Study On Fault Diagnosis System For The Wheelset Bearing Of EMU

As is known to all, rolling bearing is a kind of usual and dominant general components.Whether the rolling bearing operates well directly affects the whole function of a macine.According to statistics, rolling bearing brings about 30 percent failure of the whole rotating system which uses it. Moreover, every year train failures which are brought about by rolling bearing cause great property damage and casualties in China. Therefore, it is both theoretically and practically important for us to develop a kind of effective and efficient fault diagnosis system for rolling bearing especially without disassemble in order to save maintenance cost, avoide train operation accident and ensure the safety of train operation.In consideration that Vibration Analysis Method(VAM) is suitable for failure detections of various bearing and has relatively simple signal processing, high credibility of diagnosis and is also susceptible to feeble early failure. So it is selected to study failure diagnosis of rolling bearing. The paper mainly consists of the following aspects:Firstly, the paper introduces some common failures of rolling bearing and summarizes the cause of various failures. Based on fully understanding the vibration mechanism of rolling bearing, the different impact on vibration which created by structural properties,manufacturing deviations and the defect during running processes are contrasted and then the eigenfrequency of failure at various parts are calculated. Finally, the behaviors of amplitude and frequency under different failure modes are discussed so that it can prepare for the later failure diagnosis better.Secondly, the paper illustrates some common analysis methods for vibration signals of rolling bearing including time domain analysis, frequency domain analysis and time-frequency domain analysis. Their characters and applications in failure diagnosis are introduced in detail and their pros and cons are also summarized so that it can choose the appropriate methods to analyze the fault.Thirdly, the paper mainly introduces definition, characteristic, principle, algorithm and classification of the Wavelet Transform(WT) for showing that WT is a practical method for equipment fault diagnosis. By contrasting pros and cons of various methods in wavelet denoising, the course of wavelet denoising are explained in particular. Finally the failure diagnosis method that combines kurtosis, wavelet analysis and Hilbert envelope analysis is investigated and applied to failure diagnosis of rolling bearing to verify its effectiveness and correctness.Finally, the rolling-bearing-failure-diagnosis(RBFD) system is designed by Visual Basic6.0 and Matlab 7.0 Which includes planning functions, data collecting, signal processing diagnosing, general enquiries, test reports printing. Using the RBSD system that is introducedin this paper, whether there is failure in the bearing and where the failure is can be determined initially. The running risk which caused by bearing failure is accordingly reduced. Moreover,this system also promotes the development of the next generation RBSD system.