1
Solving the coupled partial differential equations (PDEs) that govern the dynamics of multiphysics systems is both important and challenging. Existing numerical methods such as the finite element method (FEM) are known to be computationally intensive, while machine learning techniques, like the physics-informed neural network (PINN), often falter when modeling complex systems or processes over long timescales. To overcome these limitations, we propose a new framework “Drift-Correcting Multiphysics Informed Neural Network” (DC-MPINN), specifically designed to solve coupled multiphysics problems efficiently over extended timescales–without sacrificing accuracy. This new method introduces an architecture for temporal domain decomposition that corrects drift of conserved quantities, as well as a composite loss function that allows solving coupled multiphysics problems. We demonstrate the superior performance of DC-MPINN over traditional FEM approaches in several benchmark problems. This approach represents a step forward in multiphysics computational techniques, enhancing our ability to understand and predict the behavior of physical processes across various disciplines.
RoseNNa: A performant, portable library for neural network inference with application to computational fluid dynamics
RoseNNa: A performant, portable library for neural network inference with application to computational fluid dynamics
42Accelerating Physics-Based Simulations Using End-to-End Neural Network Proxies: An Application in Oil Reservoir Modeling.
We develop a proxy model based on deep learning methods to accelerate the simulations of oil reservoirs–by three orders of magnitude–compared to industry-strength physics-based PDE solvers. This paper describes a new architectural approach to this task modeling a simulator as an end-to-end black box, accompanied by a thorough experimental evaluation on a publicly available reservoir model. We demonstrate that in a practical setting a speedup of more than 2000X can be achieved with an average sequence error of about 10% relative to the simulator. The task involves varying well locations and varying geological realizations. The end-to-end proxy model is contrasted with several baselines, including upscaling, and is shown to outperform these by two orders of magnitude. We believe the outcomes presented here are extremely promising and offer a valuable benchmark for continuing research in oil field development optimization. Due to its domain-agnostic architecture, the presented approach can be extended to many applications beyond the field of oil and gas exploration.
Read more
Copyright 2026 Cactus Communications. All rights reserved.
