超声检测中利用盲解卷积的重叠回波分离方法

An Overlapping Echo Separation Method Using Blind Deconvolution in Ultrasonic Testing

为了解决超声检测技术在被测件近表面位置容易产生回波信号重叠,导致近表面缺陷定位误差较大的问题,提出了一种基于盲解卷积的超声重叠回波分离方法。该方法首先对超声回波信号进行傅里叶逆变换和平滑处理,实现对脉冲响应函数初始化;然后分别对反射序列函数和脉冲响应函数建立凸优化解卷积模型,并使用分裂Bregman算法与交替方向乘子算法(ADMM)对两者进行交替迭代求解;最后,通过判断是否满足停止条件,获得反射序列函数与脉冲响应函数的估计值,从而实现对超声重叠回波的分离。仿真实验表明:该方法在不同强度的噪声干扰下均可以经过10次左右的交替迭代实现对重叠回波的有效分离,对噪声干扰具有良好的鲁棒性;该方法将近表面缺陷的定位误差减小到0.97%,适用于实际超声检测。

An ultrasonic overlapping echo separation method based on blind deconvolution is proposed to solve the problem of echo signal overlap in ultrasonic testing technology,which may result in large positioning error of near surface defects.Firstly,the ultrasonic echo signal is processed by inverse Fourier transform and smoothing technique to initialize the impulse response function.Then,convex optimal deconvolution models are established for a reflection sequence function and an impulse response function respectively,and the split Bregman algorithm and the alternating direction multiplier algorithm(ADMM)are alternatively used to solve the both models.Finally,the estimated values of the reflection sequence function and the impulse response function are obtained through judging whether the stop condition is satisfied,so as to realize the separation of ultrasonic overlap echo.Simulation and experimental results show that the proposed method effectively separates the overlapping echoes after 10 alternating iterations under different intensity of noise interference,and has good robustness to noise interference.This method reduces the positioning error of near surface defects to 0.97%,and is suitable for practical ultrasonic testing.