Application of Adaptive Mono-Stable Stochastic Resonance Based on Cuckoo Search for Incipient Fault Diagnosis of Rolling Element Bearings

Abstract

Bearings are the main components of the modern machinery. Poor operating environment often makes them prone to failure, which may cause significant economic losses and catastrophic disasters. Once the incipient fault appears, the machinery will not operate correctly, and even go down. Therefore, detecting their incipient fault as soon as possible is useful for bearing prognostics and health management. According to stochastic resonance theory, the weak signal can be enhanced with the assistance of proper noise in a proper nonlinear system. Therefore, stochastic resonance is suited for fault diagnosis. In this paper, an adaptive mono-stable stochastic resonance based on cuckoo search is proposed for incipient bearing fault diagnosis. The exponential type single-well system is used as the nonlinear system of the mono-stable stochastic resonance. The bearing fault signal after pre-processing is processed by the mono-stable stochastic resonance. The cuckoo search is used to search the optimal parameters of the nonlinear system. The signal-to-noise ratio is used as the evaluation of stochastic resonance. Finally, we conducted two bearing fault test to validate the effectiveness of proposed methods. The results meet the expected effect of the proposed methods.