上海交通大学学报

上海交通大学学报（自然版）

• 化学工程 • 上一篇    

基于Kriging代理模型的注塑产品翘曲优化

陈巍1，周雄辉1，张汝珍1，韩先洪1，王会凤2   

  1. (1.上海交通大学 国家模具CAD工程研究中心，上海 200030; 2.北京科技大学 材料学院，北京 100083)
  • 收稿日期:2009-05-11 修回日期:1900-01-01 出版日期:2010-04-29 发布日期:2010-04-29

Warpage Optimization for Injection Molding Based on Kriging Metamodel

CHEN Wei1，ZHOU Xionghui1，ZHANG Ruzhen1，HAN Xianhong1，WANG Huifeng2   

  1. (1.National Die & Mold CAD Engineering Center, Shanghai Jiaotong University, Shanghai 200030， China；2.School of Materials, Beijing University of Science and Technology, Beijing 100083, China)
  • Received:2009-05-11 Revised:1900-01-01 Online:2010-04-29 Published:2010-04-29

摘要： 针对传统的基于CAE的注塑产品工艺优化方法精度不高、效率低，提出了Kriging模型与自适应粒子群算法相结合的集成优化策略.Kriging模型代替CAE分析作为粒子群算法迭代过程中的适应函数，大大减少了优化算法的计算量；同时，通过在粒子群算法中引入自适应惯性权系数，加快了粒子群算法的收敛速度.算例表明，基于Kriging模型与自适应粒子群算法的优化策略可以在小样本情况下获取较高的求解精度，并通过与标准遗传算法做比较，表明该优化策略同时具有较高的计算效率.

关键词: 代理模型, 　Kriging, 注塑模具, 　翘曲变形, 　优化

Abstract: Compared with the traditional processing optimization for injection molding based on CAE, Kriging metamodel has better fitting precision for nonlinear problem. An integrated optimization strategy based on Kriging metamodel and adaptive particle swarm optimization (APSO) algorithm was constructed. Kriging metamodel was adopted to replace CAE analysis as a fitness function of particle swarm optimization(PSO) algorithm, and the computation cost can be reduced greatly. Meanwhile, adaptive inertial weight coefficient was introduced to PSO algorithm, and the convergence rate of PSO can be accelerated. It is shown that the optimization strategy based on Kriging metamodel and APSO can get higher solving accuracy, the optimization strategy also has higher computation efficiency compared with standard genetic algorithm.

中图分类号: