Predicting the permeability coefficient of polydispersed sand via coupled CFD–DEM simulations

Permeability is one of the most fundamental properties of soil, which governs many geotechnical engineering problems. However, among the published equations that predict the permeability coefficient k, the influence of a wide range of particle size distributions (PSDs) on k was not considered accurately. This research aims to quantify the influence of PSD on k and propose a prediction equation for k. Coupled computational fluid dynamics–discrete element method (CFD–DEM) simulations were conducted to reproduce the constant head permeability tests of different soil samples. The results show that among the various drag models implemented in the CFD–DEM, the Syamlal–O’Brien drag model leads to the highest accuracy in simulating sand permeability. The permeability coefficient is proportional to the square of the Sauter mean diameter of the polydispersed particle system. Therefore, the prediction equation for k considering the PSD of the particle system is proposed based on the particle gradation and the size characteristics Cu, Cc, and d10. The proposed equation is verified well by published experimental data. Additionally, the proposed equation can calculate the k of irregular calcareous sand by representing the effect of particle irregularity with the Kozeny–Carman (KC) constant.