Porosity defect detection based on FastICA-RBF during pulsed TIG welding process

Porosity is a common defect of the aluminum alloy pulsed alternating current (AC) argon tungsten-arc welding (TIG) welding, which can cause huge damage to weld quality. The spectral information which is directly derived from the optical radiation of the arc is intrinsically related to the welding defects. Aiming at the redundancy of arc spectral, this paper proposed a method of porosity defect detection based on fast independent component analysis (fastICA) and radial basis function (RBF) network. The spectral data is collected by spectrometer, and continuous spectra are removed by calculating lower envelope twice. Then fastICA is applied to extract features from selected line spectra. Finally, the porosity defect is detected by RBF network according to the mean value in period of extracted features. Experimental results show that the proposed method can be used to detect the porosity defects during aluminum alloy pulsed TIG welding process.