A GPU-based DEM model for the pebble flow study in packed bed: Simulation scheme and validation

Abstract

This study developed a GPU-based DEM (GPU-DEM) model for the pebble flow in the packed bed. The details of the GPU-DEM scheme have been illustrated. The uniform mesh neighbor-search method, unified memory addressing (UMA), and special boundary conditions were incorporated in the GPU-DEM model. Cases of the direct impact of two identical particles, the angle of repose of particles drained from a lifted hopper, and the circulating hopper to mimic a pebble flow in High-Temperature Gas-cooled Reactor (HTGR) reactors were carried out to systematically verify the applicability, computational efficiency, and accuracy of the GPU-DEM model. Related results, like the unloading speed, porosity, and velocity distribution, obtained by the GPU-DEM model have been validated too. In addition, the effects of single-precision floating-point number computation on GPU were discussed. The results demonstrated that the GPU-DEM can achieve 18– 20 times acceleration while maintaining accuracy, even when single-precision floating-point numbers were used for calculation. Therefore, it is a powerful tool to explore the particle flows in chemical engineering applications.