第42卷第5期 Vol. 42No. 5
假压技木
FORGING & STAMPING TECHNOLOGY
基于响应面模型和遗传算法的汽车发罩外板冲压
工艺参数多目标优化
胡亚男，王雷刚，黄瑶，黄忠辉（江苏大学材料科学与工程学院，江苏镇江212013）
2017年5月 May 2017
摘要：以汽车发罩外板为例，将压边力、冲压速度、凹模与板料间摩擦系数和凸模与板料间摩擦系数作为工艺参数变量，以拉延工序最大减薄率和修边工序后最大回弹量为优化目标，应用中心复合试验设计（CCD）及有限元模拟获取样本数据。由
力为1145kN，冲压速度为3480mm·s-"，凹模与板料摩擦系数为0.106，凸模与板料摩擦系数为0.13。基于优化的工艺参数指导模面回弹补偿分析并试模，研究结果表明，发罩外板实际冲压成形质量较好。
关键词：汽车发罩外板；冲压；响应面模型；NSGA-II；多目标优化 DOI: 10. 13330/j. issn. 1000-3940. 2017. 05. 031
中图分类号：TG386.1
文献标识码：A
文章编号：1000-3940（2017）05-0171-05
Multi-objectiveoptimizationonstampingprocessforautomobilehoodpanel
based on response surface model and genetic algorithm
Hu Ya'nan, Wang Leigang, Huang Yao, Huang Zhonghui
(School of Materials Seience and Engineering, Jiangsu University, Zhenjiang 212013, China)
Abstract : For automobile hood panel, taking the blank holder force, the punch velocity, the coefficient of friction between die and blank, as well as the coefficient of friction between punch and blank as processing parameter variables, and the maximum thinning rate in the drawing process and the maximum springhack in the trimming process as the optimization objectives, the sample data were obtained by the central composite design ( CCD ) method and the finite element simulation. Then, the quadratic response model was established by the test data, and the multi-objective optimization of processing parameters was achieved combining with the non-dominated sorting genetic algo rithm ( NSGA II) - Therefore, an optimized combination of process parameters was obtained with the blank holder force of 1145 kN, the punch velocity 3480 mm · s ', the coefficient of friction between die and blank of 0. 106, and the coefficient of friction between punch and blank of 0. 13. Based on the: optimized processing parameters, the springback compensation was guided, and the die tryout was con ducted. The result shows that the stamping of automobile hood panel stamping is of good forming quality-
Key words: automobile hood panel; stamping; response surface model; NSGA Il; multi-objective optimization
汽车覆盖件是车身的重要组成部分，具有尺寸大、工艺复杂和成形精度要求高等特点。汽车覆盖件成形过程中易出现拉裂、起皱和回弹等缺陷，传统的人工试模很难快速解决这些缺陷的产生[1-2] 随着有限元数值技术与优化方法的不断发展，采用
收稿日期：20161207；修订日期：20170217 基金项目：国家自然科学基金资助项目（51275216）
作者简介：胡亚男（1991-），男，硕士研究生 E-mail: huyanan_ujs@ 163. com
通讯作者：王雷刚（1963-），男，博士，教投，博士生导师 E-mail : Igrang@ ujs. edu. cn
万方数据
数值模拟代替实际试模，并结合优化方法用于优化成形参数，提高汽车覆盖件产品的成形质量，在试验和实际生产中得到验证，并广泛应用[3-5]
本文以某汽车发罩外板为例，通过三维建模软件完成工艺补充面和压料面的设计。采用有限元软件对拉延成形及修边回弹过程进行模拟，应用中心复合试验（CCD）构建工艺参数与拉延最大减薄率、修边后最大回弹量之间的二阶响应面模型，并结合非支配排序遗传算法（NSGA-II）实现多目标优化，即在零件拉延成形工序满足最大减薄率的条件下，使修边工序后的回弹量最小，基于优化后的工艺参数指导模面回弹补偿分析并进行试验验证。