Granular materials, such as powders and grains, are prevalent in both natural and industrial processes, including those in the pharmaceutical and food-processing sectors. To optimize the processing of such materials, simulation technology based on Discrete Element Method (DEM) is widely employed. DEM uses physics-based contact force and heat models to simulate interaction between individual particles. In this study, we examine the force and heat models for spherical particles and present an implementation of DEM using the node-level short-range particle simulation library, AutoPas. Furthermore, we show a straightforward implementation of non-spherical DEM by approximating arbitrary particle shapes using subspheres, utilizing md-flexible, a molecular dynamics simulator integrated within AutoPas. Furthermore, to demonstrate the effectiveness of our approach, we provide simulation results from several scenarios, particularly a rotating square tumbler and a fluidized bed. In the rotating square tumbler, we observe particle heating resulting due to frictional interactions, as well as heat conduction along the tumbler walls. These thermal effects can either amplify or oppose each other, depending on the thermal settings of the walls. In the fluidized bed, solid particles are initially packed densely. As gas particles exert significant pressure on these solids, they disrupt the alignment of the solid particles, leading to a reduction in pressure and the formation of a fluidized state. Additionally, we highlight the benefits of AutoPas’s auto-tuning capabilities in improving the efficiency of these simulations. For the rotating tumbler, AutoPas selects the Verlet Cluster Lists algorithm for neighbor identification with the AoS (Array of Structures) particle data layout, which proves more efficient than other variants, such as the Linked Cells algorithm. In the fluidized bed scenario, which involves dynamic changes in particle density, AutoPas adapts by transitioning from Linked Cells with AoS to Linked Cells with SoA, and eventually to Verlet Cluster Lists with AoS, as the particle density increases.     «

Granular materials, such as powders and grains, are prevalent in both natural and industrial processes, including those in the pharmaceutical and food-processing sectors. To optimize the processing of such materials, simulation technology based on Discrete Element Method (DEM) is widely employed. DEM uses physics-based contact force and heat models to simulate interaction between individual particles. In this study, we examine the force and heat models for spherical particles and present an impl...     »