Quantitative Detection of Laser Welding Defective Structure Based on Feature Exaction of the Pulsed Eddy Current Signal
Kuanfang He,Jiahe Liang,Jiangfeng Yong,Wenqing Shi
DOI: https://doi.org/10.1007/s11665-022-07566-x
IF: 2.3
2022-10-30
Journal of Materials Engineering and Performance
Abstract:The laser welding structure produces defect due to the disturbance of material, technology, environment and other factors. Currently, the defect of laser welding structure is mainly judged by optical, high-speed camera, infrared thermal imaging, and x-ray method. The use of pulsed eddy current (PEC) technology can achieve a noncontact and nondestructive way for welding quality monitoring. Thus, the laser welding aluminum alloy structure with the defects of pore and crack is quantitatively detected by PEC technology. The detect system is constructed to obtain the PEC signal of laser welding seam with different defects. The feature parameters of peak value, time of peak value, fundamental wave amplitude, peak and rising curvature ratio, amplitude ratio of the fundamental and third harmonic wave, and marginal spectrum peak value of the PEC signal are calculated to represent the laser welding defect-type and size quantitatively. The defect identification model based on support vector machine (SVM) is established to identify the type and the depth of the laser welding seam defect by inputting the feature parameters. Experimental results show that the identification model achieves 95.8% accuracy of the defect type and 87.5% accuracy of the depth, which provides a method for quantitative detection of laser welding defective structure.
materials science, multidisciplinary