Surface Discharges Identification of 10kV Solid Insulation Cabinet Based on Energy Characteristics Extraction of Audio Signal

In order to accurately monitor the insulation status of lOkV circuit breakers, a method to identify the insulation condition of 10kV solid insulation switchgears by using the feature of audio signal is proposed. By using audio capture device to obtain the audio signal of surface discharge of the circuit breaker in the dirty, damp and others conditions. The Frequency characteristics of surface discharges is extracted by using wavelet packet analysis. Firstly, establishing wavelet packet basis to obtain high frequency information and low frequency information by orthogonal decomposition. Secondly, the appropriate segmentation is divided by analyzing the spectral characteristics of the surface discharge signal. Thirdly, calculating the energy at each frequency segment. Finally, normalizing the energy of each frequency segment to establish characteristic vector, as the basis of state identification. The characteristic vectors of surface discharges sound and irrelevant sound are input to the support vector machine for classification and identification of the insulation status. The solid insulation switchgears surface discharges can be accurately determined by using wavelet packet analysis and support vector machine. It is more suitable for the surface discharges identification of equipment such as solid insulation switchgears.