期刊文献+

基于Kriging模型的注塑工艺稳健优化 被引量:5

Robust optimization of injection molding process based on Kriging model
下载PDF
导出
摘要 为降低薄壁注塑件的翘曲变形,提出了基于Kriging模型的工艺稳健优化方案。特别地,研究并提出了基于变异粒子群算法的Kriging模型建立与优化算方法,建立了注塑工艺参数和翘曲变形之间的Kriging模型。通过实验设计、CAE仿真结果数据建立起Kriging模型,并结合变异粒子群算法实现了工艺的优化设计。在此基础上,为减少工艺参数波动对注塑件质量的影响,建立了注塑成型的6 Sigma稳健优化设计方法,提高了工艺的稳健性与可靠度。以汽车出风口壳体为计算用例,结果表明,基于Kriging模型与变异粒子群算法的工艺稳健优化策略在小样本情况下能够得到较高质量的稳健优化参数组合。 To reduce the warpage of thin-walled injection molded parts,a robust optimization approach based on Kriging model was presented. In particular,the Kriging model establishment and optimization method based on the variant particle swarm optimization algorithm was studied and proposed. The Kriging model between the injection molding process parameters and the warpage deformation was established. The Kriging model was first established by experiment design and CAE simulation results,and process optimization was realized by utilizing variant particle swarm optimization algorithm. Then,on this basis,to reduce the effect the process parameter fluctuation on the quality of injection molded parts,a six Sigma robust optimization design method was adopted to improve the robustness and reliability of the process. Taking the automobile air outlet shell as the calculation case,the results show that the high quality parameter combination can be obtained under small sample conditions with the process robust optimization strategy based on the Kriging model and variant particle swarm optimization algorithm.
作者 李培健 周雄辉 柳伟 LI Pei-jian;ZHOU Xiong-hui;LIU Wei(Institute of Forming Technology and Equipment,School of Materials Science and Engineering,Shanghai Jiao Tong University,Shanghai 200030,China)
出处 《塑性工程学报》 CAS CSCD 北大核心 2019年第5期105-111,共7页 Journal of Plasticity Engineering
关键词 翘曲 KRIGING模型 变异粒子群算法 稳健优化 warpage Kriging model variant particle swarm optimization algorithm robust optimization
  • 相关文献

参考文献11

二级参考文献98

共引文献393

同被引文献30

引证文献5

二级引证文献23

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部