Research On Fault Diagnosis For Rolling Bearing Based On Correlation Analysis And Resonance Demodulation

Rolling bearing is a rotating machinery part used widely,so it is of importance in economic value and practical meaning to monitor working condition of rolling bearing in order to avoid serious accidents.The fault diagnosis of bearing is effective for preventing sudden accidents and important guarantee to make machinery system work safely.Therefore,rolling bearing is regarded as research object by processing vibration signal to solve the problem of the extraction of fault feature frequency.The main contents are as follows.Firstly,the background and studying meaning of this project are stated systematically and the current research status and progress and developing trend of off-line and on-line diagnostic method of rolling bearing and the application in extracting fault feature frequency are stated comprehensively based on theoretical analysis and engineering application.Then,the computing method of fault feature frequency is discussed according to the theory of machinery fault diagnosis.Secondly,in order to solve problems of the low conversion rate in the voltage-type operational amplifier and the non-linear diode in the traditional voltage-type resonance demodulation circuit,an improved resonance demodulator using a new Electronic component of current conveyer(CCII+)to replace the original voltage-type operational amplifier and a rectification circuit to counteract threshold voltage of diode is proposed in this paper.As a result,the improved resonance demodulator has a larger promotion in handling the accuracy of the fault signal which is high frequency or weak.Afterwards,the improved resonance demodulator was used to detect typical faults of a rolling bearing in real time on the running attrition test bench.Experiment results show that the improved resonance demodulator can not only broaden its applicable frequency range,but also ensure the accurate extraction of rolling bearing early weak faults.Thirdly,this paper proposes a multi-channel correlation adaptive resonance demodulation(MCC-ARD)method.Firstly,MCC-ARD uses redundant signal source to pick up fault information and optimize EMD efficiency by spectral kurtosis(SK).After that,the IMF components reconstruct by the reasonable choice of the cross-correlation coefficient.Finally the method realizes the fault diagnosis of rolling bearings by the reconstructed IMF envelope demodulation.Through the measured data analysis of MCC-ARD,it proves that MCC-ARD not only overcomes the defect of poor system correction,but also has higher fault information fidelity and spectrum identification rate than the traditional EMD combined with spectral kurtosis resonance demodulation method,which make fault diagnosis of rolling bearing results sharper and more accurate.Finally,by using the characteristics of cross correlation function,we can not only get rid of the dependence of human factors,but also weaken the internal noise of the selected resonance band,so that the fault characteristic frequency is more obvious.In this paper,the on-line diagnosis method based on cross-correlation detection is built in the form of analog circuits and extracts successfully fault characteristic frequency in the real-time fault diagnosis of rolling bearing.It shows that the method not only overcomes the limitations of a single source,using cross-correlation function weaken resonance with internal noise,so as to make the diagnosis with higher spectral identification rate,but also the method of analog circuit has the advantages of simple structure,easy maintenance,the real-time fault diagnosis of bearing fault method application and popularity have a certain reference value.