一种两阶段结构损伤识别方法研究  

A Study on Structural Damage Identification by a Two-stage Method

作　　者：肖辉[1] 常军[2] 

机构地区：[1]苏州建设集团,江苏苏州215000 [2]苏州科技学院土木工程学院,江苏苏州215011

出　　处：《苏州科技学院学报（工程技术版）》2011年第3期23-27,共5页Journal of Suzhou University of Science and Technology （Engineering and Technology）

基　　金：建设部研究开发资助项目(2008-K2-35)

摘　　要：采用曲率模态结合模态应变能对结构的损伤位置和损伤程度进行了识别研究。首先采用改进的曲率模态识别出结构损伤的大体位置,然后采用模态应变能方法对损伤位置的单元进行损伤程度识别。在损伤程度识别时考虑了功能强大的BP神经网络方法。最后,为了验证该方法的有效性,采用了一数值模拟案例进行分析。数值模拟结果表明:曲率模态能够有效地识别出结构的损伤位置,模态应变能能够有效地识别出单元的损伤程度。The paper studies the identification of the structural damage location curvature mode with the modal strain energy. First, the curvature mode is used to structural damage, and then, the modal strain energy method is used to identify and extent identify by combining the general location of the damage extent of damage elements. During the damage extent identification by the modal strain energy, the BP neural network is used. Finally, a numerical simulation of a simply supported beam is used to evaluate the efficiency of the method. The results show that the curvature mode can effectively identify the structural damage location, and the modal strain energy can effectively figure out the damage extent of the damage elements.

关 键 词：损伤识别 曲率模态 模态应变能 简支梁 BP神经网络 

分 类 号：U441[建筑科学—桥梁与隧道工程]