Research On Fault Diagnosis System Of Axle Box Bearings For Subway Trains

In recent years,the metro industry has developed rapidly in China and the major cities in the country have built or are building their own metro transit networks.The development of the urban rail transit mitigates traffic pressure greatly,which provides great convenience for people to travel.However,compared with convenience the safety is more important.As one of the key components of the train,whether the axle box bearing operating well directly affects the running status of the train.Due to the disadvantage of the maintenance work in the past that it couldn't detect the fault of the bearings accurately and in time,a fault diagnosis system is to be designed and developed which would be capable of timely and accurate diagnose.Firstly,the research and development of the method of rolling bearing fault diagnosis is introduced and the research content and structure of the paper is determined on the basis of the content of rolling bearing fault diagnosis.Then the paper introduces the basic structure and the common forms of failure of the axle box bearing and selects the vibration analysis as the method after the comparison of several methods of rolling bearing fault diagnosis.After that,the paper introduces the vibration mechanism of rolling bearing and explains the concept of the fault feature frequency.Then the denoising and filtering of the original signal is researched.The denoising algorithm on wavelet packet decomposition is selected.After understanding the theory of the algorithm,its advantage is highlighted by signal simulation test and comparison with Chebyshev low-pass filter.And then the effectiveness of the algorithm is validated through the denoising and filtering of the fault data of bearing.Fault feature extraction is researched in the next step.The paper mainly introduces time domain parameters analysis method and time-frequency domain analysis method.And it focus on the research of empirical mode decomposition(EMD)and others improved method like ensemble empirical mode decomposition(EEMD)and noise assisted multivariate empirical mode decomposition(NAMEMD).Next the advantage of NAMEMD is determined by signal simulation test after theory introduction.Then it is proved to be able to extract fault feature effectively by analyzing the bearing fault data using NAMEMD.Lastly,an metro train axle box bearing fault diagnosis system software is designed and developed using Microsoft Visual Studio and Microsoft SQL Server,which takes wavelet packet denoising and NAMEMD as the core.