非凸颗粒离散元模型的多孔介质渗透性能研究OA

Porous media of random packings of non-convex particles are ubiquitous in energy,geotechnics and civil engineering,of which the internal pore structure critically influences fluid transport properties.To ad-dress the limitations of traditional methods in predicting permeability performance,this study proposes an inte-grated analysis framework linking grain-scale parameters to macroscopic permeability.First,3D non-convex particle models are generated via a parametric contraction-deformation method.Next,high-density non-convex random packings are efficiently simulated using a GPU-accelerated discrete element method coupled with level-set contact detection algorithm.Finally,the voxelized pore model is solved based on the Navier-Stokes equations to reveal the influence mechanism of particle morphology on absolute permeability.Results show that higher particle non-convexity and surface complexity significantly degrade pore connectivity and permeability.Packings of quasi-spherical particles with uniformly distributed contraction centers exhibit better permeability than those with randomly distributed ones.Additionally,a moderately broad particle size distribution enhances pore network transport.This work provides robust theoretical basis and numerical tools for the performance optimization of irregular-particle porous media.