intelligent data validation and reconciliation for the

5th Annual Global Refining Summit 2011May 18 ‐ 20, 2011 – Rotterdam, Netherlands

Intelligent Data Validation and Reconciliation Solutions for the f dRefining Industry

Mihaela Cristina Popescu, Head of Performance

Dr. Roberto Linares, Vice President, Visiant Pimsoft

Mihaela Cristina Popescu, Head of Performance Control, Arpechim Refinery, OMV Petrom

Topics

• Business Performance Aspects in Refining

• Refining Case• Refining Case

• Production Accounting and Data Reconciliation

(PADR) Requirements(PADR) Requirements

• Powerful Data Reconciliation Concepts

• Customer Benefits Using Sigmafine• Customer Benefits Using Sigmafine

• Yield Accounting Management Presented by

OMV PetromOMV Petrom

• Conclusion

© Visiant Pimsoft Inc. 2011 All rights reserved.

2Global Refining SummitRotterdam, May 18, 2011

Business Performance Aspects in Refining

• Operationalo Yield, Recovery, Quality Targets, y, Q y go Production Rateso Process Constraints (e.g., environmental, physical limitations)

• Planning• Planningo Best Operational Parameters (e.g., LPs)o Comparison of Planned vs. Actual

• Financialso Material Losso Cost Allocationo Cost Allocationo Inventory Positionso Taxes (e.g., domestic vs. foreign)



Market Drivers

• Environmental (GHG)

• Financial (Sarbanes Oxley Act)• Financial (Sarbanes Oxley Act)

• Yield Accounting StandardsAPI Standardo API Standard

• Refining Margins

• SecuritySecurity

The ideal platform for data validation has to comply with emerging regulations in the refining industry.



Refining Case

• Source of InformationProcess Historiano Process Historian

o Oil Movement System (OMS)

o Vessel Positioning System (VPS)o Vessel Positioning System (VPS)

o LIMS

• Business Ruleso VCF according to API 2004

o Ownership Tracking

o Foreign and Domestic Tracking

o Mass Balance



Refining Case

• IntegrationAutomatic Data Collectiono Automatic Data Collection

o Integration of OMS and VPS data into Production Accounting System

• Reportingo Excel‐based for Production Accountants

o Formal and/or Web‐based for Refinery Users

o Automatic Reporting for Measured Information



Solution Map

Data Flow for Refining Case

HistorianExcel

Sigmafine

OMS

Reports

Vessel Positioning

Aggregation/Integration SQL

AccessReport Logic

Web-based ReportsRules

System



Production Accounting Solution IntegrationThe Bridge Between Business and Process DataThe Bridge Between Business and Process Data

Process Data Lab Data MovementsRecords

Planning Data

SchedulingInformation


8Global Refining Summit Rotterdam, May 18, 2011

Records Data Information

Solvability and RedundancyMaximizing the Power of Your MeasurementsMaximizing the Power of Your Measurements

Sigmafine performs flows solvability analysis.• R The flow is redundant, so it is both measured and solvable.• NR The flow is measured but not redundant so the measurement must be trusted• NR The flow is measured but not redundant, so the measurement must be trusted. • S The flow is not measured, but it is still solvable.• NS The flow is not solvable.

S SR R

NR

S SR R

NRNR

S SR R

NRNRThe more redundancyThe more redundancy

SS SS

The more redundancy the better. This way, more values may be cross‐checked to increase accuracy

The more redundancy the better. This way, more values may be cross‐checked to increase accuracy

SNR R

S

S R

S

S R

Sincrease accuracy. increase accuracy.



SNR RS RS R

Solvability and RedundancyMaximizing the Power of Your MeasurementsMaximizing the Power of Your Measurements

Sigmafine performs flows solvability analysis.• R The flow is redundant, so it is both measured and solvable.• NR The flow is measured but not redundant so the measurement must be trusted• NR The flow is measured but not redundant, so the measurement must be trusted. • S The flow is not measured, but it is still solvable.• NS The flow is not solvable.

S SR R

NRR

S SR R

NR

S SR R

NR * By adding a meter * By adding a meter

S*R SS

to the flow, the flow becomes redundant. As a result, other metered flows also

to the flow, the flow becomes redundant. As a result, other metered flows also

SR R

S

S R

S

S R

S become redundant. become redundant.



SR RS RS R

Reconciliation Finding the Best Fit for Your Model and DataFinding the Best Fit for Your Model and Data

54.2

98 498 4 98.498.4

44.2



Optimal Utilization of Existing Meters

Non Redundant Network

Accurate Meter

Redundant Network

LessAccurate Meter

Less Accurate

Meter



Error DetectionWhich Network Can Detect a Measurement Error?Which Network Can Detect a Measurement Error?

Non Redundant Network. There is no possibility of detection

Accurate meter possibility of detectionmeter failure

Redundant Network. An imbalance can be detecteddetected



Meter PerformanceError IdentificationError Identification

Pure Gas EC0F1230.AD

C0F1230.SF

B2F1003.AD1400

16001750

1649.4

Not Reconciled

400 5

C0F1230.AD

C0F1230.SF

B2F1003.AD

Pure Gas W20F0101.AD

20F0101.SF

12F1003.AD1400

16001750

1571.4

Not Reconciled

422 08

20F0101.AD

20F0101.SF

12F1003.AD

B2F1003.SF

B2F2003.AD

B2F2003.SF

F0F1013B.AD

F0F1013B.SF400

600

800

1000

1200400.5

Not Reconciled

402.47

Not Reconciled

19.939

Not Reconciled

B2F1003.SF

B2F2003.AD

B2F2003.SF

F0F1013B.AD

F0F1013B.SF

12F1003.SF

12F2003.AD

12F2003.SF

F0F1013A.AD

F0F1013A.SF400

600

800

1000

1200422.08

Not Reconciled

428.68

Not Reconciled

0.

Not Reconciled

12F1003.SF

12F2003.AD

12F2003.SF

F0F1013A.AD

F0F1013A.SF

Bias

02-Sep-01 00:00:00 11-Dec-01 14:02:54100.59 Day(s)

200

011-Nov-01 00:00:00 11-Dec-01 14:02:5630.59 Day(s)

200

0

SLO EastC0F1109M.AD

ton/h C0F1109M.SF200

25089.787C0F1109M.AD

ton/h C0F1109M.SF

SLO West20F1109.AD

m3n/h 20F1109.SF200

25099.74520F1109.AD

m3n/h 20F1109.SF

Drift

ton/h C0F1116M.AD

ton/h C0F1116M.SF

ton/h C9F1004.AD

3/h

100

150

200 Not Reconciled

76.712

Not Reconciled

181.66

ton/h C0F1116M.AD

ton/h C0F1116M.SF

ton/h C9F1004.AD

3/h

m3/h 20F1116.AD

m3/h 20F1116.SF

m3/h 29F1004.AD

3/h

100

150

200 Not Reconciled

89.789

Not Reconciled

192.7

m3/h 20F1116.AD

m3/h 20F1116.SF

m3/h 29F1004.AD

3/h

02-Sep-01 00:00:00 11-Dec-01 14:03:04100.59 Day(s)

m3/h C9F1004.SF

m3/h

50

0

Not Reconciled

m3/h C9F1004.SF

m3/h

02-Sep-01 00:00:00 11-Dec-01 14:03:05100.59 Day(s)

m3/h 29F1004.SF

m3/h

50

0

Not Reconciled

m3/h 29F1004.SF

m3/h

Meter Error



Optimum Maintenance of Refinery MetersTaking Action to Improve Meter PerformanceTaking Action to Improve Meter Performance



Improving the Overall PictureHigher Confidence in Refinery DataHigher Confidence in Refinery Data

30%Overall DX1 Trend ( Q4 2008 - Q1 2010)

20%

25%

10%

15%

20%

5%

10%

0%



Oil Loss MonitoringBefore and After Sigmafine

Error Elimination

Before and After Sigmafine

Error Elimination

Meter ImprovementsMeter Improvements

Better Better DecisionsDecisionsDecisionsDecisions



Oil Loss MonitoringBefore and After SigmafineBefore and After Sigmafine

Loss Accountability

345

Loss Accountability

Sigmafine available

0123

(%)

-3-2-1

n-04

r-04

i-04

l-04

p-04

v-04

n-05

r-05

i-05

l-05

p-05

v-05

n-06

r-06

i-06

l-06

p-06

v-06

n-07

r-07

i-07

jan-

mar mai- jul-

sep- nov

jan-

mar mai- jul-

sep- nov

jan-

mar mai- jul-

sep- nov

jan-

mar mai-



Oil Loss MonitoringBefore and After SigmafineBefore and After Sigmafine

Mass Balance Before SigmafineFeb/01 - Jan/02

(Loss % Cumulative = 1.09%)

2 0%

4.0%

6.0%

8.0%

10.0%

-6.0%

-4.0%

-2.0%

0.0%

2.0%

Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan

Loss % 3 Month Rolling Loss % Cumulative

Oil loss (%) After Sigmafine ImplementationSF Losses

1.08

1.001.20

Loss % 3 Month Rolling Loss % Cumulative

0.230.12

0.480.30

0.53

0.000.200.400.600.80

%

Ben

chm

ark

Jun/02 Jul/02 Aug/02 Sep/02 Oct/02 Nov/02

SF monthly SF cumulative



Custody Transfer Error DetectionA Significant Bias Causes Financial LossA Significant Bias Causes Financial Loss



Accounting Fuel Consumption Optimization of Energy Use Requires InformationOptimization of Energy Use Requires Information

4,500

5,000

Fuel Consumption, Oct 2010

3,000

3,500

4,000 Sigmafine

2,000

2,500

,

500

1,000

1,500

Legacy‐

1‐Oct

2‐Oct

3‐Oct

4‐Oct

5‐Oct

6‐Oct

7‐Oct

8‐Oct

9‐Oct

10‐Oct

11‐Oct

12‐Oct

13‐Oct

14‐Oct

15‐Oct

16‐Oct

17‐Oct

18‐Oct

19‐Oct

20‐Oct

21‐Oct

22‐Oct

23‐Oct

24‐Oct

25‐Oct

26‐Oct

27‐Oct

28‐Oct

29‐Oct

System



Monitoring Refinery PerformanceValidation of KPIsValidation of KPIs

Running Plan KPIs

Current/Shift/Running PlanReconciled vs. Theoretical Yields



Performance MonitoringValidating Yield DataValidating Yield Data

ReconciledYields

TheoreticalYields



Increasing Yield ProfitabilityPerformance Management ImprovementPerformance Management Improvement

• Example of diesel yield improvementso Analyzing homogeneous data related to “sweet” crudes:o Analyzing homogeneous data related to sweet crudes:

• 2003 actual average diesel yield 37,4%

• 2004 actual average diesel yield 39,6%

l d l ld d d b ho Average actual diesel yield increased due both to investments on plants and to performance management.

o Historical analysis on diesel yield showed that the average increase d f b hdue to performance management amount was about 0.8% in the period 2003‐2004.

o This led to a profit increase estimated in US$ 1 million per year.



Low Yield Detection After Catalyst “Improvement”Unmasking Hidden ProblemsUnmasking Hidden Problems

Data Quality Improvement: ProfitsMass balance yield - FCC (Fluidized Catalitic Cracking Unit)

110 *Data reconciliation group identified LCO flow meter fault

*During this period it was believed

100

105

yiel

d (%

)

g pthat LCO yield had increased.

Flow meters maintenance

90

95

Mas

s ba

lanc

e

Unit shut downDifferent catalyst loaded

Flow meters maintenance required by reconciliation group

**LCO flow meter repaired**Actual LCO yield equal to the value

80

85

28-01-2005 17-02-2005 09-03-2005 29-03-2005 18-04-2005 08-05-2005 28-05-2005 17-06-2005 07-07-2005 27-07-2005

M Actual LCO yield equal to the value previously to the catalyst change

**Lower light HC and more decanted Lower light HC and more decanted oil oil produced = produced = Less profitsLess profits



28-01-2005 17-02-2005 09-03-2005 29-03-2005 18-04-2005 08-05-2005 28-05-2005 17-06-2005 07-07-2005 27-07-2005

Date

Low Yield Detection After Catalyst “Improvement”Unmasking Hidden ProblemsUnmasking Hidden Problems

Data Quality Improvement: Profits

Eliminated loss corresponding toUS$ 23.5 million per year

M a s s b a la n c e y ie ld - F C C ( F lu id iz e d C a t a l i t ic C r a c k in g U n i t )

1 1 0 *Data reconciliation group identified LCO flow meter fault*During this period it was believed that

$ p y

9 5

1 0 0

1 0 5

bala

nce

yiel

d (%

)

During this period it was believed that LCO yield had increased.

Flow meters maintenance required by reconciliation group

**

8 0

8 5

9 0

2 8 - 0 1 - 2 0 0 5 1 7 - 0 2 - 2 0 0 5 0 9 - 0 3 - 2 0 0 5 2 9 - 0 3 - 2 0 0 5 1 8 - 0 4 - 2 0 0 5 0 8 - 0 5 - 2 0 0 5 2 8 - 0 5 - 2 0 0 5 1 7 - 0 6 - 2 0 0 5 0 7 - 0 7 - 2 0 0 5 2 7 - 0 7 - 2 0 0 5

D a t e

Mas

s Unit shut downDifferent catalyst loaded

**LCO flow meter repaired**Actual LCO yield equal to the value

previously to the catalyst change**Lower light HC and more decanted

oil produced = Less profits



D a t e

Reconciliation Coupled with OptimizationValidating Blending OptimizationValidating Blending Optimization

C t I t ti

Data Quality Improvement: ProfitsCost Integration

with Reconciled DataOptimizer Result

62,77 US$/m3 63,84 US$/m3

1,00 US$/m3 difference

x 15 000 m3

Measurement ProblemHydrotreated Diesel Flow

15.000,00US$/batch

x 15.000 m3

600.000,00US$/month

7.200.000,00US$/year

x 40batches/month

x 12month/year



Detecting Operational ErrorsFinding Undetected ProblemsFinding Undetected Problems

V l t i t i t l f th ti l it t t l

Data Quality Improvement: ProfitsVolumetric entrainment losses from the aromatic removal unit, total

amounts by month from Jun-2006 to may-2007.

900

1000

Financial losses of the aromatic removal unit, total amount by month from Jun-2006 to may-2007.

160000

Data reconciliation group detected

600

700

800

900

nmen

t, m

3

100000

120000

140000

S$

Data reconciliation group detected rubber solvent entrainment from solvent unit to gasoline unit.

300

400

500

olve

nt e

ntra

in

40000

60000

80000

Loss

, US

0

100

200

13-04-2006 12-06-2006 11-08-2006 10-10-2006 09-12-2006 07-02-2007 08-04-2007 07-06-2007

S

0

20000

13-04-2006 12-06-2006 11-08-2006 10-10-2006 09-12-2006 07-02-2007 08-04-2007 07-06-2007



DateDate





900

1000


160000Potential profit depends on the

600

700

800

900

nmen

t, m

3

100000

120000

140000

S$

Potential profit depends on the difference between rubber solvent and gasoline prices …

300

400

500

olve

nt e

ntra

in

40000

60000

80000

Loss

, US

0

100

200

13-04-2006 12-06-2006 11-08-2006 10-10-2006 09-12-2006 07-02-2007 08-04-2007 07-06-2007

S

0

20000

13-04-2006 12-06-2006 11-08-2006 10-10-2006 09-12-2006 07-02-2007 08-04-2007 07-06-2007



DateDate





900

1000


160000Total solvent entrainment: 14 % of feed

600

700

800

900

nmen

t, m

3

100000

120000

140000

S$

Total solvent entrainment: 14 % of feed Total annual losses: US$ 0.7 million

300

400

500

olve

nt e

ntra

in

40000

60000

80000

Loss

, US

0

100

200

13-04-2006 12-06-2006 11-08-2006 10-10-2006 09-12-2006 07-02-2007 08-04-2007 07-06-2007

S

0

20000

13-04-2006 12-06-2006 11-08-2006 10-10-2006 09-12-2006 07-02-2007 08-04-2007 07-06-2007



DateDate

Loss of Capacity Versus Nominal CapacityBias Detection of a MeterBias Detection of a Meter

Loss of processing capacityUS$ 25 7M/year

Carga U283A

5600

5800

US$ 25.7M/year

5200

5400

5600

essa

da (m

3)

4600

4800

5000

Car

ga P

roce

4400

4600

01-0

9-07

15-0

9-07

29-0

9-07

13-1

0-07

27-1

0-07

10-1

1-07

24-1

1-07

08-1

2-07

22-1

2-07

05-0

1-08

19-0

1-08

02-0

2-08

16-0

2-08

01-0

3-08

15-0

3-08

29-0

3-08

12-0

4-08

26-0

4-08

10-0

5-08

24-0

5-08

07-0

6-08

21-0

6-08

05-0

7-08

19-0

7-08

02-0

8-08

16-0

8-08

30-0

8-08

13-0

9-08

27-0

9-08

C did C R ili d



Carga medida Carga Reconciliada

Customer Presentation

Yield Accounting Management

by Mihaela Cristina Popescuby ae a s a opescu

OMV PETROM



ConclusionsSatisfying Global Refinery Needs in the Field of PADRSatisfying Global Refinery Needs in the Field of PADR

• Sigmafine‐based solutions

Territories

North AmericaSigmafine based solutions deliver significant benefits to refining customers worldwide.

Europe & Russia

Asia & Pacific

Latin America

• Solutions have to be adaptive, keeping the customer requirements in mind

6%

Middle East & Africa

requirements in mind.

• Sigmafine technologies are designed to satisfy current and

28%17%

designed to satisfy current and future business requirements.

28%

21%



Thank you!Thank you!

Mihaela Christina Popescu, [email protected] Roberto Linares roberto linares@visiant comDr. Roberto Linares, [email protected]

34 © Visiant Pimsoft Inc. 2011All rights reserved.

Global Refining Summit Rotterdam, May 18, 2011

intelligent data validation and reconciliation for the

Documents