Research on feature enhancement method of weak fault signal of rotating machinery based on adaptive stochastic resonance

Aiming at the problem that the traditional filtering method will filter out some useful signals when extracting weak fault features of rotating machinery, resulting in the loss of characteristic signals, a method for extracting weak fault features based on sin-cosine algorithm (SCA) is proposed. Combining the sensitivity of kurtosis to impact signals and correlation coefficients to interference noise, the paper proposes a new stochastic resonance (SR) performance evaluation index-weighted power spectrum kurtosis (WPSK), which solves the shortcoming of the traditional evaluation index that the fault frequency needs to be known in advance. The structural parameters of the SR are optimized by the SCA to improve the “resonance” effect. The SCA-based SR method is applied to the weak feature extraction of faulty bearings and compared with the SR model of particle swarm optimization, the results show that when the bearing inner-race fails, the value of WPSK increases by 33.5 %, and when the outer-race fails, the value of WPSK increases by 44.1 %.