Reconstructing anisotropic porous media from thin section images

Abstract Parameters like sediment’s origin, transport distances or next influencing factors such as cementation or dissolution, cause almost all natural porous media to exhibit some degree of anisotropy. Using the X-ray CT scanning approach to provide a virtual environment for the calculation of such properties has its own issues, including its low resolution, which gives an incomplete description of internal structure, along with its relatively high costs. Porous media can be reconstructed with the help of statistical techniques from existing high-resolution thin section images. In this work, in this article, a micro-CT file available on the Imperial College website is first selected, which has been obtained from a disordered anisotropic sandstone core sample. specific statistical algorithms are applied to the top and side images of the core sample, which represent the input images for the Multiple-Point Statistics (MPS) technique. After that, to construct artificial anisotropic media models, Optimization algorithms are used to polarize the phase occurrence potentials in the principal directions, which are obtained from the two thin-section images. Results show that applying optimization algorithms to polarize occurrence potential functions can effectively develop artificial porous media that exhibit tensor properties closely matching those calculated from high-resolution micro-CT images of the actual sandstone core sample, such as permeability and thermal conductivity tensors.