A Novel Noise-Aided Fault Feature Extraction Using Stochastic Resonance in a Nonlinear System and Its Application

The fault features of rolling bearings under time-varying speed conditions (TVSCs) are often submerged in strong noise. For one thing, the system parameters of stochastic resonance (SR) need to match the change fault feature frequency of nonstationary signals, which makes the process very complex. Order analysis (OA) converts the nonstationary signals into a stationary angular domain signal, and the fault feature order does not change with time in the order spectrum. Because the system parameters only need to match a single order of stationary angular domain signals, the process becomes simpler and more efficient. For another, the dual-tree complex wavelet packet transform (DTCWPT) decomposes the angle-domain signal into a series of frequency band components. We select the frequency band containing more fault information optimal frequency band (OFB) using statistical complexity measures (SCMs). This operation suppresses the interference of strong noise on the output effect of SR. In addition, we also demonstrate that SCM has better stability than other indices (kurtosis, Gini index, and harmonic noise ratio) under different noise intensities. Finally, a piecewise tri-stable SR system is constructed to effectively avoid the saturation problem and to improve the performance of the system in signal enhancement. Through an adaptive frequency-shifted SR method, we effectively extract the fault feature information of rolling bearings under TVSCs. The robustness of the proposed method is verified by theoretical research, simulation, and the collected experimental data.