Research For Rolling Bearing Fault Diagnosis System Based On Cyclostationary

Rolling bearing is one of the widely used parts in machinery and equipment.Its working condition and use environment affect production quality and safety of the whole machinery and equipment.Especially today modern industry is the development of intelligent,it is very important practical significance to research rolling bearing on the condition monitoring and fault diagnosis.Due to signal of the rolling bearing vibration has cyclostationarity,In cyclostationary research bearing fault diagnosis is a feasible method.On the basis of cyclostationary theory,this paper introduces adaptive filtering performance of the empirical mode decomposition method.The detailed method of calculation is given.Using the EMD analysis and cyclostationary,this paper proposes a improved algorithm based on cyclic autocorrelation function.It enhances the ability of fault recognition.According to the definition and related principle of second-order cyclic statistics,by the calculation method of the cross-correlation function of FFT.Researching the cyclic autocorrelation function has influence on the fault feature extraction.After The experimental signal is filtered on the EMD,then through it calculates the difference in value of failure of inner ring,outer ring and rolling element characteristic frequency points.Select the IMF components of the biggest fault difference for cyclic autocorrelation function analysis.And comparing with the method of vibration signal analysis.Cyclic autocorrelation function and the EMD method is proven to be practical in the rolling bearing fault diagnosis.Found this method have the effectiveness of feature extraction and the noise reduction of the bearing vibration signal.This paper designed and developed system of acquisition and processing of the vibration signal in the virtual instrument.The rolling bearing fault signal data were tested in MATLAB and it compared with method of vibration signal analysis.And it experiment by the method for bearing fault diagnosis.As the experimental results suggest.This method not only can improve the performance of noise reduction,but also it can improve the ability of fault feature extraction and diagnose fault effectively.So it is great practical value for the fault diagnosis method to combine the EMD with cyclostationary.