Fault Diagnosis Of Rolling Bearing Based On Self Correlation Analysis

In the field of industry,as one of the key parts of rotating machinery,rolling bearings have a great influence on the safe operation of the system.Therefore,it is important to find out the faults of bearings and maintain in time.The fault detection method based on vibration signal is divided into off-line detection,such as STFT,Winger-Ville,blind source separation,wavelet transform,HHT,and online detection,such as based on FPGA and DSP digital circuit testing,Resonance demodulation analog circuit detection,etc.This paper makes a series of studies based on the self correlation analysis of rolling bearing condition monitoring and fault diagnosis,the main contents are as follows:Firstly,this paper expound the research background and significance and the technical development at home and abroad of rolling bearing condition monitoring and fault diagnosis.The structure,fault types and causes of the rolling bearing are explained in this paper,and reasonable experimental equipment is selected for the vibration signal,which provides a good experimental basis for the research object.Secondly,this paper proposes a channel self correlation and cyclic spectrum analysis(CSC-CSA)of off-line rolling bearing fault diagnosis method.The method makes research and improvement of traditional empirical mode decomposition(EMD).The self correlation analysis based on correlation coefficient and the method of fault feature frequency extraction based on cyclic spectrum analysis are added.This method uses the correlation coefficient to select optimum intrinsic mode function(IMF),and then uses the optimal intrinsic mode function to reconstruct vibration signal,finally analysis the reconstructed signal by cyclic spectrum to find fault feature extraction frequency.The experimental results show that the reconstructed signal by this method not only has the high SNR,but also the frequency of the fault features extracted accurately.Thirdly,this paper proposes self correlation function and cyclic spectrum analysis(SCF-CSA)based on CSC-CSA.The method uses self correlation function instead of self correlation coefficient to do self correlation analysis,so that the self correlation analysis is not only a selection index,but a real process of noise reduction.SCF-CSA also uses spectral kurtosis to determine the parameters of the filter,avoiding interference by human factors in determining the parameters of the algorithm.Through the test and analysis of the real data from the running test table,it is proved that the method not only has a good noise reduction effect,but also improves the diagnosis accuracy.The third and fourth chapters of this paper test the working condition data of real EMU bearings based on the high-speed rolling bearing fault simulation test bench,and achieve good diagnostic results.Finally,according to the theory of self correlation function,this paper proposes an on-line fault diagnosis method for rolling bearings based on self correlation detection.This method uses correlation detection circuit and converts some modules into current mode circuit,which improves the speed and accuracy of the whole diagnosis circuit.Then a filter circuit with adjustable parameters based on current mode is designed,and the diagnosis circuit is further improved.The correctness of the method and the effectiveness of the real-time fault diagnosis of rolling bearings are verified by the measured data and the software aided analysis.On line fault diagnosis method of rolling bearing based on self correlation detection makes the research contents not only including the research of off-line diagnosis method,but also building into an online diagnosis system in the form of analog circuits by the off-line diagnosis principle,which provides a valuable reference for the compatibility of large-scale diagnostic systems.