基于自由响应信号与BP神经网络的结构损伤程度识别研究
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
提出了基于自由响应信号与BP神经网络的结构损伤程度识别方法:首先采用随机减量技术提取自由响应信号,利用AR模型进行动态建模,计算结构损伤前后的模型系数之间的欧式距离;其后以结构损伤前后的欧式距离作为特征向量,输入BP神经网络确定结构的损伤程度。四层海洋平台结构的数值模拟表明,该方法是可行的,且具有一定的抗噪声能力。
The method for structural damage degree identification based on free response signal and BP neural network is put forward:Firstly,the random decrement technique is used to obtain the free response signal and the AR model is adopted to create the dynamic model and calculate the Euclidean distance between model coefficients before and after structural damage;Secondly,the Euclidean distance before and after structural damage is used as the characteristic vector,which is put into BP neural network to determine the degree of structural damage.The numerical simulation of a four-floor offshore platform shows that the proposed method is feasible and has a certain noise resisting ability.
引文
[1]何浩祥,闫维明,马华,王卓.结构健康监测系统设计标准化评述与展望[J].地震工程与工程振动,2008,(4):154~159.
    [2]史文海,李正农,陈联盟.随机减量技术的研究现状及进展[J].防灾减灾工程学报,2008,(S):52~57.
    [3]刘彬,丁桦,时忠明.随机减量函数触发条件分析[J].振动与冲击,2007,(7):2731.
    [4]曹树谦,张文德,萧龙翔.结构振动模态分析[M].天津:天津大学出版社,2001.
    [5]吴令红等.基于时间序列的结构损伤检测[J].煤矿机电,2009,(4):71~75.
    [6]董长虹.MATLAB神经网络与应用[M].北京:国防工业出版社,2007.
    [7]于菲.基于时间序列分析与神经网络的结构损伤识别研究[D].青岛:青岛理工大学,2011.
    [8]苏金明,等.MATLAB工具箱应用[M].北京:电子工业出版社,2004.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心