上海交通大学学报(自然版) ›› 2011, Vol. 45 ›› Issue (11): 1626-1631.

• 机械工程 • 上一篇    下一篇

基于神经网络的回转球面磨削精度过程控制

许开州,魏臣隽,柴运东,许黎明,胡德金   

  1. (上海交通大学 机械与动力工程学院, 上海 200240)
  • 收稿日期:2010-10-27 出版日期:2011-11-30 发布日期:2011-11-30
  • 基金资助:

    国家自然科学基金资助项目(51075273)

Investigation of Sphere Grinding Process Control on the Basis of Neural Network

  • Received:2010-10-27 Online:2011-11-30 Published:2011-11-30

摘要: 介绍了回转球面精密磨削系统,在研究回转球面精密磨削原理和机床结构的基础上,分析了球体形状误差的产生原因,利用BP神经网络拟合进给量、磨削区段和砂轮主轴电流三者之间的非线性关系,通过网络预测调整不同磨削区段的进给量控制电流恒定从而达到恒功率磨削.

关键词: 神经网络, 球面磨削, 球度误差

Abstract: A precision sphere grinding system was introduced. On the basis of mechanism of sphere grinding system and grinder structure, the cause of sphere form error was analyzed. BP neural network was used to fit the nonlinear relationship between feed rate, grinding section and spindle current. And  the current was kept stable by adjusting feed rate previously at different grinding section, therefore, constant power grinding was obtained.

Key words: neural network, sphere grinding, sphericity error

中图分类号: