LAUSR

Pore characteristics play central roles in the transport of water and gas through granular soils, which is relevant to a wide range of geotechnical and geoenvironmental applications. In this study, a numerical procedure combining the discrete element method and pore network model is presented to relate the physically representative pore structure to the particle geometric features, and rapidly predict the macroscopic flow properties of granular assemblies. This numerical procedure is applied to investigate the effects of particle gradation and geometry on the pore characteristics and water retention curves of granular soils under isotropic compression. The results indicate that a larger uniformity coefficient ( C u ) of particle diameters results in a lower porosity. The porosity first decreases and then increases as the aspect ratio (AR) of the particles increases from 1.00 to 2.50 and reaches the minimum at AR = 1.75. As C u ranges from 1.00 to 2.16, the standard deviations and mean values of pore geometry and connectivity parameters, including pore and throat diameters, pore spacings, and coordination numbers, present linear correlations. The logarithm of air entry pressure decreases linearly with the increasing logarithm of mean pore diameter in response to the variations in both C u and AR. A higher pore diameter variance induced by a smaller C u produces a lower fitting parameter m in the van Genuchten equation for water retention curves.