Application of the improved airspace wavelet denoising algorithm in friction welding ultrasonic testing signals

摘要

In the process of ultrasonic flaw detection to friction welded joints, a number of weak-binding defects are difficult to be correctly detected due to noise pollution. This article studies the semi-soft threshold denoising method of spatial correlation coefficient based on the traditional spatial correlation coefficient de-noising method and threshold de-noising method, and come up with an improved airspace de-noising algorithm, which is used in the denoising of the defect signals. By comparing with hard thresholding, soft-threshold method and the traditional airspace algorithm on SNR and root mean square error, the detecting signals denoised by improved algorithm become more straight, and which better keep the sharp signal peak part of the curve about the original signals. So the identify characteristics of defect echo become more obvious, and we can get a smaller mean square error and the improved SNR of the reconstruction signals, which has the good de-noising effect and will lay a good foundation for the classification of defects next step.