An adaptive-gridding lattice Boltzmann method with linked-list data structure for two-dimensional viscous flows

摘要

An adaptive mesh refinement technique for lattice Boltzmann method (LBM) is proposed in this paper. It combines hierarchical linked-list data structure and the LBM calculation. Based on uniform meshes, the adaptive algorithm refines the meshes by constructing the linked-lists of nodes, cells and levels for mesh levels refined. To guarantee the stability of numerical scheme, quadratic bubble function for the nodal momentum is used to interpolate in the LBM calculation of different mesh levels. For the flows of relatively higher Re, large Eddy simulation (LES) is adopted to solve turbulence problems. Because of the use of adaptive technique, the computational time can be cut and accurate flow field information can be captured. OpenMP parallel for linked-list data structure is used to improve computational efficiency. To verify the present method, flows over circular cylinder at Re = 40, 300, 500, 1,000 and 3,900 and NACA0012 airfoil at Re = 105 for AOA = 4° are simulated.