The Research On Fault Diagnosis And Prediction Of Servo System On CNC Milling Machine

CNC machine tool’s occupancy rate continues to increase in the manufacturing equipment and it is the key equipment in the production. So when fault comes out and can not be timely and correctly diagnosed and maintained, it will bring great economic loss. As an important part of the CNC machine, any part of the servo system fault will affect the processing precision and efficiency in the continuous production process, and it may damage the whole machine tools, even affect the normal operation of entire production line, which could cause heavy economic loss. Timely and accurately determine servo system’s fault can reduce the production cost, improve the production efficiency and reliability. With the support of national natural science foundation project, Study of Online Fault Diagnosis and Optimum Maintenance of the Dynamic System with Environmental Impact, and the Project of the Department of Education, Jilin Province, Study of Fault Diagnosis and Prediction Technology of Precision Electromechanical Systems, we study deeply the fault diagnosis and prediction of the CNC milling machine’s servo system.In order to realize fault diagnosis and forecast of the CNC machine precisely, this paper makes CNC milling machine as the research object, and it is a typical representative of the CNC machine tools. The fault mechanism of the major parts of the numerical control machine tool servo system is analyzed deeply, and the major parts include transmission mechanism, drive system and detection device.Based on fault mechanism of CNC milling machine servo system, the bearing fault diagnosis is realized using wavelet method, and the motor fault is diagnosed based on extended Kalman filter method. Because of the complexity of the servo drive system and the limitations of various fault diagnosis methods, the problem of fault diagnosis can not be solved by one diagnosis method.Belief rule-base inference methodology using the evidential reasoning approach(RIMER) is put forward, and it effectively use all the uncertainty of quota information and qualitative knowledge, and it can realize the fault diagnosis of complex mechanical and electrical system. RIMER mainly includes the expression of knowledge and knowledge reasoning. The knowledge expression is achieved through the BRB expert system, and knowledge reasoning is realized through the evidence reasoning (ER) algorithm. BRB system consists of a series of confidence rules, which in essence is a kind of expert system, and it can effectively use various types of information to set up the nonlinear model between input and output. Finally the optimization model of the BRB expert system is proposed.According to the faults characteristics of CNC milling machine, combined with the above work, the fault and forecast CNC milling machine servo system is studied deeply. The machine working table fault is diagnosed using Belief rule-base method. Based on the Belief rule-base inference methodology, this paper proposes a fault prediction model, and this model can realize comprehensive fault prediction, using multiple characteristic features of the half quantitative information. Based on the model, the ball screw fault of servo system is predicted. The simulation experiment shows the method can make full use of various uncertain information, and improve diagnosis and the forecast accuracy.Based on LabVIEW, the system software of fault diagnosis and prediction of servo system is developed to carry out diagnosis and prediction functions. The experiment and practical application demonstrates that the designed fault diagnosis and prediction has high accuracy in the fault diagnosis and detection, and the system can run stably and reliably.