基于偏最小二乘回归方法的CO2焊飞溅模型
Spatter model of CO2 arc welding based on partial least-square regression method
-
摘要: 在分析短路过渡飞溅形成机理的基础上,以过渡周期为基本单元,采用统计分析方法从焊接过程电信号中提取特征参数,基于变异系数的分析方法反映了过程的均匀程度。由于特征参数间存在较强的相关性,采用偏最小二乘回归方法建立飞溅模型,实现飞溅量对特征参数的回归。试验结果表明,基于偏最小二乘方法的回归模型有较高的实用性和一定的精度,为研究CO2气体保护焊飞溅问题提供了新的途径。Abstract: According to the mechanism analysis of short-circuit process, characteristic parameters are extracted from electrical signals to welding process by statistical analysis method with transfer period as the basic analysis unit, in which the process steadiness is reflected by variation coefficient. Partial least-square regression method is suitable because of the close correlation among variables. The experimental result shows that the spatter model is practical which is a new method for analysis of spatter.