锻压技术

汽车覆盖件拉深过程中的压边力预测

英文标题：Prediction on blank holder force of automobile covering parts in deep drawing process

作者：樊浩森胡建华白雪刘运展

单位：武汉理工大学

分类号：TG386

出版年,卷(期):页码：2017,42(7):43-48

摘要：

结合数值模拟与人工神经网络技术研究了汽车内覆盖件承载地板在冲压成形中压边力的预测。将板料模型零件导入到Dynaform中进行网格划分并对其拉深过程进行模拟仿真，结合正交试验获取不同参数条件下最佳压边力的数据样本，然后运用Matlab软件中的GRNN神经网络工具箱对数据进行训练学习，采用训练好的神经网络对板料成形过程中的压边力进行预测，获得了板料拉深过程中的压边力变化曲线。通过预测结果和模拟结果对比，预测误差在10%以内。将预测的曲线对零件模拟仿真，结果显示零件最大减薄率在25%以内，并对板料进行实际冲压验证。结果显示成形效果良好，无起皱、破裂缺陷，符合实际生产的要求，说明GRNN神经网络可以用于零件冲压过程中压边力的预测。

The prediction on blank holder force of bearing floor for automobile covering parts in the stamping process was studied combining numerical simulation with artificial neural network technique. First, the model part was imported into Dynaform for mesh generation and the drawing process was studied, and data samples of the optimal blank holder force under different parameter conditions were obtained by the orthogonal test. Then, the data were trained and learned by GRNN neural network toolbox in the Matlab software, and the blank holder force was predicted by the trained neural network. Therefore, the changing curve of blank holder force in the sheet metal drawing process was obtained, and the prediction deviation is within 10% comparing with simulation results. Furthermore, the drawing process of parts was simulated based on the predicted curve, and the maximum reduction ratio is displayed within 25%. Finally,the actual stamping test for sheet metal was conducted, and the forming effect met the actual production requirements without wrinkles and rupture defects. It indicates that GRNN neural network can be applied to predict blank holder force in the stamping process.

基金项目：

华中科技大学材料成形与模具技术国家重点实验室开放基金课题(P2015-01）

作者简介：樊浩森（1991-），男，硕士研究生，E-mail：fhs584695878@qq.com；通讯作者: 胡建华 (1966-)，男，博士，副教授，E-mail: hujianhua@whut.edu.cn

参考文献：

[1]曹传剑，曹爱霞，宋慧，等. 基于变压边力的汽车翼子板成形[J]. 锻压技术，2016，41(10):149-151.

Cao C J, Cao A X, Song H, et al. Automobile fender forming based on variable blank holder force[J]. Forging & Stamping Technology, 2016, 41(10): 149-151.

[2]王晖，卫瑞元，袁军涛，等. 基于变压边力的盒形件拉深研究[J].机械设计与制造，2010，(7): 95 -96.

Wang H, Wei R Y, Yuan J T, et al. The research of box-drawing based on variable blank-holder force[J]. Machinery Design & Manufacture, 2010, (7): 95 -96.

[3]罗昱，胡建华，陈昶,等，不同压边力加载模式对盒形件拉深成形的影响[J]. 锻压技术，2016，41(7):7-10.

Luo Y, Hu J H, Chen C, et al. Influence of different blank holder force loading modes on rectangular box drawing[J]. Forging & Stamping Technology, 2016, 41(7):7-10.

[4]饶进军，包忠诩，黄菊花. 人工神经网络技术及其在板料成形智能化中的应用[J]. 塑性工程学报，2002，9(1):17-21.

Rao J J, Bao Z Y, Huang J H. Artificial neural networks technology and its applications for intelligentization of sheet metal forming[J]. Journal of Plasticity Engineering, 2002, 9(1):17-21.

[5]郭斌，孟令启，杜勇，等. 基于GRNN神经网络的中厚板轧机厚度预测[J]. 中南大学学报:自然科学版，2011，42(4):960-965.

Guo B, Meng L Q, Du Y, et al. Thickness prediction of medium plate mill based on GRNN neural network[J]. Journal of Central South University: Science and Technology, 2011, 42(4):960-965.

[6]智会强，牛坤，田亮,等. BP网络和RBF网络在函数逼近领域内的比较研究[J]. 科技通报，2005, 21(2):193-197.

Zhi H Q, Niu K, Tian L, et al. A comparative study on BP network and RBF network in function approximation[J]. Bulletin of Science and Technology, 2005, 21(2):193-197.

[7]何灿焜. 基于人工神经网络的压边力预测技术研究[J]. 锻压装备与制造技术，2006，41(5):86-89.

He C K. Study on blank holding force forecast based on artificial neural network [J]. China Metalforming Equipment & Manufacturing Technology，2006，41(5):86-89.

[8]李奇涵，王红强，刘海静，等. 基于BP神经网络矩形盒件拉深成形变压边力的预测[J]. 锻压技术，2015，40(11):27-31.

Li Q H, Wang H Q, Liu H J, et al. Prediction of variable blank-holder force of rectangular box in deep drawing forming based on BP-neural network[J]. Forging & Stamping Technology, 2015, 40(11): 27-31.

[9]陈炜，周宏超，徐雪来，等. 盒形件变压边力拉深成形研究[J]. 锻压技术，2014，39(11):20-24.

Chen W, Zhou H C, Xu X L, et al. Research on variable blank holder force deep drawing for rectangular box parts[J]. Forging & Stamping Technology, 2014, 39(11):20-24.

[10]李琳，陈忠家，李奇，等. 盒形件分区变压边力拉深研究[J]. 锻压技术, 2016，41(2):25-29.

Li L, Chen Z J, Li Q, et al. Research on box-drawing based on variable blank-holder force[J]. Forging & Stamping Technology, 2016, 41(2):25-29.

[11]张晓斌，孙宇，代珊. 基于径向基神经网络杯形件拉深成形变压边力预测技术研究[J]. 机械设计，2007，24(8):36-38.

Zhang X B, Sun Y, Dai S. Study on the prediction technology of variable blank-holding force for deep drawing forming of cup shaped parts based on radial basis neural network[J]. Mechanical Design, 2007,24(8):36-38.

[12]田银. 基于RBF神经网络的变压边力优化研究[D]. 成都：西南交通大学，2015.

Tian Y. Research on the Optimization of Variable Blank Holder Force Based on RBF Neural Network[D]. Chengdu: Southwest Jiaotong University, 2015.

[13]Kamali K. Application of a hybrid GA-BP optimized neural network for springback estimation in sheet metal forming process[A]. Joint Congress on Fuzzy and Intelligent Systems[C]. Iran, 2007.

[14]Zcan B, Alpaslan. Artificial neural networks for the cost estimation of stamping dies[J]. Neural Computing and Applications, 2014, 25(3):717-726.

[15]赵军，马瑞，王凤琴，等. 盒形件拉深智能化控制关键技术研究[J]. 塑性工程学报，2005，12(3):67-70.

Zhao J, Ma R, Wang F Q, et al. Study on the key technologies of intelligent control for rectangular box deep drawing[J]. Journal of Plasticity Engineering, 2005, 12(3):67-70.

服务与反馈：

【文章下载】【加入收藏】