Fault Diagnosis Of Cnc Machine Tool Spindle Servo System

Fault diagnosis is a multiple technology, which involves some subjects, such as modern control theory, signal process and pattern identify, computer science, artificial intelligence, electronic technology, mathematic statistics etc. It makes the modern manufacturing system more and more complicated with the high-speed development of the microelectronics and computer technology, that means the requirements of the major equipment-the Computerized Numerical Control (CNC) machine in manufacturing industry is of a higher level. And the servo system of spindle is an important part of the CNC machine. So it has an extraordinary sense to study on fault diagnosis of spindle servo system of CNC machine.This paper combined the Fuzzy Logic with Artificial Neural Network to produce a new structure of the neural network:Fuzzy Neural Network, FNN. FNN combined each advantage of fuzzy logic and neural network, it's not only imitate mankind's logical thinking and describe the fuzzy problems, but also imitate the functions of human brain's neural units and has the strong learning ability. The fault diagnosis on spindle servo system of CNC machine has superiority applied with FNN.The thesis analyzed the faults on the spindle servo system, and selected 7 couples of representative fault appearance and cause to sum up 28 couples of the fault samples. Then training the fault samples based on Levenberg-Marquardt algorithm of improved BP neural network, simulated with MATLAB, the results show the validity of the fault diagnosis on the servo system of CNC machine by using improved BP neural network.The article developed the fault diagnosis Expert System based on Fuzzy Neural Network with MATLAB GUI. The system combined the advantages of Fuzzy Neural Network and Expert System. It not only makes the knowledge library reduced, but also accelerates the reasoning speed. And It has good fault tolerance. The system has kindly Human Machine Interface (HMI) with good interaction.Finally, the thesis inquired into the further work on the fault diagnosis system.