Research On Fault Diagnosis Of Rolling Bearing Based On The Bivariate Empirical Mode Decomposition

Rotating machinery is widely used in machinery, electricity, transportation, metallurgy and other fields, which proves that the rotating machinery plays an important role in the development of human society. The mechanical system will vibrate inevitably in operation. When fault occurs, serious vibration will cause accident, which results in great losses. Therefore, it is necessary to study the fault diagnosis of rotating machinery. As the key part in rotating machine, rolling bearing plays an irreplaceable role in the mechanical system. In this paper,rolling bearing is taken as the research object. The fault of rolling bearing is diagnosed with the method of BEMD. The research contents of the paper is as follows.(1) According to the characteristics of fault signals of rolling bearing, the energy operator demodulation method based on BEMD is proposed. Firstly, the rolling bearing fault signals are decomposed with BEMD, and some single component signals are obtained. Then single component signals are enveloped with the energy operator. Based on this, the amplitude and frequency spectrum information is obtained with envelope of the signal. The simulated signals and the measured data are analyzed by this method. Compared with only using the energy operator envelope demodulation, the results show that the method of combining BEMD with energy operator can be more effective when extract the fault characteristic frequency.(2) Fault recognition based on SVM and relative BEMD energy and fault recognition based on relative BEMD energy and SOM are studied. First of all, bearing fault signal is decomposed by BEMD with the relative energy of each band as the feature vector. And then put SVM and SOM as the classifier to judge the fault type and fault degree. The method has very high recognition rate, which has certain engineering application value.(3) In view of the bearing vibration signal submerged in noise easily and having the characteristics of nonlinear, non-stationary, incipient fault detection method of rolling bearing based on chaos and BEMD is proposed. Firstly, the bearing signals are decomposed with BEMD. The intrinsic mode function containing the most bearing fault information is found. And then the phase diagram trajectory changes of boundary values using the bifurcation diagram of the Duffing oscillator are found. The internal excitation frequency is set to bearing fault characteristic frequency to observe the change of chaotic phase trajectory to detect bearing fault information. The fault signals are detected to be present with lyapunov exponents. This method is simple, fast, and has good practical value in engineering.