首页|汽车后风挡下横梁热冲压工艺的二次成型研究

汽车后风挡下横梁热冲压工艺的二次成型研究

扫码查看
为提升汽车后风挡下横梁热冲压工艺的二次成型效果,提出一种基于带精英策略的快速非支配排序遗传算法(NSGA-Ⅱ)优化梯度提升回归树(GBDT)的汽车后风挡下横梁热冲压工艺的二次成型方法.首先,采用拉丁超立方抽样技术对汽车后风挡横梁热冲压工艺参数进行区间划分;然后通过GBDT算法建立数学代理模型;最后以成型件最大增厚率和最大薄减率为优化目标,通过NSGA-Ⅱ多目标优化算法对二次成型的工艺参数进行优化.实验结果表明,GBDT模型的最大增厚率和最大减薄率的决定系数R2分别为0.9573和0.9712,均方误差分别为0.0019和0.0034.由此说明,该模型的收敛速度较快,收敛精度高.测试发现,采用NSGA-Ⅱ多目标优化算法预测输出的最大增厚率和最大减薄率分别取值为3.5%和6.41%,与真实值十分接近.综合分析可知,提出的方法能够实现热冲压工艺的二次成型优化,提升了二次成型效果,具有实际应用意义.
Study on secondary forming of rear windshield crossbeam
In order to improve the secondary forming effect of the thermal stamping process of the rear windshield beam under the vehicle,a secondary forming method for the thermal stamping process of the rear windshield beam with the rapid nondominated sorting genetic algorithm Ⅱ(NSGA-Ⅱ)optimal gradient boosting decision tree(GBDT)is proposed.Firstly,the latin hypercube sampling technique is used to divide the process parameters of the rear windshield beam;then the mathematical proxy model is established through the GBDT algorithm;finally,the maximum thickening rate and the NSGA-Ⅱ multi-objective optimization algorithm.The experimental results show that the determination coefficient R2 of the maximum thickening rate and the maximum thinning rate of the GBDT model is 0.957 3 and 0.971 2,with a mean square error of 0.001 9 and 0.003 4,respectively.This shows that the model has a faster convergence rate and a high convergence accuracy.The test found that the maximum thickening rate and maximum thinning rate of the NSGA-Ⅱ multi-objective optimization algorithm were 3.5%and 6.41%,respectively,which were very close to the true value.Comprehensive analysis shows that the proposed method can realize the secondary forming optimization of hot stamping process,improve the secondary forming effect,has practical application significance.

automotive rear windshieldhot stamping processsecondary forminggradient boosting decisisn tree(GBDT)nsndominated sorting genetic al gorithm Ⅱ(NSGA-Ⅱ)

同晶

展开 >

渭南职业技术学院,陕西 渭南 714000

汽车后风挡 热冲压工艺 二次成型 梯度提升回归树(GBDT) 非支配排序遗传算法(NSGA-Ⅱ)

2024

模具技术
上海交通大学

模具技术

CSTPCD
影响因子:0.219
ISSN:1001-4934
年,卷(期):2024.(6)