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

Abstract

Abstract 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.