基于Kriging代理模型和遗传算法的注塑件翘曲优化  

Warpage optimization for injection molding based on Kriging model and genetic algorithms

作　　者：周香[1] 陈文琳[1] 王晓花[1] 李伟[1] 

机构地区：[1]合肥工业大学材料科学与工程学院,合肥230009

出　　处：《塑性工程学报》2015年第2期142-147,共6页Journal of Plasticity Engineering

摘　　要：翘曲变形对具有配合关系的两塑件的装配有严重影响。文章针对传统优化方法提出一种高效、精准的优化方法。基于Kriging代理模型,通过实验设计,建立工艺参数与翘曲值之间近似的函数关系;应用遗传算法,对工艺参数进行全局寻优得到翘曲值最小的工艺参数组合;在优化设计的基础上,研究各工艺参数对塑件翘曲值的影响趋势。结果表明,注射时间和保压压力对其影响较大,同时Kriging模型中的相关函数参数能够反映各工艺参数对翘曲值的非线性程度。The warpage of plastic parts has serious effect on their assembly .In this paper ,an efficient and accurate optimization method is proposed to replace the traditional optimization methods .Based on Kriging model ,the approximate relationships be‐tween processing parameters and warpage values are established through design of experiment ,and the genetic algorithms are ap‐plied to the global optimization to get the best processing parameters when the warpage value is minimum .Finally ,on the basis of the optimization design ,the effects of processing parameters on the warpage values of plastic parts are discussed .The results show that the injection time and the holding pressure have great influence on the warpage ,and the parameter of correlation func‐tion in the Kriging model can reflect the nonlinearity degree of various processing parameters on the warpage values .

关 键 词：KRIGING模型 遗传算法 翘曲变形 优化设计 

分 类 号：TQ320.66[化学工程—合成树脂塑料工业]