Stochastic generation of virtual porous media using a pseudo-crystallization approach

A new algorithm is proposed to generate virtual porous media based on predefined porosity and particle size distribution (PSD). Synthetic packings of different porosities, sphericities, and PSD curves are numerically constructed using the proposed approach. Based on these virtual porous material empirical correlations are tuned for estimation of permeability and formation factor as functions of porosity and median grain size. The proposed procedure is inspired from crystallization concept addressing the capability of dealing with tight media with no porosity limitation and the flexibility to easily control the grain sphericity. The post-processing results, including estimation of permeability, formation factor and electrical resistivity, are compared with experimental data. Besides, a sensitivity analysis is conducted to elucidate the impact of grain irregularity (sphericity) on permeability.