A strategy to determine DEM parameters for spherical and non-spherical particles
In Discrete element method (DEM) simulations the choice of appropriate contact parameters is significant to obtain reasonable results. Particularly, for the determination of DEM parameters for non-spherical particles a general straightforward procedure is not available.
Therefore, in a first step of the investigation here, methods to obtain the friction and restitution coefficients experimentally for single particles [Polyoxymethylene (POM) spheres and quartz gravel] will be introduced. In the following, these predetermined DEM coefficients are used as initial values for the adjustment of bulk simulations to respective experiments. In the DEM simulations, the quartz gravel particles are represented by non-spherical particles approximated by clustered spheres. The best fit approximation of the non-spherical particles is performed automatically by a genetic algorithm.
In order to optimize the sliding and rolling friction coefficients for DEM simulations, the static and dynamic angle of repose are determined from granular piles obtained by slump tests and rotating drum experiments, respectively. Additionally, a vibrating plate is used to obtain the dynamic bed height which is mainly influenced by the coefficient of restitution.
The adjustment of the results of the bulk simulations to the experiments is conducted automatically by an optimization tool based on a genetic algorithm. The obtained contact parameters are later used to perform batch-screening DEM simulations and lead to accurate results. This underlines the applicability of the in parts automated strategy to obtain DEM parameters for particulate processes like screening.
Keywords: parameter determination. automatic adjustment of DEM parameters, arbitrary shaped particles, screening.