Abstract:Strict industry demands necessitate better penetration classification in gas tungsten arc welding of aluminum alloys. An innovative classification method, namely time-frequency spectrogram based convolutional neural network (TF-CNN), has been developed in this article. The logarithmic time-frequency spectrograms are used to interpret the raw arc sound data in the time-frequency domain and served as the input of an optimized convolutional neural network (CNN) model for welding quality classification. The leaky rectified linear unit activation function and root mean square prop optimizer are utilized to improve the property of CNN. Effective arc sound features are extracted automatically by the CNN model. The proposed method is experimentally verified to recognize four penetration states with an average accuracy of 98.2%, which significantly outperforms the four typical comparison models. The generation mechanism and characteristics of arc sound are discussed as the adaptive feature extraction of TF-CNN. These studies expand the application scope of CNN to intelligent welding and achieve notable effects.

A Novel Convolutional Neural Network Based on Time–Frequency Spectrogram of Arc Sound and Its Application on GTAW Penetration Classification