A novel structure based on stochastic resonance for fault diagnosis of bearing

For the sake of detecting the faults of bearing in the incipient stage, the efficient structure for fault diagnosis is necessary. However, the fault characteristics related to the fault diagnosis are extremely weak, and so they are difficult to be exacted. Stochastic resonance of the nonlinear system is a novel topic in the field of fault diagnosis, which can enhance the weak signal, and finally determine the fault types. Signal-to-noise ratio (SNR) is usually employed to induce the occurrence of stochastic resonance for fault diagnosis. While it may be also induce the coherence resonance, and the output can mistake the fault type. In this paper, a novel measuring index based on autocorrelation function is proposed to induce the stochastic resonance, and avoid the occurrence of coherence resonance. The measuring index is called as autocorrelation function harmonic to noise ratio index(AFHNR), and the structure for fault diagnosis based on AFHNR and stochastic resonance (Shorted as AFHNRSR) is successfully examined by the signals from numerical simulation and experimental rig.