等离子熔射过程数据挖掘技术及智能控制策略  

Data mining technology and intelligent control strategy in plasma spraying process

作　　者：夏卫生[1,2] 杨云珍[3] 张海鸥[2] 王桂兰[1] 

机构地区：[1]华中科技大学材料成形与模具技术国家重点实验室,武汉430074 [2]华中科技大学数字制造装备与技术国家重点实验室,武汉430074 [3]武汉理工大学汽车工程学院,武汉430070

出　　处：《焊接学报》2010年第5期33-36,共4页Transactions of The China Welding Institution

基　　金：国家863高技术研究发展计划资助项目(2007AA04Z142);国家自然科学基金资助项目(50675081);中国博士后科学基金资助项目(20080440940)

摘　　要：结合图像处理与模式识别技术提取等离子射流的形态特征,采用射流狭长度对其进行表征.采用数据挖掘技术将等离子弧电流、电压参数和射流狭长度特征进行聚类分析,构造对等离子熔射过程中射流发生子系统进行智能化控制的知识库.根据涂层的残余应力、孔隙率和硬度等主要性能参数以及微观组织结构对分类结果进行质量评判,生成工艺参数到最佳分类的推理规则集,从而实现了对工艺参数—射流形态—涂层性能之间关联关系的研究.基于规则库监督多模型自适应控制系统结构,提出了一种等离子射流发生子系统的智能化控制策略.The morphological feature of plasma jet was obtained by digital image processing and pattern recognition,and characterized by the aspect ratio.To develop a knowledge base for the intelligent control of plasma spraying process,the cluste-ring analysis was carried out between spraying parameters and the feature of plasma jet based on data mining technology.The quali-ty of clusters was evaluated by the properties(residual stress,porosity and hardness)and microstructure of the coatings,and the inference rule sets from processing parameters to the high quality cluster were created.Hence,the investigation of the rela-tionship between the processing parameters,plasma jet and coat-ing performance was realized.Finally,an intelligent control approach for the sub-system of plasma jet generation was presented in this paper based on the rule sets supervision and multi-model adaptive control system structure.

关 键 词：等离子熔射 射流形态 智能控制 数据挖掘 聚类分析 

分 类 号：TG174.442[金属学及工艺—金属表面处理]