Feature extraction by enhanced time–frequency analysis method based on Vold-Kalman filter

The time–frequency analysis method can extend a one-dimensional signal to a two-dimensional time–frequency plane, revealing the signal's time-varying characteristics. The time–frequency representation (TFR) obtained by the time–frequency postprocessing algorithm has the characteristics of energy aggregation and high resolution. The generalized S-synchroextracting transform (GS-SET) stands out for its strong adaptability. However, this method cannot obtain effective information when analyzing multicomponent complex signals. We propose an enhanced time–frequency analysis method to solve this problem. First, the multicomponent complex signal is decomposed into multiple mono-component signals by the Vold-Kalman time-varying filtering technique. Second, these signals are processed by the GS-SET method. Last, the obtained TFRs are linearly superimposed to obtain the results of the enhanced method. The simulated signal verifies that the proposed method can effectively represent its time-varying characteristics. The experimental signal of the rolling bearing verifies the practicability of this method.