Research On Fault Diagnosis Of CNC Servo Turret Based On Timing Sequence Division

The CNC servo turret is one of the key components of the CNC lathe.The level of its working state directly affects the processing quality of the workpiece.Frequent failures lead to a significant reduction in production efficiency and also reduce the social reputation of the product.Therefore,carry out research on the fault diagnosis technology of the CNC servo turret,deeply analyze the typical fault characteristics of the CNC servo turret,realize the identification of its abnormal state,the rapid fault locating and the rapid typical fault diagnosis,which provides a theoretical basis for the level promotion of predictive maintenance and reliability of the product,and has certain theoretical research significance and engineering application value for improving the product level.This paper takes a domestic CNC servo turret as the research object,and combines with the needs of the research on national key science and technology special projects,and carries out a research on method of the CNC servo turret fault diagnosis.Focusing on the timing characteristics of CNC servo turret work,this paper proposes a method of dividing the timing sequence based on its work cycle,and explores the signal processing and feature extraction methods by time-domain statistical analysis and EEMD analysis in different timing sequence.Based on this,this paper proposes a typical fault diagnosis method of CNC servo turret based on the PSO-SVM and FSVM is established.Finally,the validity of the model is verified through experiments.The specific research contents are as follows:(1)Determination of typical failure of CNC servo turret: According to the structure and working principle of a domestic CNC servo turret,its system is divided into 5 subsystems according to the principles of space area,independent function and modular design.Based on the subsystem division,by the field and bench test data of the CNC servo turret,the RPN method was used to determine the typical fault and location.(2)Research on the characteristic extraction method of the typical fault signal of the CNC servo turret based on the timing sequence division: through the analysis and determination of the typical failure and working characteristics of the CNC servo turret in the previous chapter,the vibration,oil pressure,current and proximity switch signals are determined as fault feature extraction signal;according to the timing characteristics of the CNC servo turret,a complete working cycle is divided into the timing sequence T1-turret release process,timing sequence T2-knife indexing process,timing sequence T3-locking process and timing sequence T4-turret cutting process.According to the time-domain characteristics and time-frequency characteristics of the signals in different timing sequence,the state feature vectors of different timing sequence are determined.(3)Research on fault diagnosis methods based on different timing sequence: According to the SVM theory and the overall high-dimensional small sample characteristics of CNC servo turret fault data,RBF kernel function is selected to build its core theoretical model of fault diagnosis.Different support vector machine optimization methods are selected according to the sample characteristics of each timing sequence,which develops the study respectively that the fault diagnosis methods based on FSVM theory for T1 and T4 timing sequence and the fault diagnosis methods based on PSO-SVM theory for T1 and T3 timing sequence.By synthesizing the fault diagnosis methods of 4 timing sequence,it realizes the fault diagnosis of the complete working cycle of the CNC servo turret.(4)Research on CNC servo turret typical fault diagnosis experimental: on the basis of the test bench of the Key Laboratory of Reliability of CNC Equipment in the Machinery Industry,a status information collection and fault diagnosis analysis system of CNC servo turret was established,to develop research about the typical fault experimental of the CNC servo turret.Research.FSVM and SVM were used to verify and compare the fault diagnosis of the timing sequence T1 and T3 of the CNC servo turret.The diagnostic accuracy of the FSVM model is better than that of the SVM model.The PSO-SVM and SVM were used to verify and compare the fault diagnosis of the timing sequence T2 and T4 of the CNC servo turret.The results proved that the former method had better diagnostic accuracy,and also verified the effectiveness of the division of the CNC servo turret working cycle.