Multi-sensor network for industrial steel plate structure monitoring via time reversal ultrasonic guided wave
Multi-sensor network for industrial steel plate structure monitoring via time reversal ultrasonic guided wave
复制标题
通过时间反转超声导波监测工业金属板结构的多传感器网络
DOI：
10.1016/j.measurement.2019.107345
复制
发表时间：
2020
期刊：
影响因子：
5.6
通讯作者：
Zhang Haiyan
中科院分区：
文献类型：
--
作者：
Zhang Hanfei;Cao Shuhao;Ma Shiwei;Lu Yu;Xiong Hanyu;Xia Qingwei;Liu Yanyan;Zhang Haiyan
主动超声导波被认为是工业钢板结构健康监测最有效、最重要的在线损伤监测方法。针对从基线信号中提取损伤散射信号实用性差的问题，基于超声导波时间反演理论和概率统计原理，提出一种无基线时间反演概率成像算法（BFTRPI算法）。首先，利用信号时间反演和波源自适应聚焦机制，通过多传感器网络消除兰姆波的色散效应。其次，将重建兰姆波信号与原始激励信号之间的能量特征差异系数作为损伤因子。最后，利用损伤与传感器-执行器通道的直接路径之间的相对距离来调整权重分布函数。将传感器网络中各传感器路径的加权概率值映射到检测区域的离散坐标上，构建这些离散坐标上出现的损伤的概率成像，进行多损伤成像和定位。铝板实验表明，该方法能够清晰分离无基线信号的损伤散射信号，实现对原始激励信号的聚焦，获得结构的损伤概率图像。与传统的概率分布函数法相比，具有更好的成像精度和成像质量。 BFTRPI算法的定位精度明显高于传统算法。 BFTRPI算法估计的位置更接近实际损伤位置。传统算法的损伤和伪影的定位面积均高于BFTRPI算法。该方法能够准确识别和定位铝板的多处损伤，验证了该方法的有效性，具有一定的工程应用价值。
Active ultrasonic guided wave is considered to be the most effective and important on-line damage monitoring method for industrial steel plates structure health monitoring. Aiming at the problem of poor practicability of extracting damage scattering signal from baseline signal, a baseline-free time reversal probabilistic imaging algorithm (BFTRPI algorithm) is proposed in this paper based on ultrasound guided waves time reversal theory and probability statistics principle. Firstly, the time reversal of signal and the adaptive focusing mechanism of wave source are used to eliminate the dispersion effect of Lamb wave by the multi-sensor network. Secondly, the energy characteristic difference coefficient between the reconstructed Lamb wave signal and the original excitation signal, is taken as the damage factor. Finally, the relative distance between the damage and the direct path of the sensor-actuator channel is used to adjust the weight distribution function. The weighted probabilistic values of each sensor path in the sensor network are mapped to the discrete coordinates in the detection area, and the probabilistic imaging of the damage appearing on these discrete coordinates is constructed for multiple damages imaging and location. Experiments on aluminium sheets show that the method can clearly separate the damage scattering signals without baseline signals, achieve the focus of the original excitation signal, and obtain the damage probability images of structure. Compared to traditional probability distribution function method, it has better imaging accuracy and imaging quality. The positioning accuracy of the BFTRPI algorithm is obviously higher than that of traditional algorithm. The estimated position by BFTRPI algorithm is more close to the actual damage position. The location area of damage and artifact for traditional algorithm are both higher than that of BFTRPI algorithm. This method can accurately identify and locate multiple damages in aluminium sheet, which verifies the validity of this method and has certain engineering application value.