中文摘要：

将小波变换与奇异值分解（SVD）相结合,利用基于小波变换的能量分布函数,进行小波系数矩阵的奇异值分解,使结构损伤突变信号的奇异点放大,提高信号的分辨率和信噪比。根据信号突变时刻小波系数出现模极大值及其在不同尺度上的传播特性,可有效地检测出信号的奇异性特征,从而对结构损伤特征信号进行时域定位。通过对一试验简支梁损伤前后加速度信号的奇异性进行对比分析,验证了该方法的有效性和可行性。

英文摘要：

Wavelet transform is combined with singular value decomposition （SVD） filtering, and energy distribution function based on wavelet transform is used in SVD of wavelet coefficient matrix, the odd and break points of damnification signals are amplified, the resolving power and the signal noise ratio are improved. When the signals change abruptly, the modulus maximum of wavelet coefficients appear. According to the characteristics of transmitting on different scales of signal modulus maximum, mutant points of input signals can be detected effectively. Accordingly, the damnification characteristic signals on time domain can be positioned. In this paper, vibration acceleration signal singularity before and after damnification of a simply supported beam are analyzed. The method is validated by the singularity analysis of the experimental beam.