A dual-grid approach to speed up large-scale CFD-DEM simulations

Abstract

With the development of heterogeneous computing, the discrete element method (DEM) and computational fluid dynamics (CFD) coupling has become an effective approach to simulate large-scale particle–fluid systems. With various models and algorithms accelerating the DEM part significantly, the CFD part becomes an efficiency bottleneck, especially for large-scale simulations. A dual-grid approach is proposed to address this issue. On the fine grid, the fluid velocity and particle–fluid interaction are solved without compromising accuracy in capturing flow details, and the pressure is solved on the coarse grid to ensure mass conservation to boost the CFD part. This approach is validated by simulating several gas–solid fluidized beds, a performance demonstrating speedup of 2.6 for low workload and 10.1 for high workload.