Effect of scale-varying fractional-order stochastic resonance by simulation and its application in bearing diagnosis

Bearing is among the most widely used components in rotating machinery.Its failure can cause serious economic losses or even disasters.However,the fault-induced impulses are weak especially for the early failure.As to the bearing fault diagnosis,a novel bearing diagnosis method based on scale-varying fractional-order stochastic resonance(SFrSR)is proposed.Signal-to-noise ratio of the SFrSR output is regarded as the criterion for evaluating the stochastic resonance(SR)output.In the proposed method,by selecting the proper parameters(integration step H,amplitude gain K and fractional-orderα)of SFrSR,the weak fault-induced impulses,the noise and the potential can be matched with each other.An optimal fractional-order dynamic system can be generated.To verify the proposed SFrSR,numerical tests and application verification are conducted in comparison with the traditional scale-varying first-order SR(SFiSR).The results prove that the parameters H,K andαaffect the SFrSR effect seriously and the proposed SFrSR can enhance the weak signal while suppressing the noise.The SFrSR is more effective for bearing fault diagnosis than SFiSR.