Research on Instantaneous Parameter Extraction of Partial Discharge Signal Based on Grey Clustering Algorithm

The identification therapy of instantaneous characteristic parameters is widely used because of its inherent advantages. The instantaneous feature transformation is sensitive to noise interference, but the actual communication signals all contain noise without exception. If the instantaneous parameters are directly obtained by the analytical signal method, it will lack physical meaning and be distorted. Aiming at this phenomenon, this paper studies the extraction of instantaneous parameters of partial discharge signals based on grey clustering algorithm. Because of its good time-frequency localization, grey clustering algorithm is widely used in the extraction and detection of instantaneous parameters of partial discharge signals. The instantaneous frequency characteristics of partial discharge signals can reflect the frequency variation of signals at different times. Compared with the traditional time-frequency analysis method, the time-frequency analysis method in the grey clustering algorithm has the advantages of high frequency resolution, strong noise resistance and no cross-term interference, which makes it possible to accurately extract the instantaneous frequency of signals by improving the grey clustering algorithm. The simulation results show that the time interval between the two shocks is 0.13 s after the original signal is decomposed by grey clustering algorithm. Obviously, the frequency of each main impact sequence is completely consistent with the rotating frequency of the rotor.