首页 | 官方网站   微博 | 高级检索  
     

基于Kriging近似模型的轿车翼子板冲压工艺优化
引用本文:于磊,徐栋恺,陈军,张健. 基于Kriging近似模型的轿车翼子板冲压工艺优化[J]. 模具技术, 2012, 0(2): 1-5
作者姓名:于磊  徐栋恺  陈军  张健
作者单位:1. 上海交通大学 模具CAD国家工程研究中心,上海,200030
2. 一汽集团模具制造有限公司,吉林长春,130011
基金项目:国家重大科技专项课题资助项目(2011zx04016-051),国家重大科技专项课题资助项目(2010zx04014-072)
摘    要:近似模型法是目前最有代表性的高效优化方法。因基于多项式的近似模型很难满足板料成形的高度非线性,所以基于Kriging近似模型法,以翼子板的各拉深筋单位长度上所受的拉深筋阻力及压边力为设计变量,以成形后板料的最小厚度为响应量,建立近似模型。通过自适应模拟退火算法寻优,并约束最小厚度的变化范围,提高优化效率。通过优化,成形后板料的最小厚度从0.521mm增加至0.544mm,消除了局部破裂趋势,有效地改善了翼子板的成形质量。研究表明,Kriging近似模型法对板料成形有较高的精度。

关 键 词:Kriging近似模型  板料冲压成形  非线性优化

Research on process optimization of fender based on Kriging surrogate model
YU Lei , XU Dong-kai , CHEN Jun , ZHANG Jian. Research on process optimization of fender based on Kriging surrogate model[J]. Die and Mould Technology, 2012, 0(2): 1-5
Authors:YU Lei    XU Dong-kai    CHEN Jun    ZHANG Jian
Affiliation:YU Lei,XU Dong-kai,CHEN Jun,ZHANG Jian
Abstract:Surrogate model is the most representative efficient optimization method.Since the surrogate model based on polynomial is hard to satisfy the high non-linearity of the sheet forming,in this paper,based on Kriging approximation model,taking the blank holder force and the drawbead forces per unit length of drawbeads for the fender as design variables and minimum sheet thickness after forming as the response variable,the surrogate model was constructed.The self-adaptive simulated annealing algorithm was adopted to search the optimum solutions,and the variation range of the minimum thickness was restricted to improve the optimization efficiency.After optimization,the minimum sheet thickness after forming is increased from 0.521 mm to 0.544 mm,the local burst trend is eliminated and the forming quality of the fender is improved effectively.Research shows that Kriging model has high prediction accuracy for sheet metal forming.
Keywords:Kriging surrogate model  sheet metal forming  non-linearity optimization
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号