激光焊熔池红外辐射特征支持向量机回归分析  

Analysis of molten pool infrared radiation characteristics based on support vector machine during laser welding

作　　者：刘英英[1] 高向东[1] 

机构地区：[1]广东工业大学机电工程学院,广东广州510006

出　　处：《电焊机》2014年第3期93-98,共6页Electric Welding Machine

基　　金：国家自然科学基金(51175095);广东省自然科学基金(10251009001000001;9151009001000020);高等学校博士学科点专项科研基金(20104420110001)资助的课题

摘　　要：大功率光纤激光焊接过程中,熔池红外辐射蕴含着丰富的焊接质量信息。以大功率光纤激光对接焊304不锈钢板为试验对象,运用近红外高速摄像机获取焊接熔池动态热像。定义并提取熔池宽度、匙孔面积、匙孔周长和匙孔质心横、纵坐标,作为熔池特征参数,运用支持向量机建立熔池特征参数和焊缝宽度的回归模型,并通过网格寻优和粒子群算法优化支持向量机参数。试验表明,所建立的支持向量回归机能够较好地融合熔池特征,预测焊缝宽度,从而为自动监控大功率光纤激光焊接质量提供试验依据。During the high-power fiber laser welding,molten pool infrared radiation contains abundant information reflecting the details of the welding quality.In the high-power fiber laser welding of a type 304 austenitic stainless plate a continue laser power 10 kW,an infrared sensitive video camera was used to capture molten pool dynamic images.Molten pool width,keyhole area,keyhole Perimeter and keyhole centroid were defined as molten pool characteristic parameters.A support vector machine regression model of molten pool characteristic parameters and weld width was set up, and grid search method and particle swarm optimization were used to optimize experimental parameter c and g.Experimental result shows that the established support vector machine regression model could fuse molten pool characteristic information effectively and predict the weld width ,thus providing experimental evidences for automatic evaluation of high power fiber laser welding quality.

关 键 词：大功率激光焊接 熔池热像 焊缝宽度 支持向量机 

分 类 号：TG456.7[金属学及工艺—焊接]