基于Kriging代理模型和遗传算法的塑件品质优化  

Quality Optimization of Plastic Parts Based on Kriging Agent Model and Genetic Algorithm

作　　者：傅建钢[1] FU Jian-gang(School of Electromechanical Engineering&Transportation,Shaoxing Vocational&Technical College,Shaoxing 312000,China)

机构地区：[1]绍兴职业技术学院机电工程与交通学院,浙江绍兴312000

出　　处：《塑料科技》2020年第5期88-91,共4页Plastics Science and Technology

基　　金：浙江省教育厅科研项目(Y201430415);绍兴市高等教育教学改革重点课题(SXSJG201809);浙江省教育科学规划课题(2019SCG143)。

摘　　要：采用Kriging代理模型和遗传算法对产品成型工艺参数进行优化。根据成型工艺对产品的影响情况,选取6个工艺参数作为设计变量,使用拉丁超立方随机产生16种设计方案。通过数值模拟得到对应的收缩率和翘曲值,建立代理模型,采用遗传算法进行寻优,得到最优设计变量组合。研究结果表明,模流分析结果与代理模型预测值具有很好的一致性,两者误差小于5.1%。优化后的翘曲值降低了62.7%,收缩率降低了22.9%,产品品质得到明显改善。The Kriging agent model and genetic algorithm were used to optimize the process parameters of product forming.According to the influence of forming process on product,six process parameters were selected as design variables.Sixteen design schemes were randomly generated by Latin hypercube.The corresponding shrinkage and warpage values were obtained by numerical simulation,and genetic algorithms were used to find the optimal combination of optimal design variables.The research results showed that the results of the mold flow analysis and the predicted value of the agent model were in good agreement,and the error between the two was less than 5.1%.The optimized warpage value was reduced by 62.7%,the shrinkage was reduced by 22.9%.And the product quality was significantly improved.

关 键 词：代理模型 工艺参数 遗传算法 模流分析 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TQ320.63[自动化与计算机技术—控制科学与工程]