基于曲率模态和小波奇异性的结构损伤识别

Structural Damage Detection Based on Analysis of Curvature Mode and Wavelet Singularity

传统的傅立叶变换适合确定一个函数奇异性的整体性质,而难以确定奇异点在空间的位置及分布情况。小波变换具有在时域和频域局部放大的性质,在工程中获得广泛应用。由于曲率模态具有较高的灵敏度,仅仅需要较低阶的模态信息就可获得很好的识别效果。本文据此提出了采用损伤前后曲率模态残差小波变换系数方法对结构损伤进行识别,通过该小波变换系数的分布情况来确定结构的损伤指标。为验证该方法的有效性,通过选用具有线性相位的双正交样条小波,对一梁结构进行了数值模拟。结果表明,采用该方法不但对单一损伤而且对多损伤均能有效地识别出结构的损伤位置。

Traditional Fourier transformation is adapt to detect the whole character of signal singularity, and it becomes difficult in identifying the exact location of those singular points and distributed situation. As a method of signal analysis in time and frequency domains, the wavelet transformation （WT） is widely applied in engineering. The approach of curvature mode is highly sensitive to damage, and be used satisfactorily to detect the damage by a few lower modes. A method of damage detection was proposed by determining WT coefficient due to the residual difference of curvature modes, and used to identify the damage location in structure. As an example, a simply supported beam model was chosen to demonstrate the method on the basis of a bi-orthogonal spline wavelet. The calculational result shows that the present method can be adopted to detect the single and several locations of damage in structures.