行李箱内板拉深压边力曲线的优化  

Optimization of blank-holder force curve for the deep drawing of trunk lid inner panel

作　　者：陈淑婉[1] 蒋敏 詹艳然[3] 黄胜[1] 

机构地区：[1]运城学院机电工程系,山西运城044000 [2]重庆宇杰汽车设计有限公司,重庆400020 [3]福州大学机械工程及自动化学院,福建福州350108

出　　处：《制造技术与机床》2014年第1期163-168,共6页Manufacturing Technology & Machine Tool

基　　金：福建省自然科学基金资助项目(2008J0153)

摘　　要：对新型高强度相变诱发塑性钢TRIP600钢板拉深成形的行李箱内板进行研究,将获得的变压边力成形窗口分为若干段,通过数值模拟的方法得到行李箱内板成形质量与各段压边力之间的正交试验数据,经极差分析确定正交试验优化方案;以正交试验数据为训练样本,通过BP神经网络建立成形质量与各段压边力之间的非线性映射关系,并以此关系作为多目标遗传算法的适应度函数进行遗传算法优化,获得一组Pareto最优解集,实现了对行李箱内板成形窗口内压边力曲线的优化。优化结果表明,相比于正交试验优选方案,采用遗传算法和神经网络相结合的方法得到的优化方案成形零件时,能较大程度地提高行李箱内板的成形质量。The acquired variable blankholder force( VBHF) forming window is divided into a number of parts in the deep drawing process for a trunk lid inner panel made of a new type of highstrength steel TRIP600. Through numerical simulation,orthogonal test dates of each part' s VBHF and the forming quality are got,and from which a primary optimal VBHF curve is obtained. Taking the orthogonal test dates to the training samples,a non linear mapping function from each part's VBHF to the forming quality is constructed within BP neural networks in order to obtain the objective function that are necessary using multiobjective genetic algorithm( NSGAII),and it has got a group of optimized Pareto solutions,in which a set of VBHF in trunk lid inner panel forming process is selected according to specific requirements. Optimization results show that: in contrast to the optimum proposal of orthogonal test,the trunk lid inner panel gets a better forming result using the optimal VBHF curve obtained by BP neural networks and NSGAII.

关 键 词：行李箱内板 TRIP600 高强钢板 变压边力曲线 神经网络 遗传算法 

分 类 号：TG386[金属学及工艺—金属压力加工]