Research On Fault Diagnosis Of Rolling Element Bearing Based On Low-order Cyclostationary Theory

Rolling element bearing is one of the most widely used machine parts,and the most susceptible machine parts.The normal operation of which directly affects the running state of the equipment,the damage of the rolling element bearing will increase the vibration and noise,decrease the accuracy of the equipment,and even cause damage to equipment and stop working,so it is of great practical significance for early fault diagnosis of the rolling element bearing.During the operation of rolling element bearings,their physical structure and operation characteristics determine their periodicity and randomness.When the bearing fails,the vibration signal presents a certain period of time variation,and exhibits certain cyclostationary characteristics.The cyclostationary theory reveals the running status of the bearing from its impact periodic,and reflects the fault characteristics through the cycle frequency.Due to the large computation of high-order cyclostationary,it produces spectral redundancy and other defects.Therefore,this paper mainly extracts the fault characteristics of the rolling element bearing based on the low-order cyclostationary theory.(1)The basic theory of low-order cyclostationary is introduced.Describing the definition and classification of the cyclostationary signals,the paper focuses on the second-order cyclostationary statistics,cyclic autocorrelation function,the spectral density function,the physical meaning of cyclic spectrum,smoothed cyclic periodogram and the definition of the spectral coherence function.Finally the demodulation performance of the second-order cyclostationary are analyzed by the amplitude modulation and frequency modulation simulation signals.(2)Research on the cyclostationarity of the rolling element bearing.Its vibration characteristics,failure forms and fault characteristic frequencies of rolling element bearings are introduced.This paper applies its cyclostationary simulation model to the inner ring,outer ring and rolling element of the rolling element bearing to simulate and analyze.(3)A feature extraction method based on Independent Component Analysis(ICA)and second-order cyclostationary is proposed.Firstly,the definition,mathematical model and basic properties of ICA are described.Secondly,the extraction method of ICA and second-order cyclostationary are studied,the simulation results show that the combination of the two methods is effective in extracting fault feature information of the rolling element bearing.(4)Experimental verification and analysis.The effectiveness of the ICA method and the second-order cyclostationary feature extraction method in the fault diagnosis of the rolling element bearing of the gearbox is validated by the diagnostics experimental bench.In the analysis of the faulty inner ring,outer ring failure of the bearing in the gearbox,the vibration signal is firstly pretreated with ICA,The second-order cyclostationary characteristics of the preprocessed signal are analyzed,compared with the second-order cyclostationary property processing without pretreatment directly,it can extract the fault characteristic frequency of the rolling element bearing well,and then the validity and accuracy of this method are verified.(5)The paper concludes and summarizes the full text and looks forward to the next step.