决策树算法在焊接工艺设计中的应用  

Application and research of decision tree algorithm in welding technology design

作　　者：钱学君[1] 胡小建[1] 

机构地区：[1]合肥工业大学材料科学与工程学院,安徽合肥230009

出　　处：《焊接技术》2009年第6期5-8,共4页Welding Technology

基　　金：国家自然科学基金项目(70631003;60736026);安徽省自然科学基金项目(070416241)

摘　　要：在焊接工艺设计过程中,因其信息量较大,人工无法对大量参数进行科学分析,也就无法获得可能潜在的信息和规则。本文在经典ID3算法的基础上,研究了决策树的改进算法C4.5及其剪枝方法,并首次将决策树C4.5算法应用于焊接领域,针对具有代表性的汽车车架焊接工艺设计过程,综合分析了各种工艺参数对焊接气孔和焊接变形的影响,为合理制定焊接工艺提供了参考,避免了焊接缺陷的产生。Because of the large amount of information during the welding design, people could not analyze the parameters by themselves, the potential and valuable rules could not be discovered. This article briefly introduced the improved algorithm C4.5 and its pruning method based on the typical algorithm of decision tree ID3, and first applied the decision tree algorithm C4.5 to welding field. In allusion to the representational process of welding design for the automobile frame, this paper synthetically analyzed the effect of various parameters on welding air vent and welding transfiguration. The decision tree could be used to contrive welding design, avoiding the appearance of welding disfigurement.

关 键 词：焊接工艺设计 汽车车架 决策树 C4.5算法 

分 类 号：TG457.21[金属学及工艺—焊接] TP274[自动化与计算机技术—检测技术与自动化装置]