Title:A comprehensive study of coupled LBM-DEM with immersed moving boundary

Abstract:A systematic study is carried out on a fully resolved fluid-particle model which couples the Lattice Boltzmann Method (LBM) and the Discrete Element Method (DEM) using an immersed moving boundary technique. Similar algorithms have been reported in the past decade, however, the roles of major model parameters are yet to be fully understood. To examine various numerical errors, a series of benchmark cases with a wide range of Reynolds number are performed, starting from a single stationary particle to multiple moving particles. It is found that for flow with low and intermediate Reynolds numbers, 20 fluid cells per one particle diameter are necessary to achieve sufficient accuracy (within 5%). For a flow with high Reynolds number, a turbulence model shall be incorporated so that the effects of unresolved small eddies can be captured in an accurate and efficient manner. Besides, the LBM-DEM results are also sensitive to the relaxation time, especially when the spatial resolution is inadequate. A large relaxation time can introduce additional diffusion of fluid momentum into the fluid-particle system, leading to weakened hydrodynamic interactions. By choosing a small relaxation time greater than the lower limit 0.5, a small fluid compressibility error and a strong coupling between fluids and particles can be achieved, at the cost of computational effort. The test cases also demonstrate the capability of LBM-DEM to describe the rheology of particle suspensions by capturing the pore-scale hydrodynamic interactions. Finally, a guideline for quickly establishing a high-quality LBM-DEM model is provided.