Flaw classification in ultrasonic guided waves signal using Wavelet Transform and PNN classifier

Abstract

In this paper, a synthesized algorithm for flaw classification in ultrasonic guided waves signal is presented, in which Wavelet Transform is utilized in the process of noise suppression and envelop extraction, and the Probabilistic Neural Network (PNN) is introduced for flaw classification of the ultrasonic testing signal. Besides, in the process of feature extraction, the necessity of attenuation correction and feature selection of ultrasonic signal is discussed. The comparison of the performances of PNN and BP classifier is made in the last chapter, which demonstrates that the performance of flaw classification is significantly improved by the synthesized algorithm presented in this paper.

Compatibility and challenges in machine learning approach for structural crack assessment

“…Although the case of four features indicates no substantial enhances in comparison with the case of a single feature, there is no assurance the optimal number of features was found because the other combinations were not considered. Similar work using the probabilistic neural network has been undertaken by Zhang et al 54 Among the six features extracted, four were chosen as the most discriminatory. Once again, no explanation was given for relying on which criterion, this feature, was founded.…”

Increased range of ultrasonic guided wave testing of overhead transmission line cables using dispersion compensation

“…Compensation for the effects of attenuation may be made using gðtÞ ¼ĝðtÞ expða tÞ; ð4Þ [31]. For a single wave mode and the i th propagation time (or distance), dispersion compensation can be performed in the frequency domain using…”