基于响应面与遗传算法的汽车前围板拉延工艺参数优化  

Optimization on drawing process parameters for automobile front panel based on response surface and genetic algorithm

作　　者：陈青山 潘书华 黄瑶[1] 王雷刚[1] Chen Qingshan;Pan Shuhua;Huang Yao;Wang Leigang(School of Materials Science and Engineering,Jiangsu University,Zhenjiang 212013,China;Yangzhou Hongyun Vehicle Industry Co.,Ltd.,Yangzhou 225200,China)

机构地区：[1]江苏大学材料科学与工程学院,江苏镇江212013 [2]扬州宏运车业有限公司,江苏扬州225200

出　　处：《锻压技术》2023年第8期72-77,共6页Forging & Stamping Technology

基　　金：国家自然科学基金资助项目(51775249)。

摘　　要：以某皮卡汽车前围板为目标,研究型面复杂零件成形后全局质量上的控制,避免局部开裂与起皱。基于二阶多项式响应面结合遗传算法进行优化设计。首先,以最大减薄率和成形安全区比率为优化目标,采用灵敏度分析法筛选对材料拉延阶段影响较大的关键工艺参数组合;其次,通过Box-Behnken试验设计在样本空间内获取少量且分布合理的样本点,并采用AutoForm软件模拟拉延+切边工序得到的各样本点的成形质量结果;然后,以二阶多项式拟合设计变量与优化目标的响应面模型,并检验其精度。最后,通过遗传算法寻优工艺参数组合并将其用于指导试生产,最终测量结果表明,零件的最薄厚度为0.64 mm,整体拉延充分,无起皱与开裂现象。For the front panel of a pickup truck,the overall quality control of parts with complex shapes after forming was studied to avoid local cracking and wrinkling,and based on second-order polynomial response surface combined with genetic algorithm,the optimal design was conducted.Firstly,taking the maximum thinning rate and the forming safety zone ratio as the optimization objectives,the key process parameter combination that had a greater impact on the material drawing stage was selected by the sensitivity analysis method.Secondly,the less sample points with reasonable distribution in the sample space were acquired by Box-Behnken experiment design,and the drawing+trimming process was simulated by software AutoForm to obtain the forming quality results of each sampling point.Furthermore,the response surface model of design variables and optimization objectives was fitted by the second-order polynomial,and its accuracy was tested.Finally,the combination of process parameters was optimized by the genetic algorithm to guide the trial production.The final measurement results show that the minimum thickness of the part is 0.64 mm,and the overall drawing is sufficient without wrinkling and cracking.

关 键 词：汽车前围板 冲压工艺 灵敏度分析 二阶多项式响应面 多目标优化 遗传算法 

分 类 号：TG386.3[金属学及工艺—金属压力加工]