An advanced fully-implicit solver for heterogeneous porous media based on foam-extend

TL;DR

This paper introduces coupledMatrixFoam, a robust, fully implicit OpenFOAM-based solver for simulating multiphase flows in heterogeneous porous media, significantly improving stability, efficiency, and scalability for complex physical interactions.

Contribution

It develops and validates a new coupled, fully implicit solver based on foam-extend 5.0, enhancing simulation robustness and computational efficiency for heterogeneous porous media.

Findings

Enables larger time steps in simulations with complex physics.

Demonstrates improved numerical stability over existing solvers.

Shows scalability suitable for high-performance computing environments.

Abstract

The study of multiphase flows in porous media is fundamental to various fields, including oil recovery, CO2 sequestration, hydrogeology, and others. Accurate predictions of fluid behavior in these systems can enhance process efficiency while mitigating environmental and health risks. In this context, numerical simulation serves as a powerful tool for anticipating complex flow phenomena and optimizing engineering strategies. Both commercial simulators and open-source softwares, such as the porousMultiphaseFoam repository based on the OpenFOAM framework, have been developed to model these processes. However, simulating heterogeneous porous media with complex porosity and permeability distributions poses significant numerical challenges. To address these, we introduce coupledMatrixFoam, a new OpenFOAM-based solver designed for enhanced robustness and numerical stability. coupledMatrixFoam integrates the Eulerian multi-fluid formulation for phase fractions with Darcy's law for porous media flow, employing a fully implicit, block-coupled solution for the governing equations. Built on foam-extend 5.0, it leverages the latest fvBlockMatrix developments to improve computational efficiency. This approach enables a significant increase in time step sizes, particularly in cases involving capillary pressure effects and other complex physical interactions. This work details the development, implementation, and validation of coupledMatrixFoam, including comparisons with porousMultiphaseFoam to assess performance improvements. Additionally, a scalability analysis is conducted, demonstrating the solver's suitability for high-performance computing (HPC) applications, which are essential for large-scale, real-world simulations.