Reconstructing Permeability and Saturation in Reservoir Simulation Using Diffusion PDE Models

This study explores the application of a diffusion partial differential equation (PDE) model for reservoir simulation, particularly aimed at reconstructing permeability and saturation fields within a saline aquifer. Focusing on pairs of input permeability and output saturation, the model is trained to capture the underlying dynamics governing fluid flow in porous media. Post-training, the model is capable of inferring or recovering the complete permeability and saturation distributions when provided with limited vertical pixel data of permeability and saturation. This approach offers a novel pathway for enhancing the resolution of subsurface characteristics, contributing to more accurate predictions in reservoir engineering and carbon storage simulations.