基于GA-BP模型与NSGA-Ⅱ算法的冲压成形工艺参数优化方法  

Stamping process parameter optimization method based on GA-BP model and NSGA-Ⅱalgorithm

作　　者：陈岁繁 王世杰 戴光明 王树飞 李其朋 何雄华 CHEN Suifan;WANG Shijie;DAI Guangming;WANG Shufei;LI Qipeng;HE Xionghua(School of Mechanical and Energy Engineering,Zhejiang University of Science and Technology,Hangzhou 31000,China;Wuxi MingshiJunzhi Medical Technology Co.,Ltd.,Wuxi 214000,China;Yongkang Didi Technology Co.,Ltd.,Jinhua 321300,China)

机构地区：[1]浙江科技大学机械与能源工程学院,杭州310000 [2]无锡鸣石峻致医疗科技有限公司,无锡214000 [3]永康市迪迪科技有限公司,金华321300

出　　处：《中国有色金属学报》2024年第7期2330-2342,共13页The Chinese Journal of Nonferrous Metals

基　　金：国家重点研发计划资助项目(SQ2020YFF0423771);浙江省“领雁”研发攻关计划项目(2024C04037)。

摘　　要：为了解决异形、薄壁特征零件在冲压成形过程中严重起皱、成形不足及能耗过高等问题,创新地建立了以压边力与摩擦因数为设计变量,以最大厚度、厚度减薄率和压边能耗为优化目标的GA-BP神经网络优化模型,并应用非支配排序遗传(NSGA-Ⅱ)算法求解,获得了冲压成形过程中的最优工艺参数组合。结果表明,通过有限元仿真获得了最优工艺参数组合下的最大厚度、厚度减薄率和压边能耗分别为0.269 mm、15.8%、8720 J,分别改善了0.74%、3.06%和11.32%,有效地改善了异形薄壁零件的严重起皱、成形不足及能耗过高此等问题。In order to solve the problems of serious wrinkling,insufficient forming and high energy consumption in the stamping process of special-shaped and thin-walled parts,a GA-BP neural network optimization model with blank holder force and friction coefficient as design variables and maximum thickness,thickness reduction rate and blank holder energy consumption as optimization objectives was innovatively established,and it was solved by non-dominated sorting genetic(NSGA-Ⅱ)algorithm,and the optimal combination of process parameters in the stamping process was obtained.The results show that the maximum thickness,thickness reduction rate and blank holder energy consumption under the optimal combination of process parameters are 0.269 mm,15.8%and 8720 J,respectively,which are improved by 0.74%,3.06%and 11.32%,respectively,effectively solving the problems of serious wrinkling,insufficient forming and high energy consumption of special-shaped thin-walled parts.

关 键 词：冲压 变压边力 GA-BP优化网络 NSGA-Ⅱ算法 

分 类 号：TG386[金属学及工艺—金属压力加工]