并行有限元网格生成方法及其应用

摘要

大型工程数值仿真中，在前处理阶段需要生成千万甚至亿量级的网格，传统的串行网格生成方法由于内存和时间的限制，难以处理如此规模的网格。针对此问题，本文提出了一种大规模网格并行生成方法。首先基于推进波前法对几何模型进行初始体网格划分，接着利用图论理论进行区域分解，并通过表面单元恢复保持其几何精度，然后通过分裂法进行网格的并行生成。将所述方法应用到实际大型工程数值仿真前处理阶段，结果表明所述方法可以获得较好的并行效率，同时所产生的网格质量可以满足后续计算需要。%In the numerical simulation of large‐scale engineering project ,ten or even hundred millions of grids are needed at the pre‐processing stage .But the traditional serial grid generation method cannot handle this scale of mesh because of the bottle‐neck of time and memory .Aiming at this problem ,a parallel large‐scale mesh generation method is studied and proposed .First ,the initial grids are created on the geometry model using the advancing front mesh generation method .Then ,the domain decomposition is executed by using the graph theory .And the geometry accuracy is retained through surface recovery . At last the spilt method is used for the parallel mesh generation .The proposed method is utilized at the pre‐processing stage of the simulation of practical major engineering projects .The results show that this method can obtain good parallel efficiency ,and the quality of the mesh can satisfy the requirement of following calculation .