Improving the bearing fault diagnosis efficiency by the adaptive stochastic resonance in a new nonlinear system

It is a challenging task to detect the weak character signal in the noisy background. The stochastic resonance (SR) method has been wildly adopted recently because it can not only reduce the noise, but also enhance the weak feature information simultaneously. However, the traditional bistable model for SR is not perfect. So, this paper presents a new model with periodic potential to induce the adaptive SR. In the new model, based on the adaptive SR theory, the system parameters are simultaneously optimized by the improved artificial fish swarm algorithm. Meanwhile, the improved signal-to-noise ratio (ISNR) is set as the evaluation index. When the ISNR reaches a maximum, the output is optimal. In order to eliminate interference to obtain more useful information, the signals are preprocessed by Hilbert transform and High-pass filter before being input to the adaptive SR system. To verify the effectiveness of the proposed method, both numerical simulation and the vibration signal of the rolling element bearing from the lab experimental are adopted. Both of the results indicate that the adaptive SR model proposed shows better performance in weak character signals detection than the traditional adaptive SR in the bistable model. Meanwhile, the experimental signals with different working conditions are also processed by the new method. The results show that the method proposed could be more widely applied.