Differentiable multiphase flow model for physics-informed machine learning in reservoir pressure management

2025-08-26

AI Summary

Accurate subsurface reservoir pressure control is hindered by geological strong heterogeneity and the prohibitive computational cost of high-fidelity multiphase flow simulations, rendering large-scale uncertainty quantification infeasible. To address this, we propose a physics-informed machine learning framework that tightly couples a differentiable multiphase flow simulator—built upon DPFEHM—with a convolutional neural network, explicitly embedding transient multiphase physical equations into the training process. We adopt a transfer learning strategy: pretraining on computationally inexpensive single-phase steady-state simulations, followed by fine-tuning on multiphase dynamic scenarios. Our method achieves high-accuracy fluid extraction policy prediction using fewer than 3,000 full-physics multiphase simulations—reducing computational cost by three orders of magnitude compared to state-of-the-art approaches requiring millions of simulations. This enables real-time pressure management and significantly enhances generalizability across realistic injection–production configurations.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Sampling/Simulation-based Search
• Multiagent Systems: Multiagent Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Accurate subsurface reservoir pressure control is extremely challenging due to geological heterogeneity and multiphase fluid-flow dynamics. Predicting behavior in this setting relies on high-fidelity physics-based simulations that are computationally expensive. Yet, the uncertain, heterogeneous properties that control these flows make it necessary to perform many of these expensive simulations, which is often prohibitive. To address these challenges, we introduce a physics-informed machine learning workflow that couples a fully differentiable multiphase flow simulator, which is implemented in the DPFEHM framework with a convolutional neural network (CNN). The CNN learns to predict fluid extraction rates from heterogeneous permeability fields to enforce pressure limits at critical reservoir locations. By incorporating transient multiphase flow physics into the training process, our method enables more practical and accurate predictions for realistic injection-extraction scenarios compare to previous works. To speed up training, we pretrain the model on single-phase, steady-state simulations and then fine-tune it on full multiphase scenarios, which dramatically reduces the computational cost. We demonstrate that high-accuracy training can be achieved with fewer than three thousand full-physics multiphase flow simulations -- compared to previous estimates requiring up to ten million. This drastic reduction in the number of simulations is achieved by leveraging transfer learning from much less expensive single-phase simulations.

Problem

• Accurate reservoir pressure control is challenging due to geological heterogeneity
• High-fidelity physics simulations are computationally expensive and prohibitive
• Uncertain heterogeneous properties require many expensive multiphase flow simulations

Innovation

• Differentiable multiphase flow simulator coupled with CNN
• Transfer learning from single-phase to multiphase simulations
• Physics-informed training with fewer than 3000 simulations