(170j) An Adaptive Lattice Boltzmann Method for the Simulation of Gas-Liquid Flows

The Lattice Boltzmann Method (LBM) has proved to be an effective numerical technique for multiphase flow simulations. However, the original LBM algorithm is restricted to uniform grid size throughout the entire computation domain, while in many practical situations the use of different grid sizes in different regions is more desirable. In this study, the LBM based on the pseudo-potential approach proposed by Shan and Chen is coupled with the adaptive mesh refinement (AMR) in order to achieve both high accuracy and low computation cost. The resolution of the grid in different parts of the computation domain changes dynamically, and the region near the gas-liquid interface is always discretized on the finest grid level. This newly developed AMR-LBM algorithm is described in detail, with emphasis on the interpolation/averaging process at the interface between different grid refinement levels. Simulations of a single bubble rising in a viscous liquid under different conditions are carried out, and the numerical results are compared against experimental results in the literature to validate the code. To prove the advantage of the new algorithm, the AMR-LBM results are compared with traditional LBM which runs on uniform grids. The improved resolution near the gas-liquid interface results in a more accurate bubble shape as well as enhanced numerical stability. It is expected that the AMR-LBM algorithm will greatly enhance the applicability of the LBM in realistic gas-liquid flows that occur in industrial applications.