Critical appraisal of pore network models to simulate fluid flow through assemblies of spherical particles

Abstract

Coupled numerical models considering fluid flow and particle movement enable fundamental analyses of a variety of phenomena in geomechanics including seepage-induced instabilities. Amongst the various CFD (Computational Fluid Dynamics)-DEM (Discrete Element Method) coupled frameworks which have been proposed, Pore Network Models (PNMs) have the potential to simulate fluid flow in granular materials accurately with a low computational cost to enable simulations on Representative Volume Elements (RVEs). However, the current models of the local conductance between the connected pores are very simple, limiting the accuracy of PNMs. This study develops novel local conductance models by detailed analysis of existing analytical studies of fluid flow through different 3D lattice packings of uniform spheres. The performance of these new models relative to existing, simpler models is demonstrated using CFD simulations in which the flow in the pore space of random assemblies of polydisperse spheres is accurately resolved. The analyses show that the new models proposed here can more accurately predict the local and global permeabilities of specimens with a wide range of void ratios and polydispersities. These models do not require any optimisation via merging pores so that they can efficiently simulate systems with an evolving pore space topology.