汽车覆盖件拉深过程中的压边力预测

发布时间：2018-04-10 16:30

本文选题：压边力预测 + 冲压　；参考：《锻压技术》2017年07期

【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。
[Abstract]:Combined with numerical simulation and artificial neural network technology, the prediction of blank holder force in stamping forming of automobile inner panel bearing floor is studied.The sheet metal model parts are imported into Dynaform for mesh division, and the drawing process is simulated and simulated. The data samples of the best blank holding force under different parameters are obtained by orthogonal test.Then the GRNN neural network toolbox in Matlab software is used to train and study the data, and the trained neural network is used to predict the blank holder force in sheet metal forming process, and the change curve of blank holder force in sheet metal drawing process is obtained.By comparing the prediction results with the simulation results, the prediction error is less than 10%.The predicted curves are simulated to the parts, the results show that the maximum thinning rate of the parts is less than 25%, and the actual stamping verification of the sheet metal is carried out.The results show that the forming effect is good, no wrinkle and fracture defect, which meets the requirements of actual production. It shows that GRNN neural network can be used to predict the blank holding force in the stamping process of parts.
【作者单位】：武汉理工大学材料科学与工程学院;
【基金】：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)
【分类号】：TG386.32;U466

【相似文献】