某汽车前隔板拉延成形工艺参数多目标优化

Multi-objective optimization of drawing process parameters for an automolile front bulkhead

为解决某汽车前隔板拉延成形中产生的起皱和破裂问题,提出了一种基于拉丁超立方抽样、Kriging模型及多目标遗传算法的优化方法。首先,将拉延成形的5个工艺参数作为设计变量,将定义的破裂程度、起皱程度及安全区域内单元面积占比作为评价成形质量的指标,并使用拉丁超立方抽样和数值仿真技术构建优化样本数据;其次,使用Kriging模型对样本数据进行非线性逼近,并用NSGA-II多目标遗传算法对逼近的Kriging响应模型进行优化,得到了最优解集,即4个拉延筋阻力系数分别为0.2952、0.3475、0.2303和0.2300,压边力为537.3425 kN;然后,使用数值仿真验证了优化策略的有效性;最后,应用该最优工艺参数进行生产试制,得到了表面质量良好、无破裂及起皱缺陷的汽车前隔板零件;应用基于试验设计和Kriging模型的多目标优化方法能够控制成形质量、减少试模次数、降低生产成本。

To solve the problems of wrinkling and cracking in the drawing of an automobile front bulkhead,an optimization method based on Latin hypercube sampling,Kriging model and multi-objective genetic algorithm was proposed.Firstly,the five process parameters of drawing were taken as the design variables,the defined fracture degree,wrinkle degree and the proportion of unit area in the safety area were taken as the indicators to evaluate the forming quality,and Latin hypercube sampling and numerical simulation techniques were used to build the optimized sample data.Secondly,the Kriging model was used for nonlinear approximation of sample data,and NSGA-II multiobjective genetic algorithm was used to optimize the approximated Kriging response model,and the optimal solution set was obtained,that is,the four drawbead resistance coefficients are 0.2952,0.3475,0.2303 and 0.2300,respectively,and the blank holder force is 537.3425 kN.Then,numerical simulation was used to verify the effectiveness of the optimization strategy.Finally,the optimal process parameters were used for trial production,and the automobile front bulkhead parts with good surface quality,no crack and wrinkle defects were obtained.The application of multi-objective optimization method based on test design and Kriging model can control the forming quality,reduce the number of die tests and the production cost.