Multi-stable stochastic resonance and its application research on mechanical fault diagnosis

Abstract

It is difficult to extract the fault features of a rotating machine via vibration analysis due to interference from background noise. Stochastic resonance (SR), used as a method of utilising noise to amplify weak signals in nonlinear dynamical systems, can detect weak signals overwhelmed in the noise. However, the detection effect of current SR methods is still unsatisfactory. To further increase the output signal-to-noise ratio (SNR) and improve the detection effect of SR, the present study proposes an improved SR method with a multi-stable model for identifying the defect-induced rotating machine faults by analysing the influence relationship between the resonance model and the resonance effect. Due to the structural characteristics of three potential wells and two barriers, the proposed resonance model can not only further amplify weak signals, but also convert into a monostable model, a bistable model or a tristable model. This result is achieved by adjusting system parameters and thus obtaining a better matching of the input signals and resonance models. Therefore, the multi-stable SR method, combined with the characteristics of the multi-stable model, can both increase the output SNR and improve the detection effect and also detect the low SNR signals and enhance the processing capability of SR for weak signals. Finally, the proposed method is applied to a gearbox fault diagnosis in a rolling mill in which two local faults located in the big gear and the pinion, respectively, are found successfully. It can be concluded that multi-stable SR method has practical value in engineering.