Ultrasonic aluminum weld testing method based on the wavelet transform and a neural classifier

This paper proposes an ultrasonic non-destructive weld testing method based on the wavelet transform and a neural network classifier. The use of Lamb waves generated by an electromagnetic acoustic transducer (EMAT) as a probe allows us to test metallic welds; in this work, the case of aluminum weld is treated. We explain how we proceed to do the feature extraction by using a method of analysis based on the wavelet transform of the ultrasonic testing signals; we propose a classification process of the features based on a neural classifier to interpret the results in terms of weld quality. The aim of this complete process of analysis and classification of the NDT ultrasonic signals is to lead to an automated system of weld or structure testing. Results of real-world ultrasonic Lamb waves signals analysis and classification for an aluminum weld are presented; these proved the feasibility of the proposed method.