Modeling and Monitoring World Class Oil and Gas Assets

Transcription

1 Modeling and Monitoring World Class Oil and Gas Assets Presented by Ken Startz Adv. Sr. Business Analyst

2 Marathon Oil Facts & Figures Established in 1887 Market Cap: ~ $21 billion (as of 10/1/12) Revenues: $14.7 billion Income from Continuing Operations*: $1.7 billion Net Income*: $2.95 billion Capital Expenditures**: $3.4 billion Net Oil Production Sold: 219,000 BPD Net Natural Gas Production Sold: 866 million CFD Net Synthetic Crude Oil Sold: 43,000 BPD Net Proved Liquid Hydrocarbon, Natural Gas and Synthetic Crude Oil Reserves: 1.8 billion BOE Employees: ~3,300 Headquartered in Houston, Texas *Marathon s downstream business was spun-off on June 30, 2011, and has been reported as discontinued operations. **Includes accruals. 2

3 Background Oil and gas exploration and / or production in ten countries. PI System installations: Four in upstream (Houston, MEGPL, EGLNG, and Norway) My first PI System implementation was at Robinson, Illinois Refinery in First OSIsoft Salesman was Pat Kennedy 3

4 Business Challenges No method to retrieve time-series data for upstream assets Requirements to gather data from six different control systems (including system with no existing interface) Embrace standards such as thin client using SharePoint Integrate with Digital Oilfield initiatives such as Viewpoint (Mash-ups from different Upstream Sources) Retrieve data from JV partners Advanced computation capability to highlight problem areas and prioritize workflows Provide knowledge management and knowledge transfer to newer technical professionals Integration with modeling applications such as Kappa Diamant, Kappa Saphir, HYSYS, ECLIPSE. 4

5 Contents Modeling: Equatorial Guinea Modeling: Droshky (Gulf of Mexico) Monitoring and Alerting: Mimm s Creek Plunger Lift Well Diagnostic Tools 5

6 Real-Time Data Infrastructure for Production Operations NAPO North Dakota Wyoming Colorado Oklahoma Gulf of Mexico Texas Gulf of Mexico IPO - EG MEGPL Alba B2 MEGPL Gas Plant EG LNG Outside Operated IPO - Norway WonderWare DCS Intellution ifix SCADA Aker Solutions MCS Experion DCS DeltaV DCS Industrial Evolution Alvheim Industrial SQL ihistorian PhD PI PI Kongsberg/Rapid OPC Interface PI Data Publisher OPC Interface PI PhD interface PI to PI PI to PI PI ` ` ` PI Central Historian RTWebParts Server (MOSS 2007) ACE Server HES RTWebParts DataLink Reliability RTWebParts DataLink Wells/Reservoir RTWebParts DataLink PI ProcessBook Kappa Diamant 6

7 Marathon Equatorial Guinea Ownership in Alba Block Marathon 63.25% (65.21% W.I.) Noble 33.75% (34.79% W.I.) GEPetrol 3.00% (Carry) Ownership in Alba LPG Plant: Marathon 52.17% (65.21% W.I.) Noble 27.83% (34.79% W.I.) SONAGAS 20.00% (Repaid Carry) Alba Field is located 18 miles NW of Bioko Island, Equatorial Guinea in Rio Del Ray Basin. Water Depth ~250 ft 7

8 Alba Field Is all about the scale Alba 9 Well and Marathon Oil Tower Houston, Texas Alba 9 Marathon Oil Tower ,600 acres ~ 14,923 soccer fields! 8

9 Bioko Island Equatorial Guinea

10 Finished Platforms B2 Platform B2 Platform C Platform Pipelines Landfall 10

11 MEGPL Gas Plant 11

12 Housing Complex 12

13 Equatorial Guinea Before E G DCS Network Offshore Platform A Honeywell Experion Offshore Platform B-B2 Honeywell Experion Offshore Platform C Honeywell Experion Gas Plant Control System Honeywell Experion Honeywell PhD Honeywell PhD One PC Worldwide PHD Access 13

14 Equatorial Guinea After Offshore Platform A * E G DCS Network Offshore Platform B * Offshore Platform C * Gas Plant Control Room * GE Turbine Compressors ( 2 exist for MEGPL ) GE Power Generators ( 2 exist for MEGPL ) Honeywell PhD PI API Node Honeywell PhD PI API Node PI System Echo PI System Echo PI to PI Interface MEGPL PI DMZ Network Segment PI to PI Interface MEGPL PI System - DataLink - ProcessBook US Based Central PI System ( Plano, TX ) Business Network Central PI System DataLink (Excel) PI WebParts (SharePoint) HYSYS Simulation Software Kappa Diamant / Saphir Modelling Package 14

15 Equatorial Guinea - Benefits Worldwide viewing of EG data Interface to HYSYS simulation package Interface to Kappa modeling packages Exports to GE and other service companies Time stamp matching of offshore and on-shore data This data is being used extensively by Engineers in EG and in the US to de-bottleneck process problems in the Gas Plant 15

16 Droshky - Background The field is 100 percent Marathon (WI). The sub-sea wells tie back to DOR s Bullwinkle Platform, Gulf of Mexico. The tie back is through an 18-mile pipeline. The majority of the data is being acquired from Aker Solutions subsea control system. Topsides data is extracted from DOR s PI system, coming from Industrial Evolution. First production was July, Downhole calculations by Baker-Hughes DHTP Calibration server on the platform. Re-calibrations performed monthly by Baker-Hughes in Houston office. 16

17 Subsea Facilities DOR Bullwinkle Host Platform Well Control Umbilical Dual 8 Insulated Flowlines 75,000 BPD Design Rate 4 Subsea Wells w/ 10 completions Flowline Manifolds 17

18 GC65 Bullwinkle GC110 GC113 Rocky Angus GC155 Manatee GC158 Brutus GC199 Noble Lorien GC243 Nexen Aspen GC200 Eagle Troika GC244 Droshky GC201 18

19 Umbilical Methanol HP Hydraulic Power Electrical Power Asphaltene / Paraffin Inhibitor Chemical / Scale Inhibitor Spare Annulus Monitoring LP Hydraulic Power Communication 18.5-mile extension cord that connects the tree control pods to host platform Allows monitoring and control of the wells 19

20 Droshky PI WebParts Graphic 20

21 Droshky Before Aker Solutions MCS (sub-sea data) Baker Hughes Flow-Calculator DHFM Remote Monitoring VIC Client Lafayette and Houston 2

22 2 2 Droshky After (Heavily Integrated!!) Neptune Gas Plant Monthly Gas Sales Allocations MMS Production Reporting SAP Entry (IBM) Production Accounting (Tulsa, OK) Monthly Allocations (IBM Manually enters into SAP) SPL Incorporated Monthly Production Allocations (Pearland, TX) PI System (DOR) Bullwinkle Control System (Intellution) Water Depth ~ 3,000 ft TOW Database Production Allocation and Reporting (Houston, TX) Production Analysis Saphir / Topaze Calculated Tags 150 Facility Tags 87 PI System (MOC) Real time Collection and Reporting (Plano, TX) Kappa Diamant Database (high frequency pressure and rate data) Subsea Tags 788 Shell Provides Daily Allocations (MOC Operators Manually enter into TOW) Aker Solutions MCS (subsea controls) Remote Monitoring Vic Client (Lafayette and Houston) Raw Data Subsea Controls Baker -Hughes SDA Flow-calculator DHFM Offshore DOR Controlled MOC Controlled

23 Droshky - Benefits Savings of $5K per user for displays vs. VIC client Daily input deck for Eclipse models saving vs. hand-typing by engineers. Automatic daily entry into production allocation system (TOW) One source of the truth using PI OLEDB COM connectors into TOW. Monthly updates for re-calibration of Baker-Hughes downhole data using PI Universal File Loader (UFL). PI System is considered Gold Database. Sub-surface modeling with automatic channel created to Kappa Modeling Suite. 23

24 PI WebParts Home Page 24

25 Typical PI WebPart Trends 25

26 Engineering Models Used ECLIPSE: Sub-surface reservoir modeling for multiphase flow through porous media. (Entire Field) Kappa Diamant/Saphir: Pressure transient analysis. Used in the early stages of a well to evaluate properties in the close vicinity to the well. (One well at a time) HYSYS: Topsides modeling. Usually used for offshore platforms. Maximize fluid recovery to optimize surface production facilities. 26

27 Simulation Using Eclipse Reservoir Simulation Data Collection & Interpretation Data Integration & Geologic Modelling Upscaling & Reservoir Simulation Errors, inconsistencies, uncertainty? Fine scale static 3D model(s) Mathematical model(s) for dynamic flow behaviour 27

28 Benefits (Modeling) Project value estimates were based on 0.1% of worldwide production, based on greater well equipment reliability, and well characterization for future wells. 28

29 Reservoir Engineer Quote A new well drilled and completed in Deepwater Gulf of Mexico can cost up to $125 million. Proper reservoir characterization includes flow and pressure build-up data down to the six-second level, in order to avoid wasting capital on unproductive wells. The PI System will be instrumental in capturing this data. 29

30 Executive Management Quote The Droshky project s execution and delivery was outstanding, and our teams working with our partners in IT advanced our capabilities in the use of real time surveillance techniques with their application, said Dave Roberts, Chief Operating Officer (COO), Marathon Oil. We have set the future course for better and faster data collection and interpretation for our producing assets with this success. 30

31 Mimm s Creek Plunger Lift Diagnostic Tool Operators are overwhelmed with tasks (operate by exception) Requested by the Technology Services group (R & D function) Result of a six-month study to optimize production 31

32 Mimms Creek POC Over 500 Natural Gas Wells spread out across East Texas 32

33 Plunger Lift Basics Plunger is dropped through fluid in production tubing with bypass valve open Flow control valve (FCV) is opened and gas is injected below the plunger. This forces the bypass valve to close and the plunger to rise, carrying with it any fluid. Plunger is caught at top while fluid is produced. After a designated amount of time, FCV is closed, the plunger is dropped and the cycle repeats. Controller at surface orchestrates all of this. 33

34 34

35 35

36 7-Day Status of All Wells 36

37 7-Day Status Individual Well 37

38 Well Status at Time of Error 3

39 Exporting Well Errors to Excel 3

40 Resulting Spreadsheet 40

41 Error Detection Pilot A new well-state tag needed to be added for each well. The states are: Well Shut-In, Well Transition, Well Flowing. PI ACE kicks off the Well State Determination process every time a new value is written to the Valve Position Indicator tag. The remaining PI ACE calculations run based on the value of the Well-State tag. A new digital state tag was also added for each well to store current and historical error flags. Current Design: Errors go to Report or PI WebParts Display Future Design: PI Notifications 41

42 Architecture ROC AutoSOL SCADA Server ROC Link Proficy Server PI Pub. PI ACE Server PI Server SharePoint Server 42

43 Benefits Reduce Lost Production Reduce Maintenance Opportunity to focus on higher value items Operators are better able to prioritize their daily routine Fewer problems slip through the cracks Conservative Estimate 43

44 Conclusions Utilizing OSIsoft s Real-Time PI System Infrastructure Standardize Operational Data Separation of Consumers Application Platform Utilizing Microsoft s SharePoint Platform Visibility to Anyone Easily Configured / Flexibility One Version of the Truth 44