Study On Fault Diagnosis Method For High Speed Traction Motor Bearings Based On Wavelet Analysis

With the appearance of high speed trains in recent years, the working conditions of high speed traction motor is more strictly inspected.The situation monitoring and fault diagnosis technology of the mechanical equipment overseas have come to the stage of application. Howerver, in our country, it needs to develop rapidly not only in fault diagnosis theory and methods but also in the engineering practice and development of monitoring diagnosis production, although the fault diagnosis has developed more than 20 years. Wavelet and Hilbert transform, Wavelet packet transform, EMD decomposition method used for the diagnosis of high speed tracking motor bearings in time domain and frequency domain are employed in the dissertation. The software platform is developed by using LabVIEW and Matlab.(1) The method of diagnosis of high speed traction motor bearings via Wavelet and Hilbert transform was studied.Firstly, Daubechies wavelet was applied to decompose the vibration signal of the fault bearings, and then the characteristic fault frequency imformation was extracted to determine the type of axial linear crack faults effectively.(2) Wavelet packet was applied in the fault diagnosis of high speed traction motor rolling bearings.The optimum plot method according to energy after wavelet packet decomposition was proposed. Threshold of wavelet packet is the sum of average and variance value, which is used to signal de-noise and reconstruct. The wavelet package was selected to do the Fast Fourier Transform and then characteristic frequency imformation was extracted to determine the type of axial linear crack faults simply.(3) Combination of Wavelet packet and EMD decomposition, the new method was explored and researched.Wavelet packet and EMD decomposition were applied to fault diagnosis of high speed traction motor bearings. Firstly, Wavelet packet was used to de-noise, EMD was used to decompose and then the IMFs with biggest correlation coefficients were selected to decomposition again with Wavelet packet and the optimum packet was selected for FFT. At last, the type of axial linear crack fault in frequency domain was determined. Although this method can be used to diagnose the fault character, too much time cost to the balance decomposition, so difficult to be combined to software. (4) Test platform construction and experiment research.Experiment verification has done via wavelet analysis after data acquisition in the test platform. Wavelet and Hilbert transform, Wavelet packet transform, EMD decomposition method were used to do the diagnosis application of fault bearings.And then the most effective diagnosis method—Wavelet packet method was founded.(5) The software platform is developed by using LabVIEW and Matlab.Taking the excellence of Matlab data process and applying yulewalk multi-band pass filter to get the characteristic frequency. And get the software of fault diagnosis. Finally,.the type of fault can be diagnosed.