EUVIMEDEUVIMED entdecken
Crossref

Application of machine learning technique in water, chemical, and thermal flooding: review paper

Rihab A. Deabl, Mohammed S. Al-Jawad

SOCAR Proceedings · 2026 · (2) · 60-78

DOI: 10.5510/ogp20260201196 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SOCAR Proceedings, 2026. Verfasst von Rihab A. Deabl und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article aims to provide literature for the development of new technologies in enhanced oil recovery (EOR), with a focus on reservoir modelling using machine learning (ML). The inaccuracy and inefficiency of traditional physics-based numerical simulations necessitate the development of faster and more intelligent tools. This review serves as a comprehensive resource on applied ML approaches for reservoir modelling. This research classifies the previous artificial intelligence (AI) research in EOR into three methods: water flooding (WF), chemical enhanced oil recovery (CEOR), and thermal flooding. The comprehensive classification is based on the algorithm used, dataset, purpose, inputs, results, and evaluation for each method in each paper. A novel method for simulating dynamic fluid distributions in WF, the Conditional Deep Convolutional Generative Adversarial Network (CDC-GAN) significantly lowers computational costs while handling complex nonlinear relationships. Particle Swarm Optimization (PSO) combined with Bayesian Random Forest (BRF) provides reliable proxy modelling and optimization, also the Echo State Network (ESN) improves prediction accuracy, however it requires high-quality historical data. While Adaptive Neuro-Fuzzy Inference Systems (ANFIS) efficiently handle uncertainties, Least Square Support Vector Machines (LSSVM) and Artificial Neural Networks (ANNs) demonstrate predictive capabilities for nonlinear relationships in the CEOR. This review emphasizes the role of Reinforcement Learning (RL) in thermal, along with the incorporation of Principal Component Analysis (PCA) and clustering techniques for improved data interpretation. Consequently, this study presents a comprehensive analysis of AI techniques in Enhanced Oil Recovery (EOR) from 2009 to 2024, offering researchers and technical experts insights for future investigations. Keywords: Bayesian random forest; particle swarm optimisation; least square support vector machine; ANNs; reinforcement learning.

Zitieren

APA

Deabl, R. A., & Al-Jawad, M. S. (2026). Application of machine learning technique in water, chemical, and thermal flooding: review paper. SOCAR Proceedings, 60-78. https://doi.org/10.5510/ogp20260201196

Vancouver

Deabl RA, Al-Jawad MS. Application of machine learning technique in water, chemical, and thermal flooding: review paper. SOCAR Proceedings. 2026:60-78. doi:10.5510/ogp20260201196

BibTeX
@article{Deabl2026,
  author = {Deabl, Rihab A. and Al-Jawad, Mohammed S.},
  title = {Application of machine learning technique in water, chemical, and thermal flooding: review paper},
  journal = {SOCAR Proceedings},
  year = {2026},
  number = {2},
  pages = {60--78},
  doi = {10.5510/ogp20260201196}
}
Kostenlos starten →