Approach for Generating Discrete-Element Models of Geomaterials Capturing Statistical Characteristics of Particle Shapes and Gradations

Geomaterials used in infrastructure, e.g., crushed stone, gravel, and sand, exhibit distinct irregularity and diversity in particle shapes and broad particle size distributions (PSDs) but with statistical characteristics. Accurately reproducing both PSD and particle-shape statistics in discrete-element models remains challenging when irregular particle geometries are involved. To bridge this gap, this paper proposes a model generation approach that constructs a shape-matched particle library according to a prescribed distribution of a selected shape indicator and enforces PSD compliance through a numerically based sieving algorithm coupled with an improved particle scaling method. The sieving algorithm evaluates particle passability using minimum projection dimensions and incorporates flat-particle effects through a flatness ratio and two passing probability parameters. The scaling method determines a feasible scaling factor via a cyclic search and applies centroid-based isotropic scaling to preserve particle shape. Experimental sieving verification and numerical examples demonstrate that the proposed approach can simultaneously match a target PSD and a target shape-indicator distribution. The proposed approach effectively enhances the accuracy of discrete-element models in particle modeling and holds wide potential applications in the modeling of geomaterials.