LAUSR

Abstract This study presents a novel Arcs-based discrete element method (ArcDEM) for efficient simulation of realistic granules with arbitrary convex and concave 2D particle outlines. In the proposed ArcDEM, a series of computational geometry algorithms are first developed to identify the convex corners and concave troughs of an arbitrary-shaped particle outline. Then, the circle-growing technique and the least squares method are combined to establish the Arcs-based particle that can represent the whole particle outline with multi-connected inward and outward arcs. Next, a new algorithm for efficient overlapping detection and precise contact resolution is developed for the Arcs-based particles. Finally, the ArcDEM is developed as a simulation tool with several implemented contact force laws and particle motions solved by an explicit time integration. To validate the feasibility and efficiency of the proposed ArcDEM, several numerical examples are performed, including (1) random allocation of non-overlapping irregular particles, (2) random packing of super-elliptical particles of different aspect ratios and blockiness, and (3) simulation of repose angle and biaxial compression tests of realistic rock particles with various roundness . The developed ArcDEM tool shows a powerful capability of numerically investigating the macro- and micromechanical properties of realistic convex and/or concave granular particles.