首页 | 本学科首页   官方微博 | 高级检索  
     检索      

不规则结构有限元分析的并行处理及其在YH-1上的实现
引用本文:周树荃,邓绍忠.不规则结构有限元分析的并行处理及其在YH-1上的实现[J].南京航空航天大学学报,1992(1).
作者姓名:周树荃  邓绍忠
作者单位:南京航空学院数理力学系 (周树荃),南京航空学院数理力学系(邓绍忠)
摘    要:在结构分析中,有限元方法是一个重要的数值方法。近些年来,关于单元刚度矩阵计算与总刚度矩阵合成的并行处理问题,A.K.Noor,梁维泰等人已进行了富有成效的工作,但他们的工作都是面向几何形状规则的结构分析问题。本文结合YH—1机的特点,对不规则结构分析问题,在提出单元分组技术的基础上,给出了一个单元刚度矩阵计算的并行算法ESVC,和变带宽存储格式下一个总刚度矩阵合成的并行算法ESVS。通过在YH—1并行机上对实例的计算结果表明:当同时计算的单元数r取120时,加速比s可达9.5,且r愈大,s就愈高。

关 键 词:计算数学  计算力学  有限元法  并行处理  结构分析

Parallel Processing of Finite Element Analysis for Irregular Structures and Its Implementation on YH-1 Computers
Zhou Shuguan Deng Shaozhong.Parallel Processing of Finite Element Analysis for Irregular Structures and Its Implementation on YH-1 Computers[J].Journal of Nanjing University of Aeronautics & Astronautics,1992(1).
Authors:Zhou Shuguan Deng Shaozhong
Abstract:Finite element method is an important numerical method for structural analysis. In recent years, A. K. Noor, Liang Weitai, et al have done some effective researches on parallel processing for computing elements stiffness matrices and assembling the total stiffness matrix, but these researches are aimed at regular structural analysis problems. In this paper, taking aim at the problems of irregular structural analysis, we put forward the method for parting elements, and, then, raise a parallel computing elements stiffness matrices method called ESVC and a parallel assembly method of the total stiffness matrix called ESVS, when we adopt variable bandwidth row storage. Computational results on YH-1 parallel computer for a model structural analysis problem show that when the number of elements in once parallel computations r is 120, the ratio speed-up of the execution time using serial algorithms to the execution time using parallel algorithms achieves 9. 5; and the bigger the r , the greater the s .
Keywords:computational mathematics  computational mechanics  finite element method  parallel processing  structural analysis
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号