基于神经网络和有限元法的“V”型件翻边成形性能研究  

Research on Flanging of V-Shaped Sheet Metal Part Based on Artificial Neural Network and Finite Element Method

作　　者：罗应兵[1] 董俊[1] 李大永[1] 张少睿[1] 彭颖红[1] 

机构地区：[1]上海交通大学机械与动力工程学院知识工程研究所,上海200030

出　　处：《机械科学与技术》2004年第6期693-695,731,共4页Mechanical Science and Technology for Aerospace Engineering

基　　金：上海市科技启明星科技跟踪计划项目(01QMH1411);上海市博士后基金 2 0 0 1资助

摘　　要：翻边是一种重要的金属成形工艺 ,曲面翻边的成形性能与几何尺寸及材料参数等因素有很大关系。本文将有限元模拟与人工智能相结合 ,建立了基于改进的BP神经网络的“V”型件翻边成形性能与几何尺寸及材料参数关系的数学模型。与单纯采用有限元法相比 ,该模型具有计算时间短和易使用的优点。与数值模拟和实验结果的比较证实了本模型的有效性。Flanges are a very common feature of sheet metal parts, it is found that the geometry and material have great effect on the deformation in the curved flange. Based on the integration of ameliorated BP neural network and finite element method, a model for prediction of the flanging of V shaped sheet metal parts was presented. Compared with the traditional finite element method, the model have the advantages of little computation time and convenience. The result of simulation and experiment show that the model is effective and feasible.

关 键 词：有限元 BP神经网络 “V”型件翻边 

分 类 号：TG302[金属学及工艺—金属压力加工]