摘要
基于BP神经网络与遗传算法建立了多目标优化模型,对拼焊板拉深质量评价指标进行详细分析,提出了一种改进的拼焊板拉深成形多目标优化函数。将所建立的BP-GA模型用于某车型中柱拼焊板拉深成形,将有限元分析结果作为训练样本,得到模型最大厚度、最小厚度(拼焊板两侧),以及焊缝移动最大误差3.23%、6.16%、0.47%、2.64%、6.65%。利用遗传算法寻找最佳的工艺参数,实验证明,基于BP-GA的拼焊板拉深成形工艺优化模型,能够对生产实践提供有效指导。
A multi-objective optimization model was established based on BP neural network and genetic algorithm,and welded blanks drawing quality evaluation was discussed.An improved multi-objective optimization function for optimizing the welded blanks drawing technology was proposed.The BP-GA model was used in a vehicle B-pillar drawing process of welded blanks.The maximum error of the model for the maximum thickness,the minimum thicknesses at both sides of the welded blanks,and the displacement of the weld line were 3.23%,6.16%,0.47%,2.64%,6.65% respectively after trained with FE results.The optimal parameters were obtained by genetic algorithm computation.The results show that the model for optimizing welded blanks drawing process based on BP-GA can provide effective guidance for production.
出处
《塑性工程学报》
CAS
CSCD
北大核心
2011年第3期48-52,共5页
Journal of Plasticity Engineering
基金
科学技术部2008年国际科技合作与交流专项资助项目(2008DFB50020)
关键词
BP神经网络
遗传算法
拼焊板
工艺优化
板料拉深
BP neural network
genetic algorithm
welded blanks
process optimization
sheet metal drawing
