基于熵权与响应面模型的DPVC工艺多目标优化  

Multi-objective optimization for DPVC process based on entropy-weight and RSM

作　　者：张海光[1] 张壮雅[1] 万福平 刘媛媛[1] 胡庆夕[1] 

机构地区：[1]上海大学快速制造工程中心上海市机械自动化及机器人重点实验室,上海200444

出　　处：《计算机集成制造系统》2014年第8期1887-1895,共9页Computer Integrated Manufacturing Systems

基　　金：中国博士后科学基金资助项目(2011M500755);上海市优秀青年教师选拔专项基金资助项目(B.37-0109-08-004);上海市重点实验室自主课题资助项目(ZZ0803)~~

摘　　要：为解决真空差压注型成形工艺参数优化问题,提出基于熵值权重法、响应面模型及多种群的改进遗传算法的真空差压注型工艺参数优化方法。采用正交试验设计和数值模拟方法获取试验样本数据,引入熵值权重法,从数理角度确定各项质量指标对产品综合质量的影响权重;结合中心复合实验设计和响应面模型,建立真空差压注型产品的综合质量评价模型;针对传统遗传算法的不足,提出一种基于多种群的改进遗传算法,从而完成对产品综合质量评价模型的优化求解;在自主开发的差压式数字控制V500-ND原型机上,通过对某摩托车前车灯灯罩的工艺参数优化实验,验证了所提方法的有效性,为真空差压注型成形工艺过程的制定提供了一种可行的工艺优化方法。To solve the forming process parameter optimization of Differential Pressure Vacuum Casting （DPVC）, a DPVC process parameter optimization method was proposed based on entropy weight method, response surface model and multi-population improved genetic algorithm. The sample data were obtained through orthogonal experi- ment and numerical simulation methods, and the weight of each index on product quality was determined with entro- py weight method. Based on center composite design and response surface model, the comprehensive evaluating model for product quality was built. Aiming at the defect of traditional genetic algorithm, an improved genetic algo- rithm based on multi-population was proposed to optimize the solution of established evaluation model. On self de- veloped V500-ND prototype machine, the effectiveness of proposed method was verified by process optimization test of headlamp in one brand of motorcycle, which could provide the feasible process optimization for DPVC forming process set.

关 键 词：真空差压注型 工艺参数优化 熵值权重法 响应面模型 多种群改进遗传算法 

分 类 号：TQ20[化学工程—有机化工]