Adaptive Modulation Stochastic Resonance of Two-Dimensional Asymmetric Bistable System and its Application in Bearing Fault Diagnosis

Abstract

A two-dimensional asymmetric bistable stochastic resonance (TABSR) system model is proposed, and the parameter optimization is performed adaptively by genetic algorithm (GA). In essence, the model is coupled by a monostable model with an asymmetric piecewise-linear model by means of linear coupling. The monostable model is used as the control system, and the asymmetric piecewise-linear model is used as the controlled system. A set of parameters of the controlled system is firstly obtained, so the TABSR system model can achieve better performance by adjusting the parameters and coupling coefficients of the control system. The output signal-to-noise ratio (SNR) formula of the two-dimensional model is derived and compared with the asymmetric piecewise-linear stochastic resonance (APSR) model. Then the effects of system parameters and coupling coefficients of the control system on system performance are explored. Through numerical simulation, the correctness of the theoretical analysis is verified, and the system is proved to have high theoretical research value. Finally, TABSR is applied to industrial bearing fault diagnosis to recognize weak fault signals which shows the practical application value and the comprehensive performance comparison with APSR is conducted to validate the advancement of TABSR.