Research on the commonness and dissimilarity of group machine tools based on BP and SAE algorithms

作     者：Yuan, Xuwu Hao, Caihua Gu, Jiarong Mao, Xinyong Liu, Hongqi Zhang, Jie Jiang, Xuchu Zhu, Hongguang 

作者机构：Huazhong Univ Sci & Technol Sch Mech Sci & Engn Wuhan 430074 Hubei Peoples R China Wuhan Intelligent Equipment Ind Technol Inst Co L Wuhan 430074 Hubei Peoples R China Huazhong Univ Sci & Technol Sch Energy & Power Engn Wuhan 430074 Hubei Peoples R China Zhongnan Univ Econ & Law Sch Stat & Math Wuhan 430074 Hubei Peoples R China CSIC 722 Res & Dev Inst Wuhan 430079 Peoples R China 

出 版 物：《COMPUTERS & INDUSTRIAL ENGINEERING》 (计算机与工业工程)

年 卷 期：2021年第158卷

核心收录：

学科分类：1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：National Natural Science Foundation of China Major scientific and technological innovation projects in Shandong Province [2019JZZY010442] Natural Science Foundation of Hubei Province, China [2020CFB180] 

主　　题：Group machine tools Machine tool dynamics Deep learning algorithm 

摘      要：The use of group machine tools has become one of the important production methods for large-scale manufacturing of high-precision parts. Due to the impact and wear of the group machine tool for long-term machining, the dynamic characteristics of the machine tool structure change significantly, resulting in a decrease in product quality;at the same time, the random errors in manufacturing and assembly make the group machine tool structure dynamics more different. Studying the dynamic differences of group machine tools will help improve the processing of group machine tools. This paper proposes a deep learning algorithm based on Back-propagation (BP) neural network and Stacked Auto Encoder (SAE) network to study the dynamics of the machine tool structure under static and dynamic conditions. The BP neural network is used to train the characteristics of some machine tools, and the dynamic characteristics of other groups of machine tools are predicted by changing the parameters. Then SAE network is used to find out the influence rule of group machine tool structure, changing on group machine tool dynamic characteristics. Experiments have proved that this method has certain significance for the study of dynamic characteristics of group machine tools.