard meter markings.

These requirements were based on field experience andthe lessons learned from other metering technologies.

Meter Requirements: Ultrasonic Transducers

Ultrasonic transducers are not common pipeline devicesand many operators are unfamiliar with their properties.A.G.A.-9 includes clear directions to the manufacturerfor the specification, marking, and testing of transducerpairs. These instructions are valuable because they willalert users as to the pertinent information that may affectthe performance of the meter. A.G.A.-9 also requires thattransducers be manufactured so that they may beexchanged and requires instructions for the exchangeprocess.

Meter Requirements: Electronics

Much discussion was given to the issue of electronicsand its evolution with time. The goal of the committeewas to require electronics which were well tested anddocumented, but to allow improvements without placinga larger than necessary burden on the manufacturer. Thisidea is evident throughout the document but is especiallyrelevant in the electronics and firmware sections.

The electronics output section includes two suggestedtypes, serial and frequency, along with a list of others.Serial communication is suggested because theultrasonic meter is clearly a very “smart” instrument andmuch of its usefulness relies on the internal informationcontained in the meter. The frequency output is a


convenient option, especially in locations which areconfigured for turbine meter inputs.

A.G.A.-9 also mentions analog outputs, directionindicator, a low-flow cutoff, and volume accumulators.

Meter Requirements: Computer Programs

Since ultrasonic meters are electronic, the computerprograms and information contained in the electronicsof the meter are extremely important. A.G.A.-9 requiresthat it be possible to interrogate the meter and determineits calibration parameters. It also requires that the meterbe securable, so that accidental or undetectable changescan be prevented.

Alarms and diagnostic functions are clearly addressedunder the computer programs heading. These sectionswere difficult to compose because of the subtledifferences associated with every different pathconfiguration imaginable. The data that is required is ofthree main types; velocity, gas speed-of-sound andelectronic failure. The velocity data is to indicate flowprofile irregularities or velocity range exceptions and tocalculate volume rate from average velocity times area.The speed-of-sound data is to be used as a diagnostictool to check for erroneous transit time measurementerrors. Other information is required to judge the qualityof the data such as “% of accepted pulses.”

Performance Requirements

The heart of A.G.A.-9 is contained in the PerformanceRequirements section. A.G.A.-9 separates ultrasonicmeters into two categories; smaller than 12” and meterswhich are 12” and larger. The division was created toallow looser performance requirements for smallermeters where tolerances are more difficult to maintain.The flow regime is also divided into regions. Essentiallythere are two regions, one low flow region and one highflow region. The flowrate dividing them is called thetransition flowrate (Qt). Manufacturers are to provide thenumerical values for minimum, maximum, and transitionflowrates. There is a requirement that the maximum valuebe at least ten times greater than the transition flowrate.

The maximum error allowable for an ultrasonic flow meteris ±0.7% for large meters and ±1.0% for small meters.This error expands to ±1.4% below the transitionflowrate. Within the error bands, the error curve for anyindividual meter may not span more than 0.7%, or onehalf the height of the error bounds for large meters. Thisis the linearity specification written in terms of the errorcurve. The repeatability of the meters must be ±0.2%for the higher velocity range and is doubled for the lower.These limits specify the performance of the meter priorto the application of any flow calibration adjustment, orin other words are a dry calibration requirement. A.G.A.-9was written in this fashion to provide a very clear andlogical picture in which to view a meter’s performancebased on readily available data, the error curve. The drycalibration requirement itself was deemed necessary to

discourage any haphazard construction of meters withthe intention of “correcting” them in the final stagesthrough flow calibration.

Individual Meter Testing Requirements

Individual meters are to be tested to strict tolerances forleaks and imperfections. A.G.A.-9 also specifies a Zero-flow Verification Test and a Flow-Calibration Testprocedure (although a flow-calibration is not required).These requirements were written mainly for consistency.After flow calibration, the user is given any number ofoptions for adjustment (within the dry calibration limitsdescribed above), however the flow-weighted mean errormethod is suggested. More sophisticated linearizationtechniques are also allowed.

Installation Requirements

A.G.A.-9 was written from the perspective of experiencedgas measurement experts however each person freelyadmitted that they were still learning. This is evident andfactually stated in the sections on installation requirements.Rather than specify numerical values for up- and down-stream pipe diameters, A.G.A.-9 requires test-supportedrecommendations from manufacturers of ultrasonic flowfrom an installation effect (both with and without a flowconditioner) or the meters. These recommendations cantake one of two forms. The manufacturer may define anup- and down-stream meter configuration which will notbe biased by more than 0.3% manufacturer may specifythe flow profile deviation which will not bias the output ofthe meter by more than 0.3%. The user is also cautionedthat protrusions, internal surface condition, thermowellposition, valve noise, and flow conditioners may influencethe meter’s performance characteristics.

A G A 9 YEAR 2001 UPDATE

This document has been published exactly three yearsas of this writing. Currently a committee task group hasidentified sixteen (16) areas to be addressed. Tasks havebeen assigned to various committee members.

These include studying and recommending more explicitinstructions for meter tube lengths, roughness andauxiliary taps, speed of sound field testing accuracyrecommendations, piping elbow, flow conditioner useand basically sharpening upareas of generalities.

The reader of the A.G.A.-9document is well advised tospend the time to read and followexamples in the appendixsection. There is some very goodtechnical information that willincrease the viewer’s generalknowledge of UltraSonic meters.

Charles W. Derr


PROPER TESTING OF ODORANT CONCENTRATION LEVELSPaul D. Wehnert

Heath Consultants Incorporated9030 Monroe Road, Houston, TX 77061

ABSTRACT

Proper odorant monitoring is required to keep naturalgas utilities under compliance with federal and stateregulations. These monitoring requirements are generallyhandled through a combination of events including;injection rate calculations, customer complaint calls,routine service personnel tests, odor concentration testsand chromatographic analysis. In the world today it iscritical to have appropriate documentation to supportproof that proper odorization of natural gas is occurring.This process will ultimately protect the public andhopefully keep us all from litigation.

REQUIREMENTS FOR ODORIZATION

Odorization of natural gas is regulated under FederalDepartment of Transportation (DOT) Title 49 Part 192.625.The code basically states “a combustible gas in adistribution line must contain a natural odorant or beodorized so that at a concentration in air of one-fifth thelower explosive limit, the gas must be readily detectableby a person with a normal sense of smell.”

This regulation since it was first introduced has alwaysled to considerable discussion in regards to exactly whatis a person with a normal sense of smell? We as humansall have varying abilities through our olfactory senses todetect odors. It has been proven in scientific studies thatage, gender, physical ailments such as allergies andcigarette smoking all effect one’s ability to detect odor.This has left us with a very “qualitative” means of detectingodorant through the use of the nose and a testing devicethat provides us the gas in air mixture. Many then ask thequestion why not use a more “quantitative” means toprovide us with the odorant concentration?

We have odorizers that can provide us extremely accuratevolumes of odorant being put into the gas distributionsystem. We also have extremely accurate means ofmeasurement to determine gas volumes downstream ofthe odorizer. This then becomes a very simplemathematical computation of odorant volume versesnatural gas measurement. This sounds all well and goodbut we must certainly remember that there are many otherfactors which affect the quantity and quality of the odorantthat can be completely out of our control.

FACTORS WHICH AFFECT ODOR QUANTITY

We talk about accurate means of computing volumes ofodorant put into the natural gas distribution system. Howare we then able to determine that the odorizer isfunctioning properly at all times? We have many typesof odorizers that employ several means of dispensingodorant into the distribution system. Many of thesesystems are affected by contaminants in the odorizerand are we able to identify when this is occurring? Anatural gas company must certainly determine whichtype of odorizer; injection, bypass or wick that is bestfor each particular piping application.

We are in the age of industry deregulation and openaccess within our natural gas transmission pipingnetworks. We are now able to get natural gas fromnumerous geographic locations including; the Gulf ofMexico, West Texas, Oklahoma, Western Canada andthe Plains States to name a few. We once knewconsistently well where we were getting our natural gasand what quality and natural occurring odorants werecontained in this source. We now have a “blend” whichcan certainly affect chemical reactions with differentodorant blending in the pipeline.

This open access and “blending” now allows for variousgas quality issues including the formation of distillatesin the pipeline that can literally absorb odorant from thenatural gas stream. Steps must be taken to insure thatwhen the formation of liquids is occurring that they areremoved from the system.

Other factors involve pipe wall absorption in the case ofnewly installed plastic pipe. Generally higher concen-trations of odorant are added during the initial com-missioning of a new pipeline to in affect “pickle” the line.

Internal corrosion of steel pipelines can produce internalcontaminants that can react chemically through oxidationand certainly affect odorant concentration.

FACTORS WHICH AFFECT ODOR QUALITY

We have mentioned physical ailments in the case ofallergies and smoking. These conditions can certainlyaffect one’s olfactory senses and their ability to detectthe smell of odorant.


It is certainly possible for natural gas to travel throughthe soil from pipeline leaks and have odorant removedby soil absorption. A gas company must determine theappropriate blend of odorant for their particulargeographic location.

The presence of external odors within the dwelling suchas cooking, perfumes and cleaning products cancertainly cause a masking and or a distraction situationand not allow an individual to differentiate the smell ofodorant.

ODORANT CONCENTRATION INSTRUMENTS

The current regulation for odorant concentration testingis primarily met with the use of electronic instruments.These instruments all employ the use of the human noseas stated in the regulation to determine the gas in airmixture at which an individual can detect the smell ofodorant. There are currently three (3) instrumentsavailable for this use.

BACHARACH ODOROMETER

This instrument is manufactured by the BacharachInstrument Company and employs the use of a flowmeter with glass and steel floats where conversions witha calibration chart for gas density and concentration arerequired.

HEATH CONSULTANTS ODORATOR

This instrument is manufactured by Heath ConsultantsIncorporated and employs the use of solid-stateelectronics for the digital display of gas in air mixtures.

YZ INDUSTRIES DTEX

This instrument is manufactured by YZ Industries andemploys microprocessor-based electronics and internaldata logging of gas in air mixtures.

Regardless of the type of portable electronic instrumentthat is utilized, it is imperative that when conductingodorant concentration “sniff” tests that the gas companyemployees are fully trained and experienced in the useof the instrument. The operator must be familiar with theoperating manual to insure that they are followingmanufacturers operating procedures. Many gascompanies conduct annual testing of their employeesto insure that they are familiar with the testing deviceand that they in fact can detect the smell of odorant.One such example would be to present the testing deviceto each gas company employee on an annual basis andallow them to run a test. This not only provides trainingthat the employee understands the use of the particularinstrument but also determines each employee’s abilityto detect the odorant. As mentioned earlier just like withthe public we to will have employees with varying abilitiesin their olfactory senses which must be known anddocumented. We must also follow the manufacturer’s

recommendation in regards to the calibration of eachinstrument to insure that the device is maintained andfunctioning properly.

ODORANT MONITORING PROGRAM

A comprehensive odorant monitoring program involvesseveral other pieces of information besides the odorant“sniff” test with the odorant concentration instrument.Yes, this is the requirement but most natural gascompanies employ other means to insure that properodorant is maintained in the distribution system.

Accurate records should be maintained on odorantinjection rates and along with measurement records wecan determine odorant levels in relation to gas volume.It is also important to keep complete records in relationto odorizer inspections to document proof of properlymaintained and functioning equipment.

The tracking of customer leak calls to central dispatch isextremely important. A natural gas company generallyhas system averages throughout the year of daily leakcalls. This is a direct result of how well your odorizationprogram is working. We will always have leaks withinthe distribution system, in customers’ homes, pilot lightsand in the street and when this occurs the public mustbe able to detect the odor and make the call. In the eventthese averages increase could signify that the odorantis being put in the system at a more substantial rate thannormal. In the event the leak call averages drop couldsignify that there may be problems at the odorizer or inthe piping network to initiate further action.

The simplest verification that natural gas has an odor isgenerally done by the customer service technician ondaily routine calls. A simple box checked as “yes” or“no” on the service form that odorant can be detectedat an appliance or meter set in the case of a change-out. This will not verify that the odorant is detected atthe appropriate concentration but it will signify that itdoes have an odor either absent, weak or strong.

The odor concentration meter tests will be performedon a periodic basis throughout the year and documentedwith the appropriate forms. We must remember that themore random tests that are conducted throughout thedistribution system the better informed we will becomeon the effectiveness of our odorization program.

The use of “quantitative” analysis instrumentation suchas titrators, analyzers and chromatographs for thechemical analysis is another vital step in odorantmonitoring. These instruments provide for real-timedeterminations of total sulfur and in many casesindividual mercaptan and sulfide component levels.

A combination of all the mentioned items will provide anatural gas company adequate records on the successof their odorization program. We must remember thatconditions are continually changing and we must be


aware of the occurrences within our system. We need toanalyze each piece of information and act accordinglywhen one or a combination of items looks out of thenormal. Investigations must them be carried forward andsolutions provided to insure adequate odorization.

CHROMATOGRAPHIC ANALYSIS

The use of titrators, analyzers and chromatographs areseveral methods employed for quantitative sulfuranalysis. A variety of detectors are used including leadacetate tapes, chemiluminescence, flame photometricand electrochemical technologies. These detectortechnologies provide for total sulfur calculations and inmany cases for complete component separation. Theseconcentrations can generally be displayed in a varietyof forms from grains, parts per million and pounds ofodorant per gas volume. These instruments can beconfigured for laboratory use where samples are broughtin or placed directly on the pipeline for real-timecalculations. A number of communications packages areavailable for transfer of information directly to acentralized gas control. These “quantitative” methodsof determining actual odorant concentrations in the gasstream does not meet the Federal requirement forodorant reporting under DOT 192.625. It does however,provide another piece of information in terms ofevaluating the overall effectiveness of the odorizationprogram.

A variety of manufacturer’s including but not limited toothe following are available:

Del Mar - Sulfur Lab 2000

This instrument is manufactured by Del Mar Scientificand have various models employing lead acetate tapeand GC columns for the detection of total sulfur and sulfurspecies.

Barton - OM-10

This instrument is manufactured by Barton Instrumentsand has various models employing an electrolyticanalyzer / titration cell for the detection of total sulfurs.

OdorEyes Systems - Accu/Line

This instrument is manufactured by OdorEyes Systemsand employs the use of an electrochemical sensor tomeasure total sulfurs.

Galvanic Applied Sciences Inc. - Model 801W,Model 902, Model 801P

This instrument is manufactured by Galvanic AppliedSciences Inc. and have various models employing leadacetate tape technology for the determination of H2Slevels and alternate readings between total sulfurs andH2S.

Houston Atlas - Scout H2S Analyzer

This instrument is manufactured by Houston Atlas andutilizes lead acetate tape technology for thedetermination of H2S levels.

Scintrex - OVD-229

This instrument is manufactured by Scintrex and utilizesthe electrochemical cell in a portable application for thedetection of individual sulfur components.

Applied Automation - Process Gas Chromatograph

Applied Automation offers a number of models of gaschromatographs for the detection of sulfur components.

Ionics - Sievers 355 Sulfur Chemiluminescence chromato-graph for analysis of sulfur compounds.

Heath Consultants Incorporated / Chromatosud -AirmoMedor

The original MEDOR was developed by Gaz de Francein the late 1970’s and manufactured by HeathConsultants/Gastech through the 1980’s and early1990’s. Most recently Heath Consultants has combinedefforts with Chromatosud the original manufacturer ofthe MEDOR in France to distribute and upgrade existingunits in the United States. The original MEDOR in theUnited States has gone through many transformationsutilizing interface operating systems from Hewlett-Packard, Spectra-Physics and Perkin-Elmer. HeathConsultants / Chromatosud now offer a new “windows”based system utilizing airmoVISTA software. This allowsupgrades of existing MEDOR technology with theaddition of an electronic interface module, airmoVISTAsoftware and the appropriate communications packagesoftware. The airmoMEDOR continues to provide theuser a complete breakdown of individual mercaptan andsulfide components utilizing an electrochemical detectioncell in a chromic acid solution.

These are primarily the more common titrators, analyzersand chromatographs that are commonly seen in themarketplace. We must remember that there are a widevariety of manufacturer’s that custom configureinstrumentation for the detection of sulfur relatedcompounds for various pipeline, petrochemical andrefinery applications. Regardless of the manufacturer, theinformation derived from this type of instrumentationprovides yet another piece of the puzzle to insure thatproper odorization is occurring.

CONCLUSIONS

We can now see that the huge task of insuring that aproper odorization program has been implemented andmaintained involves information gathering from a numberof sources. There is actually no “one” piece of information


that solely allows us to see the effectiveness of ourodorization program but rather involves a combination of“pieces” to complete the puzzle. In the world today wemust pay critical attention to our odorization programs toprotect life, property and insure complete PUBLICSAFETY. We must remember that odorant in the pipelineis the public’s primary leak detector and without thiswarning our public could be in serious danger in the eventa leak goes unnoticed. A well documented and maintainedprogram will certainly help us in the event of litigation.

REFERENCES

1. Bull, David, “Instruments for Odorization Monitoring”,presented at the AGA 1992 Distribution andTransmission Conference. Revised and updated byGeorge Lomax 1993.

2. Department of Transportation, Pipeline SafetyRegulations - Part 191 & 192.

3. Bacharach Odorometer - Product Bulletin

4. Heath Consultants Odorator - Product Bulletin

5. YZ Industries DTEX - Product Bulletin

6. Del Mar - Product Bulletin

7. Barton - Product Bulletin

8. OdorEyes Systems - Product Bulletin

9. Galvanic Applied Sciences Inc., Product Bulletin

10. Houston Atlas - Product Bulletin

11. Scintrex - Product Bulletin

12. Applied Automation - Product Bulletin

13. Ionics Sievers – www.ionicsinstruments.com

14. Heath / Chromatosud - Product Bulletin

Paul Wehnert


PIPELINE SAFETY IMPROVEMENT ACTand OPERATOR QUALIFICATION

Written by Jesus Ramos; Presented by Tom R. CheneyTexas Gas Service

1301 South Mopac Expressway, #400, Austin, Texas 78746

INTRODUCTION

The Qualification of Pipeline Personnel (OQ) rule requirespipeline Operators to develop and maintain a writtenqualification program for individuals performing coveredtasks on pipeline facilities. The qualification rule intentsto ensure a qualified pipeline work force and to reducethe probability and consequence of incidents caused byhuman error. The final rule created new subparts in thegas and hazardous liquid pipeline safety regulations. OQestablished qualification requirements for individualsperforming covered tasks, and amended certain trainingrequirements in the hazardous liquid regulations. The OQfinal rule was developed through a negotiation processand was effective on August 26, 1999. The rule requiredall individuals performing covered tasks to be qualifiedby October 28, 2002. The Operators made a great effortand investment to comply with the OQ rule. Operatorscreated their OQ program, implemented its processes,generated evaluation records for employees andcontractors, and began to breathe a relieved sigh.However, President Bush signed into law the PipelineImprovement Act on December 17, 2002.

PIPELINE IMPROVEMENT ACT

The new Act, HR 3609, has thirty-three sections mostlyaddressing pipeline related topics such as One-CallNotification Programs, Safety Orders, Penalties, NationalTransportation Safety Broad (NTSB) SafetyRecommendations, Pipeline Qualification Programs, andmany other related subject matter such as Risk Analysisand Integrity Management Programs for Gas Pipelines.The Pipeline Improvement Act of 2002 amends Title 49of the United State Code. Briefly, in one section OPSrequested a provision for the Secretary of Transportationto have authority to order an Operator of a facility totake corrective action if the Secretary decides that apotential safety-related condition exists. Another sectionrequires Research of Special Program Administration(RSPA) and Office of Pipeline Safety (OPS) to respondto recommendations received from the NTSB within 90days from receipt of such recommendations. Suchresponses shall state the intentions of OPS with respectto the recommendations and shall state the timetablefor completing the procedures and reasons for refusalsto so. The responses shall be made available to thepublic. The OPS is required to submit an annual reportdescribing each recommendation received and the OPSresponse to each recommendation for the previous year.

Most importantly, Section 13 (60131-Verification ofPipeline Qualification Programs) expects the Secretaryof Transportation to require Operators of pipelinesfacilities to develop qualification programs for theirpersonnel who perform covered tasks as define in theCode of Federal Regulations Part 192 Subpart N or 195Subpart G. This section also requires the Secretary tohave in place standards and criteria for such qualificationprograms, including a method for examining or testingthe qualifications of individuals who perform coveredtasks. Such methods may include written examination,oral examination, on-the-job training, simulations,observation during on-the-job performance, and otherforms of assessment. The method may not be limited toobservation of on-the-job performance, except withrespect to tasks where the Secretary has determinedspecifically that such observation is the best method ofexamining or testing qualifications. Further, the Secretarymust ensure that the results of any such on-the-jobperformance observations are documented in writing.The Secretary may waive or modify requirements if notinconsistent with pipeline safety. The Secretary isrequired to verify each Operator’s qualification program,including modifications to previously verified programs.In the event the Secretary fails to establish standardsand criteria as set forth in this section, pipeline facilityOperators are required to develop and implementqualification programs based on the requirements of thissection. The Secretary is required to report to Congresswithin five years on the status and results of personnelqualification regulations. Finally, this section also requiresa pilot program to be established for the certification ofindividuals who operate computer-based systems forcontrolling the operation of pipelines. The pilot programseeks the participation of three pipeline facilities. Pipelineimprovement revolves on the standards and criteriadesigned by OPS for inspecting OQ Programs and theOQ established methods developed by the Operatorsto comply with OQ Rule.

VIEW POINTS

A performance rule requires inspection of the approachesthrough which the Operator expects to achieveimprovements to its pipeline system. The OperatorQualification Rule was designed as a performance rulewith some limited prescriptive requirements. Inspectionagainst performance rule provisions is different frominspection of a purely prescriptive rule. A performancerule provides flexibility in how Operators evaluate, justifyand change their practices to satisfy the rule’s intentwithin their unique operating environment. However, such


changes will not immediately manifest themselves inrecognizable changes in performance, and performanceresults will likely vary significantly from Operator toOperator depending on factors such as the extent ofmanagement involvement, the maturity of Operatorpractices, and the commitment of individuals performingcovered tasks. The ultimate proof of the effectiveness ofOperator OQ programs will be through a continuingreview of performance trends. However, regulatorybodies cannot await performance results to demonstrateOperator program effectiveness. Therefore, OQ ruleimplementation inspection must include not onlyevaluation of compliance with its prescriptive provisions,but also, evaluation of program completeness,anticipated effectiveness of the documented approachesdesigned to qualify individuals, and to ensure they remainqualified. OPS expect to undertake a joint effort withindustry to clarify the ingredients in a successful OQprogram (criteria) and to document examples of practicesthat satisfy these criteria (benchmarks). Protocolquestions are designed to support regulatory explorationof Operators’ approaches used to address the OQ Rulerequirements. The protocols will assist to discover thecriteria and benchmarks in Operator OQ programs. Thequestions will outline the ‘thinking trail’ used by theOperator to develop and implement OQ programs.

At least three public meetings, as many workshops, andnumerous phone conferences were planned for a jointeffort between OPS, states, the pipeline industry, andthe public to describe approaches to inspectingOperators against provisions in the OQ rule, that is, toestablish standards and criteria with benchmarks. Timeperiods were considered between meetings for eachinterested segment to group and discuss concerns anddevelop positions addressing the approaches to theissues and inspection methods from each segment’sviewpoint. During the first of several public meetingsbetween the pipeline industry and OPS to discuss‘standards and criteria’, OPS revealed its position byindicating that the review of incident and accident historyfor the pipeline industry illustrated that Operator errorshave contributed significantly to these events. OPSfurther stated that the Qualification of Operator Personnelrule is “intended to eliminate Operator errors and therebyeliminate incidents and accidents to which Operatorerrors contribute.” OQ was designed to allow Operatorsalmost total flexibility in their approach to addressindividual qualification. The rule is characterized as aperformance type rule, but contained no measures bywhich trends in performance could be monitored.Therefore, OPS’ approach to inspecting compliance withthe rule’s provisions must ascertain that the safetyconcerns initiating the rule are directly confronted. OPSis “rigorously inspecting compliance with the rule’sprescriptive requirements and evaluating the approachesOperators have used to satisfy these requirements.” Theapproach OPS has developed to inspect OQ Programsmakes use of a set of inspection protocols. Theseprotocols are simply questions organized into a line ofinquiry that seeks to examine how the Operator meetsOQ provisions. Assorted questions are designed to

support regulatory exploration of the Operator’smethodology utilized to address the OQ rulerequirements. OPS have developed other newcompliance tools to enforce OQ.

Several tools are needed to provide an option in promotingimprovement at Operators who are making a strong effortto address the letter and spirit of OQ, but have not yetcompleted development of a fully satisfactory Program.One catalyst to change is the “Notice of Area ofRecommended Improvement” (NARI). If the Operator’sprocedures and processes required by the rule are notadequate, but the Operator has demonstrated anunderstanding and appreciation of what it would take toproduce adequate procedures and has indicated acommitment to make such improvements, the newcompliance tool NARI will be used. If the Operatorprocedures and practices required by the rule are notadequate and the Operator has demonstrated littleunderstanding of what it would take to produce adequateprocedures (or no willingness to do so), then a “Notice ofAmendment” (NOA) will be pursued. If there is clear non-compliance with rule requirements that cannot be easilyremedied by the Operator and which indicate a lack ofserious intent to comply with the rule objectives, a Noticeof Probable Violation (NOPV) will be pursued. OPSunderstand that the OQ rule, associated inspectionprotocols, and other enforcement tools significantly raisethe bar for pipeline safety.

OPS is developing inspection protocols both to improvethe communication of regulatory expectations with statesand the industry, and to support improved consistencyof inspections conducted by various regulatory groups.The inspection protocols provoked many issues, eachin different areas and were introduced and discussedduring the January 2003 San Antonio public meeting.Following is the regulatory perspective on major topicsto be addressed. This perspective, which has beendiscussed among regulatory members of the OQDevelopmental Team including representatives from fivestates and each of the five OPS regions, has beenprepared to support discussions. OPS intent is for thenumbered topics below to be discussed by selectedindividuals representing regulatory and industryperspectives. Further, the thirteen issues have beenlabeled into three impact categories: High Impact Issues,Medium Impact Issues, and Low Impact Issues.

HIGH IMPACT ISSUES

Scope of OQ Inspections: Should inspections go beyondevaluation of compliance with prescriptive requirementsof the Rule? Regulators cannot await performance trendsto show whether Operator programs are working. OPSneeds to examine compliance with the prescriptiverequirements as well as with the set of requirementsimplied as necessary by the rule. An example of animplied requirement is the need for a method to identifyindividuals who may have contributed to an incident/accident through performance of a covered task.Operators need a way to characterize incident causes


and to identify individuals who performed covered tasksthat may have contributed to the incident/accident.Inspections against provisions in the OQ Rule mustinclude evaluation of the approach Operators take tosatisfy those provisions.

Evaluation of Knowledge, Skills, and Abilities (KSAs):Should evaluation leading to qualification considerknowledge, skills and physical ability? The Rule requiresthat individuals be evaluated for their ability to performcovered tasks. Each covered task needs to be evaluatedin terms of the knowledge, skills and physical abilityrequired to successfully accomplish the covered taskperformance. The importance of KSAs characteristics willvary from task to task, but the evaluation process needsto consider the need for each. Some concern has beenexpressed regarding any reference in the protocols to theuse of training to support qualification of Operators.Although these concerns correctly reflect that training isnot required by the Rule, they ignore the fact that trainingis a major means to the end of qualifying individuals. Inaddition, the recently passed amendment to the PipelineSafety Act states that one of the minimum elements of aqualification program is “a program to provide training,as appropriate, to ensure that individuals performingcovered tasks have the necessary knowledge and skillsto perform the tasks in a manner that ensures the safeoperation of pipeline facilities.” Protocols currently explorethe Operators’ use of training in their approach toqualification and reevaluation. OPS need to understandthe role of training use to judge the approach adequacybeing taken by individual Operators. OPS have addressedthese comments by exploring the role of training usingprotocol questions directed at Rule provisions.

Re-evaluation Intervals: How should re-evaluationintervals be supported and justified? Regulatoryperspective on re-evaluation interval is that they caninitially be based on precedents from other regulatoryagencies. Ultimately, however, Operators must eitherestablished conservative intervals or institute othermeans to monitor trend performance resulting fromintervals selected. If longer intervals are desired,performance trends must be used to adjust conservativeintervals as appropriate. Concerns have been expressedon the extent of justification needed to establishreevaluation intervals, and on the need for monitoring toevaluate the appropriateness of the intervals selected.Initial industry practice seems to support a three to fiveyear reevaluation interval based on reference to otherregulatory bodies practices. No justification for theseintervals based on actual data has yet been identified.In addition, little formal monitoring of the effectivenessof Operator performance is being planned to substantiateintervals selected. Also, no significant activity has beenidentified to differentiate reevaluation intervals of differenttasks based on task criticality, frequency of performanceor risk associated with the task. Differentiation ofreevaluation interval is rare even though the NTSB haspublicly supported a one-year reevaluation interval forSCADA Operators. OPS’ position, in the absence ofsubstantive data on the impact of different reevaluation

intervals on Operator performance, performancemonitoring is needed to verify whatever intervals areselected. Such monitoring may be used to selectivelylengthen the reevaluation intervals beyond thosecurrently envisioned should the data support such adecision. However, the Industry has commented thattracking performance of every task by every qualifiedindividual is burdensome and not practical. While suchtracking may be difficult, several rule provisions requirethat, at a minimum, the Operator must be able to identifypersons who performed covered tasks that may havecontributed to an incident/accident. Therefore, the ruleimplies the need for some monitoring of performance.

Maintenance versus New Construction: How should wedistinguish between maintenance and new constructionin defining covered tasks? New construction tasks arenot currently covered by the rule. The OPS perspectiveis that tasks involving replacement of existing equipment(e.g., replacement in kind of a corroded pipe segment)should be covered. Additionally, tasks performed on theright-of-way should be covered. Generally, if the pipelineis to serve a new end user and/or the pipeline is in adifferent direction, then, it is new construction. There are,however, some ambiguities about how to differentiatebetween O&M tasks and those associated with newconstruction. The ambiguities are related both to thetasks themselves and to the location where the tasksare carried out. The Tasks: Initial inspections haverevealed significant differences in the definition of O&Mtasks among Operators as well as between Operatorsand regulators. Often Operators restrict the definition ofO&M tasks to those performed on an existing portion ofa pipeline. Under this interpretation, in-kind replacementof an existing pipe section necessitated by severecorrosion would be considered new construction, andrelated tasks would not be covered. However, this is nota reasonable interpretation from OPS’ viewpoint. Thedefinition OPS is using include, as covered tasks workto replace pipe segments if the pipe capacity segmentsare maintained and service is not expanded. TaskLocation: Initial industry comments object to theinterpretation that tasks performed on equipment thathas been disconnected from the pipeline may be eithercovered or not depending on whether the task isperformed on the right-of-way or at a facility remote fromthe right-of-way. The industry‘s position is that once anequipment piece has been disconnected from thepipeline, any subsequent work is not covered. OPS’position is that tasks performed on the right-of-way arecovered because the right-of-way is part of the “pipelinefacility”. To be consistent with the spirit and intent of therule, qualified individuals should perform maintenanceor repair work performed on equipment pieces that areintended to be part of the pipeline facility. In the finalanalysis the debate may not be about task or tasklocation. It may simply be that the greater good, safety,and the greater number, Operator customers, should bethe foundation for this deliberation.

Treatment of Emergency Response: Does the rule coveremergency response tasks, if not, what are its bounds?


Citing the rule’s Preamble, the Industry has stated thatany activity associated with emergency response shouldbe excluded from the OQ Rule’s scope. This positiondocumented in the rule’s Preamble is apparently basedon the differences in treatment of emergency responsefor liquid and gas pipelines in the regulations. Excludingemergency response activities may also be supportedby the fact that the congressional mandate for OQrulemaking omitted the need to include emergencyresponse tasks. The OPS perspective is the Rule’spreamble inappropriately excludes emergency responsetasks from coverage by the rule. Emergency responseactions are included in the Operation & Maintenanceregulation sections, §192.615 and §195.402(e). Theactions that an Operator takes in responding to anemergency condition are operating and maintenancetype actions, therefore, should be considered coveredtasks. It is inconsistent to allow covered tasks to beperformed by non-qualified individuals under emergencyconditions, and then require individuals to be qualifiedto do those same tasks during normal O&M activities.The OPS position is that Operators should anticipate thepossible occurrence of emergencies and makecontingency provisions by qualifying individuals toaddress these conditions. Some concern has beenexpressed about unnecessarily constraining Operatorindividuals from responding to emergencies, as well asextending the scope of the rule to include tasks that areonly performed during emergencies (e.g., fire fighting).

MEDIUM IMPACT ISSUES

Additional Covered Tasks: Is pipeline excavation a coveredtask? Pipeline excavation is a major source of accidents/incidents, it is included as an O&M task, and it should becovered. Clearly, if regulators don’t have jurisdiction overan Operator excavating a pipeline, they cannot requirethem to qualify their workers. OPS note the requirementsof §195.442 and §192.614(a) to protect their facilities.Coverage by the rule of at least one task that hashistorically contributed significantly to severe pipelineaccidents is presently being debated. OPS believespipeline excavation is included as an O&M requirement inthe regulations and that its significance should dictate thatit be covered by the OQ program. Operators that are tryingto address the spirit of the rule have included excavationas a covered task. If this task were considered to beexcluded from coverage by the rule, then it would eitherneed to be specifically added to the rule by supplementaryrulemaking, or incorporated by reference to the IntegrityManagement Rules, or supported by an industry standard.Future inspections may identify other tasks that aredebatable or clearly need to be added to the covered tasklist to satisfy the intent and spirit of the Rule. As suchtasks are identified consideration will be given to requiringtheir coverage under the OQ Rule through the use ofsupplementary rulemaking.

Extent of Documentation: What OQ records must bedeveloped and maintained by Operators? The rulerequires the Operators to keep a minimum of fourrecords. During the inspection process, additional

records, such as those referenced by the Operator’s OQProgram, may need to be evaluated to verify compliancewith rule provisions. Such records may include evaluationmethods and root cause analysis. Documentation shouldinclude decision-making processes involved indevelopment of the Operator’s program (e.g., coveredtask identification, subsequent qualification interval,method to incorporate newly recognized AOCs, andtraining requirements for an individual no longer qualifiedto perform a covered task).

Abnormal Operating Conditions - AOCs: Should the listof AOCs be dynamic? There is a difference between OPSand the industry position on abnormal operatingconditions. Operators want it clear that the AOCs listingused in qualifying individuals cannot be comprehensive,and must be limited to AOCs the Operator can reasonablyanticipate the individual will encounter while performingthe covered task. The rule requires that both generic andtask-specific AOCs be developed and used in evaluatingindividuals to perform covered tasks. Developing acomplete list of AOCs is not possible, and an industrystandard list might not be appropriate for some Operators.OPS agree that AOCs should be those conditions to whicha qualified individual can recognize and react toappropriately. Developing a set of AOCs is certainly anevolutionary process. The AOCs list should be dynamicand Operators need a means to incorporate newlyrecognized AOCs in the set used in qualifying individuals.OPS expect practices to be in place, perhaps, part of theOperator’s efforts is to identify and evaluate “near misses,”to recognize newly identified AOCs. The practices shouldalso include these AOCs in continuing skill developmentand evaluation activities. Such an effort should ultimatelylead to an increasingly comprehensive AOCs set andindividuals who are better prepared to address AOCs.

Treatment of Training: Should training practices beevaluated during OQ inspections? While not explicitlyrequired by the OQ Rule, training is key to implementingmany steps in the OQ Rule. Inspection of evaluationmethod effectiveness used to satisfy requirements of theRule must include the role of training in the Operator’sprogram. Some concern has been expressed regardingany reference in the protocols to the use of training tosupport qualification of Operators. Although theseconcerns correctly reflect that training is not required bythe Rule, they ignore the fact that training is a majormeans to the end of qualifying individuals. In addition,the recently passed amendment to the PipelineImprovement Act states that one of the minimumelements of a qualification program is “a program toprovide training, as appropriate, to ensure that individualsperforming covered tasks have the necessary knowledgeand skills to perform the tasks in a manner that ensuresthe safe operation of pipeline facilities.” Protocolscurrently explore the Operators’ use of training in theirapproach to qualification and reevaluation. Regulators’need to understand the role and use of training to judgethe adequacy of the approach being taken by individualOperators. OPS have addressed these comments byexploring the role of training using protocol questions


directed at Rule provisions. There is a relationshipbetween KSA and the training issue. KSAs may be thedetails of an effective training program with appropriatesubject content. This issue may be used to further defineknowledge factors to determine if training iscomprehensive to prevent human error.

LOW IMPACT ISSUES

Criteria for Small Operators: Will small Operator OQPrograms be subject to the same criteria as largeOperators? Protocol review has revealed that mostquestions apply to both large and small Operators. Thesame criteria will apply to all. However, the practices usedby small Operators to address rule requirements areexpected to be significantly different from those usedby large Operators. The complexity of operations maybe key; smaller Operators do not confront the same dailycomplex issues the larger Operator encounter.

Direction and Observation of Non-Qualified Individuals:Is guidance needed to support supervisors in determininghow many non-qualified individuals can be directed andobserved by one qualified person? Different tasks can bedirected and observed differently. In defining span ofcontrol, consideration must be given to time available torecognize and correct errors. Guidance is needed to avoidunnecessarily burdening supervisors. Operators inspectedto date have left the establishment of limitations on thequalified persons’ ability to direct and observe non-qualified persons in covered task performance to thediscretion of the individuals in the field. Only very generalguidance has been provided, consistent with the generalprovisions in the rule itself. OPS believes clear guidanceis needed to support decision-making in the field on theappropriate limits on span of control of qualified personsdirecting and observing non-qualified persons performingcovered tasks. The limits on span of control will likely varywith the complexity and risk associated with coveredtasks. Another consideration to address is when thequalified individual halts the work to correct or direct anon-qualified individual; will the qualified person havedirect connection with the other non-qualified individualsperforming the same or different covered task in the workvicinity?

Noteworthy Practices: Should regulators play a role inthe identification and communication of “NoteworthyPractices”? Noteworthy practices should aid in improvingefficiency and effectiveness in satisfying requirements.Recognition and communication of these practices isall Operators’ best interest. Additionally, such practicesrepresent good examples of how to address the rulerequirements. Industry commented that noteworthypractices should not be assembled and published sincethey may not be applicable or relevant to all Operators,and since they may become the defacto standardsagainst which program acceptability is judged. OPSbelieve the assembly and communication of noteworthypractices is valuable because it will support developmentof more efficient and effective ways to minimize incidentsand accidents caused by Operator error. Experience with

the Common Ground initiative shows that identificationand broad communication of noteworthy practices doesnot necessarily produce defacto standards. One way toaddress this issue is through the development of one ormore consensus standards in which examples ofpractices found to be suitable are documented. Thisshould be supplemented either by frequent revisions ofthe consensus standards to reflect newly identifiedpractices; or by a continuing effort, perhaps sponsoredjointly by industry and regulatory bodies, for identifying,documenting, and communicating noteworthy practicesdeveloped by Operators.

Persons Contributing to an Incident or Accident: ShouldOperators have documented means to identify a coveredtask whose performance may have contributed to anincident/accident along with individuals who performedthese tasks? OPS differ with the industry on whethermethods and documentation are needed to supportidentifying individuals who performed covered tasks thatmay have contributed to incidents or accidents.Regulators will look for documentation to support theinvestigative trail. Such documentation should bereferenced in Operator practices used to investigate thecauses of incidents. Prudent Operators must know theorigin and cause of an accident/incident in order to preventsimilar accident/incident reoccurrence. Documentation ofpractices to support this knowledge is required by therule. Reference to existing practices may adequatelyaddress this need in the interim; however, improvementsin these practices may be needed.

Generally, the Industry’s position on the thirteen issues isconcurrence with the regulators; the issues need to beaddressed. Industry’s primary strategy to address the issuesraised by OPS will be to develop OQ standards, embeddedin a nationally recognized standards group (e.g. ASME, API,or other) that quantifies OQ compliance specifics.

THE PROTOCOLS

The following protocols have been written to assistfederal and state pipeline inspectors who are evaluatingOperator OQ programs. The protocols are not intendedas enforcement instruments or to provide inspectors withadditional enforcement authority, but rather are intendedto provide inspectors with a template that they can usein the course of their inspections to ensure that Operatorscomply with all OQ rule elements. The protocol’sobjective is to ensure that Operators have followed therule prescriptive requirements. This objective will beaccomplished by rigorously inspecting each Operator’srecords to ensure that all individuals performing coveredtasks on pipeline facilities are properly qualified and thatsufficient documentation is maintained for theseindividuals. Proper recordkeeping is a key OQ rulecomponent. It is, therefore, important inspectors are ableto verify records that are maintained for all individualsperforming covered tasks.

The OQ inspection form is organized around nineelements, including one for field verification. Each


element has one or more associated protocol. Eachprotocol consists of 4 boxes:

(1) A protocol number accompanied by the protocolsubject or topic;

(2) A protocol question(s) (sometimes followed by‘Verify’ statements);

(3) Guidance topics; and(4) The relevant rule language.

The protocol topics have been structured into ‘ProtocolQuestion(s)’ to guide inspectors through the OQinspection process. Each protocol question is followedby ‘Guidance Topics.’ The guidance topics listcharacteristics the inspector would typically expect tofind in an effective OQ Program, and are consistent withthe rule’s intent of the regulatory language thataccompanies each protocol. Some, all, or none of thesecharacteristics may be appropriate depending on factorsunique to each Operator’s OQ Program, pipeline assets,and unique operating environment. Operators should beprepared to demonstrate that their OQ programs addresseach of these characteristics or to describe how theirprogram will be effective in their absence.

‘Verify’ statements follow many of the protocol questions.These statements have been included because they canbe directly traced to specific rule language. Therefore,compliance with each ‘verify’ statement should beconfirmed. Many ‘verify’ statements (and protocolquestions) are followed by a parenthetical statement thatindicates that the statement or question is either‘enforceable’ or ‘non-enforceable’. If the ‘verify’statement or protocol question is listed as non-enforceable, the statement or question is not enforceableunder the rule, but is nonetheless an importantconsideration for the Operator. Finally, should theinspection process reveal violations of prescriptive rulerequirements; inspectors will take appropriateenforcement actions. Should deficiencies be identifiedin how Operators address program characteristics,inspectors will seek evidence of violations related tothese deficiencies. Significant inquiries seeking furtherinformation related to program characteristics will becommunicated to the Operator as an integral inspectionprocess component.

Element 1 - Document Program Plan, ImplementingProcedures and Qualification Criteria.

1.01 Application and Customization of ‘Off-the-ShelfPrograms’.

Does the Operator’s plan identify covered tasks and doesit specify task-specific reevaluation intervals forindividuals performing covered tasks? [Enforceable]

1.02 Contractor Qualification

Does the Operator employ contractor organizations toprovide individuals to perform covered tasks? If so, whatare the methods used to qualify these individuals and

how does the Operator ensure that contractor individualsare qualified in accordance with the Operator’s OQprogram plan?

Verify that the Operator’s written program includesprovisions that require all contractor and subcontractorindividuals be evaluated and qualified prior to performingcovered tasks, unless the covered task is performed bya non-qualified individual under the direction andobservation of a qualified individual. [Enforceable]

1.03 Management of Other Entities Performing CoveredTasks

Has the Operator’s OQ program included provisions thatrequire individuals from any other entity performingcovered task(s) on behalf of the Operator (e.g., throughmutual assistance agreements) be evaluated andqualified prior to task performance?

Verify that other entities that perform covered task(s) onbehalf of the Operator are addressed under theOperator’s OQ program and that individuals from suchother entities performing covered tasks on behalf of theOperator are evaluated and qualified consistent with theOperator’s program requirements. [Enforceable]

1.04 Training Requirements (Initial Qualification,Remedial if Initial Failure, and Reevaluation)

Does the Operator’s OQ program plan contain policy andcriteria for the use of training in initial qualification ofindividuals performing covered tasks, and are criteria inexistence for re-training and re-evaluation of individualsif qualifications are questioned? [Non-Enforceable]

1.05 Written Qualification Program

Did the Operator meet the OQ Rule requirements forestablishing a written Operator qualification program andcompleting qualification of individuals performingcovered tasks? Verify that the Operator’s writtenqualification program was established by April 27, 2001.[Enforceable]

Verify that the written qualification program identified allcovered tasks for the Operator’s operations andmaintenance functions being conducted as of October28, 2002. [Enforceable]

Verify that the written qualification program establishedan evaluation method(s) to be used in the initialqualification of individuals performing covered tasks asof October 28, 2002. [Enforceable]

Verify that all individuals performing covered tasks as ofOctober 28, 2002, and not otherwise directed orobserved by a qualified individual were qualified inaccordance with the Operator’s written qualificationprogram. [Enforceable]


Element 2 - Identify Covered Tasks and RelatedEvaluation Methods

2.01 Development of Covered Task ListHow did the Operator develop its covered task list?

Verify that the Operator applied the four-part test todetermine whether 49 CFR Part 192 or 49 CFR Part 195O&M activities applicable to the Operator are coveredtasks. [Enforceable]

Verify that the Operator has identified and documentedall applicable covered tasks. [Enforceable]

2.02 Evaluation Method(s) (Demonstration of Knowledge,Skill and Ability) and Relationship to Covered Tasks. Hasthe Operator established and documented the evaluationmethod(s) appropriate to each covered task?

Verify what evaluation method(s) has been establishedand documented for each covered task. [Enforceable]Verify that the Operator’s evaluation program ensuresthat individuals can perform assigned covered tasks.[Enforceable]

Element 3 - Identify Individuals Performing CoveredTasks

3.01 Development and Documentation of Areas ofQualification for Individuals Performing Covered TasksDoes the Operator’s program document the evaluationand qualifications of individuals performing coveredtasks, and can the qualification of individuals performingcovered tasks be verified at the job site?

Verify that the Operator’s qualification program hasdocumented the evaluation of individuals performingcovered tasks. [Enforceable]

Verify that the Operator’s qualification program hasdocumented the qualifications of individuals performingcovered tasks. [Enforceable]

3.02 Covered Task Performed by Non-QualifiedIndividual

Has the Operator established provisions to allow non-qualified individuals to perform covered tasks while beingdirected and observed by a qualified individual, and arethere restrictions and limitations placed on such activities?

Verify that the Operator’s program includes provisionsfor the performance of a covered task by a non-qualifiedindividual under the direction and observation by aqualified individual. [Enforceable]

Element 4 - Evaluate and Qualify IndividualsPerforming Covered Tasks

4.01 Role of and Approach to ‘Work PerformanceHistory Review’

Does the Operator use work performance history reviewas the sole method of qualification for individuals performingcovered tasks prior to October 26, 1999, and does theOperator’s program specify that work performance historyreview will not be used as the sole method of evaluation forqualification after October 28, 2002?

Verify that after October 28, 2002, work performancehistory is not used as a sole evaluation method.[Enforceable]

Verify that individuals beginning work on covered tasksafter October 26, 1999 have not been qualified usingwork performance history review as the sole method ofevaluation. [Enforceable]

4.02 Evaluation of Individual’s Capability to Recognizeand React to AOCs

Are all qualified individuals able to recognize and reactto AOCs?

Has the Operator evaluated and qualified individuals fortheir capability to recognize and react to AOCs?

Are the AOCs identified those that the individual mayreasonably anticipate and appropriately react to duringthe performance of the covered task?

Has the Operator established provisions forcommunicating AOCs for the purpose of qualifyingindividuals?

Verify that individuals performing covered tasks havebeen qualified in recognizing and reacting to AOCs theymay encounter in performing such tasks. [Enforceable]

Element 5 - Continued/Periodic Evaluation ofIndividuals Performing Covered Tasks

5.01 Personnel Performance Monitoring

Does the Operator’s program include provisions toevaluate an individual if the Operator has reason tobelieve the individual is no longer qualified to perform acovered task based on:

• Covered task performance by an individualcontributed to an incident or accident.

• Other factors affecting the performance ofcovered tasks.

Verify that the Operator’s program ensures evaluation ofindividuals whose performance of a covered task mayhave contributed to an incident or accident. [Enforceable]Verify that the Operator has established provisions fordetermining whether an individual is no longer qualifiedto perform a covered task, and requires reevaluation.[Enforceable]


5.02 Reevaluation Interval and Methodology forDetermining the Interval.

Has the Operator established and justified requirementsfor reevaluation of individuals performing covered tasks?Verify that the Operator has established intervals forreevaluating individuals performing covered tasks.[Enforceable]

Element 6 - Monitor Program Performance; SeekImprovement Opportunities

6.01 Program Performance and ImprovementDoes the Operator have provisions to evaluateperformance of its OQ program and implementimprovements to enhance the effectiveness of itsprogram? [Non-Enforceable]

Element 7 - Maintain Program Records

7.01 Qualification ‘Trail’ (i.e., covered task; individualperforming; evaluation method(s); continuingperformance evaluation; reevaluation interval;reevaluation records)

Does the Operator maintain records in accordance withthe requirements of 49 CFR 192, subpart N, and 49 CFR195, subpart G, for all individuals performing coveredtasks, including contractor individuals?

Verify that qualification records for all individuals performingcovered tasks include the information identified in theregulations. [Enforceable]Verify that the Operator’s programensures the retention of records of prior qualification andrecords of individuals no longer performing covered tasksfor at least five years. [Enforceable]

Verify that the Operator’s program ensures the availabilityof qualification records of individuals (employees andcontractors) currently performing covered tasks, or whohave previously performed covered tasks. [Enforceable]

Element 8 - Manage Change

8.01 Management of Changes (to Procedures, Tools,Standards, etc.)

Does the Operator’s OQ program identify how changesto procedures, tools standards and other elements usedby individuals in performing covered tasks arecommunicated to the individuals, including contractorindividuals, and how these changes are implemented inthe evaluation method(s)?

Verify that the Operator’s program identifies changes thataffect covered tasks and how those changes arecommunicated, when appropriate, to affectedindividuals. [Enforceable]

Verify that the Operator’s program identifies andincorporates changes that affect covered tasks.[Enforceable]

Verify that the Operator’s program includes provisionsfor the communication of changes (e.g., who, what,when, where, why) in the qualification program to theaffected individuals. [Enforceable]

Verify that the Operator incorporates changes into initialand subsequent evaluations. [Enforceable]Verify thatcontractors supplying individuals to perform coveredtasks for the Operator are notified of changes that affecttask performance and thereby the qualification of theseindividuals. [Enforceable]

Element 9 - Field Verification

9.01 Are field/job supervisors aware of their responsibilitiesas defined under the Operator’s OQ program?

9.02 Are the observed covered task(s) performed inaccordance with appropriate Operator-approvedprocedures, and are the procedures present at the job site?

9.03 Are the individuals performing the observed coveredtask(s) adhering to the Operator-approved proceduresas written?

9.04 Are the proper tools, techniques and processesdetailed in the Operator-approved procedures employedin the performance of the observed covered task(s)?

9.05 Are the qualifications of all individuals involved inperforming the covered task(s) verified at the job site? Isthis verification process performed as detailed in theOperator’s OQ program plan? Is this verification processapplied to employees and contractors alike?

9.06 Are the qualified individuals performing the observedcovered task(s) knowledgeable of how to recognize theapplicable AOCs and what their expected reaction tothe AOCs should be?

9.07 Are individuals not qualified to perform a coveredtask performing a covered task? If so, are the non-qualifiedindividuals being directly observed by a qualified individualin accordance with the requirements of the regulation?

9.08 How are field/job supervisors informed of changesthat affect the performance of covered tasks under theirresponsibility?

9.09 In cases where the field office is part of a subsidiaryor separate district, is implementation of OQ programrequirements consistent with other districts and theoverall organization’s OQ program?

9.10 How is performance of the covered task(s) reviewed/inspected in the field by internal auditors or third partyinspectors?

9.11 What problems have been experienced in the fieldin implementing the Operator’s OQ program? If problemshave been experienced, how have they been


communicated back to the individual responsible for theOQ program?

9.12 How are Control Center operations coordinated withremote operations that are conducted with otheroperations personnel? Are these ‘other operationspersonnel’ qualified to perform the covered tasks beingperformed?

THE FINAL MEETING UPDATES

During the last public meeting, key points werepresented. In the development of a National ConsensusStandard, it is very important both federal and staterepresentatives participate in the process with industry.The Industry has formed a committee to author a nationalOQ standard utilizing American Society of MechanicalEngineers format (ASME B31Q). The standard is toaddress all thirteen issues and it is targeted be completedby June 2004. The American Public Gas Association isauthoring a guidance model for small operators to meetOQ Rule requirements. Industry should agree on the needof small Operator guidance. Some question theguidance’s need because there should be no differencewith the compliance approach because of operator size.A deliverable from OPS was the Frequent AskedQuestions (FQAs). The intend was to aid companies whodo not have the resources to interpret and understandregulations, and the FAQs were intended to assist themto effectively and efficiently implement OQ Programs.The concern remains that the questions will greatlyinfluence enforcement techniques and restrict someOperators in same manner. To some extend, theconclusion on certain key definitions still needsdiscussion and agreement between regulators andindustry. During federal field inspections, the states havehad representatives in the protocol process, and havebeen asked by OPS to consider using the protocols inintrastate inspections. The inspections should reportfindings in a manner that can be added to a databaseavailable to all inspectors. The states, however, areindependent and may or may not use the protocols asdefined. OPS intentions are to share inspection resultswith states through a database system and is intendedto focus on findings but not on individual Operators. OPSlegal department is investigating if any constraints existwith sharing findings between federal and non-interstateagencies. The Notice of Area of RecommendedImprovement (NARI) mentioned earlier will not be usedfor in the OQ inspection process. An Operator shouldexpect a Notice of Amendment or a Notice of ProbableViolation, if warrant. Concerns continue with contractorqualifications and how the Operator accepts and reviewscontractors. Field verification inspections will be usedto validate if there are violations or issues. When themethod for performing covered tasks by contractor differfrom those methods performed by an operator employee,the Operator should review and approve that differentmethod. The industry advocates reevaluation intervalsshould be based on the risk involved in the task (a tieredapproach). A balanced approach to the subject of

reevaluation intervals is needed considering the safetyrecord, and what the payback may be for overlyconcentrating on OQ in light of all other regulatoryrequirements and competitive pressures faced byindustry. Statistics demonstrate pipeline industry fatalitiesand incidence of operator error are relatively lowindicating longer subsequent qualification intervals.

There are many web sites to gather this and updatedinformation:

w w w . s g a l i n k . o r g / f o r u m /messageview.cfm?catid=17&threadid=91 orwww.sgalink.org/fourm. Within this web site, there aremany individual’s names with e-mail addresses who havebeen directly involved in all meetings, public, audioconferences, and communicate with the public forgeneral information. Southern Gas Association offerscopies of all audio conferences as a product. This website offers copies all public meeting presentations,meeting agendas, meeting handouts and meetingsummaries.

h t t p : / / w w w. t s i . d o t . g o v / d i v i s i o n s / p i p e l i n e /pub_mtg_apr_03.htm or www.tsi.dot.govhttp://primis.rspa.dot.gov/oq/index.htmhttp://ops.dot.gov/new.htm.

Following are the associations representing industry andtheir web sites may offer other important information inthe future:

American Gas AssociationAmerican Petroleum InstituteAmerican Public Gas AssociationINGAAMidwest Energy AssociationAssociation Oil Pipe LinesNortheast Gas AssociationWestern Energy InstituteSouthern Gas Association

The final May 2003 public meeting in Washington DCwas cancelled.

Jesus Ramos


ADVANCED COMMUNICATION DESIGNBob Halford

PathTech, Ltd.Odessa, Texas

A part of the decision making process when selectingthe specific wireless communication devices for a projectshould be a complete understanding of all the challengesand needs of the project. If you have completed theTelemetry Questionnaire outlined in the Basic paper, youhave answered most of the questions necessary tounderstand the requirements needed from thecommunication devices available to select from.

Possibly the most important issue to understand andoutline is the number of times per day you will requestdata to be received and the amount of data to beacquired per these times. These are called poll timesand poll data size. If you want to poll only once a dayand only get a status report that takes only a few kilobitsof data, many options are open. This would includesystems that only receive a data response if there areissues to deal with such as high or low levels, open orshut valves, etc. These are called status changes.Devices are available to do this such as the CDPDmodems as well as the Burst Technology and Cellemetrydevices which operate on very low power and send onlyvery small packets of data. The value of these are in thesimplicity of installation and use. The backbone is takencare of by the cellular carriers. Where available, the buildout is such that getting on a system is not an issue.Wherever your cell phone works, you normally can usethe overhead part of the system for the Burst, Cellemetryor CDPD systems.

Even when deciding to use these simple systems, havinga thorough knowledge of where each site in your systemis located is of vital importance for many reasons. Now,when the project’s data requirements require a wirelessdata system capable of being polled from once a day toevery few minutes with data sizes from 100 kilobits to acouple of megs, obviously more care should be taken inselecting the data hardware device. Again, before makingthose decisions, understanding the needs andrequirements by completing the Telemetry Question areis most important if not absolutely necessary. You cannotdesign advanced systems before knowing where youwant to end up with the data and how you are going toget there. Just like the importance of a good foundationwhen building any structure, you must begin to considerthe design and installation of a wireless datacommunication system as designing a structure whichalso has to have a good foundation.

Just as you must have good blueprints to complete anystructure, you must have a blueprint of your wireless datacommunications systems before you start. It is best to

begin with the front end. How is the data going to bebrought in to a PC? Where is the PC going to be locatedthat is considered the Master Polling PC? Is the MasterPolling PC going to connect straight to the Master PollingData Radio Modem or will it remotely connect to theMaster Polling Data Radio Modem by the use of adedicated telephone line, satellite, Ethernet SerialConverter Server or CDPD? Will the data be shared withother offices, whether at that location or cities or statesaway?

Besides the known telephone lines, satellite and CDPD,Serial to Ethernet Servers are now available that caninterface between the Master Polling Radio Modem manymiles, cites or states away and allow for TCIP (Telnet)connectivity to the Master Radio Modem. You can thenget the data back over the company network or TCIPinto the Master Radio Modem which then polls all theremote radios and sends the data back through thenetwork to be shared by anyone with the right access.

We now have made the decision on how we are going toget the data and view that data at some master site orshared over many. What is next?The remote system must be designed to best maximizeresources and for the most efficient and consistantmanner in which to poll the data. How do we do this?

We must know the exact physical location of each site.Using portable GPS devices that can be purchasedanywhere is the most common method. You simply goout to each site and write down the coordinates inLatitudes and Longitudes. From now on, we will call thisthe Lats and Longs.

Having the lats and longs is just the first step. What doyou do with them? This is where the Remote SystemDesign Process Begins. Here is the outline you will useto go from basic raw data to a completed system design.To professionally complete a system design, you needtwo basic software packages. You first need a simpleMapping Software and then a Propagation or Path StudySoftware Package. The Mapping Software is about$70.00. The Path Study software is from $1,500.00 to$8,000.00.

WIRELESS RADIO MODEM DESIGN PROCESS

1. Acquire the lats and longs on a excel sheet. A latitudewill begin with the smaller of the two numbers. Example31 34 45.6. This is a Latitude. 99 45 23.6. This is aLongitude. Latitude is north and south on a map (up or


down) while Longitudes are east and west (right to left,left to right). In North American, the farther north you gothe higher the latitude number. North Dakota has alatitude of 48 degrees ….., while south Texas may be 26degrees…… . The farther west you go, the higher thelongitude number. Western California is 116 degrees to124 degrees while Maine is 67 degrees…. . You mustunderstand Lats and Longs to design wireless datasystems.

2. Use the notepad option found in the accessories ofany windows program to transpose the lats and longsinto a format that will be used to later Import the lats andlongs into a mapping software. Here is the example. Let’ssay that you acquired the lats and longs with your GPSand they look like this: 35 43 51.5 N 99 14 6.3w To typethis into the format that the Delorme Mapping Softwarecan import to the map you do type the lats and longslike this, paying careful attention to where the commas,spaces, etc. go. If it is done like this, you cannot Importthe Notepad lats and longs into the map. N35 4351.5,W99 14 6.3,SITE NAME. Whatever format you usedon your GPS to get the lats and longs is okay as long asyou get the spaces and commas correct and use the Nand W where they are located on this example.

N35 43 51.5, W99 14 6.3, BAKER COMP STATN35 38 11.0, W99 12 35.9, CUSTER COMPN35 40 28.8, W99 31 1.3, MERRICK COMP STATN35 46.80, W99 34.99, ROLL REPEATERN35.74776.W99.21572.OFFICE

EXAMPLE 2. You acquired the lats and longs in degreedecimal. 32 16.27 102 09.22. Still type just as you tookthe coordinates into Notepad like this: N32 16.27,W10209.22, SITE NAME. Again, make sure the N and W arein place correctly and use commas per the example.Nothing else will work correctly.

3. Before begging the design, first, on a separatenotepad, list the lats and longs of known or availablerepeater sites that may be used to aid in the designprocess.

4. Open the mapping software. I use Delorme Topo USA.You can purchase this at any chain electronicsbusinesses like BestBuy, Office Depot ,Circuit City, etc.,for about $70.00.

5. Select the state for the project from the discs availableand open. Use the directions to go to the geo area in thestate opened where your sites are located and zoom in.

6. Go to the Draw Option and click on the Flag. Thisbrings up a menu to use that will allow you to select anicon. Use the Yellow Box for Repeater Sites, Towers, theMaster Site, etc. These are the foundation sites for thesystem.

7. Go back to the Draw Menu and click again on theFlag and select the red flag icon. These will be the remoteslave sites you will link to the Master or through theRepeater sites to the Master.

Above: By importing the Notepad list of Slave Sites, theSlave sites (red flags) are not visible with the Repeater

Sites in Yellow.

8. Go back to the Draw Menu item and click on the \ lineicon and highlight this. You are now ready to start tryingto Link the Sites back to the Master Site. Start with thosesites close to the Master and work your way out. Youfirst left click on the Master and let the mouse buttongo. You then drag the line from the Master to the SlaveSite until you see a small highlight on the bottom of theslave site (bottom of red flag).

9. You can now view the terrain between the two sitesfor a general idea of what the link might look like, or thegeneral feasibility of the link. Once you feel the link isfeasible, you are ready to do a professional path studywith the use of the path study software. I use PathLoss.There are several affordable ones on the market.Micropath is another easy to use and affordable package.


Above: All the sites have been linked back to the masterthrough repeaters which were linked first to the master

site.

You can review both or more on the internet using asearch for radio propagation software, path studysoftware, path analysis, etc.

Above: The Delorme allows to get a preliminary view ofhow the path might look.

10. Follow the directions on data insertion into the PathStudy Software and complete a link study between theMaster and Slave icons on your mapping software(Delorme). If you have a good path using the professionalsoftware, you have a link. You will complete this processover and over again until you have a link between eachslave site to the master or through the repeater orrepeaters on the system. Many times it takes goingthrough multiple repeaters to get from the last slave siteto the master.

11. Once you have completed Path Studies on all sitesin the project, that has all links joined and linked in someway back to the master. Your project is now ready for aSystem Architecture Drawing.

Above: A path study using professional software,confirms if a good path exists and also indicates how to

point the antenna correctly. This is the Azimuth.

12. A System Architecture Drawing begins with a box atthe top of a page that represents the Master PC.Remember this may or may not be the site of where themaster radio is located. If this represents the companyNetwork site where a Master PC is located that willremotely poll the Slave through a Master remotely locatedfrom the Master PC, then the first box is the MasterPolling PC and a description of the connectivity betweenthis site and the Master Polling Radio Modem. (REFERTO APPENDIX A FOR THE ARCHITECTURE DRAWING).

13. Continue to go from the Master Polling Radio Modemout to the first radio modems or repeaters linking eachsystem change as you go. Draw arrows between theboxes to indicate how the linking system works its wayout from the start to the finish of the system. All dataradio products have unique capabilities that allow forstore and forward, etc. These options are what allow forthe more complex systems to be developed in the mostefficient and cost effective manner. Spending time inresearch here is invaluable.

You must first understand the complexity of the systemyou have now designed before you can understand ifthe product you are going to select has all the optionsthat will be required to bring back you data from thefarthest site in the system back to where the data canbe viewed. Completing the Design Process above is theonly way to completely understand these needs. Eitherthe end user or a company that specializes in data designmust complete this process for the best system availableto be completed.

14. Understanding how to program each remote slavesite is important now. The company representative thatis providing the remote communication device shouldnow come to complete a seminar and teaching process

PATHTECH, LTDLTD Ma r 23 03 BJ

PATHTECHECH, LTD

ROLLOLL REPEATER

LatLatitude 35 46 51 .90 NLonLongitude 099 34 59.70 WAziAzimuth 115 .31°Elevation 231 4 ft ASLAntenna CL 45.0 ft AGL

CUSCUSTER C COMP

LatLatitudude 35 38 11 .00 NLonLongitude 099 12 35.90 WAziAzimutmuth 295 .53°Elevatition 181 4 ft ASLAntenna CLCL 45.0 ft AGLAGL

Frequency (MHz) = 915.0K = 1.33

% F1 = 10.00

Path length ((23.24 mi)0 2 4 6 8 10 12 14 16 18 20 22

16000

17000

18000

19000

20000

21000

22000

23000

24000


Bob Halford

for your field techs. I highly suggest making up a RadioProgramming Template for whatever device selected.From the System Architecture page, complete a programpage for each radio system change from the beginningto the end of the system. Save these pages in a binderand in a File to be used whenever needed. Use theseprogram templates per each system radio type to beused in the system to pre-program the test radio andthen all radios in the field before they are installed. Thereare times, when the program is completed on site. Thetyped program page is especially important at thesetimes.

15. Once you have selected the radio modem hardwarethat has the options need to best complete your systemdesign, pre-programming and testing the Master PollingSoftware and any Network Interface or Ethernet networkservers needs to be done before the first remote radio isinstalled. This is the time to install your polling software,whether it is the Field Measurement Software of thedevice installed on your remote sites or special MMIpackages available, now is the time to make sure youcan talk to the end devices. Set up the test measurementor monitoring device you are using at the remote fieldsites where your field techs can install the product pollingsoftware or MMI interface software to a Master PollingPC. If any other interfaces are going to be used, hookthem up to the system. Install a Master Polling RadioModem to the Master Polling PC or to the Interface Serveror other device such as satellite hardware or CDPD. Havethe end device company technician set up test numbersin the end device and have your company IT person testpoll the end device until you know that all set upparameters are understood. Pin outs are seriouslyimportant to know between all devices. Baud rates,protocols, and parities must match or you will notcorrectly poll remote sites or get the data through all theinterfaces.

16. While this testing process is going on, it is the timeto make sure all repeater sites will be ready for the systemto be installed. If there is a need for antennas to be higherthan 50 ft or so, plans need to be made to ether leasetower space or contract out for towers to be installed onthe recommended repeater lat and long sites.

17. Now is the time to make sure what the engineeringsaid and what is real both agree. Go to the sites andmake sure before you do the final investment on thosetower installations or install expensive coax and antennason lease towers that no man made obstacles exist whichthe software did not see. Hidden interference issues suchas high gain paging and some cell sites issues cannotbe seen until you install the system, but make sure theobvious is taken care of.

18. The Repeater Site towers and any needed coax andantennas are now installed and tested. Test these with atest radio installed and the polling software which youwill be using. Make sure you are getting the best “link”possible. If problems exist between some individual sites,now is the time to solve this. How do we do that?

19. All good radio modems now offer options that willhelp direct the path of the data through further sub-netting programming techniques. The important thing isto plan ahead and understand the needs of the projectfrom beginning to end.

SUMMARY

The needs, requests and requirements of those usingwireless data devices are what continue to drive thecreation of the new options available for wireless datamodems. The fact is that those who have access to thebest data, have the tools needed to make timelydecisions in oil and gas that translates into profits andgrowth.


APPENDIX A:SYSTEM ARCHITECTURE

MASTER FK 2 NET ID 2142

RX 0 TX 0

BAKER CO MP SLAVE / REP FK 2 ID 2142

RX 0 TX 1

CUSTER COMP SLAVE / REP FK 5 ID 2142 REP FREQ 1 RX 0 TX 2

TRANSOK SLAVE

FK 2 ID 2142 RX 0 TX 1

BAKER SLAVES BROADBENT 7-25

BAKER 1-31 BAKER 5-31

KILLHOFFER 2-23 WALKER 2-3 FK 2 ID 2142

RX 1 TX 0

CUSTER COMP SLAVES MULLINS 1-34

BAKER 2-4 WALKER 2-35 WALKER 3-35

GOODALL 1-35 CARPENTER 1-28

FK 2 ID 2142 RX 2 TX 0

MERRICK COMP STAT SLAVE / REP FK 2 ID 2142

RX 2 TX 4

ROLL REPEATER REPEATER

FK 6 ID 2142 REP FREQ 1 RX 2 TX 6

BEALS CDPD SLAVE / REP FK 2 ID 2142

RX 6 TX 7

LOVETT COMP SLAVE

FK 2 ID 2142 RX 7 TX 0

DEAN COMP CHEYENNE COMP

SLAVE FK 2 ID 2142

RX 6 TX 0

MERRICK SLAVES SMITH 4-30

HAMMON COMP FUELDUPREE 6-36

MENNONITE 2-31 FK 2 ID 2142

RX 4 TX 0

MENNONITE 3 -31 SLAVE / REP FK 2 ID 2142

RX 4 TX 5

MENNONITE SLAVES DUPREE 5-36 SMITH 3-30

SLAVE FK 2 ID 2142

RX 5 TX 0

HUGHES CF SLAVE / REP FK 7 ID 2142 REP FREQ 1 RX 2 TX 3

HUGHES SLAVES FLICK CF

REDMOON 225 MOSELEY 125 FK 4 ID 2142

RX 3 TX 0


D.O.T. TITLE 49 REGULATIONS FOR TRANSPORTATION OFSAMPLE CONTAINERS

Tom WelkerHydrocarbon Quality Associates

13806 Placid Woods, Sugar Land, TX 77478

During my travels around the United States talking aboutsampling and sample containers, it has come to myattention that the oil and gas industry in the United Statesneeds to be a little better informed on proper handling,shipping, and transportation of sample containers of alltypes. Since everybody in the oil, gas, and chemicalindustry seems to be involved in taking samples andhandling sample containers, it behooves us tounderstand the laws and rules that govern theirtransportation.

The department of Transportation (D.O.T.) Title 49 coversthe rules and regulations for the manufacture, handling,and transportation of sample containers of all types.Whether you use specially-built sample containers, oldhomemade sample containers, old World War II oxygenbottles, gigantic sample containers, or very smallcylinders, if you are transporting those sample containersin your vehicles or you are shipping them by commoncarrier and they have hazardous materials in them, youmust be aware of the rules that govern the handling ofthose cylinders.

This paper is for information only and certainly not meantto be the last word on the use of sample cylinders as faras transportation goes. Each company’s own legaldepartment must have their own interpretation of theserules and abide by these rules. If local, state, or countryregulations are in force, they must also be adhered to.However, the D.O.T. has the final say and certainly is themost stringent in many, many cases. So, consequently,you must understand how important it is that the handlingof sample cylinders must be carefully done.

The actual purpose of D.O.T. Title 49 is to ensure thesafety of the public, as well as the people directly involvedin handling hazardous materials in the sample cylinders.Also, we put this information together so that you, as anindividual, and your company can avoid penalties thatmight be imposed by law enforcement agencies and/orthe D.O.T. inspectors themselves for impropertransportation of these sample containers.

In January 1988 the D.O.T. informed the local lawenforcement agencies that anyone who had the authorityto write a citation for travel violations or anyone in thelaw enforcement business, could also write a citationfor the improper transportation of sample containers. Thismakes the enforcement arm much more encompassingthan it ever has been before. Up until that time only theD.O.T. people themselves were concerned about thesesample containers. For your information, however, these

law enforcement people around the United Statesnormally are not aware of most of the rules that governthese sample containers, nor are they aware of the finesand penalties that can be involved; and, consequently,many people would not necessarily write you a ticket ifyou were stopped.

However, so you will understand the severity of thesituation, the fine for improper transportation of a samplecontainer can be as high as $25,000.00 per samplecylinder for the company and as much s $500.00 for thedriver of a vehicle. Now remember, this is for each samplecontainer, not for a lot, or a bunch, or a wad, or a group,or a box full, but for each sample container that isimproperly transported. Because of this, it certainlybehooves us all to understand this law and comply withit.

As an example, the compliance to D.O.T. Title 29 involvesadherence to the following sections:

1. Hazardous Materials Table 49 CFR 172.101. TheHazardous Materials Table describes what you havein the sample container and how to handle it; whetherit is a hazardous flammable liquid or a flammablegas, a poison, radioactive, and other classes.

2. With each sample container, there is a requirementthat Shipping Papers are covered in 172.200. Onthe shipping papers, there must be informationregarding the proper shipping name, the hazard class(flammable liquid, flammable gas), the U.N. numberor identification number, the quantity that is in thesample container, and as of January 1991, anemergency response number must be on theshipping papers. All of these are covered in 172.101.The emergency response number is 172.602. Theemergency response number is one of the newestadditions to the shipping papers and should be verycarefully looked at to determine its’ exactrequirements for your organization.

3. Packaging Reguirements. Each sample containerhas specific packaging requirements. Therequirements vary as to what is in the samplecontainer, how it is being transported, and what its’design and construction amount to. Those are alsocovered in 172.101. They include general specificexemptions and definitions for packaging.

4. Marking Requirements. The marking requirementsfor sample containers (flammable liquid, flammable


gas, etc.) is covered in 172.300.

5. Labeling Requirements. 172.400

6. Placarding Requirements. 172.500. At the writing ofthis paper, for general transportation of samplecontainers the way we will talk about it in thisparticular instance for individuals collecting samplesfor their own companies or small volumes ofsamples, placarding requirements have beendelayed, so you will not have to have our vehicleplacarded.

7. Federal Motorcarrier Safety Regulations 39 CFR 390-337.

In addition to these specific applications of D.O.T. Title49, if you are a if a user of a sample container, (a standardsample cylinder with 1 or 2 valves - not the special pistonstyle sample container, but the rolled-in standardcylinder), if that cylinder is over 4-1/2" in diameter and/or over 12" long, that ample container must have a reliefvalve. If you are using constant pressure samplecontainers, the CP cylinder must have a relief valve onboth ends. The relief vavles on sample containers mustconform to Compresses Gas Association publicationsthat are sample containers, sample containers must beenclosed in some kind of carrying case, box, orsomething that will protect that sample container in caseof an accident. This protection can also include capsand collars on the larger cylinders that you may beinvolved in using. That is also legal.

It is also a requirement that sample containers used inliquid service should never be filled more than 80% full.It is against the law to transport a sample container filledwith liquid completely full.

Who is affected by this? As you can tell, anyone whohandles these sample containers gives them to someoneelse to handle, offers them for shipment, or is involvedin transportation is involved in the D.O.T. Title 49compliance.

One of the most important considerations when talkingabout D.O.T. Title 49 is to understand that whenever asample container is moved off of your property on toany state, county, city, township, or local highway, youmust be in compliance with D.O.T. Title 49 and localregulations.

The following is a checklist for your complianceinformation:

1. Determine if the material that you are trying to shipis regulated.

2. Determine the quality to be shipped.

3. Determine the proper shipping name. Do not usecommon names. Get the name of the product out ofthe D.O.T. Table.

4. Determine the hazard class — flammable gas,flammable liquid, others.

5. Determine the identification number. This is the U.N.number that is also in the Table.

6. Determine the proper packaging. Remember,sometime in l992 or later there will be a requirementthat, in order to comply, a package must beapproved. At this writing there is no requirement, noris there even information on approved carrying casesfor sample containers.

7. Determine the proper marking or label and applythem to the outside of the sample container. If samplecylinders are transported individually, each samplecontainer must have an appropriate tag and/ormarking, and the proper paperwork must be filledout for each sample container.

8. Prepare the proper shipping papers. Remember,each sample container must have shipping papersfilled out on it and filled out completely. Attached tothis article you will find a typical shipping paper thatis just that. It is typical. You can design your ownshipping paper, but you must have papers that arefilled out for each sample container.

As a note of interest, if you have a product in your samplecontainer that meets more than one hazard classification;for instance, sour gas is not only a flammable gas but itis also a poison, it must have both labels on the carryingcase. It must also have that as a part of the paperwork.

Each package containing sample containers offered forshipment or transported in your own vehicle must havethe owner’s name and address on them.

The person shipping the container must furnish andattach the labels prescribed for in the shipping papers.

General requirements for transporting compressed gascylinders is covered in 173.301. Three of those specificrequirements are (1) gases shipped together must becompatible, (2) pressure in the sample container at 70ºF.shall not exceed the surface pressure of the samplecontainer, and (3) pressure in the sample container at130ºF. shall not exceed 125% of the surface pressure ofthat sample container.

Another note: It is against the law to transport or offerfor transportation sample containers filled with hazardousmaterial that are not built to D.O.T. specifications. Anysample container that is offered for transportation musthave the information that is required by the D.O.T.stamped on that sample container. If re-testing is requireda re-testing date must also be stamped on the samplecontainer.

The rules that govern the transportation of samplecontainers also apply to a sample cylinder that is madeoverseas, or a foreign-made sample cylinder. If they are


offered for transportation or transported by you in yourown vehicle, you should be very cautious to be sure thatthey meet the requirements for the use of foreigncylinders.

This information is a general overview of the regulationsand should in no way be used as an interpretation byyour company of the requirements for D.O.T. Title 49.Each company must interpret the rules for themselvesand adhere to those as best meets their needs. Since

the Title 49 regulations are updated periodically, careshould be taken to ensure that your company has thelatest revision of the rules and complies to each one ofthose rules. I am not an authority on the transportationof sample containers; however, I disseminate thisinformation for just that. Simply for your information - tomake it as easy as possible for you to comply with therules that govern the transportation and shipping of yoursamples from the field to the laboratory. If we can assistyou further, please contact us.

Tom Welker


SCADA AND TELEMETRY IN GAS TRANSMISSION SYSTEMSChris J. Smith

Invensys Production Management, The Foxboro Company38 Neponset Ave., Foxboro, MA 02035

ABSTRACT

Modern business and security imperatives coupled withrapid technological change require key new architecturalelements for SCADA systems These elements arediscussed along with more traditional block diagramfundamentals, so that the reader might better understandmigration and adaptation strategies for their transmissionpipeline operations in the new millennium.

INTRODUCTION

Some SCADA system overviews might start with ageneral block diagram of a SCADA system. I’m taking astep back here and will start with the SCADA system asit appears in the business of the operation. In the newmillennium, recognition of SCADA as an importantfunctional unit in the overall business supply chain makessense, as more and more business systems are moretightly integrated to the SCADA system. SCADA systemsprovide an operational platform to control and monitorthe pipeline. The classic SCADA architecture consistedof a number of servers and network elements designedto provide a real-time representation of the field and itsdata as well as a control platform for closure of valves,and operation of pump or compressor set point.

An important step in the understanding of SCADAsystems technology is to realize the placement of theSCADA system in the overall business and operationalstrategy of the pipeline. Pipelines are assets that needto be monitored and protected. The product transferredfrom supply to delivery point provides the economicjustification of the pipeline.

The tradition of SCADA involves continuous and report-by-exception scanning of low-level data from RTUs, FlowComputers, Gas Chromatographs and PLCs. This datais elaborated with locally and remotely hostedconfiguration information to become basic real-time,historical and measurement data for the operation of thepipeline.

So in a SCADA system all the technology required toget the data from the field, elaborate the data, store it,display it, alert the operator, and provide real timeinformation to applications is provided. The SCADAsystem is an essential component in normalizing andcentralizing data for the operational and businesssystems that depend on it as well as the basis forprotection and control of the pipeline asset over its

lifetime. At a higher level, once information is available,energy costs can be optimized relating to the actualtransfer of product in the pipeline.

This paper looks at the variety of engineering andbusiness requirements, which form the basicarchitectures of SCADA relating to gas transmissionpipelines. The paper is an overview, it will raise morequestions than it answers, however it should provide aframework for understanding and development ofknowledge in the field for those who are possiblynewcomers to the field.

FIGURE 1.Pipeline Product and SCADA Information Supply Chain

FIGURE 2.Functional Elements of SCADA Block Diagram


BLOCK DIAGRAMS

The SCADA master is set up to provide a workingoperational interface to allow operators, and mastersequence controls to create on-demand supervisorycontrols which—when transmitted to the RTUs, switchplant in or out of service and adjust operational conditionsto suit the requirements of the moment. This relativelysimple task involves the use of a wide range oftechnology and interfaces spanning large distances.Now looking at the functional block diagram in figure 2,it is generally designed to show the physical connectionand communications from the field through to the controlroom. When specifications are written and quoted, thisis generally what vendors and purchasers use to gain arapid understanding of the overall physical requirementsand communications interfaces of the SCADA system.Key elements in Figure 2 are: -

• Engineering Workstation• Operator Console• Dual Database Servers• Applications Station, Office PCs• Dual Local Area Network• Communications to field devices, PLCs, RTUs• Diverse set of field devices such as flow

computers, PLCs for compressor stations, andremote terminal units

You do not get a great sense of ‘why’. What are therequirements this block diagram satisfy? Connectivity isone; with this you can see what is connected to what.Availability is another. With this you can see, if an elementfails, which other element might take its place. Note thedual network and diversity of communications. In thiscase there is one communication path to a remoteterminal unit. There are often two, a main and backup.Communications paths are expensive, so often thebackup is high cost, possibly a dial up line. Consider analternate block diagram Figure 3 In this example a pigreceiving station. Traditional SCADA system architectureswere developed to account for significant data speedand reliability problems associated with low speed longdistance analog radio telemetry and transmission. Radiocommunications are still used today, however much moreflexible and high bandwidth solutions are available withthe use of high-speed microwave, VSAT, and fiber opticcommunications. The advent of TCP/IP protocols andtheir logical connectivity, has allowed multiple applicationrequirements to coexist in the same communicationspathway. A layered communications protocol requiresmuch more bandwidth, but in turn allows for much moreflexible implementations and diversity of technology.

So the block diagram tells us about performance,reliability, connectivity, availability and partition of functionin the control room. SCADA system designs in the 20th

century had to contend with reliability of computing andcommunications equipment and pathways. Lowbandwidth radio channels were used and thus blockdiagrams depicted solutions to the problems of the era.

Another important point is that the main functional focusis one of operations and engineering. These are the basicrequirements of SCADA. A system that will continuallyprovide and operations interface for day-to-dayoperations and protection of the pipeline and pipelinenetwork.

BUSINESS AND OPERATIONAL REQUIREMENTS

FIGURE 3.Pig Receiving Facilities Block Diagram

FIGURE 4.Example Supplier Consumer Pipeline Network

The simple pipeline network in Figure 4 represents atypical business oriented arrangement between anumber of suppliers and consumers of Gas. In this caseconnected by an undersea pipeline of about 150 miles.The Gas Transmission Company develops a SCADAsystem for the above network to satisfy the followingbusiness oriented requirements.

• Monitor, analyze quality, quantity of sales gasfrom each receiving point up to delivery point

• Monitor and control the contracted qualitythroughput (gas quantity and quality) deliveredto gas buyer (s)

• Manage hydraulic capacity to accommodate gasbuyer’s swing requirements in real time.


• Develop gas balance system for determiningdaily/weekly/monthly/quarterly/annual gasreceipts and deliveries.

From this business-oriented set of requirements, a furtherset can be developed: —

1. The software must have capability to collect datafrom Gas SCADA, Gas metering, GasChromatograph and instrument analysis.

2. From entire data, the software must produce:• Phase envelope analysis every 5(five) minutes• Hydrate formation Detection and location.• Composition tracking• Pipeline transient modeling including:

o What-if analysiso Look ahead analysiso Survival timeo Line pack calculationo Process analysiso Leak Detectiono Pipeline optimizationo Fire & safety analysiso Pipeline optimization

3. Metering and Tariff Calculations 4. Management reporting

• Daily• Weekly• Monthly• Annually

It would be reasonable to assume the block diagram inFigure 2 could be applied to help solve the businessrequirements in Figure 4. In engineering SCADA systems,there are many more architectures involving datastructures and interfaces, which would couple these twosets of requirements.

Engineers often contend with the data flow and timingbetween applications to provide a gradual value add ofinformation relating to the pipeline state and the productsentrained commercial value and composition. Engineerswould require a data flow diagram like the one in Figure5 This diagram shows the many interactions and datapaths for data in a typical Gas pipeline SCADA system.The general grouping of applications is that of Real timeSCADA applications, an Applications database, ameasurement system, a real time modeling system, anda nominations tracking system. All these systems wouldgenerally run on the database and applications serversprovided at the master station as shown in Figure 2

What Figure 5 depicts are generally the static andessential applications, often provided in the SCADAsystem and which would form the economic basis ofoperations in the system. There are many otherapplications that would be developed by the engineeringand operational teams to aid and assist their work duringthe lifetime of usage of the SCADA system. This bringsus to an important dynamic in information technology.

Constant change. How is this then to be handled in statichighly available systems?

In this paper I want to quickly bring to your attention,problems of the 21st century just starting. This will bringyour overall understanding of SCADA systems to arealistic level for understanding the systems you will beseeing and working with. The requirements of availability,reliability and connectivity of Figure 2 are still importantand provide the highest-level requirements of SCADAsystems today.

SCADA SYSTEM TECHNOLOGY LIFETIMES

The case we are building here is one where it can beseen that the development of SCADA systems is acontinuous process. The lifetime of some applicationsmay be very short; the lifetime of the pipeline may bewell over forty years. How do these two factors affectthe operation and development of SCADA over thelifetime of the pipeline?

Somewhere between the Database server and the fieldequipment is the point at which internet access andcommunications flexibility is required, which would helpto match the lifetimes of measurement equipment andfield installations and the SCADA data servers andenterprise demands.

FIGURE 5.Logical Data Flow Diagram for Gas Applications


ENTERPRISE INFORMATION DEMAND

This increasing demand for information relating to theoperational and maintenance aspects of pipelines hasoutpaced the ability to upgrade field device installationsand communications networks to supply the data. De-regulation, business mergers and acquisitions, andsupply chain optimization require a point of flexibility tobe created for supply of operational and field data. Thispoint becomes a provider of a variety of data relating tothe plant and equipment in the field, and therefore thecontrolled process. Whilst the revamp and extension ofplant equipment may evolve slowly, the pace of businessand business model change is extreme and its demandsfor data are only growing, not diminishing.

Whereas at some point in the near past, the demand forSCADA data was seen as only few clients, adoption ofthe internet has given new meaning to client server,leaving the server or data source, with now middle tierarchitectures, middleware, and web integrationtechnologies to deal with before actually reachingpossible the ‘ultimate thin client’ a voice activated controlsession on a normal telephone. There is however a clearpoint of demarcation which we can describe in terms ofindustrial electronic measurement technology and therealm of the Internet. For this purpose the Blue WaterBlue Sky Line in introduced as in Figure 7.

The rapid pace of change in Internet technology is incontrast to the longevity and amount of installed industrialcontrol and measurement equipment in the field. SCADAsystem components like RTU s and Flow computersonce installed tend to stay there and capture the stateof the art of SCADA and measurement circa theinstallation date.

SCADA designers have had to come up with ways toconnect the myriad of now almost museum gradetechnology installed since the 1970s boom in SCADAwith the newer technologies whilst still providing the high

levels of security, availability and now, operationsresponsiveness demanded by the fast changing internetfueled business environment

An important realization in SCADA is generally in traditionalarchitectures, there is an information hierarchy. Theinformation is drawn up to a master station platform. TheMTU provides the only real access point or gateway tothe information. In the discussion on security this conceptis important. Figure 8 shows what I call the SCADAspace.

FIGURE 6.Technology Lifetimes in SCADA

FIGURE 7.The Internet vs. Electronic age

FIGURE 8.Defining the SCADA Information Space

Figure 3 details key components of a pipeline SCADAarchitecture located in relation to cyber andSCADAspace. The main communications pathways arealso indicated. SCADAspace is defined as that area ofinformation and technology, which is essentially privateand not accessible to or from the public Internet.Cyberspace on the other hand is defined as the generalrealm of the public Internet and the network of computersand gateways serving it. One can see that in theSCADAspace, devices are generally fixed with low-level


data structures, communications and informationtransfer protocols. In general the industry has sufferedfew security breakdowns due to the essentially privatenature of SCADAspace.

Now you can see that the Internet is represented inFigure 8 as Cyberspace. This is because it is generallynot a space under the control of the SCADA engineer,designer, vendor, or anyone else. Whereas theSCADAspace is architecturally controlled by the pipelineoperations people. Cyberspace is not. So why is it there?Why access it? Business drivers as well as thetechnology of the Internet are forcing engineers to includeaccess to the Internet in more and more components oftheir SCADAspace. Business oriented measurement andinformation was often handled by means of paper reportsand accumulations transferred from the SCADAoperations control room to the one or two clients whoneeded it.

SECURITY DEMANDS AND STANDARDSEVOLUTION

Since the 1990s, SCADA systems and cyberspace havebecome more widespread and sophisticated. As thesecurity of infrastructure such as pipelines and utilityequipment and services is now governed in part by theDepartment of Homeland Security in the USA, thedevelopment of component strategies such as those forSCADAspace is an important link in the strategy. Allinvolved in the SCADA industry are, in part, responsiblefor its development. Initiatives associated with the newAGA Standard AGA12 are trying to capture theserequirements for pipeline infrastructure.

Market forces are shifting the reach of cyberspace furtherinto the low-level data hierarchy of SCADA. The adventof disaster recovery, web-enablement, and PDAs laptopcomputing, means that SCADA products will have ahigher degree of Internet capability and servicesexposure at their access ports. Communicationstechnologies are rapidly improving to allow a higher levelof bandwidth and quality of packet information to betransmitted to the field. All of this brings the relativelyinsecure cyberspace into stronger and more elaboratecontact with the SCADAspace.

The public communications system remained vulnerable,as did open analog communications, to a determinedattacker. Successful attacks required some insiderknowledge of the SCADA configuration and addressschemes as well as in depth knowledge of the functionand operation of the station equipment, programmablelogic controllers (PLCs) and RTUs.

Encroachment of cyberspace on the general operations,maintenance, and control of SCADA components isincreasing in response to other user demands for SCADA.New business processes now demand accessibility andvisibility of data at lower levels of the plant informationhierarchy. The proliferation of laptop computing, windows

and PDA equipment, brings permanent and intermittentcontact with cyberspace into the general operation ofSCADA and the once secure SCADAspace. Anotherdevelopment in the industry is the use of DisasterRecovery servers, and VPN channeling ofcommunications, which use the Internet as the maintransfer of information and control from site to site. Thus,the ease of use and flexibility of the Internet andcyberspace tools and techniques in SCADA operations,bring the attendant risk of security breaches and accessto “deny SCADA service” to the pipeline infrastructure.Concentration and Aggregation of data in the field

Flow computer and or Remote Terminal Unit data caneither be aggregated in the field or directly polled by amaster station. Significant advantage can be gained bythe installation of intelligent data concentrators in thefield these would provide the following capabilities: -

• Give the remote site an IP Address• Direct connect 10BaseFl, Dial-up connections• Local polling and radio control in the legacy

‘electronic age protocol’• Multi-protocol handling• Data normalization• Report by exception to multiple masters with

independent IP connections• Remote configuration and download• Brokered secure information to non controlling

client applications such as measurement andasset data handlers

• A secure access point for low level data handling[AGA12]

The application of Internet Protocol (IP) and (IPsec) baseddata concentrators with speed, protocol, and datatransformation features provides a pivotal point at whichthe legacy SCADA system components can be

FIGURE 9.Current level of Cyber-SCADAspace Interoperation

(2002)


incorporated into a more modern Internet based system.These pivotal concentration points can be drawn on theimaginary BWBSL in the communications architecture.The further into the field these devices can be placed,the more enhanced the visibility reach and security ofthe replacement SCADA system.

In this architecture, the scanning and communicationmodes below the BWBSL stay relatively fixed. Abovethe BWBSL alternate communications pathways can beestablished in real time to alternate centers and withOPEN standard protocols compatible with the Internetand standard networking equipment.

In modernizing SCADA system new data concentrationequipment placed provide the following functionalitydesired by the newer architecture.

Communications Path DiversityWhereby data trapped within a single or multi-drop path‘owned’ by a particular channel master is made availableto one or more masters using TCP/IP connections atthe data concentrator.

Information Diversity and speed matchingWhereby disparate types of information and recordsgained from multiple scanned legacy devices arenormalized and made available to higher speedcommunications networks.

Protocol DiversityWhereby legacy protocols are managed by the dataconcentrator, often in polled mode on local privatenetworks, and thence after information normalization,made available to higher-level Internet age protocols anddata messaging systems over generally leased andpublic communications networks using Open SCADAprotocols to multiple masters.

Transparency and Pass throughWhereby file transfer aspects of legacy equipment andremote configuration functions can be accommodatedby a pass through mode where the legacy protocol ishandled over supporting internet transport andconnection layers like TCP/IP.

STATION COMPUTING DEVICES

High power low cost computing technologies make itfeasible to configure data concentrators for deploymentto the field. A number of key features and benefits ofthese are described here.

Features• Dual TCP/IP connections• High performance processors• Memory > 16MB• Flash permanent memory- no moving parts• Non Volatile Memory

• Environmental durability• I/O Capability• Central database• Broker Technology• Variety of legacy communications media and

serial connectivity• Configurable protocol stack selection and

assignment• Low cost

Benefits• Enterprise access direct from the concentrator

position• Measurement system access direct to the

concentrator• Maintenance system access direct to the

concentrator• Better security of open access• Alternate control room access to data• Ability to service rapidly changing data needs

and quantities• Ability to incorporate standard or common

information models close to the legacy sourceof data thus improving data consistencybetween clients

THE DISASTER RESPONSE

Newer SCADA architectures are able to transfer responsefrom building to building, state to state, and country tocountry. The ‘disaster’ is generally defined to besomething that prevents access to, disables, or destroysthe main SCADA master station and communicationsinfrastructure.

With the Internet age, the concept of operate fromanywhere is now realized. Software operator consolesessions can be created at the master station which mayprovide for visualization on a web browser from anywhereon the Internet.

Uncertainty in the connection and speed over a publicInternet which itself may be impacted by the disastermakes the use of web based operations difficult tosustain as key steps in an emergency operations plan.This difficulty means that emergency operations plansshould be able to stand up without total reliance on webbased remote operations.

The following general themes need to be considered inregard to disaster response and recovery.

• Transfer of Master Operations on reliable leasedcommunications links.

Communications to a Control room building are probablythe most critical element for transfer. If the controlbuilding can be networked to its communicationsequipment then that equipment should be reachable froman alternate control center. The placement of intelligentIP connected data concentrators in the field at key


Chris J. Smith

locations helps provide a means of communicationstransfer. The main and alternate master stations haveindependent and simultaneously active communicationpaths to the data concentration equipment. Health andreadiness checks on the alternate communicationspathways are required to ensure an emergency as wellas restoration path capability is maintained.

• Transfer of Security to operate

With the extension of pipeline market areas by acquisitionthere is a requirement to handle a large number ofoperationally distinct areas, where individual personnelin those areas are qualified and with permission tooperate. These areas of responsibility might overlap andchange during the transfer of operations under disasterconditions.

The master station software and possibly the dataconcentration software must resolve the permissions ofits clients for the area of responsibility both before duringand after the transfer of control. With more openprotocols and brokers allowing multiple subscribers andmasters, arbitration as to the permissions is an issuethat needs to be resolved as part of the design. It is nolonger sufficient to assume that logon permissions coveroperational permissions and that operational permissionsrelate directly to servers hosting operational sessions.

• Transfer of Engineering and Configuration

With the proliferation of numerous intelligent devices inthe field, management of engineered solutions andconfigurations is becoming increasingly difficult. Theincorporation of file access and transfer to each intelligentdevice from a single repository is key to the successfulmanagement of operations philosophy and equipmentsecurity.

• Getting back to normal…

An important point for SCADA software requirements isthe ability to restore operations as well as handle thefailure. A two-step process is used where the restorationof the operational system precedes a data re-synchronization phase and then final transfer of controlback to its normal case. Care must be taken with transferof operational permissions in the sequence.

SUMMARY

This paper provides a general overview of SCADA andtelemetry and its use in the transmission system. Manybroad concepts of architecture and design areintroduced. Specific details of newer data concentrationdevices are introduced, as these newer architecturalelements can aid in decoupling the business and disasterrecovery problems from the long lifecycle field equipmentmeasurement and communications infrastructure.Understanding of these concepts will bring the readerup to date with the requirements, market directions, and

possible solutions available to the SCADA engineeringcommunity today.

AUTHORS BIOGRAPHY

Chris Smith M.E: University of NSW, Sydney, NSW,Australia. Mr Smith has been working in the SCADA Oiland Gas applications and modeling field for INVENYSsince 1990. He has an extensive technical backgroundand has held the position of SCADA system architectfor Invensys. He also has extensive experience with real-time electric power SCADA applications and modeling.Currently Chris is the Invensys Channel and Marketingmanager for Invensys SCADA and resides in Foxboro,Mass. USA Email: [email protected]: [email protected]

REFERENCES 1. C. J. Smith, Disaster Recovery in Pipelines, Article,

World Pipelines, May June 2002 2. C. J. Smith, Connection To Public Communications

Increases Danger Of Pipeline Damage FromCyberattacks, Article, Oil and Gas Journal, February2003

3. Dr A Stanford-Clark, Integrating monitoring andtelemetry devices as part of Enterprise InformationResources March 2002, Websphere MQ development,IBM Software Group

4. W. Rush – Gas Technology Institute, AGA12-1 CanReduce SCADA Cyber-Attack Risks at low cost,A presentation to the AGA Operations Conference,April 28th 2003


TERMINOLOGY USED IN INSTRUMENT ACCURACYRick Williams

Rawson & Co., Inc.Address

The purpose of this paper is to offer a brief explanationand discussion of many key terms used in describinginstrument accuracy. The terms included within thisdiscussion are most commonly used for definingperformance standards with primary sensing elementstypically used in the measurement of flow, level pressureand temperature instruments. Many of the terms usedmay apply to controllers, recorders and final controlelements. However, the focus provided herein is theprimary element device. The specific devices includetransmitters (differential pressure and temperature) andflow meters (e.g. magnetic, vortex, turbine, variable areaand positive displacement).

This paper is written for the benefit of the typical user ofinstrumentation products to include instrument engineersand technicians. A comprehensive discussion ofprecision measurement must address calibration andtraceability issues. The scope of this paper will be limitedto discussing the terms associated with the applicationof instruments rather than addressing the issues ofmaintaining accuracy. Calibration is a key issue indetermining the continuous performance of aninstrument and is worthy of discussion. However, it iscomplex enough to warrant a separate discussion.

A fundamental understanding of gas laws and the effectsof compressibility are necessary for selecting andapplying instruments used for gas measurement. This isobvious since the majority of the flow measurementdevices used in industry are volumetric, and volumechanges under actual conditions. We will begin ourdiscussion with the assumption that this baseunderstanding exists. There will be no debate regardingthe merits of direct or indirect mass measurementdevices nor will there be a comparison of the advantagesof mass measurement over volumetric. The decision touse any specific technology for a solution must beweighed upon the need for mass measurement vs.volumetric, the accuracy requirements from themeasurement and the overall installed cost of the solutionwhile taking into consideration user and industrystandards.

Many of the concepts discussed herein will be painfullyobvious to the experienced instrument user. The goal isto provide a logical discussion of most key terms andthereby offer a condensed reference guide for future use.

Accuracy is a composite statement of performance thatdefines the quality of the instrument measurement. It isthe difference between the reading of an instrument and

the true value of what is actually being measured.Accuracy is normally expressed as plus or minus apercentage of either reading, calibrated span or the fullscale of the instrument. Accuracy is one of the mostcritical factors to consider when applying an instrumentfor a given application. Accuracy is also a term that ismost misunderstood by the typical user. The reason forthis misunderstanding is due to the myriad of effects oninstrument accuracy and the ability of manufacturers tooffer varying interpretations for the expression ofinstrument accuracy. In the world of marketing this iscommonly known as creative “specmanship”.

Accuracy Statements• % Reading (% Rate)• % Calibrated Span• % Full Scale

% Input

FIGURE 1

As stated in the discussion on accuracy, the specificationmay be expressed in terms of a percentage of reading,calibrated span or full scale. As illustrated in Figure 1,the comparison is enlightening. An instrument with apercent of reading accuracy statement maintains aconstant window of error throughout the measurementrange. A device with an accuracy expressed as a percentof span or full scale possesses an ever-wideningenvelope. This makes it imperative that the user shouldtake care to operate the instrument as high as possiblein the span in order to maintain accuracy.

Absolute accuracy is a term that defines how theperformance of an instrument relates to a traceablestandard.


Composite accuracy is a term that includes the totaleffect of accuracy, linearity and repeatability on aninstrument at reference conditions.

Reproducibility is an ISA term that is the combinationof linearity, repeatability, hysteresis and drift.

Reference Accuracy is the baseline accuracy for manyinstruments. This specification does not include someof the effects that cause instrument error such astemperature and pressure effects. Reference accuracyis the percentage of error associated with the instrumentoperating within designed constraints under referenceconditions. This is the most liberal of accuracystatements and is commonly misinterpreted as abenchmark for evaluating one instrument against another.To further confuse the evaluator, this accuracy statementmay not take into consideration errors induced by theoutput mode, such as the error associated with digitalto analog conversion necessary to obtain a currentoutput. However, most reference accuracy statementsdo include the effects of linearity, hysteresis andrepeatability.

System Accuracy is a term used to define the overallaccuracy of a process measurement involving more thanone component used in series or parallel. This is anexcellent means to determine the total error induced ona process variable that will be used for recording orcontrolling purposes. Individual instrument errors mustbe calculated as an error of reading (not a percentageerror) for a given measurement point. The combinationof errors then is averaged using a statistical approachsuch as taking the square root of the sum of the squarederrors. For example, a flow instrument has an error ofplus or minus one percent of calibrated full scale. Thescale is 0-100 SCFH. The reading is 50 SCFH. The erroris plus or minus 1 SCFH. The recording or controllingdevice has an error of plus or minus one half percent ofrange. The range is 0-100 SCFH. The reading is 50 SCFH.The error is plus or minus .5 SCFH. Using the formula:(12 + .5 2)1/2 = a system accuracy of plus or minus 1.1SCFH. This is more realistic than adding the total errors(1.5 SCFH) because it is not likely that all instrumentswould output their maximum errors at the same time.

Linearity is described in ISA Standard S51.1 as thedeviation from the calibration curve of an instrument froma straight-line relationship between zero input and 100%input. Ideally this is a forty-five degree slope. Mostinstruments do not possess a linear output with respectto input under reference or actual conditions.

Depending upon the instrument, there are many variablesthat can affect linearity. For example, most differentialpressure transmitters use a sensor technology that isinherently nonlinear under reference conditions. Asprocess and ambient temperature and static pressurechange during actual conditions, there are effects onlinearity. We are not talking here about the nonlinearrelationship of flow measurement from an orifice plate.

We are talking about the actual nonlinear nature of thesensor within the transmitter. Pressure transmitters varyby class. A baseline for a pressure measurement is thecommon pressure transducer that converts a pressuresignal input into a current signal output. The limitingfactors are that the input is a control signal (typically 3-15 psig and the output is a corresponding current output(typically 4-20 mA). A pressure transmitter extends thepressure input range. The output remains mostcommonly in the form of a current output. There is theadded value of compensation for these transmitters tostabilize the output for the negative effect of temperatureon the linearity of the output. This transmitter class isreferred to as conventional. The pressure effect iscompensated by zero adjustment while under staticpressure conditions. The top of the transmitter class isreferred to as “smart” due to the added capability ofhigher performance standards and the ability to offerintelligent communications from the transmitter forconfiguration and diagnostics information. Smarttransmitters have multivariable sensors for temperatureand pressure compensation. The sensors feedinformation to microprocessors that manipulate theoutput by comparing the input process variable, ambienttemperature and static pressure (flow transmitters) witha characterization algorithm stored in an EPROM withinthe transmitter. The characterization feature of smarttransmitters offers the ability to accurately produce anoutput as the actual environmental conditions change.

Repeatability is the ability of an instrument to preciselyduplicate a measurement while operating under the sameconditions while the input signal is made in the samedirection. The input signal may be consecutivemovements from zero to one hundred percent or viceversa. Many individuals believe that repeatability is asvaluable as a high degree of accuracy. In actuality, goodaccuracy cannot be achieved without good repeatability.What causes poor repeatability? Either the instrumenthas a poor repeatability specification, the instrument hasa defect, or there are one or more influences causingthe problem with a good instrument. Some of theseinfluences are piping installation problems, cavitationfrom valves and pumps, hammering from vapor pressureproblems, temperature changes (ambient and process),noisy power or grounding and vibration.

Uncertainty and accuracy are closely related. However,uncertainty is more definitive when used in the contextof instrument and system errors. This is due to the strictguidelines and equations associated with flow instrumentstandards. Uncertainty must take into consideration theactual operating conditions. Formulas used to calculateare common in flow measurement. An example of astandard formula can be obtained from ANSI/ASMEMFC-2M-1983 titled “Measurement Uncertainty for FluidFlow in Closed Conduits”. The uncertainty revealed inthese types of calculations are important in that theytake away the focus on individual instrument errors byassigning a relative importance for each variable thatimpacts the flow calculation. These formulas introduce


two very interesting concepts. First is the concept oftwo types of errors: Bias vs. Precision. Bias error maybe obtained from the manufacturer specifications.Precision error is calculated using independentcomparison tests. This concept is worthy of discussionfor two reasons. First, bias error is commonly referred toas reference accuracy. Since instruments are not usedunder reference conditions, this accuracy statement isnot valid. The errors associated with having an effect onreference accuracy should be taken into consideration.Second, not all manufacturers use the same standard topublish specifications. Some are more conservative thanothers in order to mass-produce instruments. For a givenproduct, there may be variance in absolute accuracy fromone instrument to the next, although all the instrumentsin a particular class would meet the publishedspecification. This is obviously done with the intent toreduce manufacturing costs. Precision error calculationmay allow for evaluation of a population of like-productsin order to determine worst case and average accuracy.Taking into consideration the effects of temperature andpressure on the accuracy requires some sophisticatedtest equipment that is normally associated withindependent laboratories.

The second important concept revealed from the use ofstandard flow calculations is the use of “square root ofthe sum of the squares” for all errors after a weight hasbeen applied to the individual errors. Because of the factthat errors are expressed in terms of plus or minus, itcan be argued that no instrument is likely to indicate aworst-case scenario whereby all of the combined errorswill sum to a total error in one direction or the other. Theuse of the square root of the sum of the squares offers afair averaging of these errors. A simple example of thisanalysis is given under “system accuracy”.

Temperature effects must be broken into two categories:process temperature effects and ambient temperatureeffects. Process temperature af fects instrumentperformance in two ways. For liquid processestemperature variation has an affect on viscosity. For thoseflow instruments that are viscosity sensitive, the variationin viscosity must be understood and compensated for orthere will be an error associated with the viscosity shift.Secondly, many sensors are located in close proximity tothe process so there is the potential for an effect on thesensor with process temperature variation. In addition,for gases there is an effect on the mass calculation withrespect to temperature changes.

Ambient temperature has an effect on many sensors andelectrical components. Those instruments with an analogcurrent output typically have a higher error associatedwith ambient temperature changes than the instrumentsthat use either a pulse output or a digital communicationsoutput. An instrument specification must be scrutinizedto determine if the accuracy statement is qualified toinclude the temperature effect for the type of outputintended. Depending upon the technology, ambienttemperature may have an effect that is equal to or greater

than the reference accuracy of the instrument. Ambienttemperature may affect both zero and span of aninstrument.

Pressure effects on instrument accuracy typically applyto the static pressure effect on zero or span of adifferential pressure transmitter. For conventionaltransmitters, this error may be partially resolved byadjusting the zero while the transmitter is under operatingline pressure. If there are swings in operating pressure,then this error cannot be properly compensated for withconventional transmitters. New generation smarttransmitters offer characterization with the ability to sensepressure changes and automatically correct for the errorthereby minimizing the effect. Pressure also affects thecompressibility of gasses. Therefore, pressuremeasurement compensation is necessary for accurategas and steam flow measurement.

Resolution is the ability of the instrument to continuouslymeasure and transmit all process variable data. In digitalsystems the smallest interval that exists between twomeasurement samples defines resolution. Most currenttechnologies offer microprocessors for the benefit ofenhanced performance. The tradeoff from the use ofmicroprocessors is speed and resolution. Digitalmeasurement involves sampling of the process variable.While the process variable is being sampled, there maybe small dead spots where process data is lost. On manysmart instruments, there may be multivariable datasampled. Microprocessor instruments typically processinput information via analog to digital converters. Theresolution of these types of instruments is dependentupon the resolution of the converters. For example, atwelve bit resolution A/D converter has the resolution towithin _ of 4,096 counts or in electrical terms 1 mV for a2.048 volt input. Converters include signal conditioning,sampling, multiplexing and conversion processes andthere are accuracy and linearity specifications tied toconverter. This may or may not be incorporated into theaccuracy specification of the instrument. In addition,there are roundoff errors to consider with converters. Newtechnology enhances the ability to increase the resolutionand accuracy of converters. The fact that this resolutionerror exists is not all bad. It must be considered howeverin the application of the instrument.

Other sources of resolution errors involve the specifictechnology used. DC Pulsed magnetic flowmeterssample a portion of a process variable at a fixedfrequency typically between 7.5 Hz to 30 Hz. The lowerthe frequency equals lower resolution and thereforeinadequate response time for noisy slurry applicationsor fast acting positive displacement pumps. Vortexflowmeters interpret vortex swirls commonly using piezo-electric sensors that appear to the electronics as asquare-wave pulse input. The frequency is directlyproportional with the velocity of the fluid. As line sizeincrease each pulse represents larger volumes of flowthereby reducing resolution. Fieldbus networks offer anadvantage of transmitting multivariable data as well as


diagnostics from instruments. The data obtained fromfieldbus instruments is valuable because the data isintermixed with multivariable and diagnostic data. Sincethis data is shared with data from other instruments overthe same wires, the application fieldbus strategies mustbe weighed with the need for resolution and responsetime. Smart instruments with built-in PID algorithms helpto address this issue by truly distributing the control tothe field where the primary and final elements arecollocated.

Many current technologies involve a means of samplingdata in order to provide an output that implies realprocess conditions. Near instantaneous measurementis possible with some analog instruments but is notpossible with instruments with digital electronics. Oddlyenough, there is an explosion of accuracy statementsthat imply enhanced performance from smart (by defaultdigital) instruments. The accuracy statements are correctbut the response time and resolution of the instrumentmust be taken into consideration to qualify performance.The ability of an instrument to respond to and transmitprocess data quickly is important for some critical flowor pressure applications associated with fast-actingpositive displacement pumps, compressor control andsafety systems.

Response Time is the ability of an instrument to reactto process variable changes. This is closely related tothe terms resolution and damping. Most electronicinstruments have a delay from the moment the processvariable changes to the moment the change is indicatedby the output signal. This delay will depend upon thetype of instrument, the circuitry design (analog vs. digital),the deadband for the measurement reading, and thesetting of damping adjustment that may exist on theinstrument. The time delay in producing an output signalis typically expressed in terms of milliseconds. A veryfast response time will be less than ten milliseconds.This type of device is appropriate for fast acting pressureand flow control loops. Microprocessor instruments mayhave a response time expressed in hundreds ofmilliseconds. These instruments may be very accurateand appropriate for many flow, level and pressureapplications. A common sense approach is to comparethe measuring instrument with the controller. Comparingthe response time of the instrument with the desiredsampling and execution rate of the controller providesdata that is helpful in making a qualified decision aboutthe use of that instrument for an application. A controllerwith sampling and execution rates in excess of 500milliseconds does not necessarily need an input devicewith a response time of ten milliseconds.

Damping does not affect instrument accuracy but doeshave an effect on the quality of the control loop. Dampingis used to slow the response of the instrument to processchanges. It is necessary to use damping when there isprocess noise and/or input fluctuations that areundesirable for control. Damping is a means to averagethe process variable over time in order to stabilize the

output to the controller. For that reason the adjustmentis expressed in terms of “seconds” of damping. Dampingmay be applied to an instrument with poor installationtechnique or inappropriate application of the instrumentfor the process. For this reason, it is important torecognize when applying damping that the need is valid.

Filtering is sometimes used synonymously withdamping. There is more than one interpretation forfiltering and instrument signal. As a simplistic definition,filtering gives the instrument the ability to accept or rejectand input signal. This feature is most commonly used toignore intermittent spike input signals to the instrumentthat may not be actual changes to the process input.For example, noise may be measured by vortex flowmeters or magnetic flow meters from hard solids in aliquid flow application. The solids interfere with thecontinuous measurement of the flow signal byintroducing an error normally interpreted as a flow spike.Damping of the signal may help to alleviate the problem.However, this type of signal spike is irrelevant to themeasurement of the process. Therefore, the ability toignore the spike is advantageous.

Hysteresis: A true and accurate measurement of aprocess variable shall yield a beginning measurementand an ending measurement cycle that is equal. That is,the upward curve from zero to one hundred percent andthe downward curve from one hundred percent back tozero are identical. Any deviation from these two curvesis defined as hysteresis. Hysteresis is also linked todeadband. Since deadband affects the ability of theinstrument to react to small instrument input signals, thiseffect amplifies the hysteretic error.

Deadband: All instruments exhibit a point where somechange in process data cannot be measured. This erroris commonly expressed as a percent of reading.Deadband is simply the ability of the instrument to reactto small process variable input changes. Deadband maybe associated with resolution because it does affect theability of an instrument to measure continuous data.However, resolution addresses the ability to continuouslymeasure the process variable. Deadband only addressesthe ability to monitor minor changes from a given processvariable reading. An instrument with a deadband of plusor minus one tenth of a percent and a span of 100 psihas the inability to read small changes of plus or minus1 psi at any reading point. If a resolution of plus or minus1 psi is necessary, a tighter span or an alternatetechnology is advised. Deadband has an effect oninstrument accuracy when very small changes areexpected.

Piping effects (flow profile effects): The installation ofthe instrument has the potential for greatly affectingaccuracy. Whether the application is liquid, gas or steam,the flow profile entering the instrument typically dictatesthe ultimate performance. Reynolds number constraints,turbulence, grounding, buildup on pipe walls,degradation of orifice plates, positioning of elbows,


valves, reducers, cavitation, multiphase flow, internal pipeimperfections and other variables all play into the abilityof a specific instrument technology to measureaccurately. It is beyond the scope of this paper to addressall these effects. There have been many instrumentsblamed for incompetent measurement that have sufferedfrom incompetent installation guidelines. The guidelinesfor standard installation applications are difficult to keepup with. To complicate the issue, manufacturers vary onthe recommended installation practices for a giventechnology. This adds to the confusion for users. Thepoint is that it is reasonable to adhere to themanufacture’s recommended installation practices. Thisshould be compared with industry standards such asthose published by ISA and API.

Reynolds Number Effects: Reynolds number (Re) isdefined by formula as the velocity of the flow timesdensity times pipe diameter divided by viscosity for liquidapplications. This specification is important for flow meterapplications. A great deal of time may be used to discussRe. To keep it simple there are fundamentals that help tokeep this concept into perspective. First, there is thedifference between laminar and turbulent flow. Renumbers below 2,000 are laminar. Between 2,000-4,000the Re is transitional. Above 4,000 the Re is turbulent.Most flow meters need a Re in the turbulent region tooperate accurately. The Re calculation is relativelystraightforward. If this calculation is strictly applied andother installation requirements are adhered to, then theinstrument should be expected to perform withinspecified tolerances.

Zero Shift is caused by numerous factors. What causesthis error? The answer should be evident within thepublished product specifications. If there is zero shiftthat is outside the specifications possible causes includepower or ground induced noise or a defect within theproduct. Process and ambient temperature effects andprocess pressure changes on some devices are commonsources of problems for the instrument to reproduce azero input. Some technologies offer a feature to performauto-zeroing to minimize zero shift. Smart transmittersoffer real-time temperature and pressure compensationto help minimize the effect. The knowledge that zero driftcan affect your measurement will help to define acalibration program to correct for this problem.

Span Shift may have a negative effect on the accuracyof an instrument. If an instrument span is changed, thereneeds to be a calibration verification of the change. Theinstrument should be expected to perform withinpublished specifications providing that a calibrationdevice is used to confirm the zero and span of theinstrument. Again, on pressure transmitters, temperatureand pressure effects may cause span shift. Also,adjustment of zero alone may affect the span.

Turndown may be expressed in more than one waydepending upon the instrument technology It may beexpressed as either the ratio of the maximum

measurement capability to the minimum measurementcapability while maintaining accuracy, or as the ratio ofthe maximum span capability to the minimum spancapability while maintaining accuracy. The latter definitionnormally applies to pressure transmitters and must becarefully evaluated for a given application. Instrumentsuniversally have a point on the low end of the scale whereaccuracy degrades. Specifications must be qualified todetermine where accuracy falls off and to what degreethe degradation occurs. Turndown is a term where muchcreativity is used with “specmanship”.

Rangeability is closely linked with turndown. Mostinstrument technologies offer the ability to adjust themeasurement range. It is important to discussrangeability because the use of instruments is limited tocertain ranges in order to maintain accuracy. Instrumentsmay be over-ranged and operate reliably within designlimits. However this is often at the sacrifice of someaccuracy. The degree by which an instrument may beover-ranged depends upon the technology and themanufacturer of that technology. Under-ranging aninstrument often has negative effects on accuracy. Someaccuracy statements qualify the degree of accuracydepending upon where the instrument is operated withinthe span capabilities. A case can be made that liberaluse of range application can save the cost of ownershipby reducing the number of spares. After all, a transmitterwith a maximum range of 100 IWC and a minimum rangeof 5 IWC can be used for the 70 IWC and the 10 IWCapplications. However, the combined possible errors asstated in typical specifications will yield lesserperformance on the lower span. Prudent users will applythe appropriate range as they would use the proper toolfor the job and carefully evaluate the desired results.Other sources of error include mounting position effectof the instrument, RFI interference, power supplyinfluenced error and gross calculation errors in thecontroller. The terminology used to describe instrumentaccuracy is broad with some overlap. There is casualmisuse of some of these terms by users and sales peoplethat create confusion in the marketplace. Care must beexercised to learn the terminology and apply itappropriately.

The ability of an instrument to perform accurately isdependent upon many variables. It is the human factorin the equation that ultimately determines how well atechnology performs. Process knowledge must becoupled with the knowledge of instrument technologyand proper installation techniques to ensure success.Process operations and support personnel are living ina world of ever tightening constraints with higherperformance expectations. The tools available to helpguide the way are imbedded in the knowledge base ofthose who have been there before. A wealth of data isavailable in the form of text books and technical papersand bulletins. This author is grateful for the availability ofthese tools and acknowledges authors of those sourcesby offering a list of valuable references.


REFERENCES:

1. ABB Instrumentation Flow Products Bulletin 10E-12“Predicting Flow Rate System Accuracy” as reprintedfrom the proceedings of the 1979 ISA Symposium,William S. Buzzard.

2. “How Accurate is Accurate?”, Control Magazine,William L. Mostia Jr. PE, June 1996.

3. Flow Measurement, Practical Guides forMeasurement and Control, ISA Text, David W.Spitzer, Editor, 1996.

4. Industrial Flow Measurement, ISA Text, David W.Spitzer, 1995

5. Digital Control Devices, Equipment and Applications,ISA Text, J.A. Moore, 1986.

Rick Williams


ECONOMICS OF ELECTRONIC GAS MEASUREMENTTom R. Cheney

Anadarko Petroleum1201 Lake Robbins Drive, The Woodlands, TX 77380

INTRODUCTION

There isn’t any one who isn’t impacted by the continuousgrowth and changes in the world of technology. In today’sworld, we accept computers and the functions theyperform without question. In fact, we place our hard-earned dollars and in some cases our very lives in theircare without a second thought. Computers and electronictechnologies have greatly impacted the way work is donein the oil and gas industry. A good example of how thesechanges have impacted this business is the use ofelectronic gas measurement devices often called (EFM’s).

With the use of the EFM, and associated electronicequipment, we not only have the option to measure gasvolumes but to retrieve required data instantly andenhance production through well automation.

As in any business, before the decision is made to upgradeor change from the standard method of measurementusing the typical dry flow chart recorder to EFM, theeconomics of such a decision needs to be evaluated.

It is becoming common practice to install EFM on all newmeasurement stations. The real battle over economicjustification seems to be in the decision to replace meterstations, which already have chart recorders in place withEFM’s. You will find that the economic justification of usingEFM is often difficult to quantify.

SCOPE

Over the last few years of sitting in many hours ofmeetings trying to justify the conversion to EFM andautomation, I have concluded that there is not an off-the-shelf equation for determining this economicjustification. I feel there are five primary areas that mustbe considered when making this evaluation. This paperwill focus on these main elements. These elements are(A) Field operations and efficiencies, (B) Measurementaccuracy, (C) Initial costs of installation, (D) Future growthor expansion, and (E) Standardization.

BACKGROUND

When we finally made the jump from mercury meters todry flow chart recorders, we were impressed with howfar technology had taken us. It didn’t take long until wewere introduced to the first electronic meter that not onlytook the place of the dry flow meter, but would alsoreplace the chart and the need for chart integration.

When the EFM’s first hit the market, many werepurchased and installed for the sake of testingtechnology, rather than used for the proclaimedimprovements they brought to the table.

In time, the flaws of the electronic world surfaced — suchthings as lack of enough battery power, too small ofcomputer memory, lost data, incorrect calculations, andthe lack of field personnel with the technical knowledgeto service the new devices, to name a few. The firstreaction was, “electronic equipment doesn’t work.”Manufacturers went to work and soon came up with newdevelopments and improvements resulting mostly fromcomplaints of their users.

The current generation of EFM and the end devices havemade great improvements and are now widely acceptedby the industry as the meter of choice. This popularity isdue to the improvement in their electronics, improvedaccuracy, ease of use, as well as the capability to linkwith automation and instant data transmission.

Costs have come down; reliability has improved and thegrowing use of wellhead automation is on the rise. Thesecombined elements, along with the acceptance of EFMby the American Petroleum Institute (API 21.1), havetaken this technology from a choice by the users for thesake of testing new technology to a choice based ongood business sense.

FIELD OPERATIONS AND EFFICENCIES

In the last few years, we have seen a movement in ourbusiness to have the field operator become the jack ofall trades. In many cases, this transition has includedmeasurement duties and tasks. Let’s explore the prosand cons as to how EFM has played into these decisions.

When the field operator arrives at the production locationor measurement site, generally their task list is pretty large.Not only do we expect more of them while on each location;they are also asked to visit more locations each day.

Charts are generally changed every seven to eight days.On these chart days, the operator will normally zero themeter and replace the chart. Past experience hasindicated that if the field operator is running behind,performing the zeroing process is often put off until thenext visit or chart day. The chart that was changed isthen edited or reviewed for problems prior to being sentto the chart integrator. If a recording problem is present,then it must be corrected before the operator can leave


the location. Whenever the recording pens are changed,this creates a need to check and adjust any incorrectpen tension, time lag or pen-arch problems. Dependingon conditions, this work can be a time consuming task.

Of course, this is the problem of having a chart recorder.However, before we move on to the pros of having anEFM at these locations, keep this one point in mind. Oneof the perceived advantages to having a chart recorderis the quick ability to look at a week’s worth of data atjust one glance. This information can be a quick referencefor troubleshooting a problem or understanding the flowpatterns of a particular well. With EFM, this data is stillavailable, however, requires a series of additional stepsto view the data while on location. This step typically isdone once a month when the data is collected. What wereally have is a shift in a paradigm that has existed formany years.

If all is working right, the EFM does offer the field operatorsavings of time. With just a glance, flows and pressurescan be obtained. With the presence of remote datatransmission, this information can be collected daily andoften eliminates the need for a visit to the location. Thisone option has proven to play a big role in making theeconomic decision about electronics being used at manylocations.

When making the decision about installing EFM, takeinto account the following items:

• EFM allows you to store up to thirty-five days orlonger, of data. This can eliminate the need tovisit the location at least every chart period,which typically is every seven to eight days.

• If you tie radio telemetry to the EFM, you canretrieve the data at any time.

• If you tie in automation equipment, you canoperate the location from a remote location.

• The cost of an EFM is still more expensive thana dry flow chart recorder but is coming down inprice and has proven to be very reliable.

With all this information taken into account, it appearsthat with careful planning and the proper equipment, youcan reduce manpower and the vehicle costs associatedwith traveling to each location. These savings may beredistributed to other important activities.

The need to keep technical expertise or training up tospeed is even more important with EFM than it was withdry flow chart recorders. If a technician does not performthe procedures correctly or understand the fundamentalsof what makes this complex system work together, manydollars can be lost without someone even recognizingthe errors. Each method requires different knowledgeand skills.

MEASUREMENT ACCURACIES

Let’s face it, accuracy is only as good as the equipmentand the technician who maintains the equipment. Theequipment in today’s market is top of the line. It is commonto see EFM manufacturers advertise very preciseaccuracy. If all the other variables are maintained properly,such as orifice plates, meter tubes, samples andchromatographs, this improvement is a favorable one.

Whether you have chart recorders or EFM, the need foraccuracy remains the same. The advantage that EFM hasto offer is the ability to take readings every second ifrequired. If no future editing or recalculation is required, thedata can be downloaded directly into a data base system.This process eliminates the need for someone to interpretthe data that has already been captured on a chart.

Wells with flowing characteristics that cause a wide rangeof flow volumes can be a problem for a chart recorderand integrator. It is difficult to determine flow volumes whena chart is wiping a line from 0 to 100% of the chart rangeevery few minutes or even more often. Also, it is difficultto determine accurate volumes when the well is producingand recording very low readings on the chart. Not onlycan it be difficult for the field operator to determine a flowvolume; it can be even more difficult for the chart integratorto interpret where to trace the flow pattern.

With all things taken into consideration, the EFMeliminates some of the opportunity for the introductionof human error.

Many companies have become so confident in thesenumbers generated from the field locations that directbilling is often implemented as this data is received.

As mentioned in the previous section, many people hateto see the chart recorders go away because of theinformation they offer a field operator at a glance. Onthe other hand, speed, accuracy, and the savings ofmanpower costs, which the EFM offers is a big plus inmaking the decision of which method to choose.

INITIAL COSTS OF INSTALLATION

If you are installing the basic measurement elements fordata capture or volume calculations, you will find theinitial cost for an EFM and Chart Recorder to be aboutthe same.

A chart recorder will cost you +/- $1500.00, while anEFM with no additional I/O capabilities can be purchasedfor +/- $3000.00

First glance of these numbers would lead you to believethat there is no advantage to selecting EFM’s over a ChartRecorder. Depending on your company’s overall goalsand plans, this situation may be the case. The nextsection will lead you into some thoughts that should beexplored prior to making this decision.


FUTURE GROWTH OR EXPANSION

You should take a hard look at what your companystrategies are when it comes to future operations. Additionalmoney can be needlessly spent without proper planning.There has been stand-alone EFM purchased for a location.Later it was decided to install communications at thelocation in order to obtain the flow data daily. The EFM hadto be upgraded along with the installation of thecommunication equipment. At a later date, the decision toautomate the location was made; and again another EFMwith a larger memory and lower power demands wasinstalled. Good intentions were in place; however, the totalcost of the project turned out to be much higher than itneeded to be if proper planning had taken place.

Sometimes purchasing equipment that can easily beexpanded with minor changes may be the best and mosteconomical choice if the possibility of expansion is a reality.

There are several additional functions that may be addedto the location, all of which should be considered whenmaking the EFM hardware selection. Some of thesefunctions may include the following:

Flow Control (Maintaining flow within the range of themeter.)Plunger Lift ControlProduction OptimizationDriving a SamplerValve SwitchingLeak DetectionEmergency Shut-Downs (ESD)Liquid MeteringWellhead MonitoringDownhole monitoringCompressor Control and MonitoringPump ControlGas Quality monitoringChromatography

If expansion is a possibility, the original set up shouldalso include considerations for the following:

Communication RequirementsPower RequirementsHost PC SelectionTechnical Support

There is nothing that a good manager or field operatorhates to hear more than, “Why didn’t we think of thatbefore?” If you are making the change to EFM, you mightwant to consider spending the extra money up front ifgrowth or expansion is possible in the future. In the longrun, the cost may be lower.

STANDARDIZATION

Many companies have several different brands of EFMat their locations. This may have been a decision basedon avoiding the trap of having all your eggs in one basket.The down side to this scenario is the additional inventory

of parts and technical expertise that is required tomaintain different brands of equipment.

One economical consideration is standardizing onebrand or system to be used throughout your company.This step can simplify training, manpower utilization,parts inventory and troubleshooting problems. Also, thisprocess could give you an edge for obtaining a goodprice on the equipment because of the volume ofequipment you control. Standardization also gives yousome leverage with the manufacturer in regards tochanges or improvements that you would like to seemade to the equipment or software.

CONCLUSION

Economic justification of EFM is not black and white. Inmany cases, the dry flow chart recorder may be the mosteconomical choice. On the other hand, increasingdemands for instant reliable numbers, manpowerutilization, well optimization, and remote control orautomation of our equipment are all now a reality. It isnow common to find most large gas transmissioncompanies using EFM equipment. The focus is slowlyturning away from the decision between dry flow metersand EFM and is focussing on what additional EFMupgrades are needed to meet the future needs.

Usually there is no economical or single justification formaking the change. It often boils down to where youwant to be in the years to come.

This paper has attempted to point out the need to lookbeyond your immediate needs and look into your futureneeds when making this decision.

REFERENCES

“Fluid Flow Measurement” E.L. Upp

“Economics of Electronic Measurement”Harry J. Workmon, GPM Gas Services Company1997 Proceedings ASGMT

“Economics of Electronic Measurement” Class 172Ben Wagner, Strategic Controls Corporation

Tom R. Cheney


ELECTRONIC CALIBRATORSBetsy Murphy

M M EnterprisesFort Worth, TX

Electronic calibrators are fast becoming the benchmarkfor measurement and are replacing mechanical types ofinstruments for testing and calibration checks.

Techniques, usage, traceability requirements, andproblems are changing quickly as technology advancesin the development of these instruments. Informationconcerning these issues is often outdated by the timethe technician receives it.

Electronic calibrators use a microprocessor with digitalmeasurement. What is the difference between analogand digital? Analog is a continuous signal whereas digitalis an analog signal converted into numerical data or bitsthat computers can understand change and store. Thenumerical values are converted from the analog signalat intervals of time. As the amount of time becomesshorter and shorter between the intervals of conversion,it becomes close to impossible to distinguish it from theoriginal analog signal. Stephan Schuster chairman of theRainier Corp., summed it up as follows: “Analog is thereal world and digital is a numerical representation ofthe real world.”

Computers are the driving force behind the digitalrevolution because information must be digital to be usedon a computer. Just like any other computer, technologyis often obsolete by the time an instrument is designed,manufactured, marketed, sold. and shipped to the enduser.

ASSOCIATED TECHNIQUES

the best technique for using an electronic calibrator is tohave a solid knowledge of the mechanical aspects ofthe system it is being used on. There are someprocedures that can be implemented to save some timeand preserve the integrity of the information. Temperatureis of utmost importance for measuring pressure orcalculating flow. For temperature checks the most criticalaspect is to see the same temperature as the gas stream.One method to ensure a closer test is to fill the test wellwith a non-interfering liquid, such as a light oil, to maintaina more constant measurement. In extreme ambienttemperature differences with the gas, the measurementtemperature is lost very quickly when the test well isopened to atmosphere. The result can be an averagetemperature shown instead of the actual gastemperature.

Temperature drift plays a significant role in the operationof pressure calibrators. Most portable electronic

calibrators should remain at ambient temperature for aperiod of time before the test is performed in order toallow the sensor to adjust changes between thetemperature inside the vehicle and ambient. To avoidthis time lag, the calibrator could be placed in a locationin the vehicle that is neither cooled nor heated. Anothertechnique to help avoid temperature drift is to insulatethe tubing from the line to the instrument For lowpressures, a refrigerant type of hose can be used andfor higher pressures a material such as Armaflex couldbe placed on the outside of the tubing or hose to insulatethe gas temperature from ambient warming or cooling.

USES

Electronic calibrators are used for a variety of applicationsfrom flow measurement to safety monitors. Electronicflow meters require pressure, temperature, anddifferential. Many manufacturers are offering an all in oneinstrument and some are adding communicationcapabilities. Most offer the option of passing data throughto a computer. Fixed monitoring equipment is the mostcommon application that comes to mind for portablecalibrators, but they are also being used to test theintegrity of other portable instrumentation with lesseraccuracy. Transmitters require not only pressure ortemperature, but also electrical. The onset of calibratorswith both capabilities has become very useful for thisapplication. New portable calibrators on the marketinclude chromatographs and high flow samplers. Theuses of electronic calibrators are becoming infinite andpresent many challenges to the technician trying to keepup with changing technology while maintaining goodmechanical aptitude.

TRACEABILITY

Traceability as it applies to the natural gas industry meansthat an instrument has been tested at the NationalInstitute of Standards Technology with the numberedreport of the test on file at NIST Once an instrument hasbeen tested and has received this report number, thisinstrument can then be used to test other instrumentsthat can also become traceable to the same test reportnumber as the original instrument tested at N.I.S.T.Generally, a standards lab’ uses instruments for testingwhich have been tested against an instrument that hasbeen sent to N.I.S.T. and are not necessarily the originalones tested at N.I.S.T. with the report number assigned.

Having a definition of the term doesn’t mean that thetraceability of the instrument is understandable to the


lay person with any certainty. Accuracy statements madeby the manufacturers of electronic calibrators play animportant and significant role in determining whatinstrument to use in different applications but does notindicate the actual performance in a particular situation.An accuracy statement made by a manufacturer is thetheoretical accuracy” compiled by taking physicaluncertainties into consideration. Actual accuracy wouldinclude all of the measurement uncertainties such asoperator bias, leaks, temperature gradients, gravitychanges, barometric pressure. etc. The Federal EnergyRegulatory Commission (F has determined thatinstruments being used for test purposes must havetraceability and must be twice as accurate as the itembeing tested, for example: Electronic flow meters havea stated accuracy 0.1% for pressure, meaning that thecalibrator being used for testing should be 0.05% orbetter. Many manufacturers are offering calibrators withstated accuracies of 0.025%. It is important to note howthe manufacturer arrived at this statement. Accuraciesare determined in a laboratory environment and theselaboratory conditions are usually mentioned inconjunction with this accuracy statement of the calibrator.These conditions are sometimes impossible to duplicatein the field giving a very real potential far a much greatererror that the technician expects. Often the techniciandoesn realize an error is present unless the testinstrument is drifting or the results vary greatly fromprevious tests. Temperature is the leadIng culprit inpressure measurement errors using electroniccalibrators. Accuracy is usually stated as percent fullscale, percent span, or percent of reading. Percent fullscale and percent span are the same. Accuracystatements do not mean that the instrument has an errorequal to the stated tolerance, but rather will not have anerror greater than the stated accuracy. Example one: Ifthe full scale range is 1000 psi and the stated accuracyis percent of reading. then the reading shown would bemultiplied by 0.0005 to determine if the instrument iswithin tolerance. Then when temperature specificationsare thrown in, an instrument with a stated accuracy of0.05% and is temperature compensated from 30 to 130deg. F could be much more accurate and reliable thanan instrument with a stated accuracy of 0.02 and iscalibrated at 60 deg F without compensation. It isimportant to remember the differences In accuraciesbetween the test instrument and the instrument beingtested, testing an instrument with a stated accuracy of0.1% with an instrument with a stated accuracy of0.025% gives the potential for thinking there may be aproblem with the instrument being tested, when in realityit is well within the stated tolerance.

The instrument being tested probably would not showthe error anyway because the decimal place digitswouldn’t be available. The resolution of a calibrator isthe number of digits available on the display. Some havefour-digit display some have five digits, and some evensix digit. Testing an instrument with a four-digit displaywith an instrument that has six, gives information thatcan neither be substantiated nor used.

PROBLEMS

Just as mechanical instruments have a set of problemsconnected with use, so do electronic calibrators.Electronic calibrators will show better results inmeasurement testing than will mechanicalinstrumentation except in extreme ambient temperatureconditions. In the techniques section a few ways toovercome these problems were discussed, buttemperature extreme is a serious problem with allelectronic instruments.

Isolated sensors are almost mandatory in electroniccalibrators because of liquids in the gas. Many sensorreplacements are caused by liquid destruction rather thanoverpressure. A major problem is dust, dirt, and/or sand.This problem is very prominent in instruments withchangeable modules. Field change outs of modulescauses a buildup of grit or dust around the pins. After aperiod of time the instrument ceases to perform well ornot at all. There is no way to prevent this from happening.Careful use and module changes will prolong the illeffects. Build up along the walls of the line in valve trees,straightening vanes, and etc. are not seen by electronicinstruments. Changes in metal thickness, or orifice platesare also a factor. Volumes are calculated based on themeter tube size with at particular orifice. Electronic flowmeters are calibrated based on these calculations. Whenbuild up or corrosion/erosion changes occur, tileinstrumentation cannot compensate for these changes.A lot of lost or unaccounted for gas is shown because ofthese changes. Pipeline maintenance is very critical whenusing electronic instruments. Even though accountingprocedures may show a tremendous amount of “lug”,the truth is there is probably less now than ever beforedue to new technology showing where it is or that it’sthere at an. User friendliness is extremely important, Toomany buttons to push, buttons being to small, too difficultto program, pods not labeled correctly, these issues areat the top of the list when selecting an electroniccalibrator, If the instrument is too complicated for quick,efficient learning curves, the full potential of the purchasecannot be realized. The technician has so many differentapplications and/or instruments to keep up with, alongwith the integrity of mechanical parts, that ease ofoperation is primary to saving time and labor to get thejob done correctly and efficiently.

Measurement uncertainties are showing up in areaswhere local gravity is being implemented on dead weightand pk testers. Most all electronic calibrators arecalibrated to a national or international standard forgravity and temperature. When dw or pk testers are setto local gravity standards, the temperature used isgenerally still 80 degrees F, but the gravity difference canshow an error where there really isn’t one according tomanufacturers’ specifications. If a primary standard isused to calibrate an electronic calibrator, the primarystandard should be in a climate controlled area and in afixed position, with the gravity correction being the sameas the electronic calibrator or corrections should beapplied to the results before the measurement is


assumed correct. According to many manufacturers, thegravity reference cannot be changed on electroniccalibrators, therefore a correction factor should be usedwhen using a primary standard set for local gravity.


LOW POWER FLOW COMPUTERSGreg Phillips

Bristol Babcock Inc.2000 Governors Circle West, Suite F, Houston, TX 77092

INTRODUCTION

Flow computers, themselves, are undergoing anevolution. One challenge for most vendors will be to offera low power flow computer whose pricing approachesthat of a three variable chart recorder. Many companiesin the gas transmission, gas distribution and productionindustry; expect such a flow computer to be an evolutionfrom today’s smart transmitter technology, because ofimproved accuracy and innovation of multi-variabletransmitters. That is to say, differential pressure, staticpressure and temperature all in one transmitter.

REQUIREMENTS

The key operations, performed by low power flowcomputers include calculations, historical data storage,alarm/event logging, network communications for real-time or near-time data acquisition, trending of variablesand flow rate to include data editing capability. Each oftoday’s flow computers offers some subset of thefollowing calculations:

• AGA3: Corrected flow for orifice meters. (Orifice)1992 edition

• AGA5: Energy content (BTU)• AGA7: Corrected flow for linear meters. (Turbine,

PD)• NX19: Compressibility• AGA8: Super compressibility (Detailed and

Gross)

The low power flow computer (LPFC) of today requiresthat it be designed to provide high performance whileminimizing the overall EGM installation cost. Theseelectronic gas measurement locations vary from site tosite; in most cases commercial power is not available,thus the need for LPFC’s that can operate on directcurrent (DC) power sources.

These power sources for remote locations can be batterypower, for instance a lithium battery or alkaline battery.The lithium battery can provide longer life over a greatertemperature range, while the alkaline battery is limited.The other most commonly used are the solar panel withlead acid or gel cell battery, this power source usuallyoffered as an integral package with the LPFC includespanel, charger/regulator, battery, cabling and mountinghardware. The size and cost of the solar array will dependon the geographical location and power requirementsof the LPFC. The energy from the sun creates aphotovoltaic effect in solar panels that charges a storage

battery powering the flow computer. The output voltageis 6vdc and should provide usually 30 days of autonomy.

A typical LPFC offered by most vendors will include thefollowing:

• Microprocessor capable of performingnecessary equations and calculations.

• Capability to accept smart transmitter or analogtransducer inputs.

• The ability to store historical data, hourly, daily,event and alarm logs.

• Capable of displaying real-time as well ashistorical data via a display or handheld.

• Able to be configured via a laptop computer orterminal interface by menu selection.

• Class 1, Div. 1, explosion proof, or intrinsicallysafe certified for hazardous areas.

• DC power source, internal lithium or alkalinebattery, and solar array powered.

Figure 1 shows an “RTU-style” single-run flow computer.This configuration includes a single-board RTU that isprogrammed to perform flow calculations and interfacesto differential pressure, pressure, and temperaturetransmitters. The transmitters are included in thepackage, but mount external to the flow computer.

Optional items shown include the handheld terminal andinternal modem. The hand-held terminal functions asboth an operator interface/configurator and a datatransfer medium. This same transfer medium might alsobe performed by a laptop computer. By adding a privateline modem for telephone line communications or aswitched network modem for dial-up phonecommunication wide-area networking to these remotesites can be achieved.

FIGURE 1


Since this type of flow computer is based on a single-board RTU, it will, typically, include extra input/outputpoints. For example, the unit shown in figure 1 includeseight analog inputs (4-20madc or 1-5vdc), three of whichare interfaced to transmitters. The five remaining couldbe used to interface to additional transmitters for a multi-run meter station. This RTU can also accommodateDigital I/O capability to perform run-switching valvestatus, valve control and gas sampling Instrumentation.While the additional I/O points imply that the flowcomputer can accommodate two or more meter-runs, ifsufficient RAM (random access memory) is not availableto store additional historical data, then the processingcapability is limited. With today’s technology and mostvendors using 16 bit or 32 bit microprocessors’ datastorage has not been a problem.

PRIMARY MEASUREMENT

The orifice and the positive displacement meter or turbinehas been the two primary methods of gas measurementin the United States. Some estimates put the numberabove half a million or more. These primary measuringelements impact the LPFC’s performance greatly andmust be inspected or verified as to their compliance toaccuracy specifications.

Remembering that the flow computer itself is considereda tertiary device that is responsible for correctlycalculating flow based on information that it processesfrom secondary measuring devices such as transmittersor transducers. All together these three devices arecritical to the overall accuracy of measurement werequire.

Positive displacement meters and turbines shouldnormally be sized to operate between 60% to 90% oftheir maximum linear capacity. With proper maintenancewith respect to mechanical drives, pitting, scratches, ordeposits these inspections can normally be somewhateasily performed. On a scheduled basis these metersshould be proven to insure accuracy and credibility notonly to their mechanical but also electrical frequencyoutput integrity. For instance two types of known proversare pipe and tank type. They provide a known volumeagainst which the volume indicated by the flow metercan be compared. Both these provers are volumetric withrespect to comparison results established. With propermaintenance and proving these P.D. and turbine metersshould maintain accuracy to within plus or minus 1/4 %.Orifice meters are generally inspected to ensurecompliance with ANSI/API 2530 established tolerances.These tolerances include orifice diameter, edge width,edge sharpness, surface finish, flatness, andconcentricity. There are several defects that can causeadverse effects on the meter’s accuracy such as a bentplate, a nicked bore or rough surface. Also, if it hasresidue build up or is dirty this needs to be cleaned.

SECONDARY MEASUREMENT

Most LPFC packages include secondary and tertiarydevices: DP, pressure and temperature transmitters thatwhich are interfaced to a digital flow computer.Transmitter specification provides the accuracy ofconversion to a 1-5vdc or 4-20mA dc analog signal. Todetermine the accuracy (or really, “uncertainty” orprobable error”) introduced by all three transmitters, the“square root of the sum of the squares” method is used.For example, if the accuracy of each transmitter is 0.25%,the total probable error of the secondary stage isapproximately 0.43% (this method is accepted by theindustry even though it weighs each input DP, pressure,and temperature the same).

E = DP + P + T =0.43

FIGURE 2

Figure 2 shows an arrangement that is representative ofthe new trend in flow computers. DP and pressuretransducers not transmitters, have been integrated intothe LPFC package. That is to say that the secondarymeasuring device is mounted internally to the LPFCpackaging not external. The amplifier electronics isincluded on the computer CPU (central processing unit)board. While figure 2 represents a single run application,some vendors can introduce dual-run capabilities bysimply adding additional transmitters or one single multi-variable transmitter.

TOTAL MEASUREMENT ACCURACY

A discussion of gas measurement accuracy mustencompass all the stages of conversions and calculationsthere are in a LPFC system. Since each stage introduceserror, the overall accuracy of the LPFC depends on theaccumulated errors of all stages. The API has definedthree stages in a flow computer system as “primary,”“secondary,” and “tertiary.”


Further errors are introduced when the transmitteroutputs are interfaced to A/D’s (analog to digitalconverters). Typical accuracy for the input conversion is0.025%. Again using the square root method, the totalprobable error for all three inputs is 0.043%

E = DP + P + T =0.43

The mathematically correct way to account for thecombined uncertainty of the transmitters and analoginputs is to add them together: 0.43% to 0.043% =0.473%. If you doubt this method, an alternative is totreat each of the three inputs separately. The 0.25% forthe transmitter should be added to the 0.025% for theinput, for a total error of 0.275% per input. The squareroot method for all three is still 0.473%.

Advantages of figure 2 depend upon the strategy theparticular vendor uses to process the data from the threevariables DP, pressure and temperature. If the LPFCprocessing this data uses an A/D (Analog/Digital) converterthe accuracy of that data may represent what is stated inthe above paragraph using analog transmitter technology.If the vendor utilizes smart transmitter technology withdigitized communication of the values presented, first,there is no secondary/tertiary conversion. There is no 1-5vdc or 4-20mA dc output and no A/D conversion, asperformed by an analog input. Thus, the typical 0.025%uncertainty is eliminated. In addition, the smart transmitteraccuracy is increased. For example, if the smart transmitterhad an accuracy of 0.1%, then the square root summedfor all three variables would equal 0.17%, compared with0.473% for the aforementioned system.

E = DP + P + T =0.43

If this transmitter was to be a multi-variable three in onetransmitter such as some vendors supply today, the DPsensor is fully static pressure compensated.

TERTIARY MEASUREMENT

The tertiary stage comprises the calculations within theLPFC. It is relatively easy to accurately perform aninstantaneous AGA3 or AGA7 flow equation as mostvendors do not consider this a problem. Compared withthe input accuracy, the calculation accuracy isinsignificant. However, how often are the calculationsdone? That is the main concern of manufactures ofLPFC’s. A further issue is averaging and totalizing overtimes such as an hour and a day. While some peoplethink that calculations performed inside a flow computerare extremely accurate, the truth is precision can fall offin time-based averaging and totalizing. By using doubleprecision (64 bit) floating-point math for all averaging andtotalizing, these averages and totals are usually updatedonce per second.

In general the LPFC of today will require it execute inputsampling, alarming averaging, totalizing, PID control ifrequired, and all calculations, except AGA8, once persecond. Due to the intense calculation required by AGA8for compressibility using detailed gas composition thiscould be performed once per minute.

AUDIT TRAIL ALARM/EVENT LOG

A requirement apparent to vendors that manufactureLPFC’s is the audit trail. This is a log that will keep trackof alarms and events that occur within the LPFC system.An example of alarms that may appear in the alarm logwould be:

• System power down• System power restore• Low system power level• DP, Pressure, Temp, and Turbine High alarm High

high alarm Low alarm Low low alarm Out ofrange Rate of change Return to normal

• Power down• Power restore• Low RAM battery level

Examples of events:

• Operator sign-on (laptop)• Operator sign-off (laptop)• Low flow cut-in• Low flow cut-off• Override mode on• Override mode off• Maintenance on (calibration)• Maintenance off (calibration)• Orifice plate change• Value change of constant

(Alarm and events should have the capability of beingreported over the wide area network (WAN).

SNAPSHOT LOG

Upon certain alarms and events, some LPFC’s will notonly log an alarm message, but will store the entire listof input, flow, and configurable constant values. Thisallows the user to see the entire station status, ratherthan a single message, when an exception conditionoccurs.

INSTANTANEOUS/HISTORICAL LOG

In addition to storing alarm and event audit trailinformation a LPFC will have the capability to storecurrent as well as historical information. A LPFC willtypically store 35 days of information, the amount ofinformation stored in an hourly or daily log varies anddepends on what the vendor may offer. Information thatmay be available in the hourly, daily as well as quarterhour logs are:


• Date start• Time start• Flowing time• Average Differential pressure• Average Static pressure• Average Temperature• Average Specific gravity• Average BTU• Flow extension• Flow Rate• Energy Rate• Alarms occurred• Events occurred• Compressibility (FPV)• Average C-prime• Average CO2 (Carbon dioxide)• Average N2 (nitrogen)

These items will also include station identification, stationtag and meter identification.

COMMUNICATIONS

Requirements for LPFC’s at a minimum, is to provide aRS-232 Port capable of interfacing with a laptopcomputer or a handheld device. This interface wouldenable the user to configure, monitor, and changeparameters specific to the measurement requirementsof the LPFC. It would also provide the capability to updump or collect historical logs, such as daily, hourly andaudit/event logs.

RS-232C is a serial asynchronous communicationsstandard used to connect modems, terminals andprinters with serial interfaces. The Electronic IndustriesAssociation (EIA) developed the recommendedStandard-232 to define a serial communication interface.This standard is referred to as RS-232/ RS232-C, andRS-232-D. The “C” and “D” refers to particular versionsof the standard. Although RS232C is only specified foruse in transmission lengths up to 50 feet, it is often usedfor greater distances at lower baud rates. An additionalRS-232 port has to be made available for networkcommunication for local area networks (LAN) and widearea networks (WAN).

LPFC’s in the digital communications world are definedas DTE (data terminal equipment) devices. Modems,radio modems and other communication media areconsidered DCE (data communication equipment). Themain difference between DTE and a DCE device isdefinitions of their respective transmit and receive pins.Pins’ two and three have opposite meanings, the DTEdevice transmits to the DCE receive pin, and the DTEreceives data from the DCE transmit pin. The other pinsdefined on a DB-9, DB-15 or a DB-25 pin connectorrefers to control or handshaking signals. These are usedto control the timing between device’s for transmittingand receiving data. To better define the operation ofdigital data exchange RS-232 link signaling isaccomplished with voltages that range from +or –3vdc

to +or- 25vdc. If the voltage on the transmit or receivelines is positive (between +3vdc and +25vdc) thisrepresents a “0” bit; if the voltage is negative (between -3vdc and -25vdc), this represents a “1” bit, both withreference to the signal ground pin. RS-232communication is most used in implementing wide areanetworks where linking several LPFC’s is required overmany miles. This wide area link is accomplished via theDTE to DCE (modem/radio modem interface. RS-485on the other hand is a EIA standard for serialcommunications that uses a balanced system forsignaling and basically the same signaling voltages. TheRS-485 link can be used over fairly long distances (1000ft.) and at high baud rates such as 38.4 KB. This form ofcommunication is normally used for local area networks,offshore platforms, gas plants, etc. The link is establishedby using a single twisted pair (both transmit and receiveon the same set of wires) that is connected to each deviceand each device on the network having its own distincthardware and/or software address, (this also applies toRS-232 as well). This forms a bus topology that can bemade use of by network protocols. Since it is a verysimple and inexpensive topology, RS-485 is usedfrequently in the field connecting LPFC’s over a localarea network.

TELEMETRY

Various types of communication media are available tous today. An example of the types range from thefollowing:

• (PLM) Private line modem, communications aform of DCE that modulates over leased orprivate telephone lines.

• (SNM) Switched network modem,communication over the PSTN (public switchedtelephone network) modulate over PSTN forcellular or dial-up communication links.

• (FOM) Fiber optic modem, communication overfiber-optic cable usually 64 micron or 200 micronthickness cable.

• (Radio/Microwave, Spread Spectrum) Thesetypes of communication are capable ofextending long distances, with wirelesscapability.

INPUT/OUTPUT CAPABILITY

Additional demands have been placed on LPFC’s toprovide I/O capable to perform various functions. This I/O capability is performed by analog inputs and outputs(1-5 vdc or 4-20 mA dc) or discrete inputs and outputs(open collector or relay).

INPUTS

An example of signal inputs that may be interfaced withLPFC’s are digital or discrete open/close contactclosures.


WELL TEST

In the production industry for example a separator is usedto separate condensate, water, and oil from the well.Normally measured in barrels or tenths of barrels,accumulators register the amount of liquid passingthrough a turbine and produce frequency and or contactclosure outputs proportional to the amount of productproduced. The LPFC can register these outputs fromthese accumulators and present hourly, daily and monthlytotals.

Some LPFC’s based on a demand requested eitherlocally through a laptop computer or globally over thenetwork communication media may invoke a well test.While normal EGM calculations are occurring, a well testcan be initiated where over a predetermined amount oftime usually hours, a totalization of average DP, P, T, flowrate, total volume, condensate total, and water total canbe determined.

Contact closures from various alarms may be monitoredby the LPFC as well.

• Valve status• Security hatch• Gas level detection• Intrusion alarms• High levels• Low levels

These are just a few examples of inputs that may bemonitored by the LPFC.

OUTPUTS

An example of signal outputs that may be produced byLPFC’s are analog and digital.

VALVE CONTROL

Some LPFC’s provide analog outputs in the form ofvoltage or current, and digital output in the form ofcontact open collector or pulses. The output selectedwill depend on the actuator controlling the valve, usuallya electronic to pneumatic converter is used.

Analog output PID (Proportional, Integral, Derivative)control is a continuous signal to the valve, either 1-5vdcor 4-20madc. This output is usually determined manuallyby the user or automatic by the LPFC in that the valvemay be controlled based on a pressure variable or flowrate. This term is based on which variable is chosen tobe primary. The selection would be either pressurecontrol with flow override or flow control with pressureoverride.

Digital control uses a pulsed or intermittent voltage outputto control the valve. Both outputs analog or digital arebased on a variable chosen, pressure or flow and a setpoint at which the valve will be positioned.

SAMPLERS

Almost all LPFC manufactures offer the digital outputconfigured as a standard to pulse a gas sampler. Usinga predetermined volume of natural gas the user canconfigure a rate where which a digital output will activatea sampler. The rate or sampler activation interval is basedon the cylinder size and the amount of time it will take tofill the sample cylinder.

SUMMARY

In recent years, it has become apparent that the gasproduction, distribution and transmission companiesrequire greater accuracy and low power consumption.A rather extensive amount of gas research and standardscommittee activity is pointing to higher raw inputsampling rates, higher calculation frequencies, moreintensive calculations, more data storage, and, in general,considerably more work for the processors used in flowcomputers.

With deregulation, gas marketing and contractagreements are placing their own demands on theLPFC’s. They are now used for custody transfer. Specificdata must be available for billing and auditing. Flowcomputers must reside on communication networks toprovide data “now,” not “two weeks from now.” Thecontracts also dictate accuracy, which, in turn, dictatessampling rates of DP, pressure, temperature to includeAGA calculations.

To make matters worse for LPFC vendors, there is nouniformity in gas industry requirements. Every companyseems to have a unique need, be it the data that is stored,on what interval, how to do averaging and integration,what information is required over a communicationnetwork, the communication protocol, and so forth.

In addition, the field measurement personnel have theirrequirements. The flow computer should be easilyinstalled, calibrated, and started up. Ideally, it shouldinstall like a smart DP transmitter, not three transmittersplus a computer. The unit must also be low enough inpower consumption to make solar power or batteriesviable.

Greg Phillips


NEW TRENDS IN MEASURING NATURAL GAS FLOW RATESCharlie Harris

Honeywell Industrial Automation & Control1250 W. Sam Houston Pkwy., Houston, TX 77042

HISTORICAL PERSPECTIVE

Traditionally, the flow of natural gas has been measuredby a combination of pressure transducers, smarttransmitters, and flow computers. In the earliest typesof natural gas flow measurement, transducers andtransmitters were connected to flow computers tocalculate natural gas flow rates. In terms of the realmeasurements, these transducers and transmittersserved as the heart of flow computers. They still do innewer, smarter forms.

The pressure transducer is basically a sensor thatconverts one form of energy, pressure or mechanical, toan electrical form of energy. These early transducersprovided a low-level analog input, which limited theamount of information available to the flow computer orcontrol system.

INTRODUCTION OF NEW DEVICES

Once smart transmitters were introduced to the processindustry, they became the instruments of choice. A bonusfor the natural gas industry is that smart transmitters areable to accommodate the longer geographic distancesrequired for accurate measurement of natural gas.

A smart transmitter accepts a signal from an internaltransducer, digitizes it, compensates for temperature andpressure changes and generates a scaled signal strongenough for transmission over long distances.

Initially and in most transmitters available today,piezoresistive and capacitance technology was used. Atrue smart transmitter using piezoresistive technology ismicroprocessor-based, provides remote, digital, bi-directional communication, employs enhanced self-diagnostics, and is characterized for pressure andtemperature compensation.

Smart transmitters enable the accurate measurement ofdifferential pressure (DP) across an orifice for example,from which the flow rate is then inferred.

Unfortunately, this measurement is often flawed unlesspressure and temperature transmitters are also used.

With the introduction to the market of smart multivariabletransmitters and transducers, the mass flowmeasurement of natural gas has become much moreeconomical. With one device, the temperature, pressureand mass flow is directly calculated. This reduces the

number of field devices needed and importantly, providesa mass flow measurement that is highly accurate.

While smart transmitters could always measure absolutepressure (AP), they were only sensitive and accurateenough to be used to compensate the DP measurementfor static pressure changes. In the same way, atemperature sensor on the same silicon (dye) is used tomeasure temperature and compensate the DPmeasurement for ambient temperature changes.

With the advent of the new SMV 3000 multivariabletransmitter, a revolutionary new sensor was employed.This increased sensitivity sensor provides a highlyaccurate static pressure measurement also; a 0 - 400inch DP measurement with 0.075% of span accuracyand 0 - 1500 psia AP measurement with 0.075% of spanaccuracy. Dynamic flow compensation is supplied bycalculating in real time the discharge coefficient, gasexpansion factor and thermal expansion factor. Figure 1below shows a multivariable transmitter sensor.

FIGURE 1.Multivariable DP-AP Sensor

A multivariable sensor measures differential pressure,gauge or absolute pressure, and meter bodytemperature. The absolute pressure and meter bodytemperature measurements are also used in a processcall characterization.

Characterization involves measuring the DP at differentstatic pressures and temperatures as well as measuringthe static pressure at different temperatures. The datacollected from this characterization process is stored inthe transmitter. When a DP or AP measurement is made

PV diaphragm andsensing elements

Static PressureMeasurement

High Pressure

Meter BodyTemperatureMeasurement

LowPressure


in the field, this characterization process ensures anaccurate measurement under changing pressure andambient temperature conditions.

For reverse flow applications, this type of characterizationis important because it will allow negative DP ranges (0- negative 400" H2O). This capability for measuringreverse flows can be an important advantage in thenatural gas industry.

In addition, new applications with these instruments allowthe meter body to connect up with almost any electronicsused by flow computer manufacturers for astraightforward, uncomplicated mass flow rate solution.

ACCURATE MASS FLOW CALCULATIONS

Many flow measurement applications are inaccurate, asthere are no compensation calculations for densitychanges. The historically common use of a single variableDP transmitter to calculate volumetric flow, withoutcompensating for temperature and other effects, canprovide a highly inaccurate flow measurement.

The new trend is towards multivariable transmitters thatnot only measure DP, but temperature and pressure aswell. Using all these variables, the multivariabletransmitter is able to accurately compensate for anychanges in temperature and pressure to provide a muchmore accurate flow rate reading. Because these variablescan fluctuate considerably, the flow rates, either in massor volume, are considerably more accurate.

BENEFITS TO THE NATURAL GAS INDUSTRY

What does the increased capability of multivariabletransmitters and transducers mean for the natural gasindustry? It means that a complete mass flowmeasurement is readily available and much less costlythan ever before.

These costs savings are included in the purchase price,installation, reduced number of instruments needed, lessmaintenance due to less instruments, and the oftenincalculable benefits of having accurate flow ratesavailable for custody transfer billings, accurate processcontrol, and the like. These types of benefits can beobtained in applications such as shown in Figure 2 below.

FIGURE 2.Multivariable transmitters and flow computers are thecost-effective, mass flow measurement trend in the

natural gas industry.

With the use of a multivariable transmitters andtransducers in conjunction with flow computers, naturalgas producers and transmission companies have beenable to realize lower overall costs of ownership forelectronic flow measurements.

New multivariable transducer and transmitter technologywill continue to be introduced to the market and will boostthe efficiency of the natural gas industry, which translatesinto better business results and a greater competitiveadvantage.

RS-485Modbus RTU

Temp.


AUTOMATING GAS MEASUREMENTRichard L. Cline

Integrated Information Technologies10960 Stancliff Road, Houston, TX 77099

INTRODUCTION

Since the discovery of oil and gas and the advent ofcommercial conveniences, which use oil and gas,companies have been confronted with the need toaccurately measure the oil and gas bought and sold inthe marketplace. And, as usual, the technology availableat the time was brought to bear on the measurementprocess.

All gas companies must, of course, deal with gasmeasurement and are positioned somewhere on theautomation curve. As time moves forward, so does thetechnology. New products and measurement techniquesare constantly being offered to improve the gasmeasurement process. Unfortunately, adopting the newtechnology always brings with it a price. And the price isnot only measured in dollars, but in ever increasingdifficulty in making intelligent decisions and choices.

So how does a company, with the need to progress onthe automation curve, sort through the many optionsavailable today? The effort requires a continuingeducation process. The decision maker must understandnot only what the available technology can do for hiscompany today but must understand its future impacton the company.

THE RACE TOWARD AUTOMATION

The Gas Company cannot stand still, but must continueto push toward increased automation utilizing thecontinually changing technology. The question “ShouldI automate or not?” is no longer a valid question. Theanswer has to be “Yes” as competition and governmentregulation requires it. The new computer technologyoffers us ways to improve measurement accuracy andto reduce the amount of human resource required tomanage and complete the measurement process. Byincorporating the new technology, our cost is reduced.

So, faced with the need to automate, the more relevantquestion is “How do I design my system and choose theproper system components to meet my corporateobjectives?”

PROBLEMS FACED WITH AUTOMATION

During the course of an automation project a number ofquestions, issues, and problems will surface and haveto be addressed. How they are resolved will dictate the

ultimate success of the project. Some of the more criticalquestions are:

• What are the short term and long term corporateobjectives? Is the proposed automation stepconsistent with these objectives?

• What are the shortcomings and inefficiencies inthe company’s current measurement process?

• Is the proposed new technology wellunderstood? How closely does it match thedesired solution? How much of it is enticingchrome or dazzle which is generally of short-term value. What kind of functionality is “underthe hood” that will really help the business overthe long haul? Does the technology offer realbenefits and meaningful features to the currentmeasurement process?

• Will the technology eliminate or reduce currentinefficiencies and costs? If so, how and at whatprice?

• Is the technology both a good short-term andlong-term solution?

Most technological advances are geared towardparticular markets and may not provide the best solutionfor a particular application. This can be particularly truewith software.

No solution comes without a price. This phrase appliesheavily to the gas measurement system solution and itis imperative that the decision-maker who is responsiblefor selecting the right system components evaluate allaspects of a particular solution.

A pertinent example is the continuing rush to adapt thelatest graphical user interface (GUI) advances to SCADAor communications intensive functions at the expenseof operational functionality. Standardized point-and-clickinterface benefits are recognized and are essential inoperating today’s systems. Operators not familiar withcomputers can quickly become computer literate andlearn to operate new programs. But what about the on-going long-term price to be paid after the operator hasachieved computer literacy and familiarity and is nowlooking for functionality related to his or her day-to-daytasks? Experience may indicate that added graphic“flash” in an already user-friendly system may just create


more layers between the operator and important dataand may steal valuable “horsepower” from thecommunications-intensive system environment.

So an objective of the successful project must thereforebe to keep operational functionality at the top of thepriority list.

The project’s scope must be defined. Is the focus onprocuring a “system” solution to meet “systemrequirements or is the decision-maker focusing only onindividual components without thought toward how theywill be integrated and work together?

A system solution requires the right mix of componentsand, more importantly, an appropriate amount of systemengineering to ensure that the various components playtogether properly. Is the Company going to do the systemengineering? If not, then the level of required outsideengineering and integration services needed must befactored into the process. The required level and cost ofoutside engineering will be greatly impacted by theselection of software and hardware components andvendors.

THE CHALLENGE

Why is the planned automation path such a riskyendeavor? Because the right answers to the questionsand issues encountered during the project are not easyto come by. Several major reasons can be identified.

• The technology is changing so fast that evenvendors in the business find it hard to keep up.

• Vendors bombard the decision-maker with awealth of confusing information and complexproducts, all seemingly designed to solve everycurrent and even future problem.

• Most vendors prefer to sell components andtend to avoid the responsibility of offering a totalsystem solution as this requires a diversity ofexpertise and technologies which typically goesbeyond the vendor’s product line.

The following approach or process, if followed, will helpthe decision-maker wade through the confusion in aconstructive manner. This process leads to intelligentdecisions based on real data. Key steps in the processare:

• Remain focused on the basics, the Company’sobjectives and the desired functional solution.

• Assess and evaluate all available systemcomponents in terms of the basics and how theycontribute to the overall system solution.

• Ask vendors the right questions.

• Let the accumulated data lead you to the bestsolution.

GAS SYSTEM COMPONENTS

A technical evaluation of the components in the gasmeasurement system requires that the decision-makerunderstand fundamentally how they work and what theycontribute to the process. The following discussion notonly provides this overview, but more importantly,identifies some pertinent system related questions whichshould be asked as part of the evaluation. Assessingthe answers to these questions will lead to the bestchoices in selecting the system components.

SENSORS AND TRANSMITTERS

The computation of gas flow, i.e. flow rate, through apipeline by an electronic flow meter (EFM) or a smarttransmitter requires various measurement inputs suchas gas temperature and static pressure.

Several standard technologies have been around for anumber of years. Orifice measurement, the mostcommon, uses the principle that the gas pressuremeasured via a differential sensor behind a restriction inthe pipeline (the orifice plate), is inversely proportionalto the velocity of the gas through the orifice. Thismeasurement when compared to the static pressure infront of the orifice is used to compute the flow rate. Asecond common technique uses the number ofrevolutions of a rotary or turbine meter sensor in thepipeline to compute the flow rate.

The technology associated with pressure andtemperature sensors is, today, quite well understood.However, advances are continually surfacing in regardto design, accuracy, and cost. Traditional sensors outputa current or voltage which is proportional to the measureditem and is used by the EFM to obtain the measurement.A new trend today involves “smart” sensors ortransmitters with a digital interface. This type of sensorcan be useful for applications where the measurementsobtained from the sensor or transmitter are the onlymeasurements required from the field. The hostcommunications system can interface directly to thesensor without requiring intermediate RTU or EFMequipment.

Another new class of smart transmitter is available andcan be useful in applications such as plant automation.These transmitters actually perform the American GasAssociation Report No. 3 (AGA-3) flow rate calculationand maintain a history of hourly and daily flow. Thesesensors, however, cannot be considered functionallyredundant to EFMs until they fully implement all requiredcalculations, e.g. the AGA-8 compressibility calculation.Use of this transmitter is generally limited to applicationswhere the gas composition is relatively constant or wherethe host system can automatically download newcomposition and compressibility information. Future


benefits to gas measurement from smart transmitterswill be the improved accuracy and lower cost as theyevolve into functional replacements for the traditionalEFM.

A relatively new technology is available today and canoffer cost effective benefits in certain applications. Theultrasonic flow meter, although costly compared to theabove more common technologies, offers substantialindirect cost savings as this technology uses soundwaves to measure flow rate and therefore does notrestrict the gas flow. This can mean real savings in thecost of compressor stations on long transmissionpipelines.

System issues related to sensor technology are asfollows:

• Typical interface between the standard sensorand the EFM is a wiring interface adhering to astandard current of 4 to 20 milliamps or a voltageof 1 to 5 volts.

• As the sensor must be installed in hazardousareas, equipment must be selected whichadheres to applicable ANSI and NFPA standardsfor those areas.

ELECTRONIC FLOW METER (EFM)

An essential component of the measurement system,this computer-based device computes the flow ratebased on sensor inputs. AGA has defined thecomputations to be used for flow calculation for custodytransfer. The AGA-3 standard relates to orificemeasurement calculations and AGA-7 to turbine metercalculations. A relatively new standard, AGA-8, moreaccurately computes the compressability factor to yielda more accurate flow rate calculation.

EFM technology is as well understood as sensortechnology and does not pose much mystery to thedecision-maker. There are, however, major system-related issues which must be addressed when selectingEFM equipment:

I/O CAPACITY

The move today is toward integration of technologies toreduce costs. Flow-related data represent only a part ofthe information desired from the some field sites.Interfaces for compressor alarms, valve controllers, tanklevel sensors, and fugitive emission sensors, e.g. H2S“sniffers”, are being integrated for transmission over acommon comm link to the field office or central hostsystem.

In the past, both EFM equipment and RTUs or PLCshad to be installed at the same site to gather the requireddata. This approach also required a duplication of commmedia. In the future, the need to consolidate all

measurement into the same box will be a priority toreduce hardware and communications costs.

PHYSICAL COMM LINK SUPPORT

The system solution requires that EFM data betransmitted to the user’s office. This requires a commmedia link between the office and the field site. In general,an RS-232 asynchronous interface on the EFM will beacceptable for the communications interface; however,should the site be such that multiple “clustered” devicesneed to be interfaced in the future, then an RS-485 multi-drop interface in the EFM will be useful. A singlecommunications link could, via a data concentrator, beused to link the clustered devices to the remote office.

PROGRAMMABILITY

The selected EFM unit should offer the end-user theability to easily configure the unit.

ARCHIVAL STORAGE

The host communications software periodically accessesthe EFM to upload the archived hourly and dailyinformation. The EFM must retain ample hourly and dailyinformation to compensate for a worst-case scenariowhere the host cannot access the data for a period oftime. Most units today archive at least a month of hourlyand daily data.

SUPPORTED COMMUNICATIONS PROTOCOL

Communications to the host communications softwareis via a language or protocol. This issue is importantenough to be discussed later as a key systemcomponent.

HAZARDOUS RATING

As the equipment may need to be installed in hazardousareas, equipment must be selected which adheres toapplicable ANSI and NFPA standards for those areas.

DATA CONCENTRATION DEVICES

Technology is available today to allow multiple sitedevices to be polled locally by a data concentrator inthe field. This device can offer a number of advantagesin certain applications:

• Reduced communications costs. Interface to thehost system communications software is via asingle communications media link.

• Report-by-exception, whereby alarms arereported to the host when they are detected,can be supported by the concentrator.

• Native protocols for the field devices and for thehost can be supported. The concentrator can


gather data from the field devices using thenative protocols of the devices and report thedata to the host using the native protocol of thehost thus allowing an already in-place hostsystem to gather data from multi-vendor fielddevices.

THE COMMUNICATION PROTOCOL

The communication protocol is the language used tocommunicate with the host communications software.Supported communication protocols are an extremelyimportant issue for consideration when choosing EFMs.Unfortunately, the required system solution often involvesmulti-vendor equipment in the field with differentprotocols. The consequences of this and suggestedsolutions are discussed later while looking at the hostcommunications software component. Here are generalguidelines for the decision-maker:

• Avoid proprietary protocols. Arguments that aproprietary protocol is desirable or necessary toprovide data security are not significant enoughto offset the ultimate cost of this decision. Aclosed protocol is not necessary to provide asecure interface to EFM data. Many EFMsrequire a security password to allow login andwill reject any attempt to login without theappropriate password. The consequence of aclosed protocol is to leave the end-user withfewer options for integrating equipment frommultiple vendors.

• Acquisition of data from multi-vendor equipmentis greatly simplified if a common protocol canbe used for interface to the equipment.

• Select, when possible, equipment whichsupports a well-behaved, open protocol, e.g.MODBUS. A well-behaved protocol would beone which supports relatively small messagepackets rather than large data dumps. Reliabletransmission of data dumps over potentiallyunreliable comm links such as cellular can bevery difficult. An open protocol would be onewhich is non-proprietary, well documented, andsupported by a large number of vendors.

• Use fully implemented robust protocols ratherthat partial implementations. An EFM vendormay support a particular protocol but in realitycannot provide a robust interface to the hostsoftware using that protocol. For example, thesupport may only provide access toinstantaneous data and not archived historicaldata. Access to the historical may require thevendor’s proprietary protocol.

The industry familiar MODBUS protocol can provide anexample of these criteria. Access to EFMs is often bydialup telephone requiring a long distance phone call.

The original standard MODBUS protocol, as defined bythe Gould Modicon specification, only supports theacquisition of integer data from registers. That is, onecan only ask for “single layer” data from the device. Sometime ago it was recognized that this specification neededto be expanded to accommodate the need for floatingpoint measurement data and for “multi-layer” historicaldata. An example of an extended MODBUS definition,which is widely used today, is the non-proprietary DanielExtended MODBUS protocol. An EFM supportingMODBUS without extensions may need to be calledevery hour if hourly data is needed, whereas, an EFMsupporting the Daniel Extended MODBUS protocol canbe accessed infrequently, e.g. daily or weekly, to acquiremultiple hours of historical data. This type of access isessential where hundreds of EFMs must be contactedvia long distance phone calls.

THE COMMUNICATIONS MEDIA

Host-to-EFM links can use phone lines, cellular, radio,and satellite interfaces. Selecting the propercommunications media for the office to EFM link mustfirst be based on the type of interface required to satisfycompany objectives. Do the planned system functionsrequire that the interface be a dedicated real-time two-way interface such as for the gas control function; or,can the system needs be met by a once-a-day or once-an-hour data acquisition approach. Once the candidatemedia are defined, the decision is usually based on aconsideration of initial installation cost and on-goingoperational cost.

Some other system related issues should be considered.

Site limitations.

Does the site have cellular coverage, satellite coverage,power, available and line?

Data reliability.

How reliable is the media for data? This is primarily asissue with media such as landline telephone and cellulartelephone, which were initially designed for voicecommunications. Landline modems used in the systemshould, where possible, support error correctionalgorithms. Cellular links can suffer special problemssuch as signal fade and cell switching requiring the useof special cellular modem protocols such as MNP-10 tohandle these problems effectively.

Host limitations and issues.

How many communication ports are available in thehost? Each type of media to be used will typically requireat least one dedicated port on the host.

Can the host software handle different protocols overthe same comm port?


Can the host support different poll frequencies over thesame comm port? Can the user relax the timeouts in thehost if needed to accommodate retry delays in errorcorrecting modems?

Governmental restrictions.

Is FCC licensing required for the media as is the case for900 MHz standard radios? Can the license be obtainedand how long does it take?

HOST SYSTEM SOFTWARE

Selecting the best host system software to support thecompany objectives is perhaps the most difficult task inconfiguring the gas measurement system. Thistechnology, involving both hardware and software, ischanging daily. The decision-maker is constantly beingenticed to accept new “state-of-the-art” capabilities.

In this environment, some practical advice is to not forgetthe company’s fundamental objectives and requirementsfor the system. These basics are the major functionswhich should be provided by the host software:

• Field Communications

• Data Archival

• Data Presentation

• Data Editing

• Data Distribution

• Custom Reporting

• Applications Platform

These functions are applicable to all gas measurementsystems encompassing the smallest single PC systemwith few EFMs to those running on networks withhundreds of EFMs.

No single host software package can provide the best ineach area. Choosing the perceived best in one area mayimpose constraints in another. For example, assigningthe communications function to the GUI presentationsoftware with the most “flash” may not provide thedesired system solution to support future growth in thecommunications and applications arena.

So how does the decision-maker deal with this dilemma?A good starting point is to define explicitly the currentand future requirements in each basic area and then touse this as a yardstick in evaluating each softwarepackage. Do not hesitate to press for benchmarkinformation documenting the expected communicationsperformance once the system expands to its largestplanned size. From this evaluation, the decision-makerwill understand whether a single system package is

adequate or whether an integrated system concept isrequired.

The integrated system concept is one where the selectedhost system is really an integration of several differentsoftware packages, each dedicated to a particularsystem function. For example, communicationsprocessors, a master station server, and console stationsfor the operators can together comprise the host system.T h e c o m m u n i c a t i o n s p r o c e s s o r s h a n d l ecommunications efficiently, the master station hasresponsibility for data applications, data archiving, datareporting and data distribution functions, and theconsoles provide the user-friendly GUI interface for theoperator.

Selecting the best host components should be basedon an evaluation of available functionality, design, andperformance for each of these areas. For example, themaster station server should offer a large number andvariety of available vendor applications to do the neededdata manipulation. Selecting an integrated system froma single vendor offers the advantage of provencomponent interfaces but a possible disadvantage oflimiting functionality. Choosing to integrate componentsfrom multiple vendors allows more freedom in selectingthe best functionality but will require considerable systemengineering to integrate the system.

Seeking to answer key design-related questions in eachfunctional area will help the decision-maker identifypotential constraints and pitfalls in selecting the hostsystem.

COMMUNICATIONS

The design of this function in the gas measurementsystem is crucial and the capabilities of the selected hostsoftware in this area should be closely scrutinized.

The vendor’s operating system platform for this hostcomponent should also be closely scrutinized. Acontinuously operating, reliable and field-provenenvironment is required. Avoid initial releases of the latestand greatest versions of any operating system until bug-fixing revisions are available. A nuisance bug in theoperating system can bring down your communicationsand the only fix may be a work-around by the hostsoftware vendor.

The communications software component should, if atall possible, be assigned to a dedicated platform withunneeded options de-activated in the operating system.No potentially interfering software should be installedon the platform. Pertinent communications-relatedquestions are:

• What is the priority of the communicationfunction relative to operator keyboard activity?Can operator or network-related activity impactthe system’s ability to efficiently acquirenecessary data from the field?


• Does the system provide online communicationsanalysis tools to support troubleshooting? Doesit allow the system administrator to capture andreview bi-directional asynchronouscommunications at the port level?

• What is the practical limit on the number of EFMswith which the system can communicate? Canthe vendor provide benchmark test dataquantifying performance when communicatingwith the planned number of EFM devices?

• How does one add a new EFM device to thesystem? Is this task simple or complicated?

• Are operator-initiated communications requestshandled at higher priority than scheduledcommunications tasks?

• Can the system support multiple communicationports with the same field device protocol? Withdifferent protocols? How many ports can beused for concurrent communications?

• Can the system support concurrent interface tomultiple types of communications media?

• Can the system support different protocols todifferent types of equipment on the samecommunication port?

• Can the user fine-tune communications byconfiguring command timeouts, retries andpolling frequencies at the command and devicelevel, or only at the port level?

• Can the system support unsolicitedcommunications or just solicited master/slavecommunications? For example, can it listen forand respond to a field device calling in with areport-by-exception alarm?

• Does the system support download of date/timeand configuration data to the EFM?

• After a communications failure period, does thesystem automatically access historical dataarchives in the EFM to acquire uncollected hourlyand daily data?

• Is information from different types of field devicesintegrated and archived into a single non-vendorspecific format?

Figure II shows a multi-component project where masterstation software is responsible for collecting data from anumber of different field devices.

DATA ARCHIVAL

The system must archive all received historical EFMhourly and daily data. Here are important questions forthe decision-maker:

• Is received hourly and daily data archived in sucha way as to eliminate duplicate data should thesame data be received more than once?

• Is the data archived and reported with the date/time stamp from the EFM rather than the hostsystem time stamp?

• Can the system be configured to archive thecollected data directly to folders on a networkserver? If so, are the files treated by the systemin a network-aware fashion to allow multiplenetwork users to access the archive filesconcurrently?

DATA PRESENTATION

Some data presentation requirements will be a functionof the available system environment. Whether thisenvironment is Windows, Unix, DOS or LAN-based, theselected MMI interface software for the gas controller ormeasurement operator should provide a meaningful anduser-friendly access to current and archived EFM data.Decision-maker questions are:

• Is the operator’s access to the informationintuitive? How long is the vendor’s suggestedoperator training program? Is it reasonableconsidering program content? An unusally longprogram can indicate lack of intuitive design.

• Is the data presentation and arrangement logicaland pertinent to the operator’s day-to-dayfunction?

• How many layers, or clicks, does the operatorhave to pass through to review correlated data?

• Does the system provide a mechanism forreviewing and analyzing historical data trends,i.e. changes in measurements over time?

• How are detected alarms annunciated to theoperator? Can alarms and their annunciation beprioritized from non-critical to critical?

• Can the presentation be customized? Are toolsprovided with the system to allow the user toimplement a hierarchy of overview displays?

The user friendliness of the interface and the functionaldesign features available for interfacing with the archiveddata are areas of major importance as they define theease with which daily measurement tasks will beaccomplished.


For users of Windows presentation software who wishto access EFM data via the Microsoft DDE interface, theselected host server software must be able to act as aDDE server or to pass the data to a separate DDE serverprogram.

DATA EDITING

The selected host software must provide a mechanismfor copying the acquired EFM hourly and daily data intoan area for editing and for AGA-3, AGA-7 re-calculation.There will be times when the EFM field data is found tobe erroneous and flow rates and volumes must be re-computed in the host system. An example would be achange to a different size orifice plate and the new sizeis accidentally not updated in the EFM. Decision-makerquestions are:

• Does the editing interface prevent the user frommodifying raw field data? This is essential as thefield data must be preserved in its unalteredform.

• Is the operator interface to the archive data forcopying, editing, and re-calculation user-friendly? Although this is a subjective measure,the interface which the measurement operatoruses day-after-day should be logical and easyto use.

• Does the system in a LAN environment allowmultiple measurement personnel to workconcurrently on pre-defined subsets of thearchive data?

• Does the system provide an audit function tocheck received field archive data for consistencyand to flag data deemed incomplete orquestionable? This function is highly desirablein a system with many EFMs.

DATA DISTRIBUTION

After the measurement data has been reconciled(checked, edited, and saved), others in the organization,particularly the accounting group, need to have accessto the information. The mechanisms available in the hostsystem for distribution of this data should be evaluated.Some pertinent decision-maker questions are:

• Can the reconciled measurement data be storedon a LAN server for access by any and allauthorized persons on the network?

• Can the data be automatically transferred to theaccounting mainframe in a usable format?

• Can the data be automatically stored in acorporate database?

• Can the data be directed automatically to a website for internet access by gas brokers,marketers, and producers?

• What network interfaces are possible with thesystem? Can the system function as a TCP/IPserver to deliver archived data to other users?Can the system function as a TCP/IP client todeliver data to a server?

CUSTOM REPORTING

The host system should provide features for installingcustom reports to be printed automatically and onoperator demand. Pertinent decision-maker questionsare:

• Can the system support multiple printers, suchas one for reports and one for alarms and auditevents?

• Can the system automatically direct reports toa TCP print server on the network?

APPLICATIONS PLATFORM

The selected host system platform design should supportongoing installation of add-on applications. Ideally, theuser should be able to install both new applicationsavailable from the host vendor and custom applicationsdeveloped in-house. Some systems today offer the usera large selection of applications from the vendor’sapplication library, e.g. accounting applications tomanipulate and analyze EFM data and energy de-regulation related applications for handling gasnominations and gas marketing. These features allowthe user to enhance the system as needs change withoutvendor support.

Selecting the right host software is a difficult task. Thevariety of host system capabilities and options make thetask even more difficult.

An additional important issue relating to the overall hostsystem is that of vendor support. Decision-makerquestions are:

• How many software vendors are represented bythe proposed components of the host system?If multiple vendors are represented, then theinterfaces between the software products mustbe well defined to minimize finger pointing whena need for vendor support arises.

• What kind of support will be available after-the-sale to fix problems or implementenhancements?

In general, obtaining timely and relevant support fromvendors of commercial “off-the-shelf” software may bedifficult. If the decision-maker intends to utilize this type


of software in the measurement process, he should makesure that the software contains the necessary functionsand features to meet his current and future expectations.

On the other hand, one should expect and demandquality support from a vendor offering softwarespecifically designed for the gas measurement system.The decision-maker should ask for and check referencesregarding software performance and the quality of after-sale vendor support.

Vendors whose primary product is field hardware maybe reluctant to provide software enhancement support.Although some software packages from EFM vendorsare highly functional for their equipment, the decision-maker cannot generally count on the vendor to developdriver interfaces to competitors’ EFM equipment. Also,as this software is generally offered free or at low price,the adage “you get what you pay for” may apply. Thesevendors will typically not be eager to implementenhancements. The decision-maker should investigatesoftware sources for the host software whose specialtyand primary product is the software itself.

HOST SYSTEM ARCHITECTURE

Figure 1 shows an integrated concept gas measurementsystem configured for a LAN network platform.Component functions are as follows:

• Host Communications master stations or serversare responsible for collecting and archiving allfield EFM data. These stations can also serveas front-end communications processorsresponsible for the interface to gas control RTUsand PLCs in the field. As the network grows insize, additional master stations can be addedeasily. Measurement data is archived on the LANNetwork Server.

• Console stations are used by measurementpersonnel to reconcile the data. The reconcileddata is again stored on the LAN Server.

• SCADA gas control stations provide for gascontrol functions with communications via thecommunications servers.

Important advantages of the integrated conceptarchitecture are:

• No duplicate hardware function orcommunications media links are needed.

• Architecture is applicable to small or largenetworks.

• Architecture can combine the benefits of a high-performance master station server andcommunications platform with GUI basedplatforms for operator stations.

CONCLUSION

The task of designing, procuring, and commissioning agas measurement system is complex and tedious. Askingvendors the right questions and seeking information fromothers with experience will lead to the right system forthe decision-maker. The key components and basicfeatures required in a gas measurement system are thesame whether the system is small with a single PCplatform or is large with a LAN network. Keep the focuson the required basic system functions to be providedand evaluate the system in regard to them. This is farmore important than being influenced by the day-to-daybells and whistles offered by vendors to hock their wares.Don’t be afraid to ask to see “under the hood”.


Richard L. Cline

FIGURE 1


TRAINING FIELD MEASUREMENT PERSONNELRussel W. Treat

Gas Certification Institute, LLCP.O. Box 131525, Houston, TX 77219-1525

INTRODUCTION

Technology in the field of gas measurement and controlis constantly evolving. While many are well training inthe specific equipment used in their own company’soperation, it is important to have a solid understandingof the fundamentals and theory of operation of themechanical and physical process involved as well.Therefore, the training of field measurement techniciansis of the utmost importance. These technicians must becontinually educated in order to possess the most currentknowledge of the latest equipment, electronics,communications and metering devices on the market.Also, it is essential that this type of instruction should betaught in a controlled environment where the technicianscan learn and develop the necessary skills with the leastamount of interruptions from external sources.

REASONS & BENEFITS

Listed below are some of the most significant reasonsfor training field measurement personnel.

1. To develop a thorough understanding of gasmeasurement in order that each technician cancontribute to maximizing measurement accuracythereby directly affecting the revenue generatedthrough the sales, purchase or transportation ofnatural gas.

2. Gas measurement equipment is consistently beingupgraded to accommodate for the changes intechnology and the rule making policies required byFERC. Therefore, the technician has moreresponsibility to operate and maintain themeasurement equipment according to themanufacture’s specifications and the standardsmandated by the Commission.

3. Formal classroom training provides better control forconsistently implementation of company policy andprocedure than “On the job training” where seniortechnicians pass the “tricks of the trade” down toan apprentice, are no longer an acceptedmethodology for training gas measurementpersonnel.

4. Hands On Training in a “Live Gas” environment issuperior to classroom training only, as training isdelivered in virtually the same environment as thetechnician faces in the field.

5. Training on the most current measurement devices,electronics and software provides even the seniortechnician with the skill to properly and efficientlyimplement state of the art techniques into their dailyroles and responsibilities.

HANDS ON & LIVE GAS

The most effective method of teaching and trainingshould be done under actual operating conditions. Thistype of learning technique has a greater impact on thetechnicians because all the training utilizes equipmentthat is under line pressure and contains natural gas. Whena technician is wearing their safety equipment andworking on flow or pressure controllers in a live gasenvironment, the schooling procedures of “hands ontraining” create a realistic atmosphere where thetechnicians learns to perform their tasks under actualconditions.

It is clearly apparent that learning under authenticworking conditions in a controlled training environmenthas a definite advantage. Through this trainingexperience, technicians learn through experience todevelop the problem solving expertise that is necessaryto develop and enhance their troubleshootingtechniques. At the same time, technicians learn propersafety procedures.

In this way, newly developed skill are immediatelytransferred to daily operations for resolution of operatingand/or maintenance problems as they arise in the field.Each technician should be trained so that he or she canhandle and resolve a wide range of complex problemswhen working with gas measurement and controlequipment.

In order to be a productive employee in today’s market,it is imperative that each technician receives training ondifferent types of measurement equipment available inthe gas industry. These diversified skills and knowledgewill enable each technician to be a valuable employeewho has the ability to increase their company’s net profit.

EXAMPLE CURRICULUM

The following is an example curriculum for training ameasurement technician “from scratch” to a completeand certified measurement technician.


Course No. Description Prereq. LengthBasic

MATH 101 Principles of Gas Measurement 10 days(Basic Mathematics & Fundamental Gas Laws)

VCAL 102 Natural Gas Volume Calculations MATH 101 5 daysORIF 103 Orifice Metering VCAL 102 5 daysCHRT 104 Chart Changing

IntermediateSMPL 201 Gas Sampling 2 daysMOIS 202 Determining of Moisture Content SMPL 201 2 daysSPGR 203 Specific Gravity SMPL 201 5 daysTURB 204 Turbine Meter VCAL 102 2 daysPOSM 205 Positive Displacement Meter VCAL 102 3 daysCONT 206 Principles of Automatic Control (Controllers) 5 daysREGL 207 Control Equipment CONT 206 3.5 daysODOR 208 Natural Gas Odorization 5 daysCALO 209 Calorimeter SMPL 201 3 days

AdvancedDCEL 301 Basic DC Electricity 10 daysELTR 302 Basic Electronics DCEL 301 10 daysINST 303 Electronic Instrumentation ELTR 302 10 daysCHRM 304 Chromatograph DCEL 301

SMPL 201 3 daysFCOM 305 Basics of Flow Computer Operation & Maintenance INST 303 10 days

Any training curriculum should include testing and skillsdemonstration to assure effective learning by thestudents. In addition, it is important to note that acurriculum should be tailored to the needs of the studentpopulation to optimize investment in the program.

MATH 101 — Basic Principles of Gas Measurement

This course is designed for classroom presentation aswell as for “on the job” study. It deals with both principlesand details including hands-on training.

This course develops skills in basic mathematics. Thisability enables the student to understand and use thesimple equations encountered through the course offurther training.

Part 1 — Basic Math

A great deal of work done in the measurement of naturalgas cannot be accomplished without a generalknowledge of mathematics. The mathematical functionsnecessary to calculate areas, volumes and flow througha pipeline or an orifice plate are presented in this course.

Part 2 — Fundamental Gas Law

The absolute pressure, absolute temperature and volumeof gas are very closely linked. Changes in any of thesevariables cause changes in one or both of the others.

Therefore the behavior of the gas is reviewed in order toenable visualization of physical processes involved whenthese changes occur.

The study includes purpose and principles ofmeasurement equipment such as manometers, pressuregauges, dead weight testers and recordingthermometers.

Boyle’s Law, Charles Law, deviation from Boyle’s Law,and standard units of measurement are the particularsstudied in order to obtain a working knowledge of therelationship between pressure, temperature and volume.

VCAL 102 — Volume Calculation

In this course, the various correction factors used tocalculate gas flow through an orifice are studied in detail.Also calculations pertaining to gas flow through positiveand turbine meters are studied.

ORIF 103 — Orifice Metering

Flow measurement by means of an orifice is studied indetail in this course. Also, the theories of orificemeasurement and the physical application are presented.AGA-3 guidelines pertaining to actual dimensions oforifice meter runs and their appurtenances are reviewed.The bellows type orifice meter is also studied in detail.


This “hands-on” training course provides the participantthe opportunity to perform inspections to determine ifan orifice meter meets AGA-3 specifications. The orificemeters are also inspected, adjusted, and calibrated underactual (gas flowing) conditions in order to teach normaloperating and safety procedures to the participants.

CHRT 104 — Chart Changing

In this course the importance of good chart changingpractice is discussed. The responsibilities of a “chartchanger” are reviewed and how the work is related tothe other measurement functions. Also, the properoperation of equipment used by a chart changer isstudied and performed by each student.

SMPL 201 — Gas Sampling

Natural gas is sampled for many reasons to determinequality and quantity. Techniques for sampling must varyaccording to the type of test for which the sampling isdone. Locations of sampling points, sample size, samplepressure, when and how the sample is taken are alldependent on the desired end result. Participants willacquire knowledge of the general purpose for which asample is being taken prior to actually taking the sample.Participants will then learn industry accented methods:to transfer a representative sample from a source (usuallya pipeline) into a transporting device (usually a samplecylinder), to transport the gas from the source to the labwithout affecting the representative sample, and toremove the sample from the transporting device anddivert it to the measuring device without distorting thesample.

MOIS 202 — Determination of Moisture Content

Excessive amounts of water vapor in gas can condenseand form liquid or ice-like hydrates which inhibit the flowof gas. Even as a vapor it takes up space, which couldbe occupied by gas. Water can also combine with othercontaminates such as CO2 and H2S and form acids thatcan corrode the pipe. For these reasons, water vapormust be removed from the gas stream. This coursediscusses in detail the methods used for determiningwater vapor content and the industry standard “7 lbs ofwater per MMCF.”

SPGR 302 — Specific Gravity

By definition, specific gravity of gas is the ratio of itsdensity - or weight per volume - to the density of air. Inmeasurement work, especially when using formulas forcalculating amounts of flowing gas, the specific gravityof the gas is an important factor. This course discusseshow specific gravity is measured and used and discussesthe equipment used to determine specific gravity.

TURB 204 — Turbine Meters

Flow measurement by means of turbine metering isstudied in detail in this course. Also, the advantages and

disadvantages of using a turbine meter, along with itsoperating principles are presented. AGA-7 guidelines forturbine meter runs and their appurtenances are reviewed.

POSM 205 — Positive Displacement Meter

This course begins with understanding the principles ofa positive displacement meter and how its individualparts operate. The “Mcf” and how to read the PositiveMeter Index are discussed. Much time is devoted tolearning the best techniques for repairing positivedisplacement meters.

CONT 206 — Principles of Automatic Control(Controllers)

Automatic controllers to control pressure and flow rateare useful tools, but to use them one must understandthe basis principles of automatic control. In this course,the basic principles are given in everyday words. Basicresponses of controllers are illustrated with common,familiar devices. Simple graphs show how the measuredvariable acts under regulation by an automatic controller.“Proportional band, reset action, derivative response,offset” and other terms in the language of instrumentationare simply explained to help people who are notinstrument specialists.

REGL 207 — Control Equipment(Valves and Regulators)

This course encompasses the study of fundamental gaspressure regulation with special emphasis on theregulator’s operation. The essential elements of aregulator and function of each element are thoroughlydiscussed. The first part of the course is devoted to “self-operated” and “pilot loaded” regulators. The course thengoes into a thorough investigation of the operation anddifferent applications of expansible tube type regulators(for examples, the Grove Flexflo).

The third part of the course deals with the selection ofcontrol valves. Special attention is given to Fisherdiaphragm operated globe valves and also ball valveregulators. Split range control and valve positioners arediscussed.

This is primarily a “hands-on” course where the studentsdisassemble and reassemble regulation equipment inorder to gain complete understanding of their operation.Students will also field adjust regulators, relief valves,and control equipment on a live natural gas station.

ODOR 208 — Odorization

This course is intended to provide a completeunderstanding of natural gas odorization. Information ispresented on the many aspects of odorization includingodorant compounds, odorization equipment, testmethods, and appropriate record keeping.


In order to comply with Department of Transportation’sPart 192.625, odor level instruments must be used toassure proper concentration of odorant. Thereforeseveral odorant level test instruments are discussed andactual tests performed to provide training, utilizing thebest industry accepted methods.

CALO 209 — Calorimetry

Most natural gas is used as a fuel to produce heat or toproduce energy in some other form such as steam orelectricity. Since the heating value of natural gas variesdepending on its composition, we must have somemethod of determining the heating value of the particulargas being delivered. In today’s gas industry, themeasurement of heating value is of major importance.Correction factors for heating values are being includedin many contracts when buying, selling or transportingnatural gas.

There are several instruments on the market today thatcan be used to obtain the needed information from asample of natural gas to calculate a heating value. Thiscourse studies and elaborates on one of these - theCutler-Hammer Recording Calorimeter.

DCEL 301 — Basic DC Electricity

This course is primarily designed for the entry-levelstudent, where no sophisticated math background orprevious knowledge of electricity is assumed. Therefore,he or she will be able to learn the basic concepts thathave enabled man to harness and control DC electricity.

ELTR 302 — Basic Electronics

Electronics is a field of study that comprises manydifferent components, circuits, and systems. In theinterest of time, only those areas that affect electronicmeasurement equipment will be studied; however, otheritems may be briefly discussed. Digital electronics willbe the main emphasis of this course.

After some preliminary material is covered, a variety ofelectronic components will be studied (i.e., diodes,transistors, integrated circuits, etc.). Then, someelectronic circuits will be presented (i.e., amplifiers,oscillators, power supplies, etc.). The bulk of the coursewill be devoted to digital electronics (i.e., number system,logic circuits, counters, registers, memories, etc.). Also,analog to digital (A/D) and digital to analog (D/A)conversion will be included.

Electronic test equipment i.e.: multimeter, oscilloscope,etc.) usage will be included in the laboratory typeexperiments. Overall, this course will provide a very goodunderstanding of the workings of electronic systems.

INST 303 — Electronic Instrumentation

Measurement technology is rapidly changing and oftenin today’s environment physical variables are obtainedelectronically. With the electronic system, the variablesare measured and calculations are made instantaneously.The variable measurements (differential, temperature,etc.) are made by a variety of transmitters or transducers.These devices along with their practical applications andoperations are thoroughly reviewed.

CHRM 304 — Chromatograph

A chromatographic gas analysis provides a quantitativebreakdown of gas composition. It is therefore thepurpose of this course to provide an understanding ofthe principles involved and training in the operation ofspecial equipment used to obtain a gas analysis.

FCOM 305 — Electronic Flow Computer

This course has been prepared to teach operations andmaintenance of electronic digital field computers withspecial emphasis on gas flow computers. The contentof this course assumes understanding of basic DCelectricity, the use of volt ohm and current meters, andfamiliarity with basic circuit components. However, noknowledge of computers on the part of the participant isrequired, as this course develops a fundamentalunderstanding of an electronic computer. Specifications,flow calculations, installation, operation and maintenanceare discussed along with “hands on” training, with specialemphasis on the Bristol Model 3310 and 3330 flowcomputers.

CONCLUSION

Today’s field of gas measurement has created anenvironment where the “learning curve” is no longer a“variable” in the gas industry, but a “constant.” Traininghas evolved into a continuous learning process thatproceeds throughout one’s professional career.Technology is forever changing to accommodate the worldof gas measurement; and as a result, the technician’sexpertise must be constantly developed and promotedto a level that enables him or her to function according totoday’s standards. For, it is through these learning andtraining processes that a measurement technician intoday’s market, can build a foundation and a desire togain a greater knowledge.

Russel W. Treat


BASIC APPLICATIONS OF TELEMETERING SYSTEMSCLASS #3050Stephen R. Cree

ABB Inc., Totalflow DivisionP.O. Box 60427, Midland, TX 79711-0427

Telemetering, or transporting information, has been withman from the first days of recorded history, at first inprimitive forms such as grunts and smoke signals; morerecently [in the past hundred years] in progressivelysophisticated forms including radio and satellite systems.Harnessing electricity led to the “magic” of telephony.Telegraphs, and later telephones employ a technologyso fundamental as to be the cornerstone of the telemetryprocess.

The creation of a carrier signal is foundational to anytelemetry signaling whose information is modulated[rides] upon it. Telegraphs employed a process of usingcurrent flowing through a wire as the carrier upon whicha manually operated key shifted the transmittedfrequency in coded fashion [intelligence]. Moresophisticated systems of FSK [frequency shift keying]used electronically keyed devices.

East coast gas pipelines were among the earliest to usetelemetry up and down the miles of pipeline system. Thesystem used two “devices”, a telephone line to providethe carrier and a man to voice modulate the data[information read from gauges] to people in towns withstrange names like Philadelphia and Boston.

As radio systems were being developed using atransmitted signal of a single frequency, a techniqueknown as amplitude modulation was employed. AM is atechnique to which many owed a great debt of gratitudeincluding Bing Crosby, Jack Benny and others. Morerecently frequency modulation upon a radio carrier [FM]has been far more popular, in part due to its greater signalto noise ratio properties within its range of frequencies.

Radio frequency bands have been organized intosegments allocated by the Federal CommunicationsCommission [FCC] for various activities. AM radio 535-1600 kilocycles per second, or Kilohertz is used for radiobroadcasting by licensed stations. 50- 54 Megahertz isallocated to amateur radio operators and low bandcommercial users, while 54- 108 Mh. is used by televisionbroadcasters.

These bands, with bending wavelengths, offer long-rangecommunications capabilities which make them attractivefor such uses. All of the channels on these frequencies,when use industrially, were obtained by FCC licensing.For many years it was thought that only by licensing couldinterference issues be held to a manageable level.Therefore, much planning and time went into getting any

system on the “air”. The next question that came to theforefront when radio was being designed for data usewas “what frequency do we use for this?”

Each frequency band has it’s own advantages anddisadvantages relative to distances signals can betransmitted, susceptibility to RF interference, andattenuation by objects such as buildings, trees, barnsand hills. Additionally, the elements required to reliablysustain a suitable level of effective radiated power ERP,and the designs of receivers with adequate sensitivity orsignal to noise rejection capabilities must be considered.

Problems with each band are “seen “ in the way signalsbehave in travel and require their own unique solutions,including those used in our industry.

Circa 1935 the most pivotal device of all moderntelemetry systems was first designed. Devices that couldelectronically modulate and demodulate intelligence, ineffect convert intelligence found in analog values to digitalvalues and back again! Reliable, accurate and costeffective means of communicating! Even the sky is nolonger the limit in a sense. Modems have become thecenterpiece of all industrial communications systems,whether telephone lines, microwave, or other wirelesssystems are used.

Networking computers in local area or wide areaconfigurations also utilize modems extensively andemploy a widely used electrical connection standardknown as RS232, or in some cases, for relatively shortdistances, RS485.

Supervisory Control and Data Acquisition systems[SCADA] systems use several communications platforms– hardwire, two-way radio, telephone, cell phone[including CDPD radios] digital wireless radios,microwave, fiber optics, and satellite.

Protocol [prearranged method of communication, orconvention] must be resolved, and transmission speed[baud rate] must be agreed upon as well. Obviously thefaster you can transmit [simplex] and receive [half duplex],or send and receive simultaneously [full duplex] the moreyou can accomplish in time. Some modems can do datafile compression, and have flash memory, as opposedto the legacy read only memory [RAM] still in use. Nottoo many years ago when modems could ask howanother modem was equipped, then adjusted itselfautomatically to be compatible with its conversation


partner we called them “smart modems”, and we wereright to do so.

The 450 MHz band was among the first to be widelyused in modern oil and gas applications, largely becauseof its distance of travel attributes. After this came theuse of 800 MHz. Their shortcomings, however aresignificant. Among them are power consumption,especially at remote locations, susceptibility to R.F.interference, and timing issues relative to “request tosend” and “clear to send “ handshaking and the delaysthey require. Additionally, the band is relatively full. If youhave a license it can be useful but many repeaters costthe user recurring fees in the $10-30 dollar / month range.much like cell phone, CDPD, satellite and microwavesystems.

The diagram below is intended to give the reader anoverview of some system possibilities and configurations.Many systems demand the use of more than one typeor mode of communications, and many hybrid systemsperform quite acceptably.

licenses are available on a somewhat limited basis,depending on your area ‘s density of existing licenseholders, it takes some time to procure one. The 5 wattlicensed radio is capable of operating over significantdistance, although it is important to remember that thehigher the frequency the less “flexible “ the wavelengthsbehave. In the 900 Mhz. Band there is virtually no bendingof the waves. It is line of sight only. This is a quality ofgreat strength in many systems. There is someinterference protection, but if it does occur there seemto be reports of difficulty in getting it resolved.

In properly designed spread spectrum systems,unlicensed 1 watt radios, where each radio can be usedas a remote point, or as a repeater, another attractiveelement is seen. As the number of points of contact inthe system increase, the more cost effective the solutionbecomes. Other attributes include higher baud rates,resistance to R.F. interference, full duplex capabilities,low power requirements, tiny size transceivers, lowpolling overhead or turn on/ turn off times and most haveintegrated diagnostics that are quite sophisticated.

The frequency hopping techniques of one manufacture’sspread spectrum radio are well established, providinguncommonly reliable communications especially whenpacket protocols are provided. Communications aresecure enough to prevent enemy interception, which wasthe original spread spectrum design objective for themilitary, and rugged enough for torturous applicationssuch a NASCAR racing and natural gas measurement inWest Texas.

This frequency hopping feature allows, as is the case inthe Powder River Basin of Wyoming, literally thousandsof radios to be operated by several end users in a fewsquare mile areas. Serious user density can be realizedwhere repeater resources are often shared resulting inefficiencies previously unknown. One hundred percentreliability in telemetry systems was consideredunattainable a few years ago, but considered nominalperformance in several North American systems today.But, the kind of success seen in the Powder River basindidn’t just happened. As with any engineered system, it,by definition, requires planning and attention to detail tobring smiles to faces on “opening day.” For any given

SCADA

. . .

Direct Connect

Modem

Phone Line

Radio

Satellite

CDPD Internet

SO WHAT ELSE IS AVAILABLE?

Enter 900 Mhz. This is perhaps the most interesting andpractical radio band on the planet. It is interestingbecause of its wide spread applicability including sports[auto racing] and military [army cavalry] functions. Anarmy may march on its stomach but it achieves itsobjectives with good communications. For us, however,900 MHz. provides excellent low cost solutions to mostof our telemetry problems.

There are frequencies allocated for licensed 900 Mhz.and unlicensed 900 Mhz. Spread Spectrum radio. Bothcan be successfully deployed. It largely depends onindividual preference more than communicationsengineering.

Licensed 900 Mhz may be a good choice for your needs.The radios are licensed for 5-watt transmission, and while

Primary Users of the Data Data

User of Data Current Trend Data Other EFM AuditData Historical Data Trail Data

MonitorOperations and Daily Changes — Correct andMaintenance Operations Schedule Verify

Repairs

Billing Yes

Management Yes Yes Yes Yes

Business Planning Yes

Marketing Yes

Sales Yes

Engineering Yes Yes


new system a number of questions should be asked asa first step, and the famous “do you feel lucky? “ shouldnot be among them.

Questions that you’ll be glad you got answers forinclude but are not necessarily limited to thefollowing:

4. What are you willing to invest in a Measurement/SCADA telemetry system to get the results you need orwant? Have you and those you work with discerned thedifferences between needs and wants and what areacceptable costs for each? How much dependency onvendors are you willing to accept? That is to say, to whatextent are you willing to design a system predicated onthe viability of one central provider as opposed to leavingyour options open with regard to expansion and vendorchoice.

5. Once the field data is retrieved what will you do withit? The following diagram may provide a good startingpoint for discussing your needs.

1. What is your objective – not generally, but withsufficient specificity to know when you have completedat least the first phase of the project. Into whose handsmust what type of data be delivered and when? To whomshall be given what type of controls, if any? The chartshows some who are normally included in the plans.

2. What resources are available? A/C power, a smallbuilding, a large building, cell phone service, tower spaceavailable for rent in the older 450Mhz. Or 800 Mhz. bands,licenses, protocol requirements, the compatibility of fielddevices with telemetry choices and most important—communications knowledgeable people, either yours orthose from outside. Site survey work is nearly imperative.

3. In what environment must your system perform? Atsea there are still distance limitations, especially in900Mhz whose wavelengths provide line of sightoperation. Do you have miles and miles of only milesand miles, or do you have trees and barns and mountainsto deal with. Objects can stop; yes stop communicationsbetween two points requiring a repeater in some cases,or only an antenna height adjustment in others. How canyou know the difference? Again, site surveys are nearlyimperative.

WHY DO PATH STUDIES?

Se p 1 6 0 2

PA T H T ECH , L T D

CH IP REPEA T ER

L a t it u d e 2 8 1 9 4 3 .6 4 N

L o n g it u d e 0 9 9 4 6 5 1 .9 4 W

A z im u t h 1 8 6 .5 3 °

E le v a t io n 6 6 0 f t A SL

A n t e n n a CL 6 0 .0 f t A GL

PA L A F OX

L a t it u d e 2 8 0 6 1 2 .5 0 N

L o n g it u d e 0 9 9 4 8 3 6 .7 0 W

A z im u t h 6 .5 2 °

Ele v a t io n 7 3 8 f t A SL

A n t e n n a CL 5 0 .0 f t A GL

F re q u e n cy ( M H z ) = 9 1 5 .0

K = 1 .3 3

% F1 = 6 0 .0 0

Pa t h le n g t h ( 1 5 .6 2 m i)

0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5

E

l

e

v

a

t

i

o

n

(

f

t

)

5 5 0

6 0 0

6 5 0

7 0 0

7 5 0

8 0 0

The data is in the Field Office Computer, Now What?

POINT TO MULTIPOINT SYSTEM DESIGN


The future of telemetry systems carries an element offurther excitement, but should also be viewed with adegree of caution. Historically, there have been cases ofpremature exuberance, leaving buyers disappointed.Some of the future will likely materialize nicely, rewardingthe participants handsomely.

Stephen R. Cree

Our industry has been committed to the advancementof technology for many years. We have set new standardsof excellence in areas ranging from pipeline construction,to cat cracker operation, to hydrocarbon measurementand control systems. Often utilizing advanced softwarein remote devices and in telemetry, we get operatingefficiencies elevated to higher levels with each passingyear. Other hydrocarbon producing countries may notadmit it, but this Yankee ingenuity has been andcontinues to be home in the USA.

Credits and many thanks to Freewave Technologies, Inc,Boulder Co. and particularly to Bob Halford of PathtechLtd., Odessa, Texas.


COMMUNICATION BETWEEN OFFICE AND FIELDDuane A. Harris

CMS Energy Corp.P.O. Box 4967, Houston, TX 77210-4967

INTRODUCTION

The gas industry today is constantly changing, withincreasing demands on office and field personnel. Initiallythere was FERC (Federal Energy Regulatory Commission)Order 636 that forced the gas measurement departmentsinto the electronic age. Next came corporate slashingthat has required the gas measurement groups toperform at the same level of integrity in the measurementof gas with reductions in staff of up to 60%. Then GISB(Gas Industry Standards Board) made its way into thegas measurement department through proposedstandardization. Today hourly processing requirementswith a daily a closing schedule is knocking on the doorand has already arrived at some locations. To meet thesedemands timely communication between the office andfield employees is required. Both of these locations (fieldand office) have been impacted with increased workloadsand constant upgrades in equipment and software. Withall of this occurring, it is very easy to overlook one of thekey links to accurate measurement and that iscommunication.

By the time that a gas day has started at a meter site ona chart recorder or an RTU (Remote Transmitting Unit)until the volume has been calculated or verified in thecorporate office, anywhere from 1 to 35 days can passwith as many as 8 to 10 people handling each individualvolume record. With this many people involved coveringthat span of time, communication becomes a vital partof the measurement process.

TRAINING

In order to communicate effectively you must first havean understanding about what you are talking about.Training has become even more critical with theconsolidations and heavy turnover that numerousproduction and pipeline companies have experiencedover the last few years.

Below is a list of terms that you must be familiar with tocommunicate effectively between the office and field:

1. Plate Size2. Tube ID3. Beta Ratio4. Flange Taps vs. Pipe Taps5. Mercury/Dry/EFM Meters6. Differential Pressure Range7. Static Pressure Range

8. Temperature Range9. Actual/Square Root/Percentage Charts10. Positive Displacement/Turbine/Ultrasonic Meters11. Positive Displacement/Turbine/Ultrasonic Meter

Multipliers12. Mcf/MMcf/MMbtu/Dth13. Absolute vs. Gauge Pressure14. Specific Gravity15. BTU (British Thermal Unit)16. Inerts–CO2 and N217. Current AGA Standards18. Current Industry Standards

There have been numerous occasions that a clearunderstanding of these areas had not been attained,thereby causing many costly mistakes and corrections inthe measurement department. One example of this errorrelates to flange taps and pipe taps coded incorrectly in ameasurement system. This error can cause and errorresulting in an 8% adjustment to the volume due to lackof training in the gas measurement area.

WHO IS RESPONSIBLE

Who is responsible, is a battle today that is being foughtwith a mediocre success rate. I find it more difficult eachyear to keep track of the person who is responsible toanswer specific questions regarding key areas in the workforce today. The best way to solve this problem is todevelop your own list of names for each specific area.

Who handles:

Gas Quality Sampling Issues–Lab, MeasurementTechnicians

New Station Turn-Ons–Regulatory Affairs, MarketingOperations, Marketing Sales, Gas Control,Engineering

Chart Meter Problems–Area Technicians or ChartChanger

EFM Meter Problems–Area Technician

EFM Meter Communication Problems–AreaTechnician or Communications Technician

Ordering Charts

Etc..


CHARTS

The best way to communicate on chart based stationsis by using the chart itself. Most questions that originatefrom the office could be answered before they are askedby the measurement technician and/or chart changerby simply detailing key events of what happened on thechart.

1. Back-Flow situations should always be noted aroundthe hub of the chart.

2. Liquid in the meter should also be noted around thehub or under remarks on the back of the chart.

3. Whenever a meter is zeroed or tested and the pensare recording low or high this should be noted underthe remark’s section on the back of the chart.

4. Low flow or no flow should be noted in the remark’ssection especially if this is a station that may be hardto distinguish between the two. There is a significantdifference between low flow and zero flow.

5. Actual chart changing time (placed and removed)should be recorded on the chart.

6. Any clock problems (slow, fast, stopped) or hubproblems (loose, too tight) should be noted underremarks.

EFM

Electronic Flow Measurement (EFM) requires almostimmediate response for resolving measurement issuesbetween both the office and field locations. Typically thevolume received on an hourly basis from the field RTU isbeing posted on the internet for customers to review.Each company should have their own method in placeto resolve the issue and minimize the time effect to allinternal and external customers. One key area to identifyEFM problems is the review of an alarm report thatsummarizes errors and potential problems that haveoccurred during the previous and current gas day for allRTU’s, transmitters, and all critical volume calculationcomponents. You must rely on the raw data, audit trails,prior station history, check measurement, andinformation from the measurement technician in orderto troubleshoot potential problems. Trying tocommunicate between the field and office can be difficultat times even with the many modes of communicationthat exist between between the office phone, cell phone,pager, email and fax. Timely communication is requiredin this area to meet the demands of verified measurementdata on a daily basis. Regular mail is really a slow processwhen trying to resolve any problems with EFM. Thetelephone and E-Mail will be discussed in greater detailin the next section. The trend in the industry today is torely on a rules based software package to validate all ofthe raw electronic measured data that is received fromthe RTU’s in the field. Only the meters that due not passthe validation checks are individually reviewed foraccuracy.

TELEPHONE

Telephone communication has been a source of constantirritation between the staff in the office and field locations.Continually talking to someone’s voice-mail, to findingsomeone just to “Answer” the phone has caused thetelephone to move from being one of the mostinstrumental forms of communication to being an areaof complete frustration. One key factor in using the phoneis to keep track of the best time to call the person youare trying to reach, especially someone that you callfrequently, and set that time aside to reach them daily orweekly.

Most field employees are more easily reached early inthe morning or later in the afternoon at the field office. Itis also worth noting, the typical lunch and break timesfor office employees if they are taken at regularlyscheduled times.

Although the voice-mail systems can be ratherimpersonal, they are an excellent tool to exchangeinformation back and forth. Sometimes it takes severaldays before you can reach the individual in person. But,you can in effect communicate back and forth by leavingvoice-mail messages with detailed information includeduntil you get the problem resolved.

WRITTEN DOCUMENTATION

The written document is probably the oldest and one ofthe most effective ways to communicate. Letters canalways be archived and referred to at a later date. Writtendocumentation is a necessity for audit purposes. Theonly problem is that you generally find out that you needthe documentation after it is already too late. Planningahead is essential and knowing what key areas need tobe documented for archiving purposes will be time wellspent. This process is really too slow today for a largepercentage of internal correspondence due todistribution and mail time.

ELECTRONIC - MAIL (E-MAIL)

Electronic mail has become one of the key forms ofcommunication of today. These mail systems aresometimes interfaced into your SCADA or EFM systems,which can be extremely helpful in troubleshooting EFMproblems. Other systems are usually connected to anetwork system of their own. Today the majority ofcompanies’ internal E-Mail systems are interfaced intothe Internet. The Internet has significantly enhanced thisarea of communication by opening up to not only internalcompany employees but everyone that you are currentlydoing business with. E-Mail is basically the same as aletter except you save the distribution and mail time. Onekey advantage to this system is that you can print andarchive the message from the PC. In some cases youcan certify your message so that you will know whenthe person receives it.


ACCURATE RECORDS

There is still one more area that has not been discussedyet and is extremely important. This area is keepingaccurate records on all of the key measurement devices.These records should be kept in the office and the fieldfor verification purposes. Such as Orifice Plate Size, TubeI. D’s, Etc... Some of the new measurement systems givethe ability for the office and field to view the samedatabase for verification purposes.

CONCLUSION

In today’s changing gas industry, you must have aworking form of communication between the field andoffice. With the impact of FERC Order 636, GISB,unaccounted for gas loss, and proposed hourlyprocessing gas companies must verify and process datawith more accuracy faster than ever before. The changingenvironment that we are operating in has not seemed toslow down yet. Effective communication is a requirementin order to stay competitive in the industry. You absolutelymust have effective Communication links between theoffice and field to meet the challenge placed upon theGas Measurement area.

Duane A. Harris


OVERALL MEASUREMENT ACCURACY —DETERMINATION AND INFLUENCE

Paul J. La NasaCPL & Associates

P.O. Box 801304, Houston, TX 77280-1304

ABSTRACT

This paper presents methods for determining theuncertainty of both differential and positive meteringstations. It takes into account the type of meter, numberof meters in parallel, type of secondary instruments, andthe determination of physical properties. The paper thenrelates this information to potential influence on systembalance.

INTRODUCTION

Gas measurement uncertainty is a function of thefollowing items:

• Selection of the appropriate metering device• Correct Installation of the metering device• Proper operation and processing of the metering

information• Proper maintenance of the metering device

Understanding how measurement uncertainty applies tometering requires a basic knowledge of the terminologyand assumptions used in the calculation of measurementuncertainty.

Measurement uncertainties can be categorized as thefollowing:

Pseudo Pseudo uncertainties are potential humanerrors or those associated with themalfunction of an instrument. Onceidentified, these errors can usually becorrected and are not included in thecalculation of measurement uncertainty.

Random Random uncertainties are potentialmeasurement errors that have an equalchance of being higher or lower than thetrue value of the measured variable. If alarge number of measurements are made,the random error in the individualmeasurements cancel and the mean of themeasurements will be approximate to thecorrect value.

Systematic Systematic uncertainties are measurementerrors that are directional or contain a bias.Because these errors are directional, theydo not cancel as additional measurementsare made.

Random uncertainty reduces as the number ofmeasurements increases. However, additionalmeasurements will not reduce the systematic uncertainty.

Because the random and systematic uncertainties arecharacteristically different, the calculation of each mustbe performed independently. The combination of the twoindependently performed calculations then forms thetotal measurement uncertainty.

The elements of the random and systematic uncertaintiesare classified as either independent or dependent andmust be determined before the total measurementuncertainty can be obtained. The determination of gasmeasurement uncertainty has been and is addressed innumerous industry articles and standards publications.Three such articles and publications that were referencedin the preparation of this paper are: (1) Norman andJepson, (2) Tiemstra, Rans, and Backus, (3) AGA ReportNo.3 Part 1 — 1990. However, for the purposes ofevaluation, this paper will not concern itself with theinfluence of the interdependence of variables but willutilize the calculation procedure given in A.G.A. ReportNo. 3 (API MPMS 14.3, ANSI 2530, GPA 8185-90) Part I— 1990 to determine the orifice meter measurementuncertainty and will apply the same metrology to thepositive meter (turbine, rotary, or diaphragm meter)measurement uncertainty.

The uncertainty for a single meter run is evaluated fromthe random and systematic uncertainty of the primaryelement (orifice, turbine, rotary, or diaphragm meter) andits instrumentation. The uncertainty of the primaryelement includes the uncertainty associated with the flowcoefficient, expansion factor, diameter of the meter run,diameter of the orifice plate bore, and calibration of thepositive meters.

For an Individual meter run:

UTM = URM + USM

WhereUTM = Total meter run uncertaintyURM = Meter run random uncertainty

URM = √Σ(URi )

USM = Meter run systematic uncertainty —

USM = √Σ(USi )


The percent random uncertainty contributed by eachvariable, URi , is defined as follows:

URi = ( Xi × Ai )2

The sensitivity coefficient of each variable at the point ofevaluation, Xi, can be determined by calculating theresults for the conditions of evaluation, R, and the changein the result, DR, produced independently by theaccuracy variation of each variable at the conditions ofevaluation and substituting into the following equation:

Xi = ∆R( R )The percent accuracy of each variable at the point ofevaluation, e.g., average differential pressure, isdetermined as follows:

For variables whose accuracy is stated as function of itsfull-scale value, the percent accuracy is the value of theaccuracy at full scale, AF, divided by the value of thevariable at the conditions of evaluation, VCE.

Ai = AF × 100 (VCE)

As an example, assume that one of the variables is adifferential pressure value whose accuracy is stated as0.1% of full scale, its full scale is 100 and the point ofevaluation is 50. The percent accuracy of the variable atthe point of evaluation, Ai, would be:

Ai = 0.001 × 100

× 100 = 0.2% ( 50 )The sensitivity coefficient of the variable at the point ofevaluation, Xi, could be determined by calculating theresults, R, using the point of evaluation value of 50, thencalculating the change in results, ∆R, using the point ofevaluation value, 50, changed by 0.2%. These twonumbers would be inserted into the equation for Xi todetermine the sensitivity coefficient of the variable at thepoint of evaluation.

For variables whose accuracy is stated as a function ofreading, the percent accuracy is the same though outits range.

The systematic uncertainty of each variable is determinedusing the accuracy of the calibration standards as an estimateof its contribution to the total systematic uncertainty (USM ).

USi = ( Xi × Ai )2

The accuracy of the calibration standards (Ai) areexpressed as a percent of reading so they can besubstituted directly into the USi equation along with thesensitivity coefficients (Xi ) calculated for the appropriateelement to determine the systematic uncertaintycontribution by each variable.

The total systematic uncertainty is determined using theUSM equation.

USM = √Σ(USi )

However, since the criteria applied for the determinationof the sensitivity coefficient, can vary and is specific toan application, the sensitivity coefficients used for theorifice meter uncertainty were chosen from A.G.A. ReportNo.3 (API MPMS 14.3, ANSI 2530, GPA 8185-90), Part1 — 1990 and similar sensitivity coefficient weredeveloped for the positive meters. The use of theseparticular sensitivity coefficients can result in a smallunderstatement of the uncertainty estimates resultingfrom not accounting for the interdependence of some ofthe elements.

DIFFERENTIAL METER UNCERTAINTY

The variable elements of a gas orifice meter measurementuncertainty calculation are as follows:

Differential Pressure, dpStatic Pressure, PfFlowing Temperature, TfGas Relative Density, GrGas Compressibility Factor, Zf & Zb (Fpv)Orifice Meter Coefficient of Discharge, CdOrifice Bore Diameter, dMeter Tube inside Diameter, DExpansion Factor, YDifferential Pressure Calibrator, dpcStatic Pressure Calibrator, PfcFlowing Temperature Calibrator, TfcGas Relative Density Calibrator, Grc

To calculate the measurement uncertainty for a multiplemeter run station, the variables that are independent ona per run basis are differential pressure, static pressure,temperature, and meter run tolerances. The variablescommon to all runs in the station are the relative density(specific gravity), gas composition, and calibrationstandards.


UTS = URS + USS

Where


The total orifice meter station random uncertainty is given as:

URS = Σ URi

2




USS = √Σ(URi )2 per Run

Since there are numerous combinations of equipment,operating conditions, and calculation methods existingfor orifice metering, it is impossible to establish a singlebase line uncertainty relationship. The most practicalapproach is to provide uncertainty ranges for the mosttypical orifice metering combinations.

POSITIVE METER UNCERTAINTY (ULTRASONIC,TURBINE, ROTARY, AND DIAPHRAGM)

The variable elements of a gas positive metermeasurement uncertainty calculation are as follows:

Static Pressure, PfFlowing Temperature, TfGas Relative Density, GrGas Compressibility Factor, Zf & ZbPositive Meter Linearity, PML

Positive Meter Calibrator, PMpc

Static Pressure Calibrator, Pfc

Flowing Temperature Calibrator, Tfc

Gas Relative Density Calibrator, Grc

To calculate the measurement uncertainty for a multiplemeter run station, the variables that are independent ona per run basis are positive meter calibration or proof,static pressure, and temperature. The variables commonto all runs in the station are the relative density (specificgravity), gas composition, and calibration standards.


UTS = URS + USS

Where


The total positive meter station random uncertainty isgiven as:

URS = Σ URi

2



USS = √Σ(USM )2 per Run

Since there are numerous combinations of equipment,operating conditions, and calculation methods existingfor positive metering, it is impossible to establish a single

uncertainty relationship. The most practical approach isto provide uncertainty ranges for the most typical positivemetering combinations.

ENERGY DETERMINATION UNCERTAINTY

The measurement of total energy received or deliveredis customarily the product of the measured volume andthe heating value (Hv) per unit volume. The heating valueper unit volume is typically an inferred measurementresulting from a chromatographic analysis of arepresentative sample of the gas being received ordelivered. In addition to heating value per unit volume,relative density (specific gravity) used in the determinationof volume is also obtained from the chromatographicanalysis. Industry standards, which address theperformance of chromatographic analysis, thecalculation of heating value per unit volume, and relativedensity of a gas sample, are:

• (4)ASTM D 1945-96 (GPA 2261-95) — StandardTest Method Analysis of Natural Gas by GasChromatography

• (5)ASTM D 3588-98 (GPA 2172-96) — StandardPractice for Calculating Heat Value,Compressibility Factor, and Relative Density(Specific Gravity) of Gaseous Fuels

The industry standards, ASTM D 1945-96 (GPA 2261-95) and ASTM D 3588-98 (GPA 2172-96) provide aprecision statement for repeatability and reproducibilityas a function of the mole fraction of each component inthe gas mixture. The repeatability is the expectedprecision within a laboratory using the same equipmentand the same analyst. The reproducibility is the expectedprecision when different laboratories using differentequipment and different analysts use the same method.Tables 1 and 2 provide the given repeatability andreproducibility tolerances.

Component RepeatabilityMole % %

0 to 0.1 0.010.1 to 1.0 0.041.0 to 5.0 0.075.0 to 10 0.08Over 10 0.10

TABLE 1.ASTM D 1945-96 Precision Repeatability Criteria

Component ReproducibilityMole % %

0 to 0.1 0.020.1 to 1.0 0.071.0 to 5.0 0.105.0 to 10 0.12Over 10 0.15

TABLE 2.ASTM D 1945-96 Precision Reproducibility Criteria


The individual component reproducibility tolerances werecombined using the square root of the sum of the squaresmethod as shown in ASTM D 3588-98 (GPA 2172-96) toobtain a precision statement. This is a commonmethodology employed when determining the toleranceof calculated values containing random individualelemental tolerances. Chromatographic analysis and thecalculations of Hv and relative density performed usingindustry standards, ASTM D 1945-96 (GPA 2261-95) andASTM D 3588-98 (GPA 2172-96), will produce heatingvalue results to within ± 0.25% and relative density resultsto within 0.003 relative density units for a typical pipelinenatural gas having the following composition:

Mixture MoleComponent %Methane 96.5222Ethane 1.8186Propane 0.4596Isobutane 0.0977n-Butane 0.1007Isopentane 0.0473n-Pentane 0.0324n-Hexane 0.0664Nitrogen 0.2595Carbon Dioxide 0.5956

BTU/Ft3 1036.06 Ideal Gross Hv per Real Ft3

@14.73 & 60°F0.582 Real Relative Density

@14.73 & 60°F

Since the ASTM D 1945-96 repeatability andreproducibility criteria originated from a statisticalexamination of interlaboratory test results, it includes theinfluences of properly prepared calibration gas standards.The Hv and relative density precision values assume thatthe sampling methods and sampling systems utilizedprovide a representative sample of the flowing gas streamfor analysis.

SYSTEM BALANCE INFLUENCE

Engineering departments can use metering stationuncertainty information in selecting the type of equipmentto be use in a meter station. Equipment can be selectedto meet a system balance expectation or uncertainty. Itcan be used by gas control departments to estimatewhen the uncertainty of a meter station’s measurementis increasing. It can be used to help manage lost andunaccounted-for numbers. If all one type of equipmentis installed on the inlet and all of another type on theoutlet, the metering system may not produce the desiredsystem balance results. It can be used by maintenanceto understand on which pieces of equipment toconcentrate their efforts.

REFERENCES

1. Calculation defines uncertainty of unaccounted-forgas, Norman, R and Jepson, P., Oil & Gas Journal ReportApril 6, 1987

2. Comparison of Orifice and Turbine Meter Accuracy,Tiemsyra, P., Rans, R., and Bacus, H., American GasAssociation Distribution/Transmission Conference April,1991, Nashville, Tennessee

3. Orifice Metering of Natural Gas and Other RelatedHydrocarbon Fluids, Part 1 — 1990, General equationsand uncertainty guidelines. American Gas AssociationReport No. 3, Third Edition, Arlington, VA, October 1990.

4. Standard Test Method Analysis of Natural Gas by GasChromatography — ASTM D 1945-96 (GPA 2261-95)

5. Standard Practice for Calculating Heat Value,Compressibility Factor, and Relative Density (SpecificGravity) of Gaseous Fuels — ASTM D 3588-98 (GPA2172-96)

Paul J. LaNasa


FROM PEN TIP TO VOLUME STATEMENTDavid Pulley

Metron Gas MeasurementP.O. Box 2029, Tyler, TX 75710

INTRODUCTION

Accurate and reliable gas measurement depends on acombination of efforts from two groups of people. First,we have the field personnel. They have the responsibilityof seeing that a readable chart is produced and that allinformation pertinent to volume calculation is suppliedto our next group, which is the office personnel. Thisgroup will read the chart, apply information supplied bythe field, calculate the amount of gas delivered, andgenerate and deliver volume statements to the customer.

It must be recognized that the office group can notcalculate volumes until it receives the charts and otherrequired data from the field, and that the volumes areonly as accurate as the data received.

With this in mind, let’s look at the two separate groupsand see the team effort involved in producing accurateand reliable measurement from Pen tip to VolumeStatement.

FIELD SERVICES

To produce accountable volumes, it is imperative toreceive the correct information from the field. This isobtained from the Field Measurement Technician whosubmits an installation and/or inspection report on themetering station that he is testing or placing in service.Listed below is the data that must be furnished to theoffice group before an orifice meter chart can be readand the volume calculated accurately.

FIELD DATA REQUIRED

1. Station Name and I.D. number2. Meter tube inside diameter size3. Orifice Plate size4. Tap connection (flange or pipe)5. Differential and Static ranges of the meter6. Temperature range (if recorded) or an estimate

of flowing temperature.7. Chart rotation (7 day, 8 day, 24 hour, etc.)8. Gas Analysis for gravity, BTU and FPV

calculation9. Correct Dates and Times

It is also important to know if there are any problemswith the equipment, meter calibration or outsideinfluences that would affect accurate measurement.Some examples are as follows:

1. High or Low differential zeros2. Fast or slow clocks3. Wrong orifice size4. Orifice in backwards5. Pulsation6. Liquid in line7. Blown Bellows8. Freezing causing erratic pattern9. Etc.

If the technician notices any of the above, then it shouldbe noted on the chart or field inspection report to theattention of the office personnel.

If the information supplied to the measurement office isnot correct, or is incomplete, then reported volumes areunaccountable. Therefore, it is vital for the field personnelto work closely with the office personnel to determinewhat is needed to produce accurate volumes.

RECEIVING AND PROCESSING CHARTS

Once the measurement office receives the aboveinformation, then master files are set up for all involvedmeter stations. These master files are then updated ona regular basis from new information received via fieldreports, gas analysis, etc. Once all master files are inplace and correct and current information has beenentered, then we are ready to start processing charts.

The actual step by step procedures for processing chartsmay differ from company to company, but the basicsare generally as follows:

CHART CENSORING

1. Receive and Identify stations2. Put in processing order and verify information

against the master files.3. Censor for incorrect dates, times or erratic

patterns4. Check meter inspection reports for any

information that would effect calculation.5. Highlight any changes or notations for chart

integrator.6. Send to Integration Department

CHART INTEGRATION

The chart is placed on the integrator processor and anoperator traces the original differential and staticrecordings. The integrator count or index represents the


square root of the differential times the square root ofthe static pressure over time; the result is the chartextension:

Chart Extension = √Hw x Pf x T

Where:Chart Extension = Integrator ReadingHw = Differential RecordingPf = Absolute Pressure RecordingT = Chart Rotation

I think it is important to note here that the personnel atthis level are a vital part of the overall process. It isextremely important for them to have gas measurementknowledge and proper training. Accurate integration ofa chart is more that just tracing lines. The operator mustalso be able to recognize flow pattern irregularities suchas meter freeze, fast or slow clocks, orifice plate changes,etc. If the process breaks down at this level, the overallprocess is affected.

VOLUME CALCULATION

Once integration of the charts is completed, then theintegrator extension and field data are used to calculatea gas volume for the chart period.

VOLUME STATEMENT

The calculated volumes are censored for anydiscrepancies or unusual fluctuations in flow rate. Oncesatisfied, then a Volume Statement is generated. Thepurpose of this statement is to furnish the customer witha report on volume of gas delivered for a particularmeasurement station during a specific period of time.This statement is compared to master file informationand against any special instructions that were noted bythe field. If everything is correct, then the statement issent to the customer and the month is closed out.

CONCLUSION

As stated above, the overall measurement process is ateam effort that begins in the field before the pen tip isever placed on the chart and ends with the generationof the volume statement. Therefore, it is essential to havegood communication between the field technicians andthe office personnel to keep the overall process smoothand accurate.

David Pulley


CHART AUDITINGTom Tauer

Southern Flow CompaniesP.O. Box 66190, Houston, Texas 77266

PREFACE

Gas chart auditing can be one of the more interestingand rewarding segments of chart processing. Whileperforming a valuable service for the customer or personrequesting the audit, it can give the auditor a chance touse his or her chart expertise that could become stagnatewith only routine chart processing. The chart auditorbecomes part accountant, part detective, and partmathematician. The purpose of the chart audit is normallyto insure that the volumes have been reported asaccurately and objectively as possible. Productioncompanies and operators have a responsibility to leaseowners, state agencies, and themselves to reportvolumes that are truly representative of the amount ofgas that passed a particular point.

PREPARATION

Normally a service company would not be the one torequest an audit. The producer or transmission companywould be the one to initiate the audit by sending a letterto the company to be audited. This letter would requestall the necessary information needed to perform the audit.This would include the orifice charts, any temperatureor gravity charts, the original gas volume statement, allmeter inspection reports relevant to the time period tobe audited, and gas analysis. Also, any fast clock testswould be required if they exist. Without the ability todouble check the physical meter data, you are justrecreating volumes based on possible erroneous data.While integration errors or differences in interpretationwould show up in the volume calculations, a real probleminvolving the tube size, orifice plate size etc., would goundetected.

THE AUDIT BEGINS

Once you are satisfied that you have obtained all theinformation that you will need, you can set up the stationin your own system and begin the audit. Missing chartsshould be noted and their treatment on the suppliedvolume statement should be noted. The customer orperson requesting the audit should be notified if any ofthe requested information is unsubstantial for performinga full audit.

The normal steps in censoring a chart should be followedto insure the accuracy of the audit. This means that thefollowing possible problem areas should be observed:

1. Failure of chart to record fully — this can be theresult of chart drive failure, pens not inking, or aslow clock

2. Low or high zeros — this can be confirmed withthe supplied test report.

3. Meter being left out of service — this could bethe result of a field person leaving manifoldvalves open or a gap, such as during a metertest.

4. Meter freezing — often occurs during the coldestmonths but can occur at any time conditionsare right for hydrate formation.

5. Ruptured bellows — would be noted on testreport; could result in erratic pattern or meterrecording below zero.

6. Meter over ranging — could result inunaccounted for gas.

7. Orifice plate too large — could result in lowerthan desirable differential reading withopportunity for interpretation errors and flowbeing considered “no flow.”

8. Orifice plate damage or reversal — theinspection report should be reviewed for orificeplate size, placement, and nicks etc. A plate withthe bevel placed upstream would give you alower than appropriate differential. The flowpattern should be observed after any of theseproblems are found to determine the truedifferential.

9. Time on and off — could reveal fast or slow chartdrive and confirm overlap situation.

CHART INTEGRATION

Most companies express concern that the charts not bealtered which would include marking on the chart or usingink in your reintegration. Some companies will requestthat the charts be run “dry.” The draw back with runningthe charts “dry” would be that you would have no recordof where on the differential you integrated the charts otherthan your composite integrator number. There are somecompanies that do not considered rerunning the chartswith ink a problem.

The integrator operator should maintain his or herobjectivity in the integration based on the informationprovided and the recording as it appears on the chart.You would not want the auditor to just retrace the originalintegration. The integration portion of the audit is criticaland losing your independence would serve no one.


THE COMPARISON AND HUNT

After you have your volume calculations in hand, acomparison sheet can be written and problems ordifferences noted. A good working knowledge of thecalculation factors and how they affect the volumetricoutcome is now critical in finding any problems and theirsolutions. Following are some differences and theirapproximate affect on their calculation:

1. Incorrect orifice plate used — percentage of errorwould depend on plate variance but differenceswill be sizable.

2. Incorrect tube size used in calculations —differences will depend upon plate size and tubevariance.

3. Machine constant — this involves differentialrange, static range and rotation. The differencebetween the calculations using a 100 inch versesa 50 inch would be 41%. The same percentagewould be for 200 inch vs. a 100 inch rangespring.

4. Incorrect specific gravity — will result in about a.8 of 1% difference for every 10 point of gravity.

5. Temperature difference — an error or differenceof 10 degrees will result in a 1% difference involume.

6. Supercompressibility — every one hundredpounds of pressure will yield about a 1% Fpvcorrection factor.

7. Tap connection location — Flange vs. Pipe Taps— on just a 3 x 1 (tube and orifice) the Pipe tapconnection would result in a 7% greater volume.

The Orifice Meter Constants E-2 Handbook can beinvaluable in tracking down differences and theirpercentages. The tables for gravity, temperature, and thebasic orifice factors are all clearly listed in the E-2Handbook and percentages of error/difference can befigured from these tables.

Armed with your knowledge of the different factors, acalculator, and the E-2 Handbook you can start asystematic approach to finding any errors or differencesin the volumes.

You would want to ascertain whether or not the volumedifference is in all the chart entries or only one. If onlyone volume entry is affected, you ask what is differentabout this one. If all chart entries have a variance of thesame size this would rule out some types of errors suchas individual chart data entry or a misidentified chart.

Sometimes you will find oversights such as missing chartdates, orifice plate changes picked up on the wrong dateor charts misidentified as belonging to that particularstation. You could be looking at the June 2002 chartsand have been given the June 2001 volume statement.

Differences that arise in chart integration or interpretationwould have to justified by a realistic and objective reviewof how the charts were originally integrated and yourinterpretation of the same. Was the wide band run down

the middle when a fast clock test actually showed thatthe recording was staying near the lower side and justkicking up enough to leave a solid red band? Was analready lower than appropriate differential misread by apen width? Was an intermittent type of flow patternoverstated because the integrator operator did not returnthe pen to zero between the kicks?

THE FINAL PRODUCT

A volume comparison report is written upon completionof the audit. This can list the original reported volume,the audited volume, the MCF difference and thepercentage difference. This would be accompanied bya letter to the customer or person requesting the audit.The language should be strait forward and objective innature. It should be mentioned if the auditor believesthat further action or a correction would be in order. It isup to the customer to decide what to do with theinformation. If errors are found, they should be spelledout without finger pointing or speculation as to reasonsfor the errors.

If a meter malfunction is clear, it is reasonable to suggestalternated methods of determining the flow rate on thedays that were not representative of the actual flow. Thiscould include such alternatives as using the check metervolume or if the flow in consistent, an average of the daybefore the problem and the day after the problem wascorrected.

CONCLUSION

This report can only scratch the surface of the manydifferent circumstances that will arise when performinggas chart audits. The auditor will quickly learn that thereare as many different types and forms of volume reportingas there are companies reporting them. From an auditor’sviewpoint, the statement with the most usableinformation is the best. While come volume statementsoffer meter data, hours of flow, factors,, analysisinformation and a virtual cornucopia of information;others offer very little in the way of useful calculationinformation.

Chart auditing is critical to the parties involved not onlybecause of the economic impact but as a method ofmaintaining checks and balances within the system.

Tom Tauer


NORTH AMERICAN ENERGY STANDARS BOARDAS A CASE STUDY

Cynthia CorcoranBTU Watch, Inc.

2939 Ferndale St., Houston, TX 77098

WHY STANDARDS?

• Commercial interest: to facilitate commerce bylowering transaction (production/distribution,etc.) costs

• Governmental interest: to facilitate a competitivecommercial environment and serve thegovernmental interests of public health, safetyand preservation of competition

WHO SETS STANDARDS?

Governmental bodies• Legislatures• Federal, state, local agencies

Standards development organizations• Voluntary standards through consensus based

effort involving all interested parties• Development of consensus-based, voluntary

standards is uniquely American contribution tostandards-setting

STANDARDS-SETTING POP QUIZ:

Answer “Yes” or “No” to each of the following pairs ofstatements:

Standards stifle innovation.Standards are platforms for innovation.

Standards promote trade.Standards impede access to export markets.

Standards are for the market to decide.Standards are matters of public interest and evengeopolitical interest.

Standards foster monopolization.Standards lower barriers to market entry.

THE PARADOX OF STANDARDS:

It’s all about how you set them.• Objectives• Process

LEGAL CONSTRAINTS ON STANDARD SETTING:

• Federal antitrust law prohibits unreasonablerestraints of trade.

• Standard setting judged under “rule of reason”to determine whether the challenged conduct“promotes competition or . . . suppressescompetition.” National Society of ProfessionalEngineers v. United States, 435 U.S. 679, 691(1978).

• In part, to facilitate compliance with federalantitrust law, the standards set by manyAmerican standards development organizationsare voluntary.

PROCESS FOR STANDARD-SETTING:

• The process for setting standards mustengender credibility with industry, withgovernmental regulators and with the public atlarge in order for standards to be accepted.

American National Standards Institute (“ANSI”) Modelfor standards development

• ANSI is private sector offshoot of five engineeringsocieties and the National Bureau of Standards(now known as National Institute of Standards andTechnology) of U.S. Department of Commerce.

• ANSI accredits organizations as StandardsDevelopment Organizations (“SDO”) based ontheir adherence to ANSI’s cardinal principles ofstandards-setting.

ANSI Cardinal Principles of Standard-setting• Open. Any materially affected and interested

party has the ability to participate.• Balance and Lack of Dominance. The consensus

body shall be balanced and shall not bedominated by any single interest category ororganization.

• Due Process. All objections shall have anattempt made towards their resolution. Interestswho believe they have been treated unfairly shallhave a right to appeal.

• Consensus. More than a majority but notnecessarily unanimity.

• Voluntary. Standards are not binding unlessadopted by a governmental entity as part of acode or set of regulations.

ANSI standards development:• ANSI started with development of standards for

pipe threads in 1919.• ANSI has now accredited 268 organizations

either as SDOs, accredited committees orthrough canvassing in:


• Safety • Information technology• Fire, building codes • Energy• Manufacturing • Equipment

Standards development organizations involved withbuildings and public works:

• ASTM• American Society of Mechanical Engineeers

(“ASME”)• National Fire Protection Association• American Society of Civil Engineers• American Concrete Institute

NAESB as a case study of how an ANSI standardsdevelopment organization operates:

• NAESB started in 1994 as the Gas IndustryStandards Board to develop business practicestandards and communications and e-commerce protocols for the natural gas industry.

• NAESB expanded effective January 1, 2002 toinclude wholesale gas and electricity and retailgas and electricity.

WHAT IS NAESB?

The objects and purposes of NAESB are to propose andadopt voluntary standards and model business practicesdesigned to promote more competitive and efficientnatural gas and electric service, as such standards applyto electronic data interchange (“EDI”) record formats andcommunications protocols and related businesspractices that streamline the transactional processes ofthe natural gas and electric industries.

(NAESB Certificate of Incorporation, Art. II, Section 1.)

WHY NAESB?• Uniform standards necessary to a national

energy market• Standards lower transaction costs• Standards give regulators comfort that markets

can function effectively• NAESB provides a tested and credible process

for developing standards

HALLMARKS OF THE NAESB PROCESS:• Independent• Open• Inclusive of all affected interests• Membership-driven• Consensus-based• Balance of interests• Focus on practices, not policy

End Result = Industry Credibility

NAESB GOVERNANCE

Board of Directors — Focuses on broad policy issues,i.e., changes to Certificate/Bylaws, annual plan, budget,membership, and communications with regulators

Executive Committee — Focuses on standardsdevelopment

Membership — Ratifies selected actions of Board(changes to Certificate) and Executive Committee(adoption of standards)

ORGANIZATION OF MEMBERSHIP

Quadrants — Broad market sectors, i.e., Wholesale Gas,Wholesale Electric, Retail Gas, Retail Electric

Segments — Defined areas of interest within each broadsector — e.g., producers, distributors, marketers, end users

Quadrant & Segment Organization — Structure(See Figure 1)

FIGURE 1

Wholesale Gas — 5 Segments• End Users• Local Distribution• Pipelines• Producers• Services

Retail Gas — 4 Segments• End Users• Distributors• Service Providers• Suppliers

Wholesale Electric — 5 Segments• Transmission• Generation• Marketers/Brokers• Distribution/Load Serving Entities• End Users

Retail Electric — 4 Segments• End Users• Distributors• Service Providers• Suppliers

CRITICAL MASS FOR NEW QUADRANTS/SEGMENTS

• For representation on Board and ExecutiveCommittee, each Quadrant must have at least


(i) 40 Voting Members, (ii) four Segments, and(iii) five Voting Members in each Segment

• In addition, Board must approve Quadrant andSegment procedures and, thereafter, Quadrants/Segments must conduct elections to selectBoard and Executive Committee members

GOVERNANCE AND AUTONOMY OF QUADRANTS

NAESB as a Whole• Board of Directors• Executive Committee for triage

Quadrants• Determine Segment make-up• Each Quadrant develops standards applicable

only to its own sector• Set own pace for standards development

Balanced Voting• All Quadrants have equal voting weight at the

Board level, regardless of the number ofDirectors from each Quadrant

• All Segments have equal voting rights within aparticular Quadrant

• Board may amend Certificate or Bylaws only witha super-majority vote: 75% of the Board and atleast 40% of Directors representing eachSegment of each Quadrant

• Executive Committee (as a whole or for eachQuadrant) may adopt, modify, interpret orrescind standards only with a super-majorityvote: 67% of the EC members and at least 40%of the EC Members representing each Segment

Balanced Voting/Veto RightsBoard of Directors — 75% of Board(governance) 40% of each Segment

Executive Committee — 67% of each Quadrant(standards) 40% of each Segment

Subcommittees — >50% Balanced Across Segmentsof each Quadrant

(proposed standards) balanced voting in the segments**

** Each Segment has up to two votes apportioned equallyover those Segment attendees present; one company –one vote

Standards Development Tenets• Standards development based on an Annual

Plan• Balanced voting structure at EC and

subcommittees• Multiple opportunities for involvement and

comment• Defined process for standards development with

several tiers of voting to ensure broad basedsupport

EC VOTING RULES

Voting Rules• Business subcommittees utilize a “balanced”

segment voting process to forwardrecommendations on standards development;all Segments have equal voting strength so thatno one Segment can dominate by sending themost representatives to meetings

— Members and non-members may vote at thesubcommittee level

— Executive Committee approval requires a super-majority

— Membership ultimately ratifies all action on standards

A “DAY IN THE LIFE” OF A STANDARD(see Figure 2)

Joint Development, ReconsiderationThe NAESB structure provides flexibility for Quadrantsto jointly develop standards in two ways:

• When the EC determines from the request thatit is in their best interest to do so, they jointlydevelop the standards

• When EC Quadrants other than the Quadrant(s)to which the request was assigned determinefrom the subcommittee’s recommendedstandards or action by the affected Quadrant’sEC that such standards could apply to them,such Quadrants may ask for reconsideration.The Executive Committee as a whole may thendirect that multiple Quadrants consider thestandard

VOLUNTARY NATURE OF STANDARDS

Voluntary Standards• From NAESB’s perspective, all standards are

voluntary and may be provided to regulatoryagencies to consider as they are published

• Regulatory agencies may choose to adoptstandards

Request for Standard

Triage Subcommittee

Executive Committee

Standards DevelopmentSubcommittees

Recommendation

Industry Comment

Executive Committee

Membership Ratificationby applicable Quadrants

NAESB

BalancedVote

BalancedVote

BalancedVote

BalancedVote

BalancedVote

FIGURE 2


• NAESB will not monitor for compliance, provideperformance measures for compliance, or definesanctions for non-compliance

• NAESB will not advocate• Voluntary, non-advocacy approach is typical of

ANSI accredited Standards DevelopmentOrganizations

MILESTONES IN NAESB CREATION

• September 19, 2001 — Board adoptedcertificate changes to transform GISB to NAESBon January 1, 2002

• December 5, 2001 — Board adopted newBylaws for NAESB

• December 19, 2001 — FERC issued Order onguidance for an organization to addresswholesale electric market standards

• January 1, 2002 — GISB officially changed toNAESB

• March 7, 2002 — Board approved proceduresfor the two retail quadrants

• May 16, 2002 — FERC issued Order explainingneed for WEQ to organize under NAESB, and tocoordinate on reliability issues with NERC

• June 2002 — Elections held in Retail ElectricQuadrant to determine Board and EC members;NAESB retail quadrants prepare annual plans

• June 28, 2002 — NAESB Board meeting inSeattle; REQ is represented; Board approvesSegment procedures for REQ and extends timeuntil September 23, 2002 for REQ to be fullypopulated

• July 29, 2002 — NAESB Board approvesprocedures for WEQ.

NEXT STEPS

• September 2002 — Advisory Council meetingset for before September 23 Board meeting;finish work on organization of WEQ and RGQand hold WEQ and RGQ elections for Board andEC seats

• September 23, 2002 — Annual meeting andBoard meeting at Hunt Valley, Maryland

HOW TO FIND OUT MORE INFORMATION

Web Sites:www.naesb.org

• Bylaws, Strawman, Certificate for NAESBwww.ansi.org

• Background on ANSI

Contact InformationNAESB

• Phone — 713-356-0060• Fax — 713-356-0067• Email — [email protected]

ANSI• Phone — 212-642-6000• Fax — 212-398-0023• Email — [email protected]


REQUIREMENTS OF AN EGM EDITORR. Michael Squyres

Flow-Cal Inc.P.O. Box 58965, Houston, TX 77258-8965

ABSTRACT

The natural gas industry has adopted EGM as a meansof increasing the speed and accuracy with whichmeasurement information is obtained. This has createdthe need for an electronic data management system.These systems, if not properly designed andimplemented, could potentially render the entire processuseless. Therefore, it is essential that the system addfunctionality that complements the power of thehardware. With proper implementation, such a systemwill not only facilitate operations in today's fast paced,post-FERC 636 environment, but also will establish afoundation for meeting tomorrow's measurementchallenges.

A good EGM data editing software package shouldprovide a suite of tools to facilitate accurate and timelydata processing. It should do this in a structured, featurerich, well designed environment. The program shouldinclude functions to do the following: import the data,recognize, review, and correct anomalies report, export,as well as provide advanced ad hoc query capabilities.Other considerations should include the developer'scommitment, resources, and long term strategy, vis-à-vis electronic gas measurement, as well as industry'soverall acceptance of the package.

INTRODUCTION

The natural gas industry is currently racing to update itsdownstream data handling capabilities to keep up withthe technological advances of field automation. One areaof concentration is electronic gas measurement (EGM)editing, recalculation, and reporting programs.Fundamental to all measurement systems is the abilityto accurately measure, review, correct, and report thedata. Any weakness in this chain undermines the speedand accuracy of the system, the primary reasons forautomation. In this paper, I attempt to identify the idealattributes necessary to streamline this process.

One common thread among all successful programs isa good basic design that is easy to understand and use.

This intuitive design, when combined with gooddocumentation, helps to minimize training requirementsand reduce problems. The documentation should includehard copy, as well as on-line, context-sensitive help. Thiscan reduce mistakes and technical support requirements,fostering a more productive environment.

For installations with numerous flow computers, theability to process in batches provides faster processing.

EGM Event Flow Diagram

This requirement exists in each phase, starting with theimport process and ending with the report and exportfunctions. Without batch capabilities, the processbecomes too labor intensive and reduces the system’seffectiveness.

DATA IMPORT

The first step in operating an effective data managementsystem is to import the data. The editor must have theability to import the data from a variety of meters andmanufacturers. This allows the user to combine all thedata into a single system, and affords more options whenreplacing or acquiring additional field hardware. Importingis a critical step since any mistake made here ensuresan incorrect output, which is why clean imports thatinclude all pertinent data are an absolute necessity.

FIGURE 1.Statistical Expert System Data Scan


The import files should be binary or encrypted, asopposed to ASCII, to provide a more secure audit trail.ASCII files can be easily and undetectably altered, usingany standard word processor. Deceptively altering abinary or encrypted file is much more difficult, since itinvolves a time-consuming process of reverseengineering.

Ideally, all import routines should be written andsupported directly by the meter manufacturers. The dataimport requirements include not only the flowingparameters and meter station characteristics, but alsothe data points used in the evaluation process. Theseinclude such items as the alarm setpoints, and thecalibration and transmitter ranges.

I will not attempt to list all of these data items; however,the five data categories include: characteristics, gasquality, event and alarm data, as well as the flowingparameters. Additionally, the program should facilitatethe import of on-line gas chromatographic data, whenapplicable. In fact, every variable and value used in thecomputation of the volume and MMBTU, includingcalculation methodologies, should be imported to ensuredata integrity.

IDENTIFICATION OF POTENTIAL PROBLEMS

The second step is a means of identifying the missing orsuspect data. Locating missing data involves not only adata scan to find unaccounted for flowing parameters,but also includes an event/characteristic audit to identifyany discrepancies between a characteristic fileconstructed from logged events, and the actualcharacteristic file.

Identifying the suspect data is somewhat morecomplicated and should include a number of configurableparameters that are used in a statistical data scan. Theseparameters should be retained for each meterindividually, reflecting each meter’s unique flowcharacteristics and history (see Figure 1). This processis an electronic equivalent to chart censoring.

With conventional meters, when a field technician noticeserroneous or missing data he makes a note on the chartand possibly includes a field estimate of the neededcorrection. This procedure requires an electroniccounterpart in the data management system, sinceerroneous data can easily go undetected by normalstatistical methods.

The next step in electronic data management is reviewand editing of the data, as required. The review processshould allow the user to readily review every variable andcalculation methodology used in the volume and MMBTUdeterminations. In addition, each alarm and its associatedsetpoints should be available for data validation. This datashould be presented in a tabular environment, and alsographically for pattern recognition and visual identificationof cycles or trends (see Figure 2).

FIGURE 2.Data Graph

When the data requires editing, the user must be able tomake changes to an individual record or to a date span,in either the graphical or tabular environment. Anychanges to the data must be recorded and flagged, alongwith an associated reason for the change. However, theoriginal data should always be maintained, thus allowingthe system to produce reports based on the modifieddata, while preserving the original data for audit trailpurposes.

The user should have the option of multiple editingtechniques, depending upon the type of error. Amongthe available methods are: a cut and paste, point to point,freeform, and some means of creating shifts or offsetsfor instances of calibration error.

The technique employed should be determined by theuser, based on the type of error, information available,and flow history.

Part of the editing process involves the recalculation ofthe volume and associated MMBTU. The editing programshould allow for recalculation using the 1985 and 1992AGA-3 algorithms, with super-compressibility (Fpv)options of NX-19, AGA-8 GROSS 1, AGA-8 GROSS 2,or AGA-8 DETAIL methods. Data that has not been editedshould never be recalculated, since the originalcalculations are always more accurate.

The flow computer performs its calculations everysecond, based on live values, and accrues the volumeand MMBTU; however, the editor must recalculate basedon hourly averages. For this reason, the data manager’srecalculation routines should employ some type ofskewing technique to more closely approximate whatthe flow computer would have calculated had the errornot existed in the original data.


REPORTING

The data management system should provide for anumber of standard reports. Each of these reports shouldinclude the ability to process meters individually or inbatch. Probably the most basic of all reports is the dailyvolume statement. The user should have the option toproduce the original or a revised version. Some of theadditional standard reports include: volume summary,event report, alarm report, missing data, suspect data,and characteristics.

Although the program provides most standard reports,there are always special ad hoc reports required. Forthis reason, the data manager should provide somemeans of allowing the user to extract specific data fromthe database using some standard means, such asStructured Query Language (SOL). This allows a user tocreate unique reports based on some very specificcriteria. [For example: generate a report showing allmeters exceeding 4 ppm H2S or greater than 3% CO2for the last year, sorted in descending order, based onvolume.] By allowing this type of access by the users,custom programming for unique applications is notrequired. This means end users can remain asautonomous as necessary or desired. It also means thatno source code is required for modifications, should theeditor be a third-party product.

DATA EXPORT

Once all data has been corrected and validated, somemeans of providing the pertinent data (electronically) tothe gas accounting group and to third parties must exist.In today's environment, it is no longer acceptable

to manually re-enter existing data. Therefore, the datamanagement system must provide comprehensive,flexible data export capabilities.

The ability to provide audit data that is compliant withAGA Chapter 21 is mandatory, especially for custody

transfer applications. The package should readily supportan automated means of bundling the data in an electronicformat and/or producing it as hard copy. By automatingthis process, the user can substantially reduce the timerequired to individually identify and produce thenecessary reports and files.

Some of the more generic attributes needed for the idealsystem is network support and an archiving tool for filesize management. The system should have an easymeans of archiving and retrieving the old data since thequantity of data for EGM meters can quickly exceedstorage space. Network support is an importantconsideration, since most companies already operateon LANs/WANs and most others probably will in the nearfuture.

SUPPORT

The last and perhaps most important consideration forany mission critical data management system is technicalsupport, with regular software updates. When enteringcloseout, it is essential to have access to a technicalsupport team to assist with any unforeseen problems.The support team must have the tools to remotelydiagnose and quickly resolve problems. Routine softwareupdates are necessary, not only to correct minorproblems, but to add additional functionality andenhancements. Ongoing development ensures againstsoftware obsolescence by continuing to meet any newAGA/API standards and requirements.

CONCLUSION

A good EGM editor should facilitate quick and accuratedata processing with a suite of tools for recognizing andcorrecting anomalies. It should have complete reportingand exporting functions, with advanced ad hoc queryingcapabilities. In addition to a well designed, feature richprogram, one should feel comfor table with thedeveloper's commitment, resources, and long termstrategy.

R. Michael Squyres


GAS CONTRACTS: THEN AND NOWMark B. Fillman and Gary P. Menzel

Coastal Flow Measurement, Inc.P.O. Box 58965, Houston, TX 77258-8965

Our industry has seen tremendous progress in theaccuracy of natural gas measurement since theimplementation of electronic gas measurement (EGM) inthe 1980s. With respect to orifice measurement, thetransition from mechanical chart recorders to EGM hadan unprecedented impact on our ability to measure naturalgas and adjust to market demands throughout the country.In order to realize the benefits of EGM, gas contractsshould include measurement provisions specific to thistechnology and its downstream data managementrequirements. Furthermore, they should represent bothbuyer and seller in the most equitable manner possible.This writing discusses some of the challenges in ourindustry, both then and now, while recommendingmeasurement provisions for gas contracts.

In the era dominated by circular orifice charts, the gasmeasurement industry was often described as part art—part science. Interpreting recorded variables during chartintegration is subjective, at best. The subjective natureof this phase of the measurement process has been amajor point of contention between buyers and sellers.Few contracts addressed this matter or offered anypractical dispute resolution provisions. Consequently,whoever took responsibility for determining the gasvolume at the sales point had discretionary control overthis important data. Needless to say, buyers and sellerswere typically at odds over this measurement issue.

The most immediate benefit of electronic gasmeasurement has been the elimination of subjective

interpretation from chart integration. Despite widespreadawareness of this problem, even current contracts failto provide suitable guidelines for processing orificecharts. Since the methods for resolving these issues arefairly expensive and imprecise, we highly recommendutilizing EGM wherever custody transfer measurementis performed. EGM equipment records the flow variableswith greater resolution and accuracy than mechanicalchart recorders. Accordingly, approximately one-third ofall measurement disputes could be averted usingelectronic gas measurement.

Currently, one of the most neglected issues within theindustry is auditing. In our opinion, the party responsiblefor managing measurement data should furnish auditdata to each participant in the transaction. Thatinformation should be in a mutually acceptable formatthat allows for detailed verification of the volume.Because auditing yields an average four-to-one returnon the investment, the right to audit should be explicit inevery gas contract written.

The measurement provision of your gas contract shouldcomprise a fundamental agreement between buyer andseller to employ acceptable standards and proceduresfor measurement, sampling, auditing, and disputeresolution. Even today, most contracts do not containthe necessary language to meet these objectives. Webelieve the following provision will serve to improve thequality of gas measurement and the relationship of trustbetween buyer and seller.

2002 ASGMT SAMPLE EGM CONTRACT

1.00 For purposes of this contract, volumes will be calculated and reported on a calendar month basis, from 9:00a.m. on the 1st of each month through 9:00 a.m. on the 1st of the following month.

1.01 The unit of volume for measurement of gas delivered hereunder shall be one thousand (1,000) cubic feet ofgas at a base temperature of sixty (60°F) degrees Fahrenheit and at an absolute pressure of fourteen andsixty-five hundredths pounds per square inch, absolute [14.65 psia (Texas state requirement)]. All volumesshall be measured and calculated according to the current standards prescribed in the American GasAssociation Report No. 3, Orifice Metering of Natural Gas and Other Hydrocarbon Fluids, Parts 1-4, asamended from time to time or as mutually agreed upon between the parties hereto. Any volume recalculationshall utilize a mutually agreeable technique to compensate for the inherent inaccuracy associated withcalculating volumes from averages of the flowing parameters as opposed to the actual instantaneous valuesused by the original flow computing device.

1.02 The Electronic Gas Measurement (EGM) system shall be capable of establishing an audit trail by compiling andretaining sufficient electronic data and information for the purpose of verifying daily and hourly quantities, and shallcomply with the American Petroleum Institute — Manual of Petroleum Measurement Standards, Chapter 21, Section1 — Electronic Gas Measurement (API Chapter 21), or other mutually agreeable standards. Pipeline shall preserveaudit trail information for a minimum of 4 years or the time required by any governmental agency, whichever is greater.


1.03 Upon request, the measurement data described in paragraph 1.02 of this article and calculated by the metershall be preserved and provided for auditing purposes in a mutually agreeable standard format. Standardformat shall mean the API Chapter 21 compliant electronic format provided by the manufacturer of the meteror another format for which commercially available EGM Review and Recalculation software exists and ismutually agreed upon by the parties hereto. Printed paper data is not acceptable for audit purposes. In addition,the final custody transfer volumes and MMBtus and a log of the changes made between the amounts calculatedby the meter and the settlement amounts shall be provided. Unless a currently unresolved exception exists, allvolumes shall be considered final after 3 years.

1.04 Pipeline, or its designee, shall be responsible for measurement at the point of delivery. Measurement stationshall operate within a .25 to .60 Beta Ratio range with an Electronic Gas Measurement (EGM) recording device.This device shall utilize close-mount transmitters installed on an orifice meter run whose orifice taps are in thevertical plane, along with all related equipment necessary to accurately measure the gas delivered hereunder.Close-mount shall mean that the distance between the orifice tap and the transmitter will not exceed 24inches. The differential pressure measuring range shall not exceed 300 inches of water. Alarm limits shall beset in the meter and all alarm conditions shall be logged so as to identify flow conditions which are outside theproper operating conditions of the meter. The alarm report shall be made a part of the standard audit trailincluding all alarm conditions.

The calibration and programming of and data collection from the recorder and related equipment shall bethe responsibility of Pipeline or its designee. All calibration and adjustment of this equipment shall be inaccordance with Paragraph 1.08 of this article.

1.05 Producer may, at its option and expense, install check meter equipment upstream of Pipeline’s measurementstation for checking the Pipeline’s metering equipment. Such measurement equipment shall be installed so asnot to interfere with the operation of Pipeline’s facilities and shall comply with the standards set forth in thiscontract. In addition, producer may install a check recorder on pipeline’s meter run utilizing a dual-mountmanifold system on the same orifice taps which provides independent isolation between each parties respectiverecording devices. This system shall be installed so as not to interfere with the operation of pipeline’s facility.

1.06 The temperature of the gas flowing through each of the meters at these stations shall be measured with a temperatureelement so installed as to provide an accurate measurement of the flowing temperature at the primary device.

1.07 The specific gravity, gross heating value, and composition of the gas will be determined by Pipeline by takinga representative sample once Monthly at Pipeline’s meter tube. Sampling or collection shall be performedduring the Pipeline’s scheduled meter inspection and calibration test and all samples shall be obtained andanalyzed using current Gas Processors Association (GPA) standards. These analytical results shall be appliedat the beginning of the month the sample was taken and until a subsequent representative sample is applied.

Producer shall have the right to obtain a duplicate sample of the pipeline’s gas. If a difference between thepipeline and producer’s duplicate sample exceeds one-half percent (.5%) in MMBtu, then the Producer shallhave the right to call for a retest and the analysis from the preceding period shall be used until the results can beverified or a representative sample is obtained and applied. Upon written request, pipeline or producer shallfurnish the requesting party with their natural gas sample, chromatographic gas analysis report, or any otherinformation which may be required to verify the other party’s analytical procedures or results.

1.08 Producer shall have the right to have a representative present at the time of any installing, cleaning, changing,repairing, inspecting, testing, calibrating, or adjusting done in connection with the equipment used inmeasuring gas deliveries hereunder. The records from such measurement equipment shall remain the propertyof Pipeline, or its designee, but Pipeline will make information available in accordance with Paragraph 1.03.

1.09 At least Monthly Pipeline shall calibrate or cause to be calibrated the meters and instruments used formeasurement hereunder. Pipeline shall give Producer sufficient advance notice so the Producer may, at itselection, have a representative present at such tests. For purposes of measurement hereunder, theatmospheric pressure shall be assumed to be fourteen and forty hundredths per square inch (14.40 psia)irrespective of variations in the actual atmospheric pressure from time to time. If the transmitter measuresactual atmospheric pressure, then actual atmospheric pressure shall be used.

Where unacceptable measurement differences exist and all other tests have been conducted to remedy theerror, the orifice meter run shall be inspected and tests shall be performed to verify compliance with currentAGA Committee Report No. 3, Part 2, Specification and Installation Requirements, as amended from time to


time. The expense shall be shared equally between pipeline and producer if the error is not discovered in thisprocess, but if the problem is found in the tube or its accessories, costs for the inspection, repair, and/or replacementof the tube shall be born by the pipeline unless it can be determined that the malfunction was caused by theproducer’s gas. At least once every 3 years, this inspection and test procedure shall be conducted at the pipeline’sexpense regardless of any apparent discrepancies, unless such right is waived by both parties hereto.

1.10 If, upon any test, the metering equipment in aggregate is found to be recording inaccurately by one-half percent(0.5%) or more of the correct rate in MMBtu under actual flowing conditions, registration thereof and paymentbased upon such registration shall be corrected at the rate of such inaccuracy for any period of inaccuracy thatis definitely known and/or agreed upon. In the event such period of inaccuracy is not definitely known and/oragreed upon, the adjustment will be made halfway back to the preceding test. Following any test, any meteringequipment found to be inaccurate to any degree shall be adjusted immediately to record accurately. The maximumzero-cutoff value which may be programmed for the differential transmitter reading is .25 inches.

If for any reason the meter is out of service or repair so that the quantity of gas deliveries through such metercannot be ascertained or computed from the readings thereof, the quantity of gas so delivered during theperiod when the meter is out of service or repair shall be determined on the basis of the first of the followingmethods which is feasible, as agreed between the parties hereto.

a.) By using the registration of the Producer’s check measuring equipment, if such equipment is recordingaccurately. The Producer agrees to allow the same provisions for the Pipeline to witness and audit suchmeasurement that the Pipeline allows on the custody transfer measurement equipment;

b.) By using the cumulative volumes from any field measurement source adjusting for historical differencesand/or fuel consumption between the field meters and sales. The Producer agrees to allow the sameprovisions for the Pipeline to witness and audit such measurement that the Pipeline allows on the custodytransfer measurement equipment;

c.) By correcting the error mathematically if the amount of such error is ascertainable by calibration, test, orcalculation;

d.) By estimating the quantity from the amount of deliveries during, preceding, and/or following periodswhere the delivery conditions were similar and the meters were recording accurately or;

e.) By any other method which is mutually agreeable.

1.11 Any error resulting from a source other than meter calibration shall be corrected retroactive to the date fromwhich the error began, regardless of percent error, unless other arrangements are made by mutual consentbetween Pipeline and Producer.

1.12 Any volume or energy revision made by the measurement party following initial close-out, shall not be madewithout a detailed written explanation of the revision and such changes must be mutually agreed upon betweenthe parties. Also, measurement party shall respond in writing within 30 days following receipt of any adjustmentrequest, either by making the requested volume adjustment or by stating its reasons for not doing so.

1.13 The heating value of the gas delivered hereunder shall be determined from the analysis by calculating thegross heating value of the gas in Btu per real cubic foot at a pressure base of fourteen and sixty-five hundredthspounds per square inch absolute (14.65 psia (Texas state requirement)) and a temperature base of sixty (60°F)degrees Fahrenheit with the gas assumed to be As Delivered. Such Btu shall be multiplied by each hourlyvolume in Mcf and divided by one-thousand (1,000) to yield the MMBtu for that hour.

Commercial software is available to create similar documentation customized to yourneeds and conditions. Our company recommends that you have your legal departmentapprove the language used in any commercial software purchased; however, no changeshould be made without consulting with your measurement department.

Gary P. Menzel


TRAINING OF OFFICE MEASUREMENT PERSONNELTom Cleveland

Hanover Measurement Services2100 West Loop South, Suite 1450, Houston, TX 77027

INTRODUCTION

It has long been held that the measurement function isthe “cash register” for the energy industry. Other factsthat relate to measurement are that it involves theapplication of scientific laws, knowledge of physicalproperties, application of mathematics, and facilitatedin an environment based on industry standards andcompany policies and procedures. Also, ask someonewho is not employed in the energy industry what thatperson knows about energy measurement. Chances arethat he or she will know very little. So, the energymeasurement function, while a “cash register” for theindustry and based on science, is also a unique activitythat requires specific knowledge, skills and experience.What types of unique qualifications are needed will beidentified and some methods for providing training formeasurement personnel will be discussed.

The “cash register” reference also implies a financialimplication to the organization, whether it is an internalmeasurement function of an energy company, or aservice provider that performs the measurement functionas an outsourced service offering to energy organizations.This financial implication will require that properaccounting controls be present and that all proceduresare auditable, all reported volumes are reproducible fromthe sources of raw data, and that documentation is keptthat has accompanied the data through the process ofbeing converted from raw data to reportable numbers.Those who perform these processes must understandthe importance of the financial implication and thepotential impact of measurement uncertainty to theircustomers.

MEASUREMENT OFFICE ACTIVITIES

A small percentage of the activity in the measurementoffice is somewhat routine and clerical in nature.However, the remainder is unique to the measurementfunction and can be considered a trade, or even an art,such as running a chart integrator. Since the final product,a volume statement, contains information that will betied to the financial function of the organization, and thetechnology utilized to record and calculate the numbersprovided on the volume statement is based on thescience and technology of measurement, some activitiesunique to the measurement office organization will beperformed that require some specific technicalknowledge.

The data are typically received in a raw form, such asrecorded on a paper chart, or electronically, such as EFMor gas analysis data. The raw data must be screened forvalidity, missing or erroneous data, etc. Data must beentered, either via manual data entry, or electronic import,into a measurement system. If the data are received inthe form of paper charts, they must be censored,integrated, and the data passed on to the measurementsystem. If the data are electronic, they must be in thecorrect format for import into the measurement system.Meter stations must be correctly set up in themeasurement system prior to any data import andmeasurement changes, such as orifice size changes,etc., must be entered to ensure data is handled correctlyand the results will match reality.

Practically all measurement office activities are drivenforward by a monthly business cycle. The raw data arecollected throughout the calendar month, and then finalvolume statements are delivered by a deadline in thefollowing month. Efficient processes must be in place toreceive the data and put the processes in motion to meetthe closing deadlines. Along the way, various activitieswill take place to provide validation. These activitiesrequire various amounts of skill, specific knowledge andexperience. Some require the basic knowledge of theenvironment of the recording device, such as the abilityto visually inspect the chart, marking with a pencil, thingsthat will help guide the chart integration and identifyanomalies that need the integrator’s attention. EFM datamust typically be imported into the measurement system.There are typically tools built into the import processthat flag anomalies such as missing data, out-of-rangevalues, unrecognizable digital characters, etc. Someonefamiliar with how to handle these flags is required toperform or monitor the import function.

Gas analysis information is typically applied in themeasurement system to the volume data andrecalculations are performed per some effective date.Personnel who are familiar with valid gas analysis dataranges and the proper effective dates are required toperform and monitor this function. Actions by the fieldpersonnel that could potentially change the volume valueoutcome, such as recording instrument range changesor calibrations, will trigger a notification to themeasurement office to apply the changes to the databased on an effective date. Field personnel may provideinformation to the measurement office to “alibi” problemswith the data, such as when a chart clock stops, an EFMbattery experiences low voltage, the recording gauge


lines freeze or experience leaks, and other problemsituations that require manual intervention of the data inthe measurement office. These all require somefamiliarization in the measurement office with the properprocedures to maintain audit trail requirements.

TRAINING REQUIREMENTS

As mentioned earlier, a small percentage of the activityin the measurement office is considered routine andclerical. That means there is a large percentage that isconsidered specialized or requiring a higher level of skills.To meet the requirements of the financial implication ofthe measurement work taking place, there will need tobe employees that know exactly what they are doing.There will need to be supervisory employees that takeaccountability for the office’s output and for keeping theworkload flowing through the office, meeting the deadlineneeds of the internal and external customers.

Measurement office work, whether integrating charts,manipulating volumetric and gas quality data, queryingdatabases for validation and troubleshooting purposes,due to the skill requirements, is not something a novicecan step immediately into. It is not as simple as handinga new employee a manual or instruction sheet. Theremust be experienced mentors available for newemployees to have the measurement trade passed onto them by and hopefully develop the art in those thathave the aptitude for it. The mentors, based on thefinancial reporting requirements of the organization, mustreinforce the measurement accuracy and uncertaintyrequirements for the data. It must be ingrained with thetrainee the importance of meeting closing deadlines andthe downstream commercial costs associated withhaving to make adjustments to prior periods.

TRAINING METHODS

One of the most tried and true forms of training that hasbeen around for centuries is through apprenticeship. TheWebster’s Dictionary definition is: apprentice 1.a: onebound by indenture to serve another for a prescribedperiod with a view to learning an art or trade b: onewho is learning by practical experience under skilledworkers a trade, art or calling. Another Webster’sdefinition is: art 1: skill acquired by experience, study,or observation. Still another applicable definition fromWebster’s is: trade 3.a: the business or work in whichone engages regularly b: an occupation requiringmanual or mechanical skill. The ideal situation fortraining entry-level measurement office employees is toset up an apprenticeship-like system whereby seasonedemployees that are considered experts in the variousparts of the process can work along side the newemployee. The expertise of the trainer can be passedalong to the beginner in digestible portions, without thepressure of being totally being accountable for thevolume statement accuracy. With organizational changesthat have evolved over the last 20 years as a result ofderegulation, the apprentice approach has almost

disappeared. Downsized organizations have shed manyolder, experienced workers and allowed untold numberof years of experience to walk out the door. Youngerworkers that have been hired in recent years are typicallyplaced directly into vacancies without the opportunityto develop skills and experience over a longer time.

The more routine activities of the measurement office,such as organizing charts and statements, looking overdata for obvious anomalies, filing, etc., are more easilypassed on to new employees with a minimum of traininginvolved. Manuals, including step-by-step instructions,screen-prints, examples, etc. are very effective trainingtools and are actually a collection of the knowledge andexperience gaining by others, placed in a binder thatcan be accessed by lesser-experienced employees. Themethodical ramping up of responsibilities, along with theencouragement in the trainee building his or her owninstruction manuals, is a confidence-building approachto helping make a seasoned, productive employee thatcan develop the skills needed in the measurement trade,and potentially become proficient in one or more activitiesof the process.

When a beginning employee has been given a digestibleportion of the process to take over and given continuousfeedback on performance, the confidence builds and aknowledge base builds. Once the supervisor is confidentthat a person is ready to take on additional challenges,it may be the right time to turn over more duties thatrequire additional responsibilities. The goal is for thatemployee to learn the process from one end to the other.Then, at the point that the employee is familiar and hasdeveloped the skills to handle the entire process, a singleclient, project, or other dividable portion may be handedoff to that employee. The supervisor may review the finalwork product, which is usually the volume statement,for a period of time to provide final QA/QC, along withimmediate feedback to the employee to point out thegood and the needs for improvement, to polish theemployee’s knowledge and skills.

Once employees have been exposed to the basicfundamentals of measurement office activities, theyshould be somehow exposed to field measurementactivities, which is the source of their raw data that isthe starting point of their process. Whether this exposureto field measurement is through field trips whereby theyride around with field measurement personnel, or wherefield measurement personnel come into the office toshare their trade with the office personnel, it is mostvaluable to see another perspective. Also, any exposureto the fundamentals of measurement that can beprovided, which is the same type of fundamentals thatthe field measurement personnel are taught early on, isvery valuable and will be worth the time and investmentto provide.

In addition to consideration of the specific trainingmethods, management and supervisory personnelshould continuously be considering succession planning


and the training that goes along with it. Positions ofincreasing responsibility in the measurement officeprocesses should have someone identified to bepreparing for those roles and learning the additionalduties and developing the knowledge, skills andexperience that will be required. The normal attrition foremployees in critical roles or the possibility thatsomething unforeseen may happen put a high level ofimportance to identifying successors and making surethat a sudden vacancy will not put the processes in asituation of high risk of failure.

TRAINING RESULTS

The results of proper and adequate training ofmeasurement office personnel will include the ability tomeet aggressive closing schedules routinely and withfew prior period adjustments necessary. The ability of ameasurement organization to accomplish this feat isgoing to have a significant effect on the bottom line,whether the measurement office function is an in-housemeasurement department, or a third-party serviceprovider. Employees who know what they are doing aremore comfortable with their day-to-day duties and arehappier and less stressed. Morale issues caused bystress and dependence on inexperienced or under-trained employees are a drain on the organization andlead to turnover. A churn of employees in a field such asgas measurement that takes time to develop theknowledge, skills and experience is a detrimentalsituation. As mentioned earlier, measurement is not awell-known discipline outside of the energy industry and

it is not easy to bring someone in “off of the street” thatcan step right in and contribute without some intensiveorientation and fundamentals training up front.

Well-trained employees are required for measurementoffice operations when those operations are driven byclosing deadlines, especially when the receipt scheduleof most of the raw measurement data is out of theircontrol. The processes must run efficiently and be ableto self-correct for situations such as late-delivered charts,or system malfunctions, etc. Trained employees canadapt to adverse situations, when required, whereuntrained employees may not be able to function.

CONCLUSION

Entry-level positions in most measurement organizationsthat are truly “trainee” positions are almost a thing of thepast. Positions are typically filled based on potential tolearn, so basic on-the-job training programs must be inplace. There is nothing more valuable in measurementoffices than highly seasoned, highly skilled employeesthat can pass their knowledge, skills and experience onto new employees. At the same time they are providingthis mentoring role, they are also ensuring that the finalproduct, volume statements, are meeting the stringentquality control requirements and meeting the aggressivedeadlines typically imposed.

REFERENCES

Webster’s Ninth New Collegiate Dictionary

Tom Cleveland


CONVERSION FROM VOLUME TO ENERGY MEASUREMENTRadhey S. Thakral

Duke Energy Gas TransmissionP.O. Box 1642, Houston, TX 77251

The purchase, transport, and sale of natural gas as acommodity with a specific energy value per cubic foothas transformed the natural gas industry from one of asystem based on volume measurement to a systembased on energy measurement. The following discussionwill review the evolution of natural gas industry from asystem of volume measurement to the present systemof energy measurement.

Natural gas has served as an important fuel through thecenturies. It is believed that the first commercial use ofnatural gas was by the Chinese in 900 BC. The Chineseused a system of hollow bamboo to transport the naturalgas from shallow wells. The gas was used as a fuel toextract salt from sea water. In 1816 manufactured gasfrom coal was first used in the United States to fuel gaslights. In 1821 near Fredonia, New York the first naturalgas well was drilled in the United States. William Hart, agunsmith, drilled the twenty-seven foot deep well near acreek outside Fredonia where gas bubbles had beennoticed. As a result of this well, Hart is considered thefather of the natural gas industry in the United Sates. In1843 cast iron pipe started replacing wooden pipelinesand the use of iron pipe provided the first reliable andsafe method of transporting gas to market.

As the industrial revolution progressed in the UnitedStates, the need for natural gas as a fuel also developed.After the commercial discovery of crude oil at Titusville,PA in 1859, oil was produced as a primary fuel and naturalgas was produced as a by-product. The natural gas hadlittle value as compared to oil and there were onlyrudimentary pipelines to transport natural gas to market.Crude oil, on the other hand, was sold by the woodenbarrel and could be easily transported by wagon, barge,or rail car.

In the early 1900s, with advances in drilling technologyand the invention of the rotary drill, wells could be drilledfaster and deeper. As a result, more and more gas wasdiscovered which had no market. Much of the gas wasjust flared and wasted to continue oil production.Entrepreneurs saw a business opportunity here. Theyreasoned that pipelines could be built to the gas fields,the gas bought for pennies per thousand cubic foot, andthen transported and sold for a profit. This was thebeginning of a viable new gas transmission industry.

The natural gas bought at the wellhead by pipelines wasmeasured by volume in cubic feet and was consideredto have a standard heating value range suitable formarketing. The gas was treated as a commodity

transported in a regulated pipeline environment. Typically,the federal government regulated the interstate gasmoved across state lines and the state governmentsregulated the intrastate gas moved within the state. Thegas was bought and sold by volume and the energy valueof the gas was determined only on a limited basis bycalorimeter.

In 1978 The Natural Gas Policy Act was passed byCongress as a method of leveling out the large pricedifference between the state and federally regulated gas.This act also specified that natural gas was to be boughtand sold on the basis of its actual energy content percubic foot. As a result, The Natural Act Policy Act of1978 changed the way gas was valued as a commodity.The new standard combined volume measurement andgas heating value measurement to produce an energymeasurement system for natural gas.

TERMS USED IN ENERGY MEASUREMENT

Btu (British Thermal Unit) — A Btu is defined as thequantity of heat that must be added to one pound ofpure water to raise its temperature from 58.5 degreesFahrenheit to 59.5 degrees Fahrenheit under standardpressure. A Btu is generally thought of as the heatrequired to raise the temperature of one pound of waterone degree Fahrenheit.

Standard Cubic Foot — A standard cubic foot is definedas the quantity of gas occupying a cubic foot of spaceat a specified temperature and pressure.

Total or Gross Heating Value — The total heating valueis defined as the number of British Thermal Unitsproduced by combustion of gas at constant pressure ofthe amount of gas which will occupy a volume of onecubic foot at 60 degrees Fahrenheit at the reference basepressure, with air at the same temperature and pressureas the gas, when the products of combustion are cooledto the initial temperature of the gas and air, and the watervapor formed by combustion is condensed to the liquidstate.

Net Heating Value — The net heating value is defined asthe number of British Thermal units produced by thecombustion of gas at a constant pressure of the amountof gas which will occupy a volume of one cubic foot at60 degrees Fahrenheit at the reference base pressure,with air at the same temperature and pressure as thegas, when the water vapor formed as product ofcombustion remains in the vapor state. As a result, the


Net Value is less than the Gross Heating Value by anamount equal to the heat evolved by the condensationof the water from a vapor state to a liquid state.

Saturated Gas Heating Value — The saturated gasheating value may be defined as either the Gross or Netheating Value of a standard cubic foot of gas at baseconditions which is saturated with water.

Dry Gas Heating Value — The dry gas heating value maybe defined as either Gross or Net Heating Value of astandard cubic foot of gas at base conditions whichcontains no water vapor. The Dry Gas heating Value is alarger amount than the Saturated Gas Heating Value ofa standard cubic foot of gas because the dry cubic footcontains no water vapor molecules.

Standard Base Conditions — The standard temperaturebase is 60 degrees Fahrenheit. Standard pressure istypically referred to as 14.73 PSIA.

Dekatherm — A dekatherm (dth) is a unit of heatequivalent in energy to 10 therms or 1,000,000 Btu. Onedekatherm is equivalent to 1,000 cubic feet of 1000 Btuper cubic foot gas.

EQUIPMENT USED FOR HEATING VALUEDETERMINATION

Three types of equipment used for heating valuedetermination in natural gas are the Cutler-Hammer typerecording calorimeter, the gas chromatograph, and theinferential calorimeter. The Cutlter-Hammer and theinferential calorimeter have an operating principle basedon the actual combustion of gas where as the gaschromatograph operates on a principle based on themeasurement of the molecular composition of the gas.A review of each of these three types of analyticalinstruments is as follows:

Cutler-Hammer Recording Calorimeter — The Cutler-Hammer type of calorimeter is composed of a tank unitand a recorder. Within the tank unit the actual caloricmeasurement takes place. The tank is filled with waterwhich is maintained at a constant temperature. A flow oftest gas and a separate flow of heat absorbing air arepassed through water sealed meters in the tank unit.The test gas and the heat absorbing are maintained at afixed ratio. As the test gas is burned its heat is thermallytransferred to the heat absorbing air. The water vaporformed from the combustion process is allowed tocondense to its liquid state and release its latent heat ofvaporization. As a result, all the heat produced bycombustion of the gas, including the latent heat ofvaporization of the water vapor, is a measurement of thetotal heating value of the gas. This instrument iscalibrated with a natural gas standard of certified heatingvalue.

Gas Chromatography — Gas chromatography is aphysical method of separating a gas into its componentsand then measuring the quantity of each component in

the gas mixture to determine its molecular composition.Once the molecular composition is determined theheating value of the gas is determined by calculation. Inpractice the gas chromatograph is composed of asample injection system with a regulated carrier gassupply, thermally stabilized separation columns, adetector, and a recorder. The analytical results from thegas chromatograph can be used to calculatesupercompressibility, specific gravity, gallons perthousand cubic feet of liquids, and heating value. Thegas chromatograph is calibrated using a natural gasstandard with a certified mol composition for eachcomponent in the gas.

Inferential Calorimeter — The inferential calorimeter is acombustion type instrument. In practice, the test gas isburned at it stoichiometric or perfect combustion pointto obtain an inferred or deduced heating value. Thisinstrument is calibrated with a gas standard of certifiedheating value.

ENERGY VALUE CALCULATION

When calculating the energy value, it is of utmostimportance to have the volume per cubic foot and theBtu per cubic foot on the same pressure base. Also notethat a dry Btu per cubic foot has a larger heating valuethan a saturated Btu per cubic foot and that a cubic footof gas on a higher pressure base has a larger Btu percubic foot.

Thermal Calculation Using Volume and Btu - The thermalvalue for a gas at a standard temperature and pressureis calculated by multiplying the gas volume in cubic feetby its Btu per cubic foot. Standard conditions are 60degrees and 14.73 PSIA.

Example: What is the energy value of 1,000 cubic foot ofgas with a heating value of 1050 Btu per cubic foot.

Energy value = (1,000 cf) (1050 Btu/cf)= 1,050,000 Btu= 1.05 MMBtu= 1.05 Dekatherms

Thermal Correction Factor Using Volume and Btu — Thethermal correction factor for a gas at standardtemperature and pressure is calculated by dividing themeasured heating value per cubic foot by a base of 1,000Btu per cubic foot. Standard conditions are 60 degreesand 14.73 PSIA.

Example: What is the energy value factor for a gas havinga heating value of 1050 Btu per cubic foot at a baseheating value of 1000 Btu per cubic foot.

Energy value factor = 1050 Btu/cf 1,000 Btu/cf= 1.050

Example: What is the energy value of 1,000 cubic foot ofgas which has an energy value factor of 1.050.


Btu Energy value = (1,000 cf) (1.050)= 1,050

Example; What is the energy value of 1,000 cubic feet ofgas which has an energy value factor of 1.050.

Btu Energy value = (1,000 cf) (1.050)= 1,050

Pressure Base Correction - The measured pressure basemay be different from the contract pressure base. Dueto the different pressure bases used for reporting volumeand heating value, it is usually necessary to put thesenumbers into a standard base for energy measurement.

Example: What is the heating value and pressure basecorrection factor for gas measured at a meter pressurebase of 14.65 PSIA which has a contract pressure baseof 14.73 PSIA at constant temperature.

Heating value correction factor= 14.73 PSIA = 1.0054 14.65 PSIA

Volume correction factor= 14.65 PSIA = 0.9946 14.73 PSIA

Water Vapor Correction - The heating value of naturalgas can be calculated on a saturated base, an actualbase, or a saturated base depending on contractrequirements. The saturated or the dry heating valuefactor at a pressure base of 14.73 PSIA and 60 degreesFahrenheit can be calculated from the partial pressureof water vapor, 0.2563 PSIA, at saturation. The actualheating value is first determined as if dry and then itscorrected to actual.

Example: Determine the dry and saturated heating valuecorrection factor at 14.73 PSIA and 60 degreesFahrenheit.

Dry heating value correction factor=

14.73 PSIA(14.73 PSIA - 0.2563 PSIA)= 1.0177

Saturated heating value correction factor= (14.73 PSIA - 0.2563 PSIA)

14.73 PSIA= 0.9826

Example: Determine the dry heating value of a gas havinga saturated Btu of 1000 at 14.73 PSIA and 60 degrees.

Dry heating value = (1.0177) (1000 Btu/cf) = 1018 Btu/cf

Example: Determine the actual heating value for a gashaving a 45 lbs./MMcf water vapor content. The dryheating value was 1018 at 14.73 PSIA.

Water Vapor Correction Factor (F) isF = 1 - (45 lbs) (21.0181 cf)

(MMcf) ( lbs )F = 1 - 945.815

1,000,000F = 1 - 0.000946F = 0.999054Actual heating value = (1018) (0.99905)

= 1017 Btu/cf

CONCLUSION

This discussion has covered the early history anddevelopment of the natural gas industry and the transitionof natural gas from a volume based to an energy basedcommodity. Three types of instruments used todetermine the heating value of gas were discussed. Alsothe definitions used in energy measurement werereviewed and some examples were presented of thecalculations which are typically used to convert gasvolume to an energy value.

Radhey S. Thakral


ELECTRONIC GAS MEASUREMENT AUDITINGKenneth W. Blackburn

Houston Flow Measurement, Inc.4116 Roseway, Houston, TX 77025

INTRODUCTION

It has been stated that measurement is the cash registerin the exchange of natural gas. The natural gas businessis based on the buying and selling of this commodity.Measurement is responsible for balancing the input(buying) and output (selling). Errors, on either side of thisequation, affect the balance the entire business is basedupon. Add the fact that natural gas measurement canbe extremely complicated, auditing not only becomesdesirable, but necessary. As it implies, auditing anelectronic gas meter (EGM) requires careful examinationof large amounts of data in order to verify volumes andto verify the cash register. An experienced auditor is themost valuable tool in this process. In order to maintainthe scope of this paper, a general knowledge of naturalgas measurement and EGM fundamentals will beassumed.

STANDARDS USED

It is not necessary for the auditor to be an expert, but itis very helpful for the auditor to at least be aware ofstandards used in natural gas measurement and volumecalculation.

Field specifications and calculation specifications usedare the American Gas Association (AGA) Report 3 OrificeMetering of Natural Gas or the American PetroleumInstitute (API) 14.3. This report details manufacturing,installation, operation, and calculation guidelines usedthroughout the natural gas industry. EGM manufacturersfollow guidelines or standards published by the AmericanPetroleum Institute (API). This standard is known as theAPI 21.1 Flow Measurement Using Electronic MeteringSystems — Electronic Gas Measurement. Thesestandards were developed by the natural gas industryand manufacturers in order to set minimum standardsregarding EGM and providing for EGM audit trails. Forthis paper, we will assume the meter station and datameet AGA/API specifications.

A knowledge of company policy and contract guidelineswill also be required. Policies and contracts vary widelyregarding auditing, corrections, and settlements.

AUDITING PROCEDURES

With the introduction of electronics in gas measurement,new challenges were encountered in volume auditing.

Many of the same processes used in auditing of chartrecorders also apply to EGM auditing. Parameters anddata that are applied in the calculation of volumes mustbe verified in both cases. The differences in auditing arefound in the characteristics of the devices themselves.Chart recorders provide visible records that can bereinterpreted and recalculated. EGM volume reportvariables are often averaged daily or hourly at best. Thisoften makes recalculation of EGM volumes becomeuncertain on varying flow rates.

One of the best possible tools used in auditing EGMvolumes is a check meter measuring the exact samegas. Check meter volumes and all other data can bedirectly compared to the audit station, increasing theprobability of detecting errors and increasing thelikelihood of settlement of a discrepancy. If the checkstation consists of an entirely separate meter tube andplate, most errors can be traced. If the check meter islocated on the same meter tube, error detection isreduced to the secondary devices (recording equipment)since both meters use the same primary element (i.e.,meter tube, orifice plate). If a meter is located on thesame physical orifice taps, then effectiveness of thecheck meter is reduced even further. A chart recorder,used as a check meter, often becomes helpful whendetermining errors in measurement. We oftenrecommend using a chart recorder backup, in the eventthat the EGM becomes unreliable or inoperable. If checkmeasurement is not available then a balance may beused from other meters. As in any balancing process,care must be taken to account for all gas. This also opensthe scope of the audit and possible error points to theother measuring stations included in the balance.

In summary, it is extremely valuable, from an auditperspective, to have some form of check metering inplace. Ideally, the best form of check measurement isachieved using an entirely separate station (meter tube& meter), measuring the exact same gas. Realistically,any form of check measurement is better than nothingat all. The effectiveness and reliability of the checkmeasurement, just as with the station being audited, istotally dependent on the quality of the measurement.For the check measurement to have credibility, checkstations should follow the exact same AGA/APIguidelines as applied to the audit station.

Proper meter testing, sampling, and reporting in the fieldare the most important exercises in preventing problems


from occurring in the first place. Good practices in theseareas will directly affect the outcome of volumes andthe frequency of errors found in auditing. It is also goodto witness meter tests to reduce the chance of mistakesmade by the tester. This also gives the auditor one moresource of information when problems arise.

AUDIT DATA

EGM auditing requires careful scrutiny of an array of data.In order to complete an audit, the first step is to acquireall existing information used in the determination ofvolumes. EGM data usually consist of reports generateddirectly by the EGM, field location, or office. These reportsprovide volume information and other details used involume determination. Reports differ by manufacturerand company, but most provide similar basic reports.The following are typical EGM reports and other datarequired to perform a complete audit.

• Characteristic Report — Provides details onstation name, location, calculation factors, AGAdata, analysis data, alarm data, and calibrationdata.

• Volume Report — Provides daily (or hourly)volumetric readings with averaged differentialpressure (DP), static pressure (AP), andtemperature.

• Alarm Summary — List daily alarm conditionssuch as low differential, high differential, lowstatic pressure, high static pressure, low power,and many others.

• Event Report — Details all station activity suchas meter tests, calibrations, plate changes, andany other change made to the stationparameters.

• Meter Test Reports — Along with the eventsreport, most companies complete hard copymeter test reports when tests are performed orchanges are made to the station. This isabsolutely necessary to maintain the audit trail.

• Change Reports — Some companies useseparate reports when changes are made (i.e.,plate change, meter tube change, range change,etc.).

• Witness Reports — Meter tests witness reportverifying tester’s results.

• Analysis Reports — Gas composition informationtaken from spot samples, composite samples,or on-line chromatography.

• Check Station Data — Volume and other reportsfrom any check stations available.

Once all of the related reports and data are obtained,the actual audit begins. Always verify station name andID. Volume reports should be reviewed for obvious errors,such as missing days or any abnormal data. Thecharacteristic report should be reviewed to determinethat proper composition data was used, and that all otherstation parameters are correct. This should be done foreach audit, even if the audit is on a continuous basis.Occasionally, parameters are unintentionally modified orlost due to meter problems.

Review alarm summary to detect problems in the EGM.EGM manufacturers provide daily alarm summaries thatcan be a valuable audit tool. The alarm summary detailsmany conditions of interest to the operator as well asthe auditor. If there are numerous low DP alarms, theorifice plate may be too large. If there are high DP alarms,the orifice plate is too small. These conditions may notbe corrected by orifice plate changes if parameters falloutside allowable beta ratio (orifice plate/meter tube) oruncontrollable flow conditions. Generally, the differentialshould be kept above 30% of differential scale tomaximize orifice accuracy. However, this is not alwaysobtainable due to flow characteristics of the station. Thealarm limits are reported in the characteristic report andshould be verified. The alarm summary also reports errorsin the EGM transducer signals, caused by over-rangingtransducer or other transducer trouble. Alarm summaryreveals errors in the calculation process that can becaused by invalid parameters or meter problems.

Review event reports to ensure proper calibrations andthat changes (i.e., plate changes, composition, etc.) areposted properly. Most manufacturers provide detailedevents in this report that can often explain discrepancies.Any changes made to the station should be verifiedclosely. Any time changes are made, volumeaccumulators should be reset in order to ensure thatcorrect parameters are used for calculations from thetime of the change forward. Depending on contractualtolerances and obligations, corrections or adjustmentsmay be required from meter tests. This can bedetermined from the events reports. Test and foundpoints should be recorded during each meter test toprovide this audit trail.

Volumetric data can often be obtained in electronicformat allowing for integration, recalculation, andcomparison by computer. The computer will reduce timespent on audits by automatically flagging potentialproblems. Volumes should be compared to checkstation, if available. If a problem is detected throughreviewing the volumes or by check meter comparison,more detailed research should follow. Measurementproblems and/or errors can result from a wide variety ofsources. Some problems are easily detected andcorrected, others are impossible to correct. This is whereexperience is beneficial to the auditor. For the sake ofthis paper, we will divide errors into two categories, Fieldand Office.


FIELD ERRORS

With EGM, the bulk of responsibility of the meter stationand resulting volumes is with the field personnel. Withthis added responsibility and the inherent dynamicsinvolved in measuring natural gas, most, but not all, errorscan be traced to the field. For this reason, adequatetraining and clear communication with field personnelare essential ingredients in the natural gas measurementprocess.

With any type of measurement, unique characteristicsof the measurement devices in use and the actual gasbeing measured all affect the measurement results.

Primary Element

Many errors or discrepancies are a result of operationalor physical properties. Fluids in the gas stream can renderfalse differential readings. Fluids in gauge lines can causedifferential zero shifts. Pulsation, caused by compression,can result in differential error. Gauge line error due tocompression, flow obstructions, or piping, can also causedifferential error that is sometimes difficult to detect.Freezing causes problems when liquids or hydrates formsolids in piping or gauge lines. This condition can occureven when ambient temperatures are above freezing(32°F), as long as hydrates are present in the gas stream.This freezing typically results in a shifting zero or“wandering” differential. This is often difficult to detecton EGMs unless the data is graphed or carefullyreviewed. Most EGM data will be averaged over dailyperiods, so this differential error can easily be overlooked.This is especially true when there is no checkmeasurement or balancing available.

Not only do physical properties of the gas affectmeasurement, but also flow characteristics. When flowpatterns are steady and pressures are constant, any typeof applicable gas measurement device is capable ofmeasuring gas accurately. Typically, our chart recordercheck meters will run well less than 0.5% of EGMvolumes, even with nominal fluctuations in flow patterns.When flow patterns become erratic, measurementbecomes more difficult. Initially, it is sometimes difficultto properly size the orifice plate when the differential isextremely erratic. Stations on intermitter controlled wellsare examples of this situation. These erratic flow patternsare difficult to measure and even more difficult to auditaccurately. These extremes in flow rates often makerecalculation difficult if not impossible. EGMs normallycalculate volumes once per second, but accumulatevolumes and averages on an hourly or daily basis. Thismakes recalculation of EGM data somewhat inaccurate.The best practice, in every case, is to work towardsdesigns and policies that prevent and eliminate thepossibility of these errors.

The meter tube and orifice plate are absolutely essentialelements. AGA-3 report, as mentioned previously,provides specifications for meter tubes, orifice plates,

as well as piping configurations. This is normally handledduring design and installation, or handled in the field.Meter test reports and change reports should all bereviewed to verify meter tube and orifice plateinformation. It is good practice to inspect the orifice plateand orifice seal ring each meter test. This will help findany mystery plate changes. Often errors occur whenchanges are made. Sustained changes in differential aresuspect when they occur at the time of a meter test orother meter work. Occasionally, the tester or operatormay change the orifice plate without properly reportingit or possibly replace the orifice plate backwards. Orificeplates are installed with bevel downstream. When anorifice plate is installed backwards, the differential willbe reduced. This error is difficult to measure because itdepends on the meter tube, orifice plate size, andapplicable flow rates. Most often, corrections are madeby comparing the error reading with the correctedreading.

Plate changes, as well as all other changes, areautomatically reported in the events report. Use the eventsreport to verify that the orifice plate change was recordedand applied properly. Volume accumulators should bereset any time any change is made, in order to ensurethat correct parameters are used in calculations.

Secondary Element

Always verify the meter station ID. Using incorrect dataor volumes can prove tiresome in the auditing process.The characteristics report defines all parameters usedin the calculation of volumes by the EGM. All of theseparameters must be verified as correct. With EGM, fieldpersonnel are often charged with not only testing, butalso updating plate changes, compositional data, andother calculation parameters. Often, data is enteredincorrectly or omitted entirely. Care must be taken whenheating values (Btu) are entered. The heating valueentered must be as set forth by contract (saturated, dry,or actual) and at the same pressure base as was thevolume computation. It is a common mistake to disregardpressure base for Btu resulting in an error in MMBtu.The Btu is not used in EGM calculations; however, theyare often collected with the rest of the EGM data andimported into reporting systems where MMBtu iscalculated.

EGMs typically have a zero cutoff feature, which is usedto prevent calculation of volume when there are slightzero shifts in transducers. If this option is set too highand low flow conditions exist, it is possible for the meterto ignore this low flow. Conversely, transducer zero shiftabove zero cutoff can result in calculating flow duringno flow conditions. Always verify zero cutoff anddifferential recordings, along with low DP alarms. Over-ranging of the meter is also a severe problem. If thisoccurs, volumes will be understated. Checkmeasurement or estimates must be used. Always checkhigh DP alarms in the alarm summary to verify that thedifferential is not being over-ranged.


Kenneth W. Blackburn

EGM has become very reliable; however, as with anyelectronic device, failures will occur. Transducer ortransmitter failures are usually detected during the metertest or by erroneous readings reported by the EGM.Occasionally, boards or power supplies will fail resultingin missing data. Most EGMs are battery powered withsolar charging systems. With long periods of cloudyweather, units will power down (“sleep”) in order toconserve power. Batteries also malfunction, resulting inloss of power. These situations can usually be detectedby missing data. Occasionally, meters will reporterroneous readings, especially after “sleep” periods, boardmalfunctions, lightning strikes, or other environmentaldamage. In either case, check measurement or estimatedvolumes must be used.

OFFICE ERRORS

Most EGM errors do originate in the field; however,problems do occur in the office. EGM data is collectedin the field or by remote, imported into data handlingsystems, edited, manipulated, recalculated, and sent outin reports. There are many potholes along this highwayof data.

Depending upon company policy, contract requirements,and availability, composite compositional data is often

applied and volumes recalculated in the office. Analysisor composition reports should be reviewed carefully toensure application of correct specific gravity, components,and Btu values.

Estimates are often inserted in the office. Estimatesshould come from check measurement or third parties ifpossible. Company policies regarding volume estimatesvary widely.

Incorrect volume calculations can result fromrecalculating with incorrect or incomplete data. Sinceraw unedited data is the most accurate, data should notbe recalculated unless absolutely necessary.

CONCLUSION

The business of natural gas measurement is truly acomplicated, specialized task. The fact is that mistakeswill be made no matter who is measuring, what is beingmeasured, or how it is being measured. This fact makesauditing absolutely necessary. The auditor must beknowledgeable in all aspects of measurement, from thefield to the office. The obvious goal in natural gasmeasurement should be accuracy and it is up to theauditor to verify the measurement accuracy.


INTERNET BASED MEASUREMENT MONITORING & CONTROLJeffery Perrin

eLynx TechnologiesHouston, Texas

Bringing measurement and control to the Internet givescompanies more cost effective systems for measuring,monitoring and controlling of oil and gas processes.These types of systems help companies with their currentoptimized resources by providing their employees dataanywhere and anytime.

Outsourced Internet based Supervisory Control and DataAcquisition (SCADA) systems have proven to reduce theoverall maintenance cost to companies. Instead ofcompanies relying on proprietary based systems forupgrades and enhancements, the Internet basedsystems are thin client based and rely only on java andactive X based add-ins to a standard browser, i.e. InternetExplorer 6.0. Increases in enterprise wide data has alsoproven to reduce costs due to employee efficiencies ingathering data for compiling reports for marketing andengineering departments.

SCADA HISTORY

Years ago monitoring and control of remote facilitiesmeant crews spending hours driving from one remotefacility to the next, often on a full-time basis. Once aproblem at a facility was found, they fixed it, which oftenmeant making an adjustment to a piece of equipmentsuch as a pump or a valve. One of the first steps forwardwas equipping these crews with radios so as to call intoheadquarters when a problem was located, or to call foradditional crews when needed.

Another step in automating this process was by simplyadding a monitoring and alarm system. These weretypically electro-mechanical devices at remote sites thatwould send a signal back to a control center via radio orleased telephone lines. While these systems were a stepforward, the high cost of the computing technology oftenmade them a tough sell to management on a purely cost/benefit basis. After all, once a problem was reported, acrew still needed to be dispatched to correct thedeficiency.

SCADA AND CONTROL

Most of these systems evolved into SCADA systems withcapabilities that enable remote control of facilities, notjust monitoring. This added ability to control facilities(open/close, on/off, up/down, etc.) within specificparameters via an automated system is when SCADAmade sense to corporate management and not justoperations staff.

Over time the price and size of computing powercontinued to be driven lower, and as computingstandards emerged, the market for SCADA systems grewsignificantly. Now, even small and medium sized oil andgas companies and utilities could afford these systems.Lower computing costs also opened the door to addedfunctionality, such as modeling and analysis software,at a relatively small incremental cost. One key result ofthis phenomenon is that the price for SCADA solutionsis bottoming out, but technical innovation continues ata fairly constant rate.

BENEFITS

Today the SCADA users are looking for dependability,scalability, and flexibility. One avenue for this powerfulmix that has appeared on the horizon is the use of theInternet and Web-based technology. While deliveringSCADA applications via the Internet can ultimately havethe same effect as a traditional rollout, a number ofadvantages exist in considering Web-based SCADAapplications, such as:

• Scalability: SCADA applications accessible ona corporate intranet make the informationavailable to all in the organization that needs thisdata to do their job. By providing this data viathe Internet to their desktop, the cost ofpersonnel doing their job (i.e. time to performtasks) is greatly reduced, opening the door tonew business applications for this data.

Next generation Web-based applications willoffer easy to use comprehensive configurationtools so as to allow users to design their ownWeb screen and applications. Assignedcustomer administrators have full access tothese tools to address their needs without havingto rely on an outside source.

• Cost: The cost of using standard Internettechnologies (PPP, TCP/IP networks, browsers)greatly reduces the investment in proprietaryhardware and software as typically supplied.Web-delivered SCADA can also turn a largecapital investment in communication hardwareand analytical software into an affordablemonthly expense.

• Cost Savings: Some producer companies haveexperienced up to a 25% overall cost reduction


in operating costs compared to otherconventional non-SCADA systems. Althoughthey can not quantify the exact cost savingsconnected to each system benefit derived fromautomation, they do know that most of thesavings come from these improvements in fieldoperations:o Measurement data once made available

across the Internet, offers producersmarketing, production control, and instantaccess to volumes from their wells.

o Prevention of catastrophic equipment failurefrom improved alarm monitoring.

o Reduced fuel consumption from remotecontrolling of compressor speed to maintaindesired suction pressure. Continuousremote control through Internet Web basedsystems prevents the machines fromrunning at higher rates than intended.

o Reduced downtime at field compressorsites. Prompt notification of compressorshutdowns in the field allows for correctiveaction to the site of the problem.

o Many producers experience up to 99.9%run-time out of their compressors with realtime monitoring systems provided bySCADA. Internet access allows more of theEnterprise to be involved with decisionspertaining to compressors, etc.

o Improved scheduling of preventativemaintenance. For both compressors andmetering systems offers continuousmonitoring of compressor valvetemperatures, engine cylinder temperatures,engine air/fuel ratios, fuel consumption ratesand similar operating parameters permitsgraphic display of equipment operatingtrends. Potential problems are identified forscheduled maintenance before a breakdownoccurs that requires emergency measures.

• Faster deployment: The use of industrystandard networking technologies that arefamiliar to a broader group of engineer’s meansit is easier to make in-house modifications anddeployments or find local expertise that can learnthe system.

• Enhanced production: The benefits ofefficiency are nice in terms of cost. But mostproducers find that by far the greatest benefit ofInternet based systems goes to the top line:increased revenue from rising production. For agas producer, the greatest benefit of having ahighly efficient gathering system is reduceddowntime at the metering and compressorstations. When compressors stop, thebackpressure shuts in the wells. With the quickresponse and widespread availability of theInternet, operators can respond to thecompressors problems, thus reduce downtime.

When a production well shuts in due to problemsassociated with the well, operators can bealerted to respond in a timely manner as toreinstate production.

SECURITY

Two types of security are used. First is user authenticationutilizing an assigned user name and password. Secondis data encryption for information transferred utilizing theweb pages.

• User Authentication: Users of the web-basedsystem will be assigned a username andpassword. Each user has specific rightsassigned to view reports, view trends, viewalarms and/or edit alarms. These can be specificto each well location viewed from the web site.

• Data Encryption: Data encryption has becomevery important these days as we continue toincrease the uses and expose more confidentialand valuable data to the Internet. SecureSockets Layer (SSL) protocol has become thestandard in the industry in providing secureinformation transfer through the Internet. SSLprovides a range of security services includingserver authentication, client authentication, dataintegrity and confidentiality. Most data centersuse 128-bit SSL encryption with a digitalcertificate registered through Verisign Inc.

OPERATIONAL ADVANTAGES

Given that delivering SCADA via the Internet istechnologically and economically possible, what are theoperational advantages? Consider for example, aninterstate pipeline system. With facilities equipment andcustody transfer points across literally hundreds of miles,the ability to access and manage real-time systemoperating data and customer data 24/7 is critical.Accomplishing all of this via the Web increases theavailability of this data quickly and cost-effectively tothe complete Enterprise.

A few examples demonstrate the practical applicationof these systems. Consider maintenance crews assignedto a section of pipeline. Having expedited access to anyproblems would save time and money, maintain productflow and ensure maximum revenue. A crew could accessthis information from a desktop at the office or via a hand-held device (such as a PDA) while out in the field. Thisdata can also be managed to send a page or email toon-call workers when operating parameters are outsideof predetermined levels.

Another example is tracking data at custody transferpoints. Web-based SCADA would make this valuablebusiness data available ‘round the clock’ to all requiringthis information for a variety of reporting needs. Thisinformation could also be managed to send messages


to customers or pipeline marketing staff when usagepatterns fall outside of predetermined levels. Additionalservices are available to the end user of the data.Businesses that quantify, scan, and adjust final billingnumbers can replace the usual integration departmentsor outside services. Once the measurement data isavailable on the net, many third party services can beperformed on this data in an expedient manner.

Web-based SCADA applications are a reality today.Vendors have developed a solution that is being deployedin the field across the nation on production locations.Operations and maintenance personnel are using thesystem to save time and money on operating andmaintaining their remote pumping facilities and to keepthe gas flowing at or near capacity, thus impacting thegas company’s top and bottom line performance.Operation staff can access the report any time fromvirtually any location (see Monthly Total Report below).

The ability to view production data in graphic form andset operating parameters can also be executed via theWeb (see Trend Display below).

DATA PRESENTATION

The data will be present via a standard web browserwith HTML developed web pages. These types of pagesare usually developed with standard software likeMicrosoft FrontPage or Adobe Page Mill. These softwaretools create web pages in a graphics environment andoutput HTML.

Java can also be used in the web environment to createapplications that run over the Internet, i.e. historicaltrending application.

The following is a list of operational data that will be madeavailable to the user through the user interface. Each ofthese values are available for display, trending andreporting.

• Meter Name• Date/Time stamp• Instantaneous Flow Rate (MCFD)• Previous Day Accumulated Volume (MSCF)• BTU• Differential Pressure• Static Pressure• Temperature• Battery Voltage

The following is a list of types of web pages available tothe user through the user interface:

• Overview Display (Groups)• Group Summary Display (tabular overview)• Graphic Well Display• Trending Display (per location)• Alarm Display (per location)• Monthly Total Report (per location)

Overview Display (Groups)

Group Summary Display

Graphic Well Display


Trend Display

Alarm Display

Monthly Total Report

CONCLUSION

Internet based measurement and control systems haveproven to provide a solution for more cost sensitivecompanies with personnel operating in multiple locations.

Jeffery Perrin


METHODS OF GATHERING EGM DATADennis Kline

TXU Business Services1601 Bryan St., Suite 26-034, Dallas, TX 75201-3411

INTRODUCTION

As competition, deregulation and economic healthamong utilities becomes more intense, capitalexpenditures and cost reductions become high priorities— yet operational control and system reliability are moreimportant today than ever before. Coupled with rapidlychanging technology, this paradox presents a uniquechallenge for the efficient design and operation of aremote Electronic Gas Measurement (EGM) system.

Capturing remote data and the method of transportingthe data back to a central host essentially falls into twobasic categories, or a combination of these twocategories:

• wireline — such as telephone facilities, copperbased facilities, or fiber optic cable

• wireless — such as satellite, cellular, PCS, or autility owned radio networks such as MAS orSCADA.

In this paper, we will discuss some of the advantages,efficiencies, costs and risks associated with thesetechnologies, and how each can be utilized in ElectronicGas Measurement applications.

WIRELINE TECHNOLOGIES

Copper Plant

Copper infrastructure is widespread, readily available,and easy to commission. It has been estimated that inthe United States alone, over $100 billion has beeninvested in the existing copper infrastructure. Even withaggressive alternative access deployment, most areaswill be copper-served for years to come. In urbaninstallations with minimal data requirements, the initialinstallation costs associated with copper can be far lessthan other technologies. However, for rural installations,special construction charges such as trenching can becost prohibitive.

POTS (Plain Old Telephone Service) dial up technologyis readily understood and has been used for EGMapplications for many years. Little, if any, capital expenseis required. Recurring monthly fees are typically chargedagainst operational costs and one time installation costscan be minimal. In urban areas, a simple facility order tothe incumbent Local Exchange Carrier (LEC) orCompetitive Local Exchange Carrier (CLEC) schedulesboth installation and service.

But even though copper is “everywhere,” it is rapidlyrunning out of both capacity and bandwidth. Copper cancompetently handle bandwidth speeds up to 56 kb/sand Digital Loop Carrier (DLC) technology can providetemporary relief from congestion. Yet the demand forhigher connectivity is dramatically increasing, with theimbedded copper cable plant becoming the bottleneck.Coupled with old and aging cables, copper may berapidly reaching capacity, bandwidth limitations andreliability.

Often times, remote gas measurement locations are ininaccessible environments. As such, facility costs,installation costs, reliability and mean time to repairbecome significantly more critical. Rural installations mayinvolve multiple LECs, CLECs and Inter-ExchangeCarriers (IXCs), and may include significant specialconstruction installation charges. In rural applications,where multiple LECs are often involved, the “Mom andPop” telephone companies can become your worstmaintenance nightmare.

Copper facilities typically represent the area where theutility has the least amount of control, highest recurringcosts, and least reliability. The most frequent cause ofservice disruption in a facility based network is notequipment failures — it is leased line facility outages!While cable cuts are arguably the most disruptive, otherfactors, such as wet cables, human error, equipmentproblems, or dribbling data errors can all be equallytroublesome and difficult to correct. Resolving suchimpairments can be time consuming, even in majormetropolitan areas where facility maintenance is best.

xDSL Vegetable Soup

From the general consumer perspective, xDSLtechnology is certainly the most widely recognized, withAsynchronous Digital Subscriber Line (ADSL) being themost popular. ADSL technology can typically offer datarates up to 512 kb/s, with subscription costs of $50-$100 per month. Higher bandwidths can be attained,but with significant distance limitations.

Execution and deployment, however, has not been easy.While some market projections indicate that over 70%of existing worldwide telephone lines are ADSL capable,the reality is that older copper infrastructure, bridge taps,incomplete cable records, impedance mismatches,excessive cable distances from the central office, andmaintenance issues have all contributed to bothcustomer and carrier frustrations.


For Gas Measurement applications, xDSL may be aviable option in major metropolitan that are close to theserving telephone offices, where high data rates may bedesired, or for locations where multiple EGM sites hubtogether. In rural applications, where remote gaslocations are often located, xDSL is not feasible becauseof distance limitations from the central office.

Fiber Optics

When compared to other technologies, fiber offersdistinct advantages, such as high bandwidth,insignificant distance limitations, low monthly recurringcosts for existing terminations, and no blind spots incoverage. Additionally, with an excess of available fiberin many cities around the country, the monthly leasingcosts for existing fiber terminations will continue to erode.

While fiber offers extremely high bandwidth andpotentially low monthly leasing costs, it is estimated thatless than 5% of commercial sites are actually touchedby fiber. Fiber may run down the middle of the right-of-way, but the actual entrance link into a specificcommercial location may not exist. Fiber is more suitablefor large industrial sites or dense campus locationsrequiring significant data throughput, such as hublocations used for backhaul. Though the cost of fibertermination equipment has gone down, the construction,trenching, civil engineering, right-of-way, and labor costshave all increased with inflation — making the cost ofextending fiber for most gas measurement applicationsprohibitive.

Hybrid Fiber Coax (Cable Modems)

With widely deployed cable systems already in place andhigh penetration into the urban marketplace, cable TVsystems offer the potential for high bandwidth at anaffordable price. Data rates of 1.544 Mb/s can be readilyaccommodated, with monthly fees in the $50 range.However, in the US, most cable television systems areanalog one-way systems serving the residential market,not businesses applications. Upgrading the analoginfrastructure to digital represents a major capitalinvestment to cable operators who are already facingstiff competition from the home satellite industry.

Security and privacy issues are also concerns, as cablesystems are essentially a “party line”. Residential andbusiness subscribers are on the same cable, havingpotential access to all traffic, regardless of the owner.Individual connections are maintained by using differenttime slots on the cable, but it would not be technicallydifficult to break into this rudimentary encoding. Privacyissues, such as credit card numbers and eavesdropping,require more powerful encryption methods.

Equally daunting, the cable industry is typically workingin a regulated monopoly environment. Competingtechnologies, particularly wireless technologies, are mostoften not burdened by local, municipal, or state regulatory

agencies. The cable industry must change from aregulated to an entrepreneurial technology, and operatorswill be challenged to overcome the public’s generalperception of poor customer service, as well as anexisting employee base somewhat inexperienced intelephony and data applications.

As with fiber and xDSL, Hybrid Fiber Coax solutions aretechnically and economically feasible only within majormetropolitan areas, and generally not suitable for mostgas measurement applications.

WIRELESS TECHNOLOGIES

Rising wireline facility costs coupled with the everincreasing desire for higher data speeds has necessitatedthat many utilities investigate ways to reduce existingmonthly operating costs, improve network reliability,accommodate higher data rates, and better manageinstallation schedules

Management and control of facilities in a wireless networkmust address the real-time operational elements inproviding reliable service. The principal objectives indeploying a wireless network is to provide maximum up-time, proactive restoration of network impairments andfailures, self-control of the network, and elimination ofmonthly recurring costs.

Wireless networks for gas measurement essentially fallinto two categories:

• public networks — such as Sprint PCS, AT&TWireless, Verizon Wireless, etc.

• private networks — such as utility ownedmicrowave systems, Multiple Address Systems,or SCADA.

Public Networks (Wireless Alphabet Soup )

When considering the use of PCS or Cellular systemsfor gas measurement applications, monthly subscriptionrates, footprint coverage, modem costs, and the cost ofair time all become distinguishing factors. Equallyimportant, as parity among service providers levels out,somewhat less intangible factors include: customerservice, billing accuracy, network reliability, and migrationpaths towards next generation technology.

Analog cellular technology, such as Cellular Digital PacketData (CDPD), has been used in gas measurementapplications for many years. CDPD can be economicaland somewhat easy to deploy. In many cases, a modemboard mounted in a weatherproof box with a small“button” antenna is all that is needed. While CDPD is apacketsized data technology, it resides on an analogoverlay and is therefore indelibly tied to the analog cellularnetwork.

Despite the sales and marketing pitch a local carrier maypresent, all analog technology (which includes CDPD) is


being rapidly phased out. Wireless carriers are in theprocess of implementing the next generation digitalnetworks towards an ultimate goal, which is 3G (ThirdGeneration) technology. Market data shows that 3G,which offers high speed data throughput, hassubstantially more profit margin than the voice and dataofferings available today. Thus, regardless of technologyplatform, all carriers will eventually migrate to a 3Gnetwork.

For gas measurement, this creates an enormouschallenge in developing a long term strategy and budgetAs each carrier migrates towards 3G, many questionsare left unanswered. Should a utility deploy a TDMAsolution or would a CDMA solution be more prudent ?As public carriers drive towards 3G, will the cost of eachmigration step be passed on to the end users? Will theequipment purchase today be obsolete after thetransition process? Will the coverage the carrier offerstoday be the same footprint after 3G is deployed? A longterm, carefully thought out business plan should be madebefore utilizing a public wireless carrier for EGMapplications.

Utilities should be cautious of the long term viability ofTDMA (Time Division Multiple Access). An increasingnumber of equipment manufacturers and cellular carriershave already announced their intentions to eventuallymigrate away from TDMA as they plan for 3G. Themigration path for TDMA based carriers might be: CDPD(still offered today); Global System for MobileCommunications (GSM); General Packet Radios Service(GPRS); then Enhanced Data Rates for Global Evolution(EDGE); and finally Wideband CDMA (W-CDMA). Not allof these technologies are backwards compatible -meaning that with each migration step the customer (ie:the utility) will have to retrofit or replace existing modems.Even if carriers offer some form of negotiatedconcessions (such as modem trade-in plans), the utilitywill still experience equipment and installation expenses.

While existing CDMA based carriers have a lessdisruptive 3G migration path, it will not be totally painfree. CDMA migration will typically consist of IS-95,1XRTT (already available in most locations today) andthen directly to CDMA2000. The good news is that alldevices are expected to be backward compatible. Assuch, CDMA migration is expected to be significantlyless disruptive than TDMA, and should be less costly forthe end user.

EGM applications deployed today typically require lowdata rates. In a dial up or CDPD application, it only takesa few minutes to capture and transport metered databack to the central processing point. When using CDPD,the utility usually pays a flat monthly fee for sufficient airtime to meet metering requirements. However, the highdata rates proposed by 3G represent a two-edged sword.While everyone likes the idea of high data throughput,there is a definitive cost associated with this newtechnology. The days of “all you can eat” air time for a

flat monthly rate are quickly disappearing from majorcarrier rate plans. As carriers implement 3G, rate planswill evolve with much higher minimum data requirementsas well as costs. Are significantly higher data rates reallynecessary for gas measurement applications? Are utilitieswilling to pay substantially higher monthly fees just tohave the ability to send data at 384 kb/s or 2 mb/s, whenthe actual requirement is much lower?

Unfortunately, unless a utility negotiates a strategicpartnership with a carrier, the cost of 3G service may beoverly prohibitive. The minimum data usage for 3Gservices may start far above what gas measurementtypically requires. Below are two examples of air timerates recently quoted by a major carrier (not includingthe cost of the modem):

• CDPD $8 per month, plus $0.05 per kb of data• GPRS $20 per month for 3 mb of data

With these prices, CDPD (which is going away) may becost effective, but GPRS (the replacement technology)is substantially more expensive, and a utility would payfor 3 mb of data that it would never need or use for anindividual meter. Possibly by bundling all measurementrequirements in one “bucket”, the aggregate amount ofall metering devises together may justify the cost ofservice. Carriers will undoubtedly segregate voice anddata usage separately in an attempt to increase revenue.By bundling all wireless needs under one Corporatecontract, the utility may be able to lower costs to anacceptable level. Or, perhaps as an incentive, a utilitycould negotiate for the carrier to subsidize the cost ofdata modems, such as they do today with cell phones.

Without question, when it comes to using public carriers,every utility will be faced with tough economic decisions.Prudent cost management will be critical for gasmeasurement systems in this environment.

Short Message Service (SMS)

Short message text service is packetized data sent overthe public carrier network. It is the text messaging thatyou see on your PCS cell phone, and can have 1-way or2-way capability. SMS uses only short bursts of data(typically from 160 characters to about 255 characters,depending upon the carrier). SMS text messages arenot sent on the same channels as voice or even data,but ride the control channel, which limits the availablecapacity. Typical latency delay is about 4 seconds fromthe time a message is sent to the time it is received, butthis is dependant upon available capacity on the controlchannel at the time the message is sent. Some carriersoffer “store and forward” capability, which means thatthe system will hold an undelivered message for severaldays until retrieved. SMS service can be more expensivethat CDPD, and most carriers only offer SMS as anadditional feature on top of other rate plans. As such, autility may not be able to negotiate a favorable rate planfor stand alone SMS service. As carriers migrate from


one technology to the next in an effort to reach 3G,customers will face many of the same upgrade issuesnoted earlier.

Satellites — ( LEOs, MEOs and GEOs )

Satellite technology has the ability to offer largefootprints, but with trade offs in high operating cost andlatency. Geostationary Earth Orbiting Satellites (GEOS)are parked at a specific point along the Equator and aresynchronized to track the Earth’s rotation. While GEOSoffer large coverage zones, by their very nature they mustmaintain a high altitude stationary orbit (about 25,000miles above the earth) which causes an inherent delayin the signal and is intolerable for high-speed data. Thisinherent delay, called latency, is approximately 250milliseconds roundtrip. For voice applications, using echocancellers can compensate for this delay. Dataapplications, however, suffer dramatically. For example,Transmission Control Protocol (TCP), which is widelyused for Internet data transfers such as Web files andemail, does not respond well to high latency links andvariable round trip times. When utilized over linksexhibiting these characteristics, the result is generallylow throughput and inefficient use of bandwidth.

Low Earth Orbiting Satellites (LEOS) are only a fewhundred miles above the earth. While LEOS have minimaldelay characteristics because they are not in asynchronous orbit (about 110 milliseconds roundtrip),there must be a fleet of satellites in simultaneous orbitto ensure connectivity. This fleet of satellites increasesthe overall operational costs of the satellite system, whichin turn is passed on to the individual users (ie: the utility).

Medium Earth Orbiting Satellites(MEOS) are locatedbetween 6,000 miles and 10,000 miles above the earth,and exhibit characteristics intermediate between GEOSand LEOS — that is, modest latency but higheroperational costs.

Some data oriented satellite systems provide only 1-waydata transmission; others provide bidirectionalcommunications with either symmetric or asymmetricalbandwidth. For extremely remote EGM locations, satellitetechnology can provide a viable means of retrievingmetered data, but not without a price. The cost of terminalequipment can approach $2,000, with minimum monthlyair time charges starting at $20 per month, making itcost prohibitive for many metering applications.

Multiple Address Systems (MAS)

MAS has been widely used by utilities for many years, isa field proven technology, and is readily understood.MAS is often deployed in a hierarchical topology - thatis, a remote station sends data back to a nearby Sub-Master, who in turn sends the data to a Master Station,who in turn sends the data back to the main host orcentral collection point. A typical system might include:

• Remote Terminal Units (RTUs) — located at themetered point, the remote terminal collects datainformation from the metered device, which isthen transported back to a Sub-Master or MasterStation

• Sub-Master — The Sub-Master polls andcollects data from surrounding RTUs. It will thenpass this data back to the Master Station.Because the Sub-Master talks to manysurrounding RTUs, it requires an omni-directionalantenna and a directional antenna (yagi) tocommunicate with the Master Station.

• Master — The Master station communicateswith surrounding Sub-Master stations. It mayalso poll RTUs in close vicinity to the Master.When a Master polls both Sub-Master stationsand local RTUs, it will typically do so with thesame omni-directional antenna. Thus, a Masterstation may have only one omni-directionalantenna for multi-purpose applications. Thecollected data is sent back to the centralizeddata host via telco land lines, microwave radio,Internet, Satellite or other means.

Each MAS radio has a unique address, which allows datato be polled for each individual RTU /meter. Polling datafor one individual device takes only a few milliseconds,but a dozen or so RTUs may take up to one minute. Theentire system polling time (sometimes referred to aspolling latency or delay) depends upon how many uniqueaddress are assigned throughout the entire system.Theoretically, the number of remote stationscommunicating with one Master station is unlimited.However, in real life applications, the amount ofacceptable delay to poll an entire system is often thelimiting factor.

Multiple Address Systems are fairly inexpensive to deployIt is not uncommon for the remote, sub-master andmaster radios to all be the same type and model of radios.This allows for common spares and easy maintenance.MAS radios can be purchased in the $500 to $1,000price range, and come in numerous flavors and optionsto enhance reliability, performance, maintenance andcosts.

MAS radios are available in several frequencies, bothlicensed and unlicensed, but all require near line of sight.In the Untied States, the frequency bands typically usedare:

• 450 MHz — because this band requires a FCClicense, operation is protected from interference.This band is the most “forgiving”, in that trueline-of-sight is not always mandatory. Slight pathobstructions and blockage can occur while stillmaintaining good signal continuity.


• 2.4 GHz Spread Spectrum — this band is“license free” and does not require an FCClicense. However, this band is subject tointerference from other communications devicesand should be used with great caution. In rurallocations, interference issues may be minimal.For metropolitan areas the potential forinterference significantly increases. Otherdevices operating in this same band includemicrowave ovens, cordless telephones, garagedoor openers, wireless LANs, laptop modems,etc. If interference should be encountered, thereare limited options available to mitigate theinterference and continue operating as normal.In most cases, the presence of stronginterference will mean that the site cannotcontinue to be serviced by unlicensed radios.

• 900 MHz — this band is available in licensedand unlicensed frequencies. True line of sight isalmost always required for stable operation. Thepotential for interference noted for 2.4 GHz aswell as the protection from interference notedat 450 MHz also apply.

Many utilities already have existing Multiple AddressSystems in place. When deploying EGM applications,careful consideration should be given to utilizing thesesystems. MAS typically offer robust response times, easyinstallation, easy expansion, and in the case of licensed

Dennis Kline

systems, protection from interference. MAS has little orno costs beyond the initial equipment deployment. Oncein operation, there is no further expense beyond periodicmaintenance. Other solutions, both wireline and wireless,often involve equipment and /or installation costs PLUSmonthly recurring costs. A Net Present Value projectioncan demonstrate that MAS can easily have a paybackof 12 -18 months.

SUMMARY

An investigation of these different technologies showsthat each has opportunities, challenges, advantages, anddisadvantages. Figure 1 compares a real-life examplefor cost per monitored point verses technology and time.Solutions considered for this application were: CDPD;Satellite; Telco leased lines; and MAS. In this example,installation costs were assumed to be similar for allsolutions. As shown. the cost and payback of a MultipleAddress System is 18 months.

Deployment decisions for EGM applications will includemany drivers, such as: initial equipment costs, monthlyrecurring charges, capital constraints, life expectancy,reliability, and availability. This is particularly true in theWireless Public Carrier domain, where technologydecisions made today will have significant implicationsfor the future. Utilities need to avoid locking themselvesinto a wireless platform that has a limited life expectancyor that will require painful and costly upgrades.

FIGURE 1. Cost Per Monitor Point

1st day 5 years18 months 3 years

Commercial CDPD Satellite Modem Dial-Up Leased Lines Multiple Address System

american gas

Documents

gas chromatograph

daniel measurement

orifice measurements

coastal flow measurement

b10 gas flow conditioningpage

school dates

annual school exhibitseptember

field determination