Research On COT-kurtogram Based Fault Features Extraction For Rolling Element Bearing Under Varying Speed

Rolling bearing is an important component of rotating machinery, its working state is directly related to the performance of the whole mechanical system. In order to ensure the safe operation of equipment and reduce accidents, it is very important to carry out the research on condition monitoring and fault diagnosis of rolling bearing. The speed of vibration signal produces larger fluctuation with the change of working conditions or load, vibration signal shows non-stationary and contains abundant information which can comprehensive reflect its working condition. So the accurate and effective extraction of fault feature information has become the key to fault diagnosis of rolling bearings.Based on order analysis algorithm and spectral kurtosis theory, the paper proposes the COT-kurtogram algorithm applied to extract fault feature of rolling bearing. The method can extract weak vibration signals from low frequency interference and strong background noise under variable speed and realize the status identification of rolling bearing. The main researches are listed as follows:This paper firstly outlines the rolling bearing structure, failure forms, and then research vibration mechanism, fault types and fault characteristic frequency. On the basis of above theory, order analysis technology is studied. On the basis of comprehensive introduction to spectral kurtosis theory, we analyse various calculation methods of spectral kurtosis, and compare advantages and disadvantages of them. Then As a new method of rolling bearing fault feature extraction, COT-kurtogram algorithm is proposed by combining the advantages of them.COT-kurtogram algorithm is used in simulated rolling bearing fault signal and measured rolling bearing fault signal, and compared with traditional spectral kurtosis algorithm,the result proves that the COT-kurtogram algorithm can significantly improve the accuracy of the fault feature extraction. And the method is applied in mixed rolling bearing fault signal, it can detect each fault feature effectively.Based on the software platform of Lab VIEW, the virtual measurement and diagnosis system for rolling bearing is developed. The system can realize the acquisition, processing, analysis, storage of rolling bearing vibration signal and has certain practical significance and application value.