GIS partial discharge signal feature extraction method based on Hilbert transformation

The invention requests to protect a GIS partial discharge signal feature extraction method based on Hilbert transformation. According to the method, Hilbert transformation is utilized to obtain an envelope of a GIS partial discharge signal, statistical characteristic parameters, time domain characteristic parameters and frequency domain characteristic parameters of a partial discharge pulse waveform are extracted based on the envelope, and an improved kernel principal component analysis method is adopted to carry out dimension reduction on the extracted characteristic parameters. And finally obtaining the characteristic quantity which is low in dimensionality and can correctly execute partial discharge classification. Compared with a common GIS partial discharge signal feature extraction method, the partial discharge feature quantity extracted by the GIS partial discharge signal feature extraction method is not only low in dimensionality, but also clearer in physical significance, and the partial discharge si