Modeling Welding Deviation of Rotating Arc NGW based on Support Vector Machine
Support Vector Machine (SVM) was employed to describe welding deviation of welding seam tracking in multi-layer single pass, narrow gap rotating arc gas metal arc welding (GMAW). First, work piece was designed and processed to mimic multi-layer single pass welding groove to obtain experimental data under different deviations. Second, for data preprocessing, welding current signal in each rotating circle was divided into 12 parts, and the average value of each part and the difference of the according left-right part were selected and normalization to build an input-output table for predicting the weld deviation. Then, kernel function was selected and its parameters were determined by a grid search method using cross validation approach. The SVM model was built with the train set and was validated with the test set. It showed that the model can meet the criteria of welding seam tracking. In addition, comparison between SVM model and BP neural network model was made and it showed the former has better performance because it adapt to the little sample problem and can avoid the local extreme.
Narrow gap rotating arc MAG welding SVM Welding deviation extraction
Wenhang Li Kai Gao Feng Yang Dandan Sun Jiayou Wang
School of Material Science and Engineering, Jiangsu University of Science and Technology, Zhenjiang School of Material Science and Engineering, Jiangsu University of Science and Technology, Zhenjiang
国际会议
济南
英文
1-9
2012-08-24（万方平台首次上网日期，不代表论文的发表时间）