Study Of Fault Diagnosis And Prediction Of The Inner Ring Of The Rolling Bearing Based On Fractal

With th e rapid development of machinery and equipment industry,the relevant fault diagnosis techniques become increasingly important.This paper describes the historical development of mechanical fault diagnosis,and current development of rolling bearing fault diagnosis technologies have been discussed in this paper.This paper introduces the basic structure of the rolling bearing,the main reason for failure and failure modes,and empirical formula for calculating rolling bearing fault characteristic frequency has been studied.Spectrum analysis methods and simulation methods also have been discussed.The above mentioned three methods have been used for diagnosing early weak inner bearing failure of bearing 6205-2RS JEM SKF.The results show that for the same bearing inner ring failure,the obtained fault character frequency by these three methods differ from each other.Taking into account the non-stationary characteristics of early weak bearing fault signal,pointing out it can be used to further develop fault diagnosis method based on fractal methods.For the actual sampling signals,four kinds of fractal characters recognition method have been investigated in this paper,which are based on frequency analysis,based on fractional Brownian motion,based on properties of self-similarity and self-affinity,based on wavelet coefficients.The paper also studied three specific algorithm of fractal dimension,pointed out that the fractal dimension can assist with the diagnosis of bearing failure.By using the signal discrimination method of fractal characteristics,research and found a weak early bearing failure signal has significant fractal properties.For the fault bearing vibration signal,the corresponding three fractal dimensions have been calculated,indicating that the fractal dimension as an indicator of signal complexity,can quantitatively describe the rolling bearing weak fault behavior,it also can sensitively reflect the rolling bearing faults.Fractal dimension provides a good monitoring tool for early detection of potential failure.This paper also researched rescaled range analysis(R/S analysis method),Hurst exponent algorithm,and Vn curve.Based on the existing bearing fault vibration signal data,some fault predictions have been carried out.Research shows that combining the Hurst exponent and Vn curve,depending on the size of Hurst exponent and evolution of Vn curve,it can be a good predictor of signal’s future trends in a period of time to achieve short-term or medium-term forecast failures.The methods of fractal feature judgment have been discussed in this paper,it can be found for the weak rolling bearing fault behavior it has definite fractal characters,and fractal dimensional,rescaled range analysis,Vn curve have been used to rolling bearing’s early fault diagnosis and prediction.The research methods and research work also have practical reference value for the field of fault diagnosis.