Development of Surrogate Reservoir Models (SRM) For Fast Track Analysis of Complex Reservoirs

Transcription

1 Development of Surrogate Reservoir Models (SRM) For Fast Track Analysis of Complex Reservoirs Shahab D. Mohaghegh, WVU & ISI Modavi, A., Hafez, H. Haajizadeh, M., Kenawy, M., and Guruswamy, S., Abu Dhabi Company for Onshore Oil Operations - ADCO SPE Intelligent Energy Conference, Amsterdam, The Netherlands, April

2 The Bottle-Neck Real-Time, High Frequency Data Stream Time Scale: Seconds, Minutes, Hours Full Field Models for Reservoir Simulation & Modeling. One of the major tools for integrated Reservoir Management Time Scale: Days, Months,. 2

3 Objective Developing the next generation of intelligent applications as enabling technologies in response to the needs of smart fields. Development of a Surrogate Reservoir Model (SRM) based on a Full Field Model (FFM) for a giant oil field in the Middle East. 3

4 FFM Characteristics Full Field Model Characteristics: Underlying Complex Geological Model. ECLIPSE TM 165 Horizontal Wells. Approximately 1,000,000 grid blocks. Single Run = 10 Hours on 12 CPUs. Water Injection for Pressure Maintenance. 4

5 SRM Characteristics Accurate replication of Full Field Model Results (for every well in the field): Instantaneous Water Cut Cumulative Oil Production Cumulative Water Production Ability to run in real-time. Remove the bottleneck. 5

6 Surrogate Reservoir Models A subset of a more general set of models called Surrogate Intelligent Models Real-Time Optimization Real-Time Decision Making Analysis of Uncertainty An absolute essential tool for smart fields (i-fields) 6

7 SRM an Engineering Tool Are Surrogate Reservoir Models the same as Response Surface techniques? NO. Unlike purely statistical techniques, SRMs are designed to be engineering tools. 7

8 SRM an Engineering Tool Depending on the project objectives, SRMs are developed to preserve and respond to the physics of the problem. Honoring the physics is an important validation step in the development process of SRMs. 8

9 Surrogate Reservoir Models Project Objectives: To create an accurate surrogate model that can mimic the Full Field Model. Use the surrogate model to perform Monte Carlo Simulation to quantify the uncertainties associated with the FFM. 9

10 Surrogate Reservoir Models Project Objectives: It can provide a foundation for: Analysis of Uncertainty (Monte Carlo Simulation) Real Time Optimization Real Time Decision Making Automatic History Matching 10

11 Surrogate Reservoir Models Methodology: Identify the specific objective of the Surrogate Model. Identify the necessary information needed to accomplish the objective (understand the physics). Resolve the major issue related to the curse of dimensionality Develop and validate the Surrogate Model 11

12 Methodology Identify the specific objective of the Surrogate Model. Analysis of uncertainty. Develop a bean-up schedule for the wells in the asset. Optimize the oil production from the asset (minimize left behind oil). 12

13 Very Complex Geology 13

14 Very Complex Geology 14

15 Very Complex Geology Reservoirs represented in the FFM. 15

16 Curse of Dimensionality Source of dimensionality: STATIC: Representation of reservoir properties associated with each well. DYNAMIC: Simulation runs to demonstrate well productivity. 16

17 Curse of Dimensionality, Static Representing reservoir properties for horizontal wells. 17

18 Curse of Dimensionality, Static Potential list of parameters that can be collected on a per-well basis. Parameters Used on a per well basis Latitude Deviation Horizontal Well Length Distance to Free Water Level Flowing Reference Point Cum. Oil Reference Point Distance to Nearest Producer Distance to Major Fault Longitude Azimuth Productivity Index Water Reference Point Oil Prod. Reference Point Cum. Water Reference Point Distance to Nearest Injector Distance to Minor Fault IMPORTANT NOTE: Specific objective of the surrogate model must be identified in advance. 18

19 Curse of Dimensionality, Static Potential list of parameters that can be collected on a per-grid block basis. Parameters Used on a per segment basis Mid Depth Relative Rock Ttype Initial Water Saturations Horizontal Permeability Reference Point Capillary Pressure/Saturation Function Thickness Porosity Stylolite Intensity Vertical Permeability Reference Point Reference Point IMPORTANT NOTE: Specific objective of the surrogate model must be identified in advance. 19

20 Curse of Dimensionality, Static Total number of parameters that need representation during the modeling process: 18 parameters x 40 grid block/well = parameter per well Total of 732 parameter per well Building a model with 732 parameters per well is not realistic, THE CURSE OF DIMENSIONALITY Dimensionality Reduction becomes a vital task. 20

21 Curse of Dimensionality, Dynamic Well productivity is identified through following simulation runs: All wells producing at 1500, 2500, 3500, & 4500 bpd (nominal rates) No cap on field productivity Cap the field productivity 21

22 Curse of Dimensionality, Dynamic Well productivity through following simulation runs: Step up the rates for all wells No cap on field productivity Cap the field productivity 22

23 Curse of Dimensionality In order to address the Curse of Dimensionality one must understand the behavior and contribution of each of the parameters to the process being modeled. Not a simple and straight forward task.!!! 23

24 Curse of Dimensionality To address this issue, we use ISI s Fuzzy Pattern Recognition technology. 24

25 Key Performance Indicator Parameter: Reference 25

26 Key Performance Indicator 26

27 Key Performance Indicators Potential Benefits: This analysis would confirm or dispute hypotheses and findings, purely based on the collected data: Geology Petrophysics Dynamic Data 27

28 Key Performance Indicators 28

29 Key Performance Indicators Please Note: The lower the bar, the higher the influence. 29

30 Candidate Selection 30

31 Surrogate Modeling Methodology: Divided the data into three partitions. Training dataset 40% Calibration dataset 20 % Verification dataset 40% Train, and validate the model. Run the model for the ranked candidate wells to decide on the most efficient production strategy. 31

32 Surrogate Modeling 32

33 Surrogate Modeling 33

34 Optimal Production Strategy Well Ranked No. 1 IMPORTANT NOTE: This is NOT a Response Surface SRM was run hundreds of times to generate these figures. 34

35 Optimal Production Strategy Well Ranked No. 100 IMPORTANT NOTE: This is NOT a Response Surface SRM was run hundreds of times to generate these figures. 35

36 Analysis of Uncertainty Objective: To address and analyze the uncertainties associated with the Full Field Model using Monte Carlo simulation method. 36

37 Analysis of Uncertainty Motivation: The Full Field Model is a reservoir simulator that is based on a geologic model. The geologic model is developed based on a set of measurements (logs, core analysis, seismic, ) and corresponding geological and geophysical interpretations. 37

38 Analysis of Uncertainty Motivation: Therefore, like any other reservoir simulation and modeling effort, it includes certain obvious uncertainties. One of the outcomes of this project has been the identification of a small set of reservoir parameters that essentially control the production behavior in the horizontal wells in this field (KPIs). 38

39 Analysis of Uncertainty Following are the steps involved: 1. Identify a set of key performance indicators that are most vulnerable to uncertainty. 2. Define probability distribution function for each of the performance indicators. a. Uniform distribution b. Normal (Gaussian) distribution c. Triangular distribution d. Discrete distribution 39

40 Analysis of Uncertainty Following are steps involved: 3. Run the neural network model hundreds or thousands of times using the defined probability distribution functions for the identified reservoir parameters. Performing this analysis using the actual Full Field Model is impractical. 4. Produce a probability distribution function for cumulative oil production and the water cut at different time and liquid rate cap. 40

41 Analysis of Uncertainty Following are steps involved: 5. Such results bounds to be much more reliable and therefore, more acceptable to the management or skeptics of the reservoir modeling studies. 41

42 Analysis of Uncertainty 42

43 Analysis of Uncertainty 43

44 Analysis of Uncertainty Average S Reference point in Top Layer II Value in the model = 8% Lets use a minimum of 4% and a maximum of 15% with a triangular distribution

45 Analysis of Uncertainty Average Capillary Reference point in Top Layer III Value in the model = 79 psi Lets use a minimum of 60 psi and a maximum of 100 psi with a triangular distribution

46 Analysis of Uncertainty PDF for HB001 Cumulative Oil and Cumulative Water production at the rate of 3,000 blpd cap after 20 years. 46

47 Analysis of Uncertainty Such analysis can be performed for all wells at any rate and any number of years. There is a higher probability of acceptance of the ideas for rate increase by the management, if we show that: We are aware of the uncertainties associated with our analysis. Uncertainties are being accounted for in our decision making process. 47

48 CONCLUSIONS A successful surrogate reservoir model was developed for a giant oil field in the Middle East. The surrogate model was able to accurately mimic the behavior of the actual full field flow model. 48

49 CONCLUSIONS The surrogate reservoir model would provide results in real time. The surrogate model was used successfully to analyze uncertainties associated with the full field flow model. This approach and methodology is an important and essential step toward 49

50 CONCLUSIONS Development of successful surrogate reservoir model is an important and essential step toward development of next generation of reservoir management tools that would address the needs of smart fields. 50