基于Schwarz混乱松弛法的压铸充型模拟并行分区算法  

Parallel domain decomposition algorithm based on Schwarz chaotic relaxation for die casting mold filling simulation

作　　者：王罡[1] 熊守美[1] 柳百成[1] 

机构地区：[1]清华大学机械工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2005年第5期581-584,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家"九七三"重点基础研究项目(G2000067208-3);国家"八六三"高技术项目(2001AA331030-04)

摘　　要：压力铸造充型过程数值模拟是压铸计算机辅助工程应用的一个重要方面,而计算规模和效率一直是这一模拟过程中的关键问题。为扩大求解规模,该文应用Schwarz混乱松弛的思想,设计了一种异步并行算法,并在并行机群计算系统实现。该算法将整体的求解域分解为多个子域,每个子域的计算作为子任务分配给并行计算系统内的节点,每个节点可异步求解。采用通信机制进行数据交换,实现全局收敛。这种算法能够有效利用多节点优势,计算规模随节点数呈线性增加。对气缸盖罩盖零件进行了充型模拟验证,获得了满意结果。Numerical simulations of the die-casting mold filling process is a very important aspect of die-casting computer- aided-engineering applications but the calculation scale limits effective use. The Schwarz chaotic relaxation method was used to design an asynchronous parallel algorithm to improve the solution speed. The algorithm was implemented in a parallel computer cluster calculation system, which decomposes the solution domain into many sub-domains. The tasks in each sub-domain are distributed to a node in the cluster with each node computing asynchronously. A communication algorithm controls data exchange between nodes and global convergence of the whole solution domain. Mold filling of an engine block head was simulated with satisfactory results. The results show that the algorithm takes full advantage of the nodes and that the computation speed linearly increases with increasing number of nodes.

关 键 词：压铸充型 并行计算 区域分解 Schwarz混乱松弛 

分 类 号：TG211[金属学及工艺—铸造]