Intelligentized evaluation on dynamic characteristics of CO2 arc welding power source based on key Eigen values and support vector machine

The automated evaluation on dynamic characteristic of the arc welding power source has currently emerged as a popular research topic. This technology is vital to the detection of automated welding and test on the quality of new welding machines. Gradually substituting the traditional manual detection, the technology can also release the burden of welders and lower the cost of human resource. Based on experiments, this paper has explored and then selected 9 Eigen values which are closely related to the dynamic characteristics of CO₂ arc welding power source. In accordance with the binary classification using SVM, the dynamic characteristics were further examined and next defined whether qualified or not. Results of experiments show that the method has fully achieved the intelligentized evaluation on dynamic characteristics of CO₂ arc welding power source, reaching an accuracy rate of 94.8718%, which is both theoretically and practically valuable.