基于多层网格的限制边界LBM局部加密方法

Limited Boundary LBM Local Grid Refinement Method Based on Multi-Grid

中文关键词: 格子Boltzmann方法  多层网格  局部加密  计算加速

英文关键词: lattice Boltzmann method  multi-grid  local grid refinement  accelerated calculation

中文摘要:

      多层网格方法是进行局部网格加密的方案之一，具体实施灵活简便，但网格间过多的信息交换拖慢了模型的求解速度。本文在多层网格的基础上，提出了限制边界的局部加密方法，省略了细网格补丁在演化过程中与粗网格大量的信息交换，在保留标准LBM优势的情况下，缩短了计算时间，且拥有良好的数值稳定性；经过顶盖驱动流算例的数值实验验证，表明该方法拥有较高的计算精度，在高Reynolds数下该方法的表现依然良好，可以清晰呈现出流场的各项特征，相比于标准LBM有较大的加速比。

英文摘要:

      The multi-grid method is one of the solutions for local grid refinement, which is flexible and easy to implement. However, excessive information exchange between grids slows down the solving speed of the model. On the basis of multi-grid, a local refinement method with limited boundaries was proposed, which omits a large amount of information exchange between fine grid patches and coarse grids during the evolution process. While retaining the advantages of standard LBM, it shortens the calculation time and has good numerical stability; The numerical experiment verification of the lid-driven flow example shows that this method has high computational accuracy, and its performance is still good at large Re numbers. It can clearly present various characteristics of the flow field, and has a larger acceleration ratio compared to standard LBM.