基于偏最小二乘回归方法的CO_2焊飞溅模型  

Spatter model of CO_2 arc welding based on partial least-square regression method

作　　者：蔡艳[1] 杨海澜[1] 徐忻[1] 吴毅雄[1] 

机构地区：[1]上海交通大学焊接工程研究所,上海200030

出　　处：《焊接学报》2004年第5期125-128,共4页Transactions of The China Welding Institution

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CO2According to the mechanism analysis of short-circuit process, characteristic parameters are extracted from electrical signals ot 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.

关 键 词：变异系数 特征参数 偏最小二乘回归 飞溅 CO2焊 

分 类 号：TG444[金属学及工艺—焊接]