Parallelizable Permeability Estimation of Digital Porous Media for Sandstone Using Subvolume Properties for Flow in Porous Media

Abstract

In subsurface energy extraction, permeability or conductivity is the vital parameter for quantifying fluid flow in porous media. Three-dimensional digital core technique is widely used to calculate flow parameters and to analyze the internal structure and properties of rocks. However, one major problem is its high computational cost associated with fine-scale simulation of porous media, especially for large and complex rock samples. In this study, we propose to use subvolume properties to increase computational efficiency. Specifically, we first construct digital cores of dune sand and sandstone by CT scanning technology, and divide the whole core into multiple subvolumes and calculate their permeabilities. Then, we reassemble the subvolumes and compute the permeability for the whole core. This approach may lead to underestimation as the connectivity between subvolumes could be lost. To address this issue, we divide the whole core into different-sized subvolumes, and then use curve fitting to deduce the permeability of whole rock sample via extrapolation. The results show that the proposed method has improved accuracy, and is significantly faster than simulating the whole digital core, since the computation on subvolumes can be easily parallelized. This approach provides new ideas for accurate and efficient permeability estimation for digital porous media.