Stochastic approach to simulation of interaction of multi-spherical particles

Abstract

Particle-based numerical simulation methods are widely applied in many areas. The particle shape is important factor playing decisive role in evaluation of particle’s interaction behavior on various scales. A general so called multi-sphere (MS) approach for representation of non-spherical particles by rigidly connected multi-spheres is recently explored in the Discrete Element (DEM) and applied to various shapes [1-3]. Contact detection efficiency and simplicity of implementation using sphere-to-sphere contact, is the main advantage of the multi-sphere model. Approximation the non-smooth spherical particle by applying MS models was considered in [4]. The studies performed shows that the multi-sphere method has certain limitations when used for the approximation of a non-spherical body. Moreover, increase of the approximation degree may introduce new errors itself at least on the single grain level and can be computationally inefficient. In the light of the above findings and the scattering of the initial data, application of the MS models to complex shapes in a frame of deterministic approach remains questionable. The stochastic approach is suggested in this paper to evaluate the performing of the MS models. Although our investigation is limited to normal sphere-plane contact, the concept may be easy extended to other shapes and hopefully to multi-particle problems.