中文摘要：

在基于小波分析的结构损伤识别研究中，小波函数的选择是首先要考虑的问题．选取了部分Db（Daubechies Ⅰ）族小波函数，采用小波概率神经网络方法，对一个4层钢框架进行损伤识别研究，探讨了选择不同Db族小波函数对损伤识别结果的影响．研究发现，正则性好、消失矩大等特性的小波函数其损伤识别效果最好．

英文摘要：

Several Db（Daubechies Ⅰ ） family wavelet functions are chosen in the study of a numerical model on four-story steel frame structure by using wavelet probabilistic neural network methods, and the effect of different wavelet functions on structural damage identification is discussed. This study shows that the wavelet function with good regularity and great vanishing moments is the best one in structural damage identification.