Total solution for structural mechanics problems

被引:31
作者
Liu, GR [1 ]
Xu, YG [1 ]
Wu, ZP [1 ]
机构
[1] Natl Univ Singapore, MIT Alliance, Ctr Adv Computat Engn Sci, Dept Mech Engn, Singapore 119260, Singapore
关键词
mathematical models; inverse problems; sensitivity matrix; neural networks; structural mechanics;
D O I
10.1016/S0045-7825(01)00314-0
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
A concept of total solution for structural mechanics problems is proposed in this paper. It aims to establish a systematic approach to provide a comprehensive solution for practical structural mechanics problems, especially for traditional forward problems with incomplete input information (load, material property, boundary condition) and inverse problems with insufficient observations of effects (displacement, acceleration, stress, etc.). The approach for a total solution is to formulate such a structural mechanics problem as a parameter identification problem based on the forward solver of problem. All the unknown parameterized information in this forward model is determined through an iterative procedure of conducting alternately forward and inverse (or mixed) analyses. Algorithms of implementing inverse and mixed analyses are developed, which include a sensitivity matrix-based method and a modified adaptive neural networks method. Numerical investigations have been made to demonstrate the feasibility and validity of the proposed approach as well as the implementation algorithms. (C) 2001 Elsevier Science B.V. All rights reserved.
引用
收藏
页码:989 / 1012
页数:24
相关论文
共 39 条
  • [1] ANIKONOV YE, 1997, FORMULAS INVERSE ILL
  • [2] Anju A, 1997, INT J NUMER METH ENG, V40, P1015, DOI 10.1002/(SICI)1097-0207(19970330)40:6<1015::AID-NME101>3.0.CO
  • [3] 2-W
  • [4] On model updating using neural networks
    Atalla, MJ
    Inman, DJ
    [J]. MECHANICAL SYSTEMS AND SIGNAL PROCESSING, 1998, 12 (01) : 135 - 161
  • [5] NEURAL NETWORKS AND THEIR APPLICATIONS
    BISHOP, CM
    [J]. REVIEW OF SCIENTIFIC INSTRUMENTS, 1994, 65 (06) : 1803 - 1832
  • [6] An exact method for analysing sound reflection and transmission by anisotropic laminates submerged in fluids
    Cai, C
    Liu, GR
    Lam, KY
    [J]. APPLIED ACOUSTICS, 2000, 61 (01) : 95 - 109
  • [7] Adaptive neural networks for model updating of structures
    Chang, CC
    Chang, TYP
    Xu, YG
    [J]. SMART MATERIALS AND STRUCTURES, 2000, 9 (01) : 59 - 68
  • [8] CHANG CC, IN PRESS J SOUND VIB
  • [9] Doebling S.W., 1996, ALAMOS NATL LAB REPO
  • [10] DOEBLING SW, 1998, STATE ART STRUCTURAL