Optimal description of two-dimensional complex-shaped objects using spheropolygons

Recent advances in the morphological description of particles in granular material systems allow two-dimensional complex-shaped particles to be realistically simulated using spheropolygons, i.e., the Minkowski sum of a disk and a polygon. For identical numbers of vertices, spheropolygons achieve a better description of shapes than polygons, but require that the optimal spheroradius be determined. Here we propose a method for generating spheropolygons that optimizes the description of particle morphologies, i.e., minimizes the error images and the numbers of vertices. Because the error images of individual particles are a proxy for the accuracy of granular matter flow calculations, while the numbers of vertices are a proxy for the computational time, the method is optimally applicable to discrete element methods. We demonstrate the proposed method using pebbles, gravel, and crushed shells.