东北大学学报（自然科学版）

东北大学学报(自然科学版) ›› 2007, Vol. 28 ›› Issue (10): 1497-1500.DOI: -

• 论著 • 上一篇    下一篇

基于Neumann展开的转角信息识别近似算法

康玉梅;陈耕野;李州;   

  1. 东北大学资源与土木工程学院;东北大学资源与土木工程学院;中国市政工程西北设计研究院 辽宁沈阳110004;辽宁沈阳110004;甘肃兰州730000
  • 收稿日期:2013-06-24 修回日期:2013-06-24 出版日期:2007-10-15 发布日期:2013-06-26
  • 通讯作者: Kang, Y.-M.
  • 作者简介:-
  • 基金资助:
    国家自然科学基金资助项目(50504005)

Approximate algorithm based on Neumann expansion for rotational response identification

Kang, Yu-Mei (1); Chen, Geng-Ye (1); Li, Zhou (2)   

  1. (1) School of Resources and Civil Engineering, Northeastern University, Shenyang 110004, China; (2) China Northwest Municipal Engineering Design and Research Institute, Lanzhou 730000, China
  • Received:2013-06-24 Revised:2013-06-24 Online:2007-10-15 Published:2013-06-26
  • Contact: Kang, Y.-M.
  • About author:-
  • Supported by:
    -

RichHTML

0

PDF (PC)

565

摘要： 针对结构识别中转角信息难以测量的问题,提出了基于Neumann展开的转角信息识别问题近似算法.在结构识别中,由于各单元刚度参数未知,利用结构有限元分析中按静力凝聚的方法由平移求结构的转角是无法实现的,在结构参数差异不大的情况下,可以采用Neumann展开方法得到逆矩阵,进而得到结构的转角信息.理论分析和数值算例表明:该算法在结构参数差异不大的情况下适用,在材料参数全同的情况下是精确解答.其范数误差一般较小,而绝对量较小的分量误差较大.

关键词: Neumann展开, 转角响应, 结构识别, 静力凝聚法, 有限单元法

Abstract: Aiming at the problem that the rotational response is difficult to measure in structural identification, an approximate algorithm based on Neumann expansion theory is developed for rotational response identification. The rotational response cannot be calculated according to structural displacement response based on the static condensation algorithm in FEM analysis since the stiffness parameters are unknown in structural identification. When the difference among structural parameters are not big, the inverse matrix is available by introducing Neumann expansion to get the information on rotational response. Theoretic analysis and numerical examples are given to illustrate that the approximate algorithm based Neumann expansion is applicable when structural parameters differ not widely from each other, and an accurate solution can be obtained if the structural parameters are all the same. The norm errors of the approximate algorithm is generally small, while the smaller the absolute value is, the greater the component errors will be.

中图分类号: