Self-Learning Reservoir Management
- Luigi Saputelli (PDVSA/U. of Houston) | Michael Nikolaou (U. of Houston) | Michael J. Economides (U. of Houston)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Evaluation & Engineering
- Publication Date
- December 2005
- Document Type
- Journal Paper
- 534 - 547
- 2005. Society of Petroleum Engineers
- 5.7.2 Recovery Factors, 4.6 Natural Gas, 2.3 Completion Monitoring Systems/Intelligent Wells, 4.1.5 Processing Equipment, 5.5.2 Construction of Static Models, 5.1.2 Faults and Fracture Characterisation, 5.5.8 History Matching, 5.3.2 Multiphase Flow, 1.8 Formation Damage, 5.3.9 Steam Assisted Gravity Drainage, 5.5 Reservoir Simulation, 5.1 Reservoir Characterisation, 4.3.4 Scale, 5.4.6 Thermal Methods, 5.4.1 Waterflooding, 3.1.6 Gas Lift, 5.2.1 Phase Behavior and PVT Measurements, 5.1.1 Exploration, Development, Structural Geology, 7.1.10 Field Economic Analysis, 7.1.9 Project Economic Analysis, 5.2 Reservoir Fluid Dynamics, 5.1.5 Geologic Modeling, 5.6.4 Drillstem/Well Testing, 4.1.2 Separation and Treating, 1.2.3 Rock properties, 6.5.2 Water use, produced water discharge and disposal, 7.1.5 Portfolio Analysis, Management and Optimization, 5.6.8 Well Performance Monitoring, Inflow Performance
- 2 in the last 30 days
- 1,932 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
In this work, we present an industrial automation framework for control andoptimization of hydrocarbon-producing fields while satisfying business andphysical constraints. The all-encompassing reservoir-management problem isdecomposed into a hierarchy of decision-making problems at different timescales.
We exemplify the proposed approach through a case study on a multiple-layerreservoir with a classical waterflood problem, in which a numerical reservoirmodel is used as a virtual field. A model-predictive control (MPC) strategy isused to regulate well and field instrumentation at economically optimal setpoints determined by an overlying supervisory control level. The studydemonstrates significant reduction in water-handling costs and increased oilrecovery.
This work is a starting point for further development in automaticintelligent reservoir technologies, which capitalize on the abilities ofpermanent instrumented wells and remotely activated downhole completions.
Reservoir management today is facing remarkable challenges in optimizingprofitability while satisfying a number of constraints (physical, financial,geopolitical, and human). To optimize profitability, engineers traditionallyhave used mathematical models, field data, and domain expertise in an effort tomake decisions about the best operating scenario. To increase the opportunitiesfor profitability by greatly increasing the volume of available field data andthe number of potential operating scenarios, the industry has recently starteddeploying sophisticated hardware for remote sensing and actuation of wells andfacilities.
However, the acquisition of domain expertise about an oil field is a lengthyand often unstructured activity that cannot be undertaken easily on acontinuous basis. In addition, because of the complexity and magnitude of anall-encompassing optimization problem for an entire oil field, decisions aremade in a fragmented way for various pieces of that oil field. The lack ofintelligent software applications exacerbates the situation. As a result, thecapabilities of new sensing and actuation hardware have not been fullyrealized, making it difficult to justify the significant cost that suchhardware imparts. In fact, it is fair to say that not much can be expected froma feedback-based decision-making loop unless all elements in the loop areproperly configured, connected, and functioning. The industry state of the artis clearly far from this ideal end.
To address the above issues, we propose a fieldwide optimization and controlframework1 with the following key features:
• It uses a hierarchy of time scales to separate the levels over whichdecision making is performed, thus rendering a complex problem solvable.
• It integrates field data for continuous learning of key reservoirfeatures, based on simplified empirical models suitable for real-timeoperations.
• It continuously optimizes reservoir performance while satisfying allbusiness and physical (surface and subsurface) constraints.
• It uses an advanced feedback-control strategy, which can be implementedeasily on field controllers at the wellhead or downhole.
• Its multiscale structure can naturally host optimization levels such asmultilateral selection, well location, and portfolio optimization.
To exemplify the above framework, we develop in this work an MPC (recedinghorizon) scheme that underlies a supervisory optimization level, which predictsthe best operating points of a hydrocarbon-producing field. The resultingstructure is a self-learning and self-adaptive scheme that optimizes multiphasefluid migration in compartmentalized reservoirs while integrating downholecompletions, wellhead restrictions, and business objectives andconstraints.
To demonstrate the capabilities of the proposed approach, we develop asoftware prototype and test it on a case study using a commercial reservoir-and well-modeling environment as a virtual reservoir. Dynamic simulations showthat the proposed strategy results in significant reduction of water injectedand produced, with a simultaneous increase in overall oil recovery. For thecase study presented, the self-learning reservoir-management strategy is ableto reduce cumulative water production by almost 80% and reduce water injectionby 55%, increasing project profitability from 13 to 55%.
In the sections that follow, we first give a background of currentreservoir-management challenges as applied to continuous oilfield modeling anddecision-making processes, multivariable optimization, and automatic control.We then present our proposed approach based on petroleum-system identification,an MPC strategy, and a closed-loop linear programming-optimization level thatsearches continuously for the best operating point of the field. Finally, weoffer suggestions for further work.
|File Size||2 MB||Number of Pages||14|
1. Saputelli, L.: "Self-Learning Reservoir Management: An Application ofModel Predictive Control for Continuous Oilfield Optimization," PhDdissertation, U. of Houston, Houston (2003).
2. Saputelli, L.: "Drilling Annual Report," PDVSA internal report, Caracas(1999).
3. Saputelli, L. et al.: "Integration of Computer-AidedHigh-Intensity Design With Reservoir Exploitation of Remote and OffshoreLocations," paper SPE 64621 presented at the 2000 SPE International Oil andGas Conference and Exhibition in China, Beijing, 7-10 November.
4. Yeten B., Durlofsky, L.J., and Aziz, K.: "Optimization of Nonconventional WellType, Location, and Trajectory," paper SPE 77565 presented at the 2002 SPEAnnual Technical Conference and Exhibition, San Antonio, Texas, 29 September—2October.
5. Ramirez, W.F.: Application of Optimal Control Theory to Enhanced OilRecovery, Elsevier, New York City (1987).
6. McKie, C.J.N. et al.: "Economic Benefits From AutomatedOptimization of High Pressure Gas Usage in an Oil Production System," paperSPE 67187 presented at the 2001 SPE Production and Operations Symposium,Oklahoma City, Oklahoma, 24-27 March.
7. Sudaryanto, B.: "Optimization of Displacement Efficiency of Oil Recoveryin Porous Media Using Optimal Control Theory," PhD dissertation, U. of SouthernCalifornia, Los Angeles (1999).
8. Murphy, J.: "Matching Elegance & Brawn: The Challenges of Smart WellTechnology," Oil & Gas J. (April 1999).
9. Daneshy, A. and Donnelly, J.: "JPT Roundtable: The Funding andUptake of New Upstream Technology," JPT (March 2004) 28.
10. Nikravesh, M. et al.: "Design of Smart Wellhead Controllersfor Optimal Fluid Injection Policy and Producibility in Petroleum Reservoirs: ANeuro-Geometric Approach," paper SPE 37557 presented at the 1997 SPEInternational Thermal Operations and Heavy Oil Symposium, Bakersfield,California, 10-12 February.
11. Jansen, B. et al.: "Automatic Control of Unstable GasLifted Wells," paper SPE 56832 presented at the 1999 SPE Annual TechnicalConference and Exhibition, Houston, 3-6 October.
12. Nævdal, G. et al.: "A NewMethodology for Designing Instrumentation of Production Wells for InflowProfile Estimation," paper SPE 59301 presented at the 2000 SPE/DOE ImprovedOil Recovery Symposium, Tulsa, 3-5 April.
13. Wang, S. et al.: "OptimalControl Strategy and Experimental Investigation of Gas-Liquid CompactSeparators," paper SPE 63120 presented at the 2000 SPE Annual TechnicalConference and Exhibition, Dallas, 1-4 October.
14. Dalsmo, M., Halvorsen, E., and Slupphaug, O.: "Active Feedback Control of UnstableWells at the Brage Field," paper SPE 77650 presented at the 2002 SPE AnnualTechnical Conference and Exhibition, San Antonio, Texas, 29 September-2October.
15. Brouwer, D.R. and Jansen, J.D.: "Dynamic Optimization of WaterfloodingWith Smart Wells Using Optimal Control Theory," SPEJ (December 2004)391.
16. Brouwer, D.R. et al.: "Recovery Increase ThroughWaterflooding With Smart Well Technology," paper SPE 68979 presented at the2001 SPE European Formation Damage Conference, The Hague, 21-22 May.
17. Lee, E.B. and Markus, L.: Foundations of Optimal Control Theory,John Wiley and Sons, New York City (1967).
18. Saputelli, L. et al.: "ACritical Overview of Artificial Neural Network Applications in the Context ofContinuous Oil Field Optimization," paper SPE 77703 presented at the 2002SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 29September-2 October.
19. Skogestad, S.: "Plantwide Control: The Search for the Self-OptimizingControl Structure," J. of Process Control (2000) 10, 487.
20. Ljung, L., Gustavsson, I., and Soderstrom, T.: "Identification of LinearMultivariable Systems Operating Under Linear Feedback Control," IEEETransactions on Automatic Control (1974) 19, 836.
21. Ljung, L.: System Identification: Theory for the Users, PrenticeHall, Englewood Cliffs, New Jersey (1987).
22. Ternyik, J. IV et al.: "Virtual Measurement in Pipes: Part1—Flowing Bottomhole Pressure Under Multiphase Flow and Inclined WellboreConditions," paper SPE 30975 presented at the 1995 SPE Eastern RegionalMeeting, Morgantown, West Virginia, 18-20 September.
23. Mohaghegh, S. et al.: "Reservoir Characterization ThroughSynthetic Logs," paper SPE 65675 presented at the 2000 SPE Eastern RegionalMeeting, Morgantown, West Virginia, 17-19 September.
24. Renard, G. et al.: "SystemIdentification Approach Applied to Watercut Analysis in Waterflooded LayeredReservoirs," paper SPE 39606 presented at the 1998 SPE/DOE Improved OilRecovery Symposium, Tulsa, 19-22 April.
25. Nikolaou, M.: "When is Non-linear Dynamic modeling necessary?,"Proc., 1993 American Control Conference, San Francisco (2-4 June)1067-1071.
26. Nikolaou, M. and Hanagandi, V.: "Non-linearity Quantification and itsApplication to Non-linear System Identification," Chem. Eng. Comm.(1998) 166, 1.
27. Nikolaou, M. and Misra, P.: "Linear Control of Non-linear Processes:Recent Developments and Future Directions," Computers and ChemicalEngineering J. (2003) 27, No. 8-9, 1043.
28. Cutler, C.R. and Ramaker, B.L.: "Dynamic Matrix Control—A ComputerControl Algorithm," Proc., 1980 Joint Automatic Control Conference, SanFrancisco, WP5-B (1980).
29. Nikolaou, M.: "Model Predictive Controllers: A Critical Synthesis ofRecent Developments and Industrial Needs," Chemical Engineering Series ,Academic Press, Burlington, Massachusetts (2001).
30. Genceli, H., and Nikolaou, M.: "A New Approach to Constrained PredictiveControl with Simultaneous Model Identification," AIChE J. (1996) 42, No.10, 2857.
31. Eker, S.A. and Nikolaou, M.: "Adaptive Control Through On-LineOptimization: The MPCI Paradigm," Appl. Math. And Comp. Sci. (1999) 9,No. 1, 101.
32. Schwarm, A., Eker, S.A., and Nikolaou, M.: "MPCI: A New Approach toClosed-Loop Identification and Adaptive Control," FOCAPO, Snowbird, Utah(1998).
33. Schwarm, A. and Nikolaou, M.: "Chance Constrained Model PredictiveControl," AIChE J. (1999) 45, No. 8, 1605.
34. Cybenko, G.: "Approximations by superpositions of a sigmoidal function,"Math Control Signal Systems (1989) 2, 303.
35. Hornik, K., Stinchcombe, M., and White, H.: "Multilayer feed forwardnetworks are universal approximators," Neural Networks (1989) 2,359.
36. Moore, A.W.: "Cross-validation for detecting and preventingoverfitting," available from www-2.cs.cmu.edu/~awm/tutorials/overfit09.pdf.
37. Akaike, H.: "Information Theory and the Extension of the MaximumLikelihood Principle," Second Intl. Symposium on Information Theory ,V.N. Petrov and F. Csaki (eds.), Akailseoniai-Kiudo, Budapest (1973)267-281.
38. Liu, W. and Ramirez, W.F.: "Interactive Personal Computer OptimalControl Calculations for Steamflooding ," paper SPE 25523 available fromSPE, Richardson, Texas (1992).
39. Sudaryanto, B. and Yortsos, Y.C.: "Optimization of Displacements inPorous Media Using Rate Control," paper SPE 71509 presented at the 2001 SPEAnnual Technical Conference and Exhibition, New Orleans, 30 September-3October.
40. Holmes, J.A., Barkve, T., and Lund, O.: "Application of a Multisegment WellModel to Simulate Flow in Advanced Wells ," paper SPE 50646 presented atthe 1998 SPE European Petroleum Conference, The Hague, 20-22 October.
41. Nyhavn, F., Vassenden, F., and Singstad, P.: "Reservoir Drainage With DownholePermanent Monitoring and Control Systems. Real-Time Integration of DynamicReservoir Performance Data and Static Reservoir Model Improves ControlDecisions," paper SPE 62937 presented at the 2000 SPE Annual TechnicalConference and Exhibition, Dallas, 1-4 October.
42. Yeten, B., Durlovsky, L.J., and Aziz, K.: "Optimization of Smart WellControl," paper SPE 79031 presented at the 2002 SPE International ThermalOperations and Heavy Oil Symposium and International Horizontal Well TechnologyConference, Calgary, 4-7 November.
43. Litvak, M. et al.: "PrudhoeBay E-field Production Optimization System Based on Integrated Reservoir andFacility Simulation," paper SPE 77643 presented at the 2002 SPE AnnualTechnical Conference and Exhibition, San Antonio, Texas, 29 September-2October.
44. Queipo, N.V., Goicochea P., J.V., and Pintos, S.: "Surrogate Modeling-BasedOptimization of SAGD Processes," paper SPE 69704 presented at the 2001 SPEInternational Thermal Operations and Heavy Oil Symposium, Porlamar, MargaritaIsland, Venezuela, 12-14 March.