a practical approach to modelling lng train design for ... daniel documents/… · project thereby...

A Practical Approach to Modelling LNG Train Design for Minimizing Measurement & Allocation Uncertainty page 1DANIEL MEASUREMENT AND CONTROL WHITE PAPER

ABSTRACTOur ability to measure accurately has formed the basis of trade in all industries. As a consequence, the performance of measurement systems will affect the profitability of any business. In the Oil and Gas Industry measurement systems are used for fiscal and custody transfer purposes, for process plant operations and for plant efficiency monitoring. It is of great importance therefore that these systems are engineered, operated and maintained to industry standards and within equipment specifications.

This paper highlights the implications of measurement uncertainties on LNG product allocations and how, through a study of these uncertainties and the implementation of a Measurement Upgrade Enhancement Project, the final Allocation Uncertainty was more than halved. The study addressed each of the metering elements associated with the allocation process and identified which of these had the greatest impact (and/or required attention). The measurement upgrade addressed issues affecting key meter performance such as flow profile effects and valve noise with respect to ultrasonic measurement.

A major factor in the success of the project was the use of a Measurement Exposure Model (MEM) which determined the measurement uncertainties associated with key metering points, and the impact they had on the final LNG Product Allocation. By using the MEM to focus on the effect of these measurement uncertainties, unnecessary modifications/upgrades were avoided and savings made in the overall project.

INTRODUCTION1 MEASUREMENT AND ALLOCATIONThe following are basic questions that should be asked of any LNG Production facility:

• What does X% uncertainty of LNG measurement mean and how can this be improved?

• To what extent do the plant measurement systems affect the LNG product allocation?

• What is the minimum acceptable financial risk to Joint Venture (JV) partners on an LNG complex and how can this be improved?

These are some of the issues which have to be addressed

among JV partners for equity on a multi-train LNG complex. This situation can arise if the ownership among LNG trains and the commercial terms governing their operation are different. Thus a single LNG facility using shared resources where applicable, may have multiple trains (plants) where there can be several owners and several feed gas streams, as is the case at Atlantic LNG Company of Trinidad & Tobago.

The problem of accurately determining the LNG production of a train arises because of the following:1. The absence of reliable LNG flow measurement at the

outlet of the trains.2. Shared LNG storage at the facility.3. Recycling of LNG vapours from tank storage into the

production streams.

The first constraint is dictated by the manufacturers’ products. Reliable dynamic LNG measurement technology is actively being researched by a number of flow measurement vendors with the main challenges being the ability to accurately measure LNG leaving the trains and the ability to prove such a meter. In addition, there are no dynamic LNG flow measurement standards.

The second constraint is from a standpoint of practicality. It is good to have the option to pump LNG production to several tanks. With the cost of an LNG storage tank being in the vicinity of 250-300 million USD it is prudent to share this facility since this would strengthen the project economics. However, this complicates ownership traceability where LNG production is commingled across several storage tanks.

The third constraint is the recycling of the LNG vapours from storage, loading and LNG cooldown to production facilities means that shared production from one train/owner/shipper can be allocated to another.

As a result of the above issues, the allocation process used to determine the LNG production of a train on a multi-train facility can be very complex. This product allocation will be governed by the commercial terms and will be largely a function of mass, energy or standard volume balanced equations, for example,

www.daniel.com

A Practical Approach to Modelling LNG Train Design for Minimizing Measurement & Allocation Uncertainty

DANIEL MEASUREMENT AND CONTROL WHITE PAPER

Train Production = Inlet - Fuel Consumption-NGL Production - Train LossesIt is worth noting that static custody transfer measurement, as outlined by The International Group of LNG Importers (GIIGNL), takes place on LNG tankers at an uncertainty of ±0.2-0.3%, however what about dynamic LNG train production?

The key question then becomes, what are the uncertainties associated with these measurement points and what impact does it have on the LNG product allocation?

This paper highlights the implications of measurement uncertainties on LNG product allocations and how the use of a Measurement Exposure Model (MEM) assisted in the implementation of a Measurement Upgrade Enhancement Project thereby reducing the LNG and NGL product allocation uncertainty by an average of 2 %. In executing this measurement upgrade, the issue of valve noise and ultrasonic measurement (inlet feed gas meters) was also addressed since the correct installation and operation of inlet feed gas meters had the greatest impact on the LNG Allocation.

2 BUILDING A MEASUREMENT EXPOSURE MODEL (MEM)A Measurement Exposure Model (MEM) is a mathematical tool that monitors the facility’s measurement points and product allocation equations to determine the individual and/or collective impact of each measurement point on the allocation of products.

In order to build the MEM, an accurate assessment had to be made of the measurement systems associated with the Allocation System. This was accomplished by completing a comprehensive Audit of the metering elements feeding into the Allocation System for the LNG Pant and assessing their respective measurement uncertainties. The current operating Measurement Uncertainties of meters used for Custody and Allocation purposes were then used as an input by the Model (the MEM) to establish the final Allocation Uncertainties in order to quantify the financial “risk” to partners.

3 MEASUREMENT UNCERTAINTY ASSESSMENTIt should be appreciated that no measurement can be absolutely precise, since there are inevitable biases/inaccuracies introduced as a result of the measurement instruments chosen, their calibration and installation. Uncertainty then is an estimate of the limits to which one can expect an error to go, under a given set of conditions as part of the measurement process. Whilst the determination of measurement uncertainty is independent of the MEM, it forms an essential input into the Model, if an accurate assessment of an existing Allocation Systems Performance is required. All relevant Measurement systems are independently assessed and their Measurement Uncertainty established using proven Calculation Techniques (API and ISO) complying with the relevant Measurement System Guidelines for Design, Operation and Maintenance.

Figure 1 - Typical MEM Mimic Screen

A Practical Approach to Modelling LNG Train Design for Minimizing Measurement & Allocation Uncertainty page 3

The calculation process requires such information as:• Calibration Procedures, frequency and tolerance• Primary Measurement Element Uncertainties (Orifice,

Turbine Meter, Coriolis, USM, Etc.)

• Secondary Instrument Uncertainties (DP, Pressure, Temperature, Densitometer, Flow Computer, etc.)

• Correction Algorithms, Process Simulation Assumptions, etc.

• Laboratory Analysis Uncertainties

This information is used to calculate the individual Measurement System Uncertainties and their consequent contribution to the

overall LNG Train Metering Allocation Uncertainty (calculated separately by the MEM).Typical Custody & Allocation Measurement Systems include:

• Gas Pipeline Meters (Custody)

Figure 2 - Typical Input Screens for Uncertainty Assessment

Figure 3 - Typical Output Screens of the Uncertainty Assessment


• Train Inlet Meters (TIM) (Allocation)• Train Liquid Output Meters (Allocation)• Train Output Meters (Allocation)• Storage Tank Measurements (Stock/Allocation)• Fuel Gas Meters (Allocation• Flare Gas Meters (Allocation)• Gas Recycle Meters - where fitted (Allocation)• Tanker Loading Measurements (Custody)

Using Figure 3 Output Data, the MEM user can enter the estimated measurement uncertainty relating to the metering element Operating Condition, thus enabling the Model to simulate most working conditions and reflect the additional uncertainty in Allocation for those conditions. This can be particularly helpful where the operator needs to be cognisant of the effect that instrument malfunctions or other failures may have on the final Allocation uncertainty for the system.

4 MEASUREMENT EXPOSURE MODEL (MEM)The Measurement Uncertainty Assessment forms an essential input to the MEM. In cases where the LNG Plant is not yet designed, design estimates of measurement uncertainty can be entered into the MEM such that the Design Engineers can accurately assess their impact on the Plant Allocation. This aspect of the MEM is not to be underestimated, as it could save considerable Capital Expenditure on unnecessary design for uncritical elements of the Plant Measurement System. In the

case of an existing LNG Plant, weaknesses in the measurement systems provided can be identified, thus providing substantial justification for change, in terms of misplaced Allocation Revenue.

Typically the MEM attributes measurement uncertainties to each of the measurement points, then calculates the flow weighted impact these uncertainties will have on Plant Balance and Allocation. These calculations can be based on estimated (or measured) flow rates through the various meters. By using this technique, the MEM can

identify any “high risk” metering elements in the system and establish the overall Allocation Uncertainties.

The results are typically presented as follows:

• Measurement Uncertainty of each measurement point expressed as a percentage and in terms of product

• Measurement Uncertainty associated with the Plant Balance expressed as a percentage and in terms of product

• Measurement Uncertainty associated with the Plant Allocation expressed as a percentage and in terms of product

• Cost based Risk Analysis to contributing Fields and Owners for various operating scenario’s

Thus, once the initial uncertainty analysis has been performed on each metering system to establish the Allocation measurement point Inputs, then the MEM analysis can be performed to asses the sensitivity of these points on the Plant Balance & Allocation and the results documented.

Any trouble spots highlighted can be re-assessed in order to maximize the Plants performance. In this way, attention need only be focused on those elements that most affect the system and the reduced risk associated with the changes can be identified by the model.

Figure 4 - Typical Input & Output Screens of MEM


As can be seen from the Mimic illustrated in Section 2.0, the whole Allocation Process depended on the Train Inlet Meters (TIM) to enable back allocation of all the Atlantic LNG Plant OUTPUT products to the Gas Delivery Points (GDP), it follows that their measurement accuracy is of prime importance. The MEM demonstrated the cost of the measurement uncertainty to the Operators and their Partners by reference to a live Mimic responding to reported Allocation data.

5 MEASUREMENT PROJECTS JUSTIFICATIONThe objectives of Capital expenditure plant projects can be generally categorized into two main areas:

• Maintaining and Sustaining Production (Reliability Projects): Major upgrades and/or overhauls to plant equipment aimed at improving the cost of operation and for mitigating against equipment obsolescence.

• Throughput and Yield Projects (Optimization Projects): Equipment enhancements and upgrades geared towards increasing the production capability of the asset.

With measurement projects both of the objectives can be achieved.

As shown in the MEM, the single biggest measurement point is the inlet measurement to each of the LNG trains. In general, although measurement projects may not generally increase production it is the perception of the production that is being addressed. This perception translates into financial exposure risk for all parties. As can be seen from the MEM an improvement in the train inlet measurement uncertainty from 5% to 1% translates to an improvement in Allocation Uncertainty of 2% or in other words a financial exposure benefit of 2% to the shareholders.

Therefore, using traditional techniques for project justifications such as NPV, Pay Pack and IRR measurement projects may be justified on this basis.

As in the case of the Train Inlet Measurement Upgrade project assuming that a 4% reduction in measurement uncertainty can be practically achieved then this project should be evaluated in relation to other initiatives competing for financial resources.For example, assuming that three days of downtime on a facility are required for the measurement upgrade and the total cost of the project is 1.2 million USD, for an LNG complex with annual production target of 576 TBtu this gives an NPV (financial risk) of 444 million USD and a payback in less than 1 year with a profitability index of 17.

Assumptions: Gas Price $ 5.00 per MMBtu,Discount Factor of 10%,Capital Outlay = Loss production + Cost of project= $25.2MMProject Life cycle 20 years.

Of interest to note, is that once other projects are undertaken which increases the 576 TBtu for the facility then this project financing can be further enhanced, therefore giving economic benefits to perpetuity. Therefore it is imperative that on an LNG complex inlet measurement uncertainty be optimized.

6 TRAIN INLET MEASUREMENTAs shown from the MEM an improvement in inlet measurement meant a significant reduction in LNG allocation product uncertainty. The challenge for the project team would be to improve the Train inlet measurement to 1% or better thus improving the allocation uncertainty by 2%. The project team chose to improve on the existing single path ultrasonic meters by replacing them with multi path ultrasonic meters. Since the inlet measurement to the LNG train significantly impacted on the apportioned LNG, it had to be as close as possible to custody metering systems.

Ultrasonic meters (USM) were selected as the preferred flow measurement technology for the following reasons:

1. Endorsed by various regulatory bodies, standards and codes for custody measurement, e.g. AGA Report 9, UK Department of Trade & Industry’s Petroleum Measurement Guidelines and the Norwegian Petroleum Directorate.

2. High accuracy can be achieved in the field when properly installed and maintained.

3. High turndown ratios available for individual meters.4. Minimum pressure drop.5. Strong diagnostic built in capability.6. Low maintenance as compared to other technologies e.g.

Orifice or Turbine meters.

However, the USM can be susceptible to noise generated by control valves. This noise, typically referred to as white noise affects the performance of the meter. Also, the level of noise generated by a control valve is a function of the flow-rate, differential pressure drop, pressure and valve trim characteristics.

7 USM NOISE REDUCTION STRATEGIESMost control valve manufacturers achieve their low noise levels by pushing their audible (20Hz-20kHz) emissions out of the audible range and into a range (20-200kHz) which typically affect USM detection signals. This was illustrated during our


flow calibration of the meters where a control valve similar to one used at the inlet of the facility was installed to determine the extent of noise generated, and to see how the meter tuning (noise reduction techniques) would impact the performance.

To combat noise, most USM manufacturers will have a combination or all of the following noise reduction technologies embedded in the meter’s electronics:

• Signal Stacking• Digital Filtering• Correlation

Techniques• Statistical

Methods

Additionally installing blind tees on the piping system, locating the meters far from noise generating equipment and installation of silencers downstream will mitigate the impact of noise on the meter.

8 FLOW TESTING RESULTSAs a requirement by AGA Report 9, USM being used for custody transfer should be calibrated at a Flow Test Facility. Although the Train Inlet Meters (TIM) were not used for custody transfer they have significant impact to the allocation of LNG on a multi-train complex, thus the meters were calibrated by placing a laboratory Flow Standard Meter Bank in line with the USM being calibrated. The Flow Standard Meter Bank being used to accurately measure flow, had been calibrated using standards that are traceable to NIST.

The newly purchased 24” 4 path USM were calibrated at the following points as recommended by AGA Report 9. qmin, 0.10qmax, 0.25qmex, 0.40qmax, 0.70qmax, and qmax, together with two additional points. The meter’s performance was

checked with the control valve at 5 times the normal pressure drop in operation. The following is the sequence of the testing performed and results from Colorado Engineering Experiment Station Inc (CEESI):

Calibration of the meter using the Flow standard as a referenced

Figure 5 - Initial Calibration Results with USM and Flow Standard

Figure 6 – USM Flow Testing at CEESI


in Fig. 5 below shows the results (out of the box meter) in relation to the flow standard. A meter factor per test point is applied as a piece-wise (multi-point) linearization.

The above picture (Fig. 7) shows the valve and meter being

prepared for noise verification checks by comparing the performance of the meters with the Flow Standard Meter while throttling the control valve at various differential pressures. Fig.

8 shows the USM signal detection waveform with no noise present. Fig. 9 shows the USM signal waveform distorted by noise generated by the control valve. Fig. 10 and Fig. 11 show the application of Stacking and Digital Filtering (Noise reduction

Figure 7 - USM and Control Valve at Flow Test Facility (CEESI)

Figure 8 - Typical detection used by the USM for flow calculations(No Noise)


techniques) respectively which were required to achieve a good reading on the USM.

With Stacking Applied only, the USM was still unable to register a good reading.

Figure 9 - Noise as seen by the USM (Noise distorts the USM detection waveform)

Figure 10 - USM Noise seen with the Application of Stacking


The illustration above (Fig.12) shows the improvement of the USM Calibration against the Flow Standard Meter Bank when the control valve was throttled at various differential pressures.

9 PERFORMANCE OF USM VS CUSTODY ORIFICE MEASUREMENT SYSTEMAfter the flow calibration of the ultrasonic meters at CEESI the

next step was to execute the installation and commissioning of the meters. The meters were shipped and stored until an opportune time for installation. Having the control valve available at CEESI proved invaluable, since the noise characteristics generated by this particular valve were analyzed and this significantly minimized the commissioning time. Furthermore this addressed the concern as to whether the valve and meter in close proximity

Figure 11 - USM with Digital Filtering and Signal StackingThe USM was able to register good readings

(Calibration Results)Figure 12 – USM Calibration with Flow Standard and Control valve


could fulfil plant flow regulation and product allocation. It must be noted however, that in some of the installations, further noise tuning (custom data filtering) was required. This was readily observed on installations where the pressure drop was significantly greater than that tested at the facility.

The diagram shows that percentage deviation between the USM and the orifice measurement system is of 0.16%-0.6% with the USM reading marginally higher than the orifice measurement.

10 OTHER LNG TRAIN MEASUREMENT ISSUESAlthough, the inlet measurement had the single largest impact on the LNG train there are other areas where measurement improvements can be made to the facility. However, these must be done at the design stage thereby improving on the overall allocation process. These areas include:

1. Fuel Measurement. Installing a good single point of measurement for fuel enhances the quality of measurement that can be installed and simplifies the number of allocation points. For example on a facility with 3 LNG trains, having 1 meter per train for the fuel can replace 10 or more metering points per train. This improvement reduces the cost of operation and maintenance.

2. Marine Flare Measurement. There are a number of practical challenges with installing good flare measurement during plant operation (availability issues). Having marine flare measurement implies that losses can be applied to individual shippers accordingly if agreed commercially.

3. Vapour recovery measurement will aid in the traceability of LNG production per train.

11 MAINTENANCE OF MEASUREMENT AND ALLO-CATION UNCERTAINTY

With the inlet measurement being so critical to the allocation of LNG and NGL on a multi train complex, this point of measurement must be effectively maintained. Our experience to date has shown that the USM probes must be kept clean for good meter performance.

The piping system should be designed such that there is no accumulation of liquid upstream of the USM. Additionally, if the USM are in series with an orifice measurement system (custody), the similarity of the analytical data in the USM and the custody is of great importance to the flow registration alignment. Operating the orifice custody metering skids at higher differentials per stream also helps with the alignment of the readings registered by both systems.

The USM should also be installed with double block and bleed isolation valves so that the meter probes can be easily accessed for maintenance. Effective isolation also facilitates the easy removal of the meters for re-calibration exercises, typically every six years.

Maintenance of secondary instrumentation, temperature and pressure transmitters must also be done at regular intervals.

Figure 13 - The performance of the ultrasonic meters and the custody metering system on the facility


12 CONCLUSIONMaintenance of measurement uncertainty greatly affects the allocation of products on a multi train LNG facility. It is therefore important that all plant measurement systems be optimized for minimum uncertainty. Building of a MEM can help technical personnel on the facility quickly diagnose mis-measurement issues and plan for future capital expenditure. Design engineers would also benefit from using the MEM at the conceptual stage of a project to plan out the specifications for the measurement system.

Train inlet measurement, which is of great concern to all parties, must be properly designed, operated and maintained to ensure equity in the allocation. This paper highlighted the issues on inlet USM measurement and control valves. Although USM have gained custody appeal, their use must be planned and coordinated with the flow/pressure regulation valve into the facility, since USM are susceptible to noise induced by valves.

Ronald RobertsInstrumentation & Controls Engineer IIAtlantic LNG Company of Trinidad & TobagoPoint Fortin Trinidad & [email protected]

Justin WalterSenior Measurement ConsultantMetco Services LimitedAberdeen, [email protected]

REFERENCES:Allen Fagerlund et al, ‘Identification and Prediction of Piping System Noise,’ Noise Conference Oct 2005.

Bill Johansen & Joel Clancy, CEESI ‘Flow Calibrating Ultrasonic Gas Meters’ International School of Hydrocarbon Measurement May 2003

Charles Derr, Daniel Measurement and Control ‘Energy Measurement using Ultrasonic Meters & Gas Chromatography’ International School of Hydrocarbon Measurement May 2003

Gerrit Vermeiren and Sven Lataire, SGS ‘How Accurate is the Shipboard Custody Transfer Measurement system?’ LNG Journal July/August 2005

James E Gallagher, Savant Measurement Corporation, ‘Orifice Flowmeters and the Estimated Uncertainties in Natural Gas Service,’ International School of Hydrocarbon Measurement May 2003

Kevin Warner and Klaus Zanker, Daniel Industries, Inc ‘Noise Reduction in Ultrasonic Gas Flow Measurement’ 4th International Symposium on Fluid Measurement June 1999.

Lars Farestvedt, FMC ‘Multipath Ultrasonic Flow Meters for Gas Measurement’ International School of Hydrocarbon Measurement May 2003

Ronald H. Dieck, ‘Measurement Uncertainty Methods and Applications’ ISA 4th Edition 2007

STANDARDS:AGA Report No 9 Measurement of Gas by Multipath Ultrasonic Meters (1998)

ISO/TR 5168 Measurement of fluid flow: Evaluation of uncertainties

ISO/TR 7066-1 Assessment of uncertainty in calibration and use of flow measurement devices

ISO 13686 Natural Gas – Quality designation

ISO 14111 Natural Gas – Guidelines to traceability in analysis

ISO 14532 Natural Gas – Terminology

©2010 Daniel Measurement and Control, Inc. All Rights Reserved. Unauthorized duplication in whole or in part is prohibited. Printed in the USA. DAN-TECHNOLOGIES-A-PRACTICAL-APPROACH-TO-MODELLING-LNG-TRAIN-DESIGN-FOR-MINIMIZING-MEASUREMENT-&-ALLOCATION-UNCERTAINTY

Daniel Measurement and Control, Inc. is a wholly owned subsidiary of Emerson Electric Co., and a division of Emerson Process Management. The Daniel name and logo are registered trademarks of Daniel Industries, Inc. The Emerson logo is a registered trademark and service mark of Emerson Electric Co. All other trademarks are the property of their respective companies. The contents of this publication are presented for informational purposes only, and while every effort has been made to ensure their accuracy, they are not to be construed as warranties or guarantees, expressed or implied, regarding the products or services described herein or their use or applicability. All sales are governed by Daniel’s terms and conditions, which are available upon request. We reserve the right to modify or improve the designs or specifications of such products at any time. Daniel does not assume responsibility for the selection, use or maintenance of any product. Responsibility for proper selection, use and maintenance of any Daniel product remains solely with the purchaser and end-user.

Emerson Process ManagementDaniel Measurement and Control, Inc.www.daniel.com

North America / Latin America:HeadquartersUSA - Houston, TexasT +1.713.467.6000F +1.713.827.3880USA Toll Free 1.888.FLOW.001

Europe: Stirling, Scotland, UK T +44.1786.433400F +44.1786.433401

Middle East, Africa: Dubai, UAET +971.4.811.8100F +971.4.886.5465

Asia Pacific: Singapore T +65.6777.8211F +65.6777.0947 / 0743

a practical approach to modelling lng train design for ... daniel documents/… · project thereby...

Documents