Cyclostationary Theory In The Rolling Bearing Fault Detection

Rolling bearing is one of the most important mechanical parts in the rotating machinery. The performance of rolling bearings has direct influence on the performance of the whole machine. At the same time, rolling bearing is also one of thevulnerable parts,and bearing failure is one of the foremost causes of breakdowns in the rotating machinery. Thus, it is of great importance to detect the defect at its incipient stage and alert the operator before it develops into a catastrophic failure.The fault principle and the vibration types are analyzed in detail, which are the basises of fault diagnosis. Some facilitated criteria for the diagnosis are also introduced in this paper.Resonance demodulation theory is the most widely used fault-diagnosis technique because of the high SNR it achieves. Unfortunately,the band-pass filter used in resonance demodulation can not eliminate all the noises due to the complexity of the machinery. The output of band-pass filter includes the vibration signal caused by the fault, as well as the vibration signal regarded as noise. The noises are caused by many factors other than the fault of the rolling bearing.In the incipient stage the fault information is usually masked by the noises, and makes the diagnosis difficult.According to the cyclostationary theory, the stationary noise has no influence at non-zero cyclic frequency. In this paper,based on the resonance demodulation theory and the cyclostationary theory, as well as according to the experiments, it is verified that the cyclic spectrum slice analysis to the envelope of modulated signal performs better on the noise restraint. Then this technique is used for the bearing fault diagnosis, and based on the simulation with Matlab, the characteristic frequency of fault bearing is acquired. In this way the position of the bearing fault can be located.The last part gives a conclusion and presents some suggestions for further development.