A recurrent neural network proxy for production optimization with nonlinear constraints under geological uncertainty
Abstract/Contents
- Abstract
- In well control optimization problems, the goal is to determine the time-varying well settings that maximize an appropriate objective function, which is often specified to be the net present value (NPV) of the operation. Derivative-free optimization algorithms have been widely used in this setting because of their flexibility and ease of implementation. Because these approaches require many simulation runs, proxy models have been developed to predict objective function values for a set of inputs such as time-varying well bottom-hole pressures (BHPs). However, when nonlinear output constraints such as maximum well/field water production rate or minimum well/field oil rate are specified, the problem is more challenging because well rates as a function of time are required for constraint evaluations. In this work, we first introduce a recurrent neural network (RNN)-based proxy model to predict oil and water production and injection rates, as a function of time, for specified time-varying well BHP schedules. This proxy model is incorporated into a derivative-free (particle swarm optimization) production optimization procedure and tested for cases that include nonlinear output constraints. Next, we extend the RNN-based proxy to predict well responses over multiple realizations. This is achieved by incorporating a convolutional neural network (CNN) with the RNN. This enables us to perform robust optimization over an ensemble of geomodels. The proxy is then incorporated into a closed-loop reservoir management (CLRM) workflow, in which we successively apply history matching followed by production optimization to maximize expected NPV. Lastly, we extend the CNN-RNN proxy to estimate production and injection rates over multiple realizations of embedded discrete fracture models (EDFMs). This EDFM-CNN-RNN proxy is used as a function evaluator in robust production optimization problems.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2022; ©2022 |
| Publication date | 2022; 2022 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Kim, Yong Do |
|---|---|
| Degree supervisor | Durlofsky, Louis |
| Thesis advisor | Durlofsky, Louis |
| Thesis advisor | Horne, Roland N |
| Thesis advisor | Volkov, Oleg, 1975- |
| Degree committee member | Horne, Roland N |
| Degree committee member | Volkov, Oleg, 1975- |
| Associated with | Stanford University, Department of Energy Resources Engineering |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Yong Do Kim. |
|---|---|
| Note | Submitted to the Department of Energy Resources Engineering. |
| Thesis | Thesis Ph.D. Stanford University 2022. |
| Location | https://purl.stanford.edu/js442bt1226 |
Access conditions
- Copyright
- © 2022 by Yong Do Kim
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...