Rapid multiscale pore network modeling for drainage in tight sandstone

Abstract

Tight sandstone possesses multiscale pores. Capturing and modeling the important pore properties into a pore network for drainage is challenging and computationally expensive. We present a method for modeling of a multiscale pore network, which enables us to handle huge rock micro computed tomography (μCT) images with high computational efficiency and minimal experimental input data. This method uses micro-links to represent the upscaled transport properties of micropores unresolved by the image and connects pore clusters that have topological relationships in the image to extract a connected pore network from it. The size distribution of micro-links is optimized and determined by experimental permeability. The effects of the number distribution of micro-links on flow properties are investigated to obtain the robust parameter. The presented method is validated on a digital core and the published experimental data. This method is successfully applied to rapidly evaluate the influence of mineral components on drainage behavior in two tight sandstone rocks, which involves a large number of computations. The two μCT images both contain more than 109 voxels and are handled by a personal computer, proving the potential in rapidly, accurately, and conveniently handling huge tight sandstone μCT image for evaluating the drainage characteristics.