Application of CWT-Texture features on Transformer-PD-source Discriminatin

In this paper a new approach of feature extraction is introduced in order to classify different power transformer PD models. This approach starts with employing continuous wavelet transform as a tool to map single PD pulses into 2-dimensinal space. Then texture analysis is used to extract features. Results showed there is a high potential of complete separation of different PD models using the proposed data mining tool

Keywords:

power transformer , partial discharge , continuous wavelet transfor , texture analysis