基于BP神经网络和有限元的剪切旋压工艺辅助设计  

Process Aided Design of Shear Spinning Based on the BP Neutral Networks and FEM

作　　者：金明冬[1] 李从心[1] 张剑[1] 贺成志[1] 

机构地区：[1]上海交通大学国家模具CAD工程研究中心,上海200030

出　　处：《锻压装备与制造技术》2007年第2期99-102,共4页China Metalforming Equipment & Manufacturing Technology

摘　　要：以生产数据为学习样本,建立BP神经网络模型,实现对剪切旋压工艺参数的预测。根据预测的工艺参数对锥形件强力旋压过程进行模拟,建立有限元模拟后处理评价模型。实验证明将人工神经网络和有限元结合,用于旋压工艺的制定过程,实现旋压工艺制定的智能化,可加快旋压工艺的制定过程。A BP neutral networks model has been established in order to predict process parameter of shear spinning based on many data accumulated during manufacturing.The process of power spinning for taper has been simulated and an estimation model of finite element post-process has been established according to predicting pro- cess parameters.It is proved that combing the integration of neutral networks and finite-element in design process of power spinning can realize the intelligence of planning process and fasten the process.

关 键 词：机械制造 强力旋压 BP神经网络 有限元模拟 

分 类 号：TG335.19[金属学及工艺—金属压力加工]