采用多片压电元件(PZT)粘贴在焊接铝板结构表面,用阻抗分析仪测量PZT导纳信息,研究激励频率对PZT导纳值的影响以及不同位置PZT传感器对裂纹感知的灵敏性.构建人工神经网络,对焊缝结构的缺陷进行定位.实验结果表明:PZT传感器的导纳谱能灵敏地反映结构状态的微小变化；且不同位置PZT对裂纹感知的灵敏度不同；对于同一PZT,激励频率不同,其阻抗变化量不同；提取导纳信息中表征结构健康状况的特征量作为神经网络的输入向量,训练后的神经网络可以迅速并准确地实现焊缝结构损伤的监测和定位.%A welding aluminum plate with several PZT sensors pasted on the surface is used as the research example. The change of the piezoelectric impedance inspirited by different frequencies of the excited signal is studied by analyzing the PZT impedance spectrum. An artificial neural network is constructed to locate the crack. Experimental results show that even small changes of structural state can be detected by the impedance changes of PZT patches. For the same PZT, different frequencies of the excited signal have different impedance changes. The characteristics representing the health of structure are chosen as the input vectors of the artificial neural network. The trained neural network can monitor and position the welding cracks quickly and correctly.
展开▼
