Physics-informed machine learning with differentiable programming for heterogeneous underground reservoir pressure management

TL;DR

This paper introduces a physics-informed machine learning approach using differentiable programming to efficiently manage underground reservoir pressures, accounting for heterogeneity and uncertainty, enabling near real-time decision making.

Contribution

The authors develop a differentiable physics-based simulator integrated with neural networks to predict optimal extraction rates, significantly accelerating reservoir pressure management tasks.

Findings

The framework achieves 400,000 times faster simulation than traditional models.

It effectively manages reservoir pressures under complex heterogeneity and uncertainty.

The approach enables near real-time analysis and robust uncertainty quantification.

Abstract

Avoiding over-pressurization in subsurface reservoirs is critical for applications like CO2 sequestration and wastewater injection. Managing the pressures by controlling injection/extraction are challenging because of complex heterogeneity in the subsurface. The heterogeneity typically requires high-fidelity physics-based models to make predictions on CO$_2$ fate. Furthermore, characterizing the heterogeneity accurately is fraught with parametric uncertainty. Accounting for both, heterogeneity and uncertainty, makes this a computationally-intensive problem challenging for current reservoir simulators. To tackle this, we use differentiable programming with a full-physics model and machine learning to determine the fluid extraction rates that prevent over-pressurization at critical reservoir locations. We use DPFEHM framework, which has trustworthy physics based on the standard two-point flux finite volume discretization and is also automatically differentiable like machine learning models. Our physics-informed machine learning framework uses convolutional neural networks to learn an appropriate extraction rate based on the permeability field. We also perform a hyperparameter search to improve the model's accuracy. Training and testing scenarios are executed to evaluate the feasibility of using physics-informed machine learning to manage reservoir pressures. We constructed and tested a sufficiently accurate simulator that is 400000 times faster than the underlying physics-based simulator, allowing for near real-time analysis and robust uncertainty quantification.