高层剪切型结构的参数识别与损伤检测

PARAMETER IDENTIFICATION AND DAMAGE DETECTION OF MULTI-LAYER SHEAR-TYPE BUILDING

本文首先提出了一种基于结构测量响应和神经网络技术进行结构参数识别的方法 ,然后根据剪切型结构的特点 ,提出了该类结构的参数识别策略与损伤检测方法 ,即根据需要可一次或分次识别结构的所有参数 ,并通过对比不同时期识别的结构刚度值 ,进行结构损伤判别及识别损伤位置与损伤程度。仿真计算结果表明 ,即使测量响应存在较高的随机噪声 ,本文方法识别的结构刚度值仍具有很高的精度 ,并能准确实现结构损伤判别与定位 ,显示了该方法用于实际工程结构损伤检测的潜力。

WT5BZ]In the paper,a methodology of structural parameter identification,which is based on structural responses and neural network technique,is proposed and then is used to develop strategy of parameter identification and method of damage detection on multi_layer shear_type building.Considering the sparse characteristic of structural parameter matrix,parameters could be idetified in partition.By comparing the stiffness data acquired at various stages during structure's life,structural damage can be distinguished and localized.The method is demonstrated by an example of 5_DOFs shear_type building.Results show that the method can accurately identify structural parameters and predict the location of stiffness change for different damage cases,even if measured responses are polluted with 5～10%zero_mean random noise. [WT5HZ]