• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2020, Vol. 56 ›› Issue (1): 148-156.doi: 10.3901/JME.2020.01.148

• 数字化设计与制造 • 上一篇    下一篇

基于改进BP神经网络面向STEP-NC 2.5D制造特征的智能宏观工艺规划

张禹1,2, 李东升1, 董小野1, 王志伟1, 杨树华2, 巩亚东1   

  1. 1. 东北大学机械工程与自动化学院 沈阳 110819;
    2. 沈阳鼓风机集团股份有限公司 沈阳 110869
  • 收稿日期:2019-02-09 修回日期:2019-07-25 出版日期:2020-01-05 发布日期:2020-03-09
  • 通讯作者: 张禹(通信作者),男,1979年出生,博士后,副教授,硕士研究生导师。主要研究方向为STEP/STEP-NC,智能设计与制造,复杂装备数字化设计与制造。E-mail:zy4097534@126.com
  • 基金资助:
    国家自然科学基金(51205054)、中国博士后科学基金(2017M611245)、中央高校基本科研业务费专项资金(N180313010)、中央高校基本科研业务费专项资金(N160304009)、辽宁省自然科学基金(2019-MS-124)和东北大学博士后基金资助项目。

The Intelligent Macro-process Planning Based on an Improved BP Neural Network for STEP-NC 2.5D Manufacturing Features

ZHANG Yu1,2, LI Dongsheng1, DONG Xiaoye1, WANG Zhiwei1, YANG Shuhua2, GONG Yadong1   

  1. 1. School of Mechanical Engineering and Automation, Northeastern University, Shenyang 110819;
    2. Shenyang Blower Works Group Corporation, Shenyang 110869
  • Received:2019-02-09 Revised:2019-07-25 Online:2020-01-05 Published:2020-03-09

摘要: 智能工艺规划是实现智能制造的关键技术。面向新型NC编程数据接口国际标准STEP-NC (STEP-compliant numerical control,STEP-NC),一种基于改进BP神经网络面向STEP-NC 2.5D制造特征的智能宏观工艺规划方法被提出。在该方法中,用于STEP-NC 2.5D制造特征加工操作方法决策的集成BP神经网络模型首先被建立。考虑到BP神经网络容易陷入局部最优和收敛速度慢,将混沌算法、遗传算法和BP神经网络算法相结合得到用于加工操作方法决策的改进的BP神经网络。在此基础上,将零件的加工信息归一化处理后输入到改进的神经网络中,进而实现了STEP-NC 2.5D制造特征加工操作方法的智能生成。最后,通过实例验证了该方法的有效性和可行性。

关键词: STEP-NC, 宏观工艺规划, 加工操作方法, 改进BP神经网络

Abstract: The intelligent process planning is a key technology to realize intelligent manufacturing. For the novel NC data model called STEP-NC, an intelligent macro-process planning method based on an improved BP neural network for STEP-NC 2.5D manufacturing features is proposed. In the method, an integrated BP neural network model used for machining operation decision-making of STEP-NC 2.5D manufacturing features is firstly established. Then, considering the problems of BP neural network such as local optimum and slow convergence speed, an improved BP neural network algorithm for machining operation decision-making of STEP-NC 2.5D manufacturing features is given by combining chaotic algorithm, genetic algorithm with BP neural network. Further, corresponding machining operation methods are intelligently obtained after machining information of a part is normalized and input into the improved neural network. At the end, it has been concluded by a case that the proposed method is effective and feasible.

Key words: STEP-NC, macro-process planning, machining operation method, improved BP neural network

中图分类号: