Electrical System Design and Fault Analysis of Machine Tool Based on Automatic Control

Abstract

Automatically controlled machine tools have been used extensively in the industrial field, and fault analysis methods have garnered increasing attention. This paper first describes the software and hardware design of a machine tool and then presents a fault analysis of the machine tool. The fault types of machine tools are analyzed. A signal is obtained from a vibration sensor, the characteristic value is extracted, and the fault is analyzed using a back-propagation neural network (BPNN). The experimental results show that the BPNN yields the best performance when the structure is 8-9-8, and its recognition rate is 97.22% for different types of faults. Meanwhile, the recognition rate of naive Bayes is only 76.73%, and that of a support vector machine is only 85.55%, which is significantly lower than that of the BPNN. The results show that the BPNN is effective in fault analysis and can be promoted and applied more extensively.