Abstract: The lattice Boltzmann method (LBM) is a mesoscopic numerical simulation method developed in recent decades which has good ability of dealing with complex boundary and parallel computing.The simulations were performed for back facing step flow by LBM, and the characteristics of the back step flow under different Reynolds numbers were obtained. The research results showed that, in the case of low Reynolds number, the distance between the reattachment point and the separation point became larger as the Reynolds number increased, the value of K increased with the increase of Reynolds number, while the growth rate gradually became smaller. In the case of high Reynolds number, the distance between the reattachment point and the separation point increased with the increase of the Reynolds number, and the growth rate gradually decreased. When Re=8000, the distance between the reattachment point and the separation point tended to be stable. The value of K first became larger as the Reynolds number increased, and the increase rate became smaller as the Reynolds number increased, and eventually tended to be stable. For the back step flow under high Reynolds number, the simulation was performed by the LES⁃LBM method, which was in good agreement with the previous experimental data. It showed that the LBM⁃LES model had high accuracy for simulating the back facing step flow after high Reynolds number.

Key words: Back facing step flow, Lattice Boltzmann method, Flow law, Large eddy simulation, Reynolds number

摘要： 格子玻尔兹曼方法（LBM）是近几十年来发展的一种介观尺度的数值模拟方法，具有良好的处理复杂边界的能力和并行计算能力。利用LBM对后台阶流动进行了数值模拟，得到了不同雷诺数下后台阶流动的特性。研究结果表明，在低雷诺数情况下，台阶后再附着点与分离点的距离随着雷诺数的增加而变大；在高雷诺数情况下，台阶后再附着点与分离点的距离随着雷诺数的增加而增加，增长幅度逐渐降低，当Re=8 000时，再附着点与分离点之间的距离趋于稳定，并达到峰值；随着雷诺数的继续增大，这个距离开始减小，最终趋于一个定值。对于高雷诺数下的后台阶流动，采用LES⁃LBM的方法进行模拟，与前人的实验数据吻合得很好，表明LBM⁃LES模型对高雷诺数后台阶流动进行模拟具有较高的精确性。

关键词: 后台阶流动, 　格子玻尔兹曼方法, 　流动规律, 　大涡模拟, 　雷诺数