基于曲率模态的结构损伤识别

Structural damage identification based on curvature mode

随着国内桥梁数量已趋于饱和,更多的桥梁结构进入运维阶段,对老旧桥梁的损伤识别、对新建桥梁的健康监测已经成为亟待解决的社会问题。基于桥梁结构的动力特征响应信息,利用人工神经网络的数据分析方法,提出桥梁结构损伤识别的三步识别法:首先利用固有频率判别桥梁结构是否发生了损伤;若发生损伤,则利用曲率模态分析损伤位置;在损伤位置确定后,利用人工神经网络进一步对数据的隐含信息进行提取,识别损伤程度。研究证明人工神经网络在桥梁结构损伤识别方面有较强的适用性,可以为今后的发展提供一定参考。

As the number of Bridges in China tends to be saturated and more bridge structures enter the operation and maintenance stage,damage identification of old bridges and health monitoring of new bridges have become urgent social problems to be solved.Based on the dynamic characteristic response information of bridge structure,a three-step identification method for bridge structure damage identification is proposed by using the data analysis method of artificial neural network.If damage occurs,the curvature mode is used to analyze the damage position.After the damage location is determined,the artificial neural network is used to further extract the hidden information of the data and identify the damage degree.The research proves that the artificial neural network has strong applicability in the damage identification of bridge structure,which can provide some reference for future development.