Method of Rolling Bearing Fault Detection Based on Two-Dimensional Tri-Stable Stochastic Resonance System

Abstract

To solve the problem that it is difficult to detect the weak bearing fault signal from high-intensity noise, a method of bearing fault detection based on the two-dimensional tri-stable system (TDTS) is proposed. The analytical expression for the spectral power amplification (SPA) of the TDTS is derived using the probabilistic flow method. The stochastic resonance (SR) mechanism of the TDTS is studied with the SPA as a measure index. The fourth-order Runge–Kutta method and the genetic algorithm are utilized to carry out the numerical simulation. Apply the TDTS to the weak periodic signal detection and the bearing fault detection. In this work, the potential field structure of the TDTS is discussed. The SPA of the TDTS to the periodic driving force is derived. The influence of the coupling coefficient, the driving frequency, and the driving force amplitude on the SPA is analyzed. The performance of the TDTS and the ODATS to detect weak signals from intense noise are compared. The theoretical analysis shows that SR phenomena occur in the TDTS. As the noise intensity increases, the SPA presents a unimodal trend, that is, the noise of a specific intensity can enhance the response of the TDTS to the weak external driving force. Moreover, the enhancement effect is proportional to the amplitude of the driving force and the coupling coefficient, and inversely proportional to the driving frequency. The experimental results indicate that the TDTS can effectively detect the weak periodic signal and detect the inner and outer ring faults of the bearings. The detection effects are preferable to that of the one-dimensional asymmetric tri-stable system (ODATS). The effectiveness and advancement of the proposed method are proved.