Research On Fault Diagnosis Algorithm Of Axle Box Bearing Of EMU

Rolling bearings are mechanical joints that play an important role in rotating machinery.Under the influence of various factors in the complex environment,rolling bearings have become one of the most vulnerable parts.Its state directly affects the safe operation of the train,especially in the high-transmission EMUs.It is easy to cause major accidents when the axle box bearing replacement in the running part is not timely.Therefore,it is very necessary to study the fault diagnosis technology of the EMU rolling bearings and improve the diagnostic accuracy,the accident rate will reduce and the safe operation of the vehicle will be guaranteed.In this thesis,the axle box bearing of the EMU is taken as the research object,and the related key technologies of fault diagnosis are analyzed.The main contents and results of this thesis are as follows:(1)The basic composition of the rolling bearing is studied,and the failure mechanism,the causes,the basic forms and frequency of vibration characteristics are analyzed.The common diagnostic methods are summarized and the advantages and disadvantages of the related algorithms are analyzed;(2)Based on the classical mathematical model widely used in the industry,the vibration signals of the inner rings,outer rings and rollers of the CRH380BL EMU are simulated to prepare for verifying the feasibility of the subsequent algorithm;(3)For the signal preprocessing stage,considering the non-stationary characteristics of the fault vibration signal,a wavelet denoising method is presented,and a new wavelet threshold function denoising method is proposed.This new threshold function reduces the signal distortion and oscillation problem in the denoising process on the basis of retaining the advantages of traditional soft-hard threshold function noise reduction.And compared it with other improved threshold functions,the superiority of this method is verified;(4)For the feature extraction problem of vibration signals,a method combining improved EEMD and sample entropy is adopted.The method eliminates the false components in the EEMD,and uses the sample entropy as the characteristic value of the real components.Sample entropy can truly reflect the characteristics of the signal,thereby the different fault states of the roller bearing can be better distinguishing;(5)For the fault diagnosis classification of rolling bearings,the characteristics of the samples is analyzed and the intelligent SVM is selected.In order to improve the diagnostic accuracy,the SOA algorithm with good global performance is used to optimize the parameters of SVM.The SOA-based SVM fault diagnosis algorithm is established.Its feasibility is verified by simulation data;(6)The validity of the proposed method is verified by collecting signals of normal state,outer ring fault,inner ring fault and roller fault from the laboratory high-speed train vibration test bench.Compared with the traditional algorithm and other literature algorithms,the superiority of this algorithm is proved.