Effects of image resolution and numerical discretization on permeability evaluations

Digital Rock Physics (DRP) analysis is a widely employed technique for predicting transport parameters from 3D images of core samples. However, the effects of image resolution and spatial discretization of the DRP mesh grid have rarely been systematically studied in detail. To address this issue, we examine a generic sand pack, representing a homogeneous porous medium. This sample was imaged using X-ray micro-tomography at three different spatial resolutions (6, 3, and 1.5 microns/voxel). Permeability is then numerically evaluated by solving the Stokes flow equations using a finite volume method. The processed meshes for converged macroscopic evaluations consist of 105 million to 58 billion cells, necessitating the alternative use of a two-step upscaling method. By employing both methods, this study analyzes the respective influences of image resolution and spatial discretization. Significant effects are observed from both image resolution and spatial discretization, the analysis of which can contribute to identifying optimal strategies for enhancing the accuracy of permeability evaluation.