基于时间序列与神经网络的空间网壳结构损伤检测方法
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
针对网壳结构损伤识别中遇到的模态信息不完备、模态局部化和跃迁现象以及结构自由度巨大所带来的建模精度低等实际困难,提出了一种适用于网壳结构损伤检测的方法——基于时间序列与神经网络的结构损伤检测法.首先,利用ANSYS瞬态分析模块提取结构加速度响应;然后,通过MATLAB(Matrix Laboratory)编程软件建立时间序列自回归模型即AR(Auto-Regressive)模型并提取模型参数,从而判定结构损伤是否发生;最后,对AR模型参数作差并对其进行组合作为损伤特征向量输入到由MATLAB神经网络工具箱建立的BP(Back Propagation)神经网络,对结构损伤进行定位.以凯威特单层球面网壳作为研究对象进行数值模拟,验证该方法的有效性.
For the difficulties of the latticed shell structure damage identification,such as the modal information incomplete,mode localization and transition phenomenon and the low accuracy of modeling by structure of large degree of freedom and what not,this paper suggests a kind of method suitable for latticed shell structure damage detection based on time sequence and neural networks.By using ANSYS transient analysis module,it makes the time series autoregressive(AR)mode through MATLAB programming software and picks up its parameters so as to identify the occurrence of structural damage.It also makes the AR model parameters for difference,carries on the combination as damage characteristic vector,and inputs the BP neural network in MATLAB neural network toolbox to estimate the damage location of the structure.With a single spherical latticed shell structure as the research object,the numerical simulation is used to verify the validity of this method.
引文
[1]胡庆昌.建筑结构抗震减震与连续倒塌控制[M].北京:中国建筑工业出版社,2007.HU Qing-chang.The control of shock absorption and progressive collapse in the building[M].Beijing:China Architecture&Building Press,2007.
    [2]马高,屈文忠,陈明祥.基于时间序列的结构损伤在线诊断[J].武汉大学学报:工学版,2008,41(1):81-85.MA Gao,QU Wen-zhong,CHEN Ming-xiang.Time series-based structural on-line damage detection[J].Engineering Journal of Wuhan University,2008,41(1):81-85.
    [3]吴令红,熊晓燕.基于时间序列分析的结构损伤检测[J].煤矿机电,2009(4):71-75.WU Ling-hong,XIONG Xiao-yan.Structural damage detection based on time series analysis[J].Colliery Mechanical&Electrical Technology,2009(4):71-75.
    [4]倪博溢,萧德云.MATLAB环境下系统辨识仿真工具箱[J].系统仿真学报,2006,18(6):1493-1496.NI Bo-yi,XIAO De-yun.The identification simulation toolbox in MATLAB environment system[J].Journal of System Simulation,2006,18(6):1493-1496.
    [5]陈伟,瞿伟廉,姜正国.人工神经网络技术在结构损伤诊断中的应用[J].世界地震工程,2002,18(1):146-149.CHEN Wei,QU Wei-lian,JIANG Zheng-guo.The application of artificial neural network technology to structural damage detection[J].World Information on Earthquake Engineering,2002,18(1):146-149.
    [6]杨晓明,焦阳,郭余锁.基于BP神经网络的结构损伤检测[J].中国新技术产品,2010(16):27.YANG Xiao-ming,JIAO Yang,GUO Yu-suo.Structural damage detection based on BP neural network[J].China New Technologies and Products,2010(16):27.
    [7]张长江.基于BP神经网络的平面框架结构的刚度损伤诊断的理论研究[D].乌鲁木齐:新疆大学,2006.ZHANG Chang-jiang.Research on the stiffness of frame structure damage diagnosis based on BP neural networks of theory study[D].Urumqi:Xinjiang University,2006.
    [8]董长虹.Matlab神经网络与应用[M].北京:国防工业出版社,2007.DONG Chang-hong.Matlab neural network and applications[M].Beijing:National Defense Industry Press,2007.

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