Fast Well Control Optimization with Two-Stage Proxy Modeling
Waterflooding is one of the methods used for increased hydrocarbon production. Waterflooding optimization can be computationally prohibitive if the reservoir model or the optimization problem is complex. Hence, proxy modeling can yield a faster solution than numerical reservoir simulation. This fast...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-04-01
|
Series: | Energies |
Subjects: | |
Online Access: | https://www.mdpi.com/1996-1073/16/7/3269 |
_version_ | 1797607988180025344 |
---|---|
author | Cuthbert Shang Wui Ng Ashkan Jahanbani Ghahfarokhi Wilson Wiranda |
author_facet | Cuthbert Shang Wui Ng Ashkan Jahanbani Ghahfarokhi Wilson Wiranda |
author_sort | Cuthbert Shang Wui Ng |
collection | DOAJ |
description | Waterflooding is one of the methods used for increased hydrocarbon production. Waterflooding optimization can be computationally prohibitive if the reservoir model or the optimization problem is complex. Hence, proxy modeling can yield a faster solution than numerical reservoir simulation. This fast solution provides insights to better formulate field development plans. Due to technological advancements, machine learning increasingly contributes to the designing and building of proxy models. Thus, in this work, we have proposed the application of the two-stage proxy modeling, namely global and local components, to generate useful insights. We have established global proxy models and coupled them with optimization algorithms to produce a new database. In this paper, the machine learning technique used is a multilayer perceptron. The optimization algorithms comprise the Genetic Algorithm and the Particle Swarm Optimization. We then implemented the newly generated database to build local proxy models to yield solutions that are close to the “ground truth”. The results obtained demonstrate that conducting global and local proxy modeling can produce results with acceptable accuracy. For the optimized rate profiles, the R<sup>2</sup> metric overall exceeds 0.96. The range of Absolute Percentage Error of the local proxy models generally reduces to 0–3% as compared to the global proxy models which has a 0–5% error range. We achieved a reduction in computational time by six times as compared with optimization by only using a numerical reservoir simulator. |
first_indexed | 2024-03-11T05:38:17Z |
format | Article |
id | doaj.art-ed45d27847124cddb5f983d9fb4d2195 |
institution | Directory Open Access Journal |
issn | 1996-1073 |
language | English |
last_indexed | 2024-03-11T05:38:17Z |
publishDate | 2023-04-01 |
publisher | MDPI AG |
record_format | Article |
series | Energies |
spelling | doaj.art-ed45d27847124cddb5f983d9fb4d21952023-11-17T16:39:41ZengMDPI AGEnergies1996-10732023-04-01167326910.3390/en16073269Fast Well Control Optimization with Two-Stage Proxy ModelingCuthbert Shang Wui Ng0Ashkan Jahanbani Ghahfarokhi1Wilson Wiranda2Department of Geoscience and Petroleum, Norwegian University of Science and Technology, 7031 Trondheim, NorwayDepartment of Geoscience and Petroleum, Norwegian University of Science and Technology, 7031 Trondheim, NorwayDepartment of Geoscience and Petroleum, Norwegian University of Science and Technology, 7031 Trondheim, NorwayWaterflooding is one of the methods used for increased hydrocarbon production. Waterflooding optimization can be computationally prohibitive if the reservoir model or the optimization problem is complex. Hence, proxy modeling can yield a faster solution than numerical reservoir simulation. This fast solution provides insights to better formulate field development plans. Due to technological advancements, machine learning increasingly contributes to the designing and building of proxy models. Thus, in this work, we have proposed the application of the two-stage proxy modeling, namely global and local components, to generate useful insights. We have established global proxy models and coupled them with optimization algorithms to produce a new database. In this paper, the machine learning technique used is a multilayer perceptron. The optimization algorithms comprise the Genetic Algorithm and the Particle Swarm Optimization. We then implemented the newly generated database to build local proxy models to yield solutions that are close to the “ground truth”. The results obtained demonstrate that conducting global and local proxy modeling can produce results with acceptable accuracy. For the optimized rate profiles, the R<sup>2</sup> metric overall exceeds 0.96. The range of Absolute Percentage Error of the local proxy models generally reduces to 0–3% as compared to the global proxy models which has a 0–5% error range. We achieved a reduction in computational time by six times as compared with optimization by only using a numerical reservoir simulator.https://www.mdpi.com/1996-1073/16/7/3269global and local proxy modelingmachine learningderivative-free optimizationreservoir simulation |
spellingShingle | Cuthbert Shang Wui Ng Ashkan Jahanbani Ghahfarokhi Wilson Wiranda Fast Well Control Optimization with Two-Stage Proxy Modeling Energies global and local proxy modeling machine learning derivative-free optimization reservoir simulation |
title | Fast Well Control Optimization with Two-Stage Proxy Modeling |
title_full | Fast Well Control Optimization with Two-Stage Proxy Modeling |
title_fullStr | Fast Well Control Optimization with Two-Stage Proxy Modeling |
title_full_unstemmed | Fast Well Control Optimization with Two-Stage Proxy Modeling |
title_short | Fast Well Control Optimization with Two-Stage Proxy Modeling |
title_sort | fast well control optimization with two stage proxy modeling |
topic | global and local proxy modeling machine learning derivative-free optimization reservoir simulation |
url | https://www.mdpi.com/1996-1073/16/7/3269 |
work_keys_str_mv | AT cuthbertshangwuing fastwellcontroloptimizationwithtwostageproxymodeling AT ashkanjahanbanighahfarokhi fastwellcontroloptimizationwithtwostageproxymodeling AT wilsonwiranda fastwellcontroloptimizationwithtwostageproxymodeling |