Mechanical Fault Diagnosis Research Based On Wavelet Analysis Method

Gearbox and bearing are the most familiar components of the equipments. Its running states will affect the whole machine's function. The fault principle and the fault vibration property were analyzed in this paper. The means of vibration signal analysis and fault diagnosis of the gearbox were also introduced, and a new signal analysis method is proposed after wavelet packet de-noising.Excellent effects can not be got by using regular Fourier analyzing the non-stationary vibration signals produced by the fault gearbox. The wavelet analysis method, which can analyze the signals both in time and frequency domains, very suitable for the fault diagnosis of the gearbox.The paper takes gearbox, sliding bearing and automobile engine as research object. Reducer fault diagnosis base on wavelet envelope spectrum analysis is proposed by wavelet theory study; Reducer fault diagnosis base on wavelet packet energy monitoring is proposed by wavelet packet study; Gearbox, sliding bearing and automobile engine de-noising method base on wavelet packet transform are proposed by wavelet packet transform study; Wavelet analysis method and wavelet packet de-noising method are simulated by matlab wavelet toolbox study.A residual signal spectrum analysis and statistics analysis method is proposed after wavelet packet de-noising. Firstly, decompose the original vibration signal by wavelet packet theory, and extract the residual signal between the de-noising vibration signal and the original vibration signal. Finally, extract the trait frequency of fault signal by residual signal spectrum analyze,judge the machine run state by residual signal statistics analysis.