Novel Adaptive Search Method for Bearing Fault Frequency Using Stochastic Resonance Quantified by Amplitude-Domain Index

Abstract

Rotating machinery devices are prone to failure due to severe working conditions, whose failures will further induce other mechanical faults. Therefore, the fault diagnosis of rotating machinery is important, especially in the case of unknown fault types and fault characteristics. A novel adaptive search method is proposed for the bearing fault frequency by using stochastic resonance (SR) with general scale transformation. The amplitude-domain indices, which are independent of the specific fault frequency of vibration signal, are applied to quantify SR response. Simulated bearing fault signal is used to illustrate the good performance of these indices in evaluating SR output. Then, an adaptive search procedure for bearing fault frequency is presented in detail and verified by different vibration signals collected from the multiple working conditions. The searching results demonstrate that the proposed adaptive search method is accurate, effective, and sensitive for detecting unknown failure frequencies of rolling bearings. The proposed method might have significant application value in the condition monitoring of rolling bearings.