基于多传感器和支持向量机的GMAW焊接过程模式识别研究

Research on GMAW Welding Process Pattern Recognition Based on Multi-sensor and Support Vector Machine

GMAW(熔化极气体保护焊)是一个复杂的物理、化学过程,存在高度的复杂性和非线性性,有效的提高对其焊接过程模式识别的准确性一直是GMAW焊接过程监控的一个关键问题。文章提出一种基于多传感器和SVM(支持向量机)的焊接过程模式识别方法。通过多传感器对焊接过程中靶材力、振荡、电弧电流、电压和声压等信号同时进行采集,并进一步对其进行方差、小波以及希尔伯特黄变换;再综合焊接过程中各信号数据和训练样本的特点选取并建立多分类SVM和核函数模型,并利用对焊缝的CT断层扫描加以验证,实验结果表明该方法对焊接过程模式具有较高的准确率。

GMAW （gas-shielded metal arc welding） is a complex physical and chemical process, there is a high degree of complexity and nonlinearity. Effectively improving the accuracy of pattern recognition of its welding process has beena key issue of GMAW process monitoring. A method for pattern recognition of welding process based on multi-sensor and SVM （support vector machine） is presented in this paper. The signals of target force, oscillation, arc current, voltage and sound pressure are collected at the same time by multi-sensors, and the variance, wavelet and Hilbert-Huang Transform are carried out. According to the characteristics of signal data and training samples in the welding process, the multi-classification SVM and kernel function models are established and verified by CT tomography of welding seams. The experimental results show that the method hasa high accuracy for the welding process mode.