Computational performance of contact search during DEM simulation of hopper filling

Abstract

The Discrete Element Method (DEM) recently presents numerical tool widely used in simulation of heterogeneous solids. The method captures dynamic behaviour of solid by considering motion and interaction of material particles. The most computationally expensive procedure of the DEM is search of contacts for particles, therewith each pair of colliding particles has to be identified. To investigate and accelerate computational performance of the DEM, computational performance study of two popular cell and sorting contact search algorithms was done in the present work. Both algorithms are implemented into DEMMAT code. Series of numerical tests was conducted for the above purposes by considering distributed filling of hopper. Mono-sized and poly-dispersed assemblies of elastic spherical particles were examined and computational performance during particle flow is illustrated.