Based On Vibration Signal Of Bearing Condition Monitoring And Fault Diagnosis Methods

The bearing is a kind of very important component in the modern mechanical systems, and their work environment is very poor, making it very prone to failure during runtime. With modern mechanical systems tend to the development of large-scale, precision, high-speed, bearing failure influence the system more and more largely. Therefore, the bearing fault diagnosis technology is becoming increasingly important. As the vibration signal is directly related to bearing failure, and bearing fault diagnosis method based on vibration signal is fast, simple operation, reliability. At present, this fault diagnosis method has become one of the hot bearing fault diagnosis methods.Use the four analytical methods in the time domain analysis which are the probability density, rms, peak, kurtosis and power spectrum to analyze the vibration signals for monitoring bearing running state and determine the bearing failure according to the complex frequency of the bearing fault vibration signal.Analysis vibration signals by using demodulated resonance based on empirical mode decomposition and Hilbert spectral analysis for time-domain analysis methods can not diagnose fault types and in the band-pass filter passbands in resonance demodulation method are difficult to choose.Finally, analysis the data form Case Western Reserve University bearing fault data of center, monitor its status and diagnose the faults. Experimental results show that together kurtosis and density, rms, peak in time-domain analysis method can determine if the bearing is fault, while the resonance demodulation method can diagnose the early outer and inner fault in the three kinds of surface damage class.