Intelligent Data Validation and Reconciliation Solutions for the

Transcription

1 5 th Annual 2011 May 18 20, 2011 Rtterdam, Netherlands Intelligent Data Validatin and Recnciliatin Slutins fr the Refining Industry Mihaela Cristina Ppescu, Head f Perfrmance Cntrl, Arpechim Refinery, OMV Petrm Dr. Rbert Linares, Vice President, Visiant Pimsft

2 Tpics Business Perfrmance Aspects in Refining Refining Case Prductin Accunting and Data Recnciliatin (PADR) Requirements Pwerful Data Recnciliatin Cncepts Custmer BenefitsUsing Sigmafine Yield Accunting Management Presented by OMV Petrm Cnclusin 2

3 Business Perfrmance Aspects in Refining Operatinal Yield, Recvery, Quality Targets Prductin Rates Prcess Cnstraints (e.g., envirnmental, physical limitatins) Planning Best Operatinal Parameters (e.g., LPs) Cmparisn f Planned vs. Actual Financials Material Lss Cst Allcatin Inventry Psitins Taxes (e.g., dmestic vs. freign) 3

4 Market Drivers Envirnmental (GHG) Financial (Sarbanes Oxley Act) Yield Accunting Standards API Standard Refining Margins Security The ideal platfrm fr data validatin has t cmply with emerging regulatins in the refining industry. 4

5 Refining Case Surce f Infrmatin Prcess Histrian Oil Mvement System (OMS) Vessel Psitining System (VPS) LIMS Business Rules VCF accrding t API 2004 Ownership Tracking Freign and Dmestic Tracking Mass Balance 5

6 Refining Case Integratin Autmatic Data Cllectin Integratin f OMS and VPS data int Prductin Accunting System Reprting Excel based fr Prductin Accuntants Frmal and/r Web based fr Refinery Users Autmatic Reprting fr Measured Infrmatin 6

7 Slutin Map Data Flw fr Refining Case Histrian Sigmafine Excel Reprts OMS Vessel Psitining System Aggregatin/ Integratin Rules SQL Access Reprt Lgic Webbased Reprts 7

8 Prductin Accunting Slutin Integratin The Bridge Between Business and PrcessData Prcess Data Lab Data Mvements Planning Scheduling Recrds Data Infrmatin 8

9 Slvability and Redundancy Maximizingthe Pwer f Yur Measurements Sigmafine perfrms flws slvability analysis. R The flw is redundant, s it is bth measured and slvable. NR The flw is measured but nt redundant, s the measurement must be trusted. S The flw is nt measured, but it is still slvable. NS The flw is nt slvable. R S S R NR S NR S S The mre redundancy the better. This way, mre values may be crss checked t increase accuracy. NR S R 9

10 Slvability and Redundancy Maximizingthe Pwer f Yur Measurements Sigmafine perfrms flws slvability analysis. R The flw is redundant, s it is bth measured and slvable. NR The flw is measured but nt redundant, s the measurement must be trusted. S The flw is nt measured, but it is still slvable. NS The flw is nt slvable. R S S R R NR *R S S * By adding a meter t the flw, the flw becmes redundant. As a result, ther metered flws als becme redundant. R S R 10

11 Recnciliatin Finding the Best Fitfr Yur Mdel and Data

12 Optimal Utilizatin f Existing Meters Nn Redundant Netwrk Redundant Netwrk Accurate Meter Accurate Meter Less Accurate Meter 12

13 Errr Detectin Which Netwrk Can Detect a Measurement Errr? Accurate meter failure Nn Redundant Netwrk. There is n pssibility f detectin Redundant Netwrk. An imbalance can be detected 13

14 Meter Perfrmance Errr Identificatin Pure Gas E Bias C0F1230.AD C0F1230.SF Nt Recnciled B2F1003.AD B2F1003.SF Nt Recnciled B2F2003.AD B2F2003.SF Nt Recnciled F0F1013B.AD F0F1013B.SF Nt Recnciled Pure Gas W F0101.AD F0101.SF Nt Recnciled 12F1003.AD F1003.SF Nt Recnciled 12F2003.AD F2003.SF Nt Recnciled F0F1013A.AD 0. F0F1013A.SF Nt Recnciled Sep-01 00:00: Day(s) 11-Dec-01 14:02: Drift 0 11-Nv-01 00:00: Day(s) 11-Dec-01 14:02:56 SLO East C0F1109M.AD tn/h C0F1109M.SF Nt Recnciled tn/h C0F1116M.AD tn/h C0F1116M.SF Nt Recnciled tn/h C9F1004.AD m3/h C9F1004.SF Nt Recnciled m3/h SLO West F1109.AD m3n/h 20F1109.SF Nt Recnciled m3/h 20F1116.AD m3/h 20F1116.SF Nt Recnciled m3/h 29F1004.AD m3/h 29F1004.SF Nt Recnciled m3/h 0 02-Sep-01 00:00: Day(s) 11-Dec-01 14:03:04 Meter Errr 0 02-Sep-01 00:00: Day(s) 11-Dec-01 14:03:05 14

15 Optimum Maintenance f Refinery Meters Taking Actin t Imprve Meter Perfrmance 15

16 Imprving the Overall Picture Higher Cnfidence inrefinery Data 30% Overall DX1 Trend ( Q Q1 2010) 25% 20% 15% 10% 5% 0% 16

17 Oil Lss Mnitring BefreandAfter Sigmafine Errr Eliminatin Meter Imprvements Better Decisins 17

18 Oil Lss Mnitring Befre and After Sigmafine Lss Accuntability Sigmafine available ja n04 m ar -0 4 m ai -0 4 ju l-0 4 se p04 n v04 ja n05 m ar -0 5 m ai -0 5 ju l-0 5 se p05 n v05 ja n06 m ar -0 6 m ai -0 6 ju l-0 6 se p06 n v06 ja n07 m ar -0 7 m ai -0 7 (%) Lss Accuntability 18

19 10.0% Oil Lss Mnitring BefreandAfter Sigmafine 8.0% 6.0% 4.0% Mass Balance Befre Sigmafine Feb/01 - Jan/02 (Lss % Cumulative = 1.09%) 20% 2.0% 0.0% -2.0% -4.0% -6.0% Feb Mar Apr May Jun Jul Aug Sep Oct Nv Dec Jan Lss % 3 Mnth Rlling Lss % Cumulative Oil lss (%) After Sigmafine Implementatin Benchmark SF Lsses Jun/02 Jul/02 Aug/02 Sep/02 Oct/02 Nv/ % SF mnthly SF cumulative 19

20 Custdy Transfer Errr Detectin A Significant BiasCauses Financial Lss 20

21 Accunting Fuel Cnsumptin Optimizatinf f Energy Use Requires Infrmatin 5,000 4,500 Fuel Cnsumptin, Oct ,000 Sigmafine 3,500 3,000 2,500 2,000 1,500 1, Legacy System 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 21

22 Mnitring Refinery Perfrmance Validatin f KPIs Current/Shift/Running Plan Recnciled vs. Theretical Yields Running Plan KPIs 22

23 Perfrmance Mnitring ValidatingYield Data Recnciled Yields Theretical Yields 23

24 Increasing Yield Prfitability Perfrmance ManagementImprvement Example f diesel yield imprvements Analyzing hmgeneus data related t sweet crudes: 2003 actual average diesel yield 37,4% 2004 actual average diesel yield 39,6% Average actual diesel yield increased due bth t investments n plants and t perfrmance management. Histrical analysis n diesel yield shwed that the average increase due t perfrmance management amunt was abut 0.8% in the perid This led t a prfit increase estimated in US$ 1 millin per year. 24

25 Lw Yield Detectin After Catalyst Imprvement UnmaskingHidden Prblems Data Quality Imprvement: Prfits Mass balance yield - FCC (Fluidized Catalitic Cracking Unit) 110 *Data recnciliatin grup identified LCO flw meter fault *During this perid it was believed 105 that LCO yield had increased. yield (%) Mass balance Unit shut dwn Different catalyst laded Flw meters maintenance required by recnciliatin grup **LCO flw meter repaired **Actual LCO yield equal t the value previusly t the catalyst change **Lwer light HC and mre decanted il prduced = Less prfits Date

26 Lw Yield Detectin After Catalyst Imprvement UnmaskingHidden Prblems Data Quality Imprvement: Prfits Eliminated lss crrespnding t US$ 23.5 millin per year M ass balance yield - FCC (Fluidized Catalitic Cracking Unit) balance yield (%) Mass 110 *Data recnciliatin grup identified LCO flw meter fault 105 *During this perid it was believed that LCO yield had increased Unit shut dwn Different catalyst laded ** Flw meters maintenance required by recnciliatin grup Date **LCO flw meter repaired **Actual LCO yield equal t the value previusly t the catalyst change **Lwer light HC and mre decanted il prduced = Less prfits 26

27 Recnciliatin Cupled with Optimizatin Validating Blending Optimizatin Data Quality Imprvement: Prfits Cst Integratin ti with Recnciled Data Optimizer Result 62,77 US$/m 3 63,84 US$/m 3 1,00 US$/m 3 difference Measurement Prblem Hydrtreated Diesel Flw x m ,00 US$/batch x 40 batches/mnth ,00 US$/mnth x 12 mnth/year ,00 US$/year 27

28 Detecting Operatinal Errrs Finding Undetected Prblems Data Quality Imprvement: Prfits Vlumetric Financial entrainment lsses f the t lsses armatic frm remval the armatic unit, ttal amunt remval by unit, mnth ttal t amunts by frm mnth Jun-2006 frm t Jun-2006 may t may S lvent Lss, entrain US S$ nment, m Data recnciliatin grup detected rubber slvent entrainment frm slvent unit t gasline unit Date 28

29 Detecting Operatinal Errrs Finding Undetected Prblems Data Quality Imprvement: Prfits Vlumetric Financial entrainment lsses f the t lsses armatic frm remval the armatic unit, ttal amunt remval by unit, mnth ttal t amunts by frm mnth Jun-2006 frm t Jun-2006 may t may S lvent Lss, entrain US S$ nment, m Ptential prfit depends n the difference between rubber slvent and gasline prices Date 29

30 Detecting Operatinal Errrs Finding Undetected Prblems Data Quality Imprvement: Prfits Vlumetric Financial entrainment lsses f the t lsses armatic frm remval the armatic unit, ttal amunt remval by unit, mnth ttal t amunts by frm mnth Jun-2006 frm t Jun-2006 may t may S lvent Lss, entrain US S$ nment, m Ttal slvent entrainment: 14 % f feed Ttal annual lsses: US$ 0.7 millin Date 30

31 Lss f Capacity Versus Nminal Capacity BiasDetectin f a Meter Lss f prcessing capacity US$ 25.7M/year 5800 Carga U283A Carga Prcessada (m3) Carga medida Carga Recnciliada 31

32 Custmer Presentatin Yield Accunting Management by Mihaela a Cristina Ppescu OMV PETROM 32

33 Cnclusins SatisfyingGlbalRefinery Needs inthe FieldfPADR Sigmafine based slutins deliver significant benefits t refining custmers wrldwide. Slutins have t be adaptive, keeping the custmer requirements in mind. Sigmafine technlgies are designed t satisfy current and future business requirements. Territries Nrth America Eurpe & Russia Asia & Pacific Latin America Middle East & Africa 6% 17% 28% 21% 28% 33

34 Thank yu! Mihaela Christina Ppescu, Dr. Rbert Linares, 34