A robust index-guided GSR approach to improve the efficiency of bearing fault diagnosis

Abstract

Stochastic resonance (SR) has been widely used in bearing fault diagnosis due to the enhancement principle of energy conversion from noise to weak signal. However, the diagnosis efficiency and robustness is still challenging.
Under the mechanism of generalized stochastic resonance (GSR), in the paper we propose a scale-transformed linear oscillator (SLO), and obtain the analytical expression of fault feature amplification (FFA) to replace the numerical implementation of output signal-to-noise ratio (SNR) in the multi-parameter optimization.
It brings the substantial benefit to the reduction of time complexity in fast fault diagnosis, which is verified in both theory and simulations.
In the experimental diagnosis for some typical cases, the results show that the proposed method is valid and exhibits the superiority in diagnosis performance, efficiency, and robustness, which demonstrates that the FFA-guided GSR-SLO method has great potential in the engineering applications, especially the real-time fault diagnosis in complex operating environments.

A local spectrum enhancement‐based method and its application in incipient fault diagnosis of rotating machinery

“…Qiao and Shu 28 proposed an adaptive coupled neuron‐based SR to extract the weak fault characteristics in rolling bearings. Zhang et al 29 proposed a generalized SR to diagnose multifaults in bearings. Though the effectiveness of the SR system in enhancing weak fault characteristics has been demonstrated by the aforementioned studies, it needs to be pointed out that the improved effectiveness of an SR system is highly dependent on the selection of parameters, which are either manually tuned or preset according to one's experience.…”