Research On Fault Diafnosis Method Of EMU Axle Box Bearing Based On Vibration Analysis

As an important part of the bogie of high-speed EMU train,axle box bearing undertakes the important task of bearing and converting wheel rotation into linear motion of the train.Due to the harsh operating environment such as high cold,wind and sand,it is very easy to cause axle box bearing failure,which seriously affects the safety of the train.Therefore,it is very important to carry out fault diagnosis and timely monitoring of the axle box bearing of the train.At present,the main ways to monitor the health status of axle box bearings are temperature monitoring,sound monitoring and vibration monitoring.This thesis mainly uses the monitored vibration data to study the fault diagnosis method of axle box bearing.The main research work is as follows:(1)The vibration mechanism of axle box bearing is studied,the effects of internal and external factors of axle box bearing on vibration are summarized,and some main failure forms of bearing are analyzed.The natural frequency and fault characteristic frequency of the bearing are analyzed in detail,and the corresponding formulas are deduced.Finally,according to the analysis of bearing vibration mechanism,the corresponding numerical prevention model of axle box bearing vibration is established,which provides data support for the subsequent verification of bearing fault diagnosis method.(2)The fault test of axle box bearing of high-speed EMU train is introduced.This thesis introduces the basic structure of the test-bed,the selection of axle box bearing,the full working condition test process of axle box bearing,as well as the selection of vibration data acquisition equipment and vibration sensor.The test data under multiple working conditions are set by the control variable method,and each group of data is divided and stored.It provides strong data support for the subsequent research and verification of axle box bearing fault diagnosis method based on vibration data.(3)A bearing fault diagnosis method based on comprehensive KPC deconvolution is proposed.Firstly,the original vibration signal is decomposed by CEEMD(Complementary Ensemble Empirical Mode Decomposition),and the IMF(Intrinsic Mode Function)component is selected by kurtosis,correlation coefficient,fault impact ratio and their comprehensive index KPC.The reconstructed signal is deconvoluted by MOMEDA(Multipoint Optimal Minimum Entropy Deconvolution Adjusted)algorithm,and finally envelope analysis is carried out.Through the comparative analysis between the characteristic frequency obtained in the envelope diagram and the theoretical characteristic frequency,the fault location is finally determined.(4)An intelligent bearing fault diagnosis method based on multi-dimensional features is proposed.Based on Chapter 4,the diagnosis method extracts the time-domain features of the filtered signal,and combines the IMF feature vector obtained by EEMD method with the timedomain feature index to form the required feature vector.Then it is input into IPSO-BP(Improved Particle Swarm Optimization with Back Propagation)network,and finally the fault diagnosis and pattern recognition of bearing are realized.The effectiveness of the method is verified by the data sets of Case Western Reserve University and Southwest Jiaotong University.