摘要
利用线性预测(LPC)分析方法建立电弧声信号数学模型并提取声音频域特征,分别用电弧声的LPC预测系数及反射系数作为输入向量构造样本集,建立CO2焊接过程支持向量机模型,对焊接过程保护气体流量进行了识别和分类。测试表明:利用电弧声建立的SVM模型能对焊接气流量进行有效识别。
The LPC analysis was used to make the mathematical model of arc sound and abstract the frequency characteristics of sound.Constructed input vectors by the LPC prediction and reflectance coefficients of sound signals respectively,the classification models based on SVM have been built for pattern recognition of the shield gas flux in CO2 welding.The testing results showed that the classification capabilities of SVM models were effective.
出处
《电焊机》
2007年第4期68-70,73,共4页
Electric Welding Machine
