Identification of Different Types of Partial Discharge Sources from Acoustic Emission Signals in the Time-Frequency Representation

The subject matter of this paper is to introduce the time-frequency representation in analyzing the acoustic emission signals emitted by partial discharge sources. Three different types of partial discharge sources used to generate the acoustic emission signals during the partial discharge (PD) occurrences are created in an experimental tank filled with transformer oil. These partial discharge sources are the plain pressboard, the floating metal in the pressboard and the bubble in the pressboard. The acoustic emission (AE) signals are detected and stored as time-frequency representation, in the form of spectrogram, by utilizing the short-time Fourier transform (STFT). Finally, seven descriptors are introduced in order to extract the features from each of the spectrogram. The obtained results also confirmed the ability of the proposed technique to discriminate between different types of PD sources.