稀疏重建的兰姆波结构损伤定位

Damage localization with Lamb waves via sparse reconstruction

研究了基于稀疏重建算法的兰姆(Lamb)波损伤成像方法。利用损伤的稀疏性,对测量信号在字典中寻求其稀疏表示,通过稀疏重建方法得到定位缺陷的图像。由兰姆波信号传播模型理论,计算检测区域所有可能出现缺陷位置对应传感器接收到的散射信号波形,以此组成过完备字典;并通过梯度投影算法从字典中搜寻匹配的原子,匹配原子在字典中的位置对应于缺陷在检测区域中的位置。仿真实验采用多模态兰姆波进行损伤定位,结果表明稀疏重建方法在低噪声情况下可以实现对单个、两个缺陷的定位,和延迟求和成像方法对比,稀疏成像具有更少的噪声干扰及伪像。

This research investigated the guided wave imaging based on the sparse reconstruction techniques, looks for the sparse representation of the measured signals in the dictionary and obtains the images of the located damages with the sparse reconstruction techniques. All the possible locations of the damages in the detected area are calculated and compared with the waveform of the scattering waves to form the complete dictionary. And then the gradient project algorithm is used to search for the matched atoms in the dictionary and to match the locations of the atoms in the dictionary with the locations of the damages in the detected area. In this paper, the multi-mode Lamb waves is utilized to locate the damages, and the application of the sparse reconstruction techniques can make the localization of the single or two damages come true, which exhibit smaller spot size and fewer artificial defects than the results of the delay-and-sum imaging.