基于有限元分析的压力容器损伤阵列稀疏特征建模与定位方法

Finite element analysis based sparse array features modeling and localization method for pressure vessel damage

针对压力容器无损检测中较难检测的封头部位,提出了一种基于压电超声导波阵列的在线监测方法。首先,建立压力容器封头结构的有限元模型,利用窄带Lamb波模拟结构损伤源的散射信号;其次,将被监测区域离散成稀疏点,对每个稀疏点的回波信号提取阵列波达时间差作为导波阵列的稀疏特征信息,并计算损伤回波信号的波达时间均方根值(RMS),形成基于有限元模型的损伤稀疏特征样本库;最后,通过计算实测损伤回波信号均方根值与损伤稀疏特征样本库匹配成像,像素点最高的位置即为损伤位置。试验结果表明,模拟损伤距离定位误差为20 mm,角度定位误差为1°,与实际损伤位置较为符合,能够较准确地反映缺陷的位置。

Aiming at the difficulty of nondestructive testing of pressure vessels head part, a piezoelectric ultrasonic guided wave array is proposed for pressure vessel damage localization. Firstly, a finite element model of the head part of pressure vessel is established, and the narrow-band of Lamb wave is excited to simulate the echo of the damage. Secondly, discrete the monitored area into sparse points, the time of flight(TOF) of array signal at each sparse points is obtained as the sparse feature of the guided wave array, and calculates the root mean square(RMS) value of them to form the damage sparse feature sample library. Finally, the RMS of the measured array signal when a damage happen is calculated to match the damage sparse feature sample library for damage imaging, and the highest position of the pixel is the position of damage occurred. The test results show that the simulated damage distance positioning error is 20 mm and the angle positioning error is 1°, which is more consistent with the actual damage position and can accurately reflect the position of the defect.