Zobrazit v EDS

三维并行显式非连续变形分析接触判断与云计算研究进展.

Uloženo v:
Podrobná bibliografie
Název: 三维并行显式非连续变形分析接触判断与云计算研究进展.
Alternate Title: Research progress of contact detection and cloud computing for 3D parallel explicit discontinuous deformation analysis.
Autoři: 王熙1,2 xiwang_chn@foxmail.com, 李华明1, 武威2,3,4 weiwu@tongji.edu.cn, 朱合华2,3,4, 刘发波1, 张洪2
Zdroj: Hazard Control in Tunnelling & Underground Engineering. Sep2021, Vol. 3 Issue 3, p111-118. 8p.
Abstract (English): Contact detection and computation efficiency has always been the key problem of three-dimensional discontinuous deformation analysis(3D-PEDDA), which is one of the most powerful numerical methods. Contact detection suffers from low efficiency, contact indeterminacy, and the incapability to deal with concave blocks. The efficiency of the entire DDA program is also one of the most crucial bottlenecks. This paper introduced the multi-cover method and the last entrance plane method to deal with lacking of efficiency and indeterminacy of contact detection, as well as the local convex decomposition method for concave polyhedron. In addition, in order to improve the computational efficiency, 3D explicit DDA was adapted to unify data reading and writing mode, and parallelized by adding precompile instructions. In order to expand to high-performance cloud computing and supercomputers, the new 3D-PEDDA programs were all developed on Linux system, and preliminary experiments were carried out on a simple cloud virtual machine. Numerical examples verified the efficiency and accuracy of the newly developed parallel 3D-PEDDA, and it is promising that it can be easily extended to high-performance virtual machines or supercomputers to realize the analysis of large-scale projects. [ABSTRACT FROM AUTHOR]
Abstract (Chinese): 三维并行显式非连续变形分析方法(three dimensional parallel explicit discontinuous deformation analysis,3D-PEDDA)作为隧道与地下工程数值模拟中强大的数值方法之一,一直存在接触判断效率低、接触不确定性和难以处理凹体等问题,其整体计算效率也是瓶颈之一。介绍了多覆盖方法、最后侵入面法和局部凸分解方法,将其引进3D-PEDDA来解决接触问题。为提升整体计算效率,新的3D-PEDDA程序经过改编以统一数据读写模式,并通过添加预编译指令的方法实现并行化。新的3D-PEDDA程序全部在Linux系统上进行开发,并在简单的云端虚拟机上进行初步试验。数值算例验证了所开发的3D-PEDDA的效率、精度和运行在更高性能的云主机上的潜力. [ABSTRACT FROM AUTHOR]
Databáze: Academic Search Index
Popis
Abstrakt:Contact detection and computation efficiency has always been the key problem of three-dimensional discontinuous deformation analysis(3D-PEDDA), which is one of the most powerful numerical methods. Contact detection suffers from low efficiency, contact indeterminacy, and the incapability to deal with concave blocks. The efficiency of the entire DDA program is also one of the most crucial bottlenecks. This paper introduced the multi-cover method and the last entrance plane method to deal with lacking of efficiency and indeterminacy of contact detection, as well as the local convex decomposition method for concave polyhedron. In addition, in order to improve the computational efficiency, 3D explicit DDA was adapted to unify data reading and writing mode, and parallelized by adding precompile instructions. In order to expand to high-performance cloud computing and supercomputers, the new 3D-PEDDA programs were all developed on Linux system, and preliminary experiments were carried out on a simple cloud virtual machine. Numerical examples verified the efficiency and accuracy of the newly developed parallel 3D-PEDDA, and it is promising that it can be easily extended to high-performance virtual machines or supercomputers to realize the analysis of large-scale projects. [ABSTRACT FROM AUTHOR]
ISSN:20965052
DOI:10.19952/j.cnki.2096-5052.2021.03.12