首页> 外文期刊>Computer Methods in Applied Mechanics and Engineering >Large-scale parallel flow analysis based on free mesh method: a virtually meshless method
【24h】

Large-scale parallel flow analysis based on free mesh method: a virtually meshless method

机译:基于自由网格方法的大规模并行流分析:一种几乎无网格的方法

获取原文
获取原文并翻译 | 示例

摘要

In this paper, large-scale parallel computing based on a virtually meshless method called the 'free mesh method' for incompressible viscous flow with three-million D.O.Fs or one-million nodes is presented. By using the free-mesh method, a seamless system from the stage of nodal information input without node-element connectivity information to the stage of the system equations solver is successfully developed. The system can compute the global mass, advection, diffusion, gradient and divergence matrices node-by-node independently from only nodal information on a parallel computer. With techniques of node generation based on nodal density control and a kind of hash sorting method for search nodes, a lid-driven cavity problem having three-million D.O.Fs can be successfully solved. In the problem, the process of constructing system equations from nodal information by the free-mesh method, requires about l42 s using 64 processors of the Hitachi SR220l.
机译:在本文中,提出了基于虚拟无网格方法(称为“自由网格方法”)的大规模并行计算,该方法用于具有300万个D.O.Fs或一百万个节点的不可压缩粘性流。通过使用自由网格方法,成功开发了一个无缝系统,从没有节点元素连接信息的节点信息输入阶段到系统方程求解器阶段。该系统可以独立于并行计算机上的节点信息来逐节点计算整体质量,对流,扩散,梯度和散度矩阵。借助基于节点密度控制的节点生成技术和一种用于搜索节点的哈希排序方法,可以成功解决具有300万个D.O.Fs的盖驱动腔问题。在该问题中,使用日立SR220l的64个处理器,通过自由网格方法根据节点信息构造系统方程的过程大约需要142 s。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号