基于改进多稳态系统随机共振的轴承微弱故障诊断

To deal with the serious problem of side frequency interference in the traditional stochastic resonance method, an improved multi-stable stochastic resonance model was proposed for the weak bearing fault diagnosis in a strong noise background. In the presence of Gaussian white noise and a periodic force, the analytical expressions of mean first passage time and spectral amplification were obtained. Results revealed that there was an optimal set of system parameters to maximize the stochastic resonance effect of the improved multi-stable model. This improved multi-stable stochastic resonance model was applied to the weak fault diagnosis of inner and outer rings of bearings. Additionally, the quantum particle swarm optimization algorithm was employed to optimize the system parameters and the damping coefficient. It is shown that the proposed method could effectively identify the weak fault characteristic frequencies in a strong noise background. Compared with the traditional multistable stochastic resonance method, this method solves the problem of serious side frequency interference and raises the spectrum peak at the characteristic frequency of the output signal, greatly improving the performance of weak bearing fault diagnosis.