A new look at classification of transformer normal and abnormal currents

The paper proposes an enhanced wavelet-based feature extraction technique to classify transformer inrush currents (TIC) and transformer internal faults (TIF). The proposed tool utilizes the number of wavelet coefficients of local maxima as current signal slides into Kaiser's window. The general pattern of number of coefficients of local maxima at the first three resolutions are used to design a new automated tool for monitoring and classifying abnormal conditions in power transformers. The proposed monitoring technique is evaluated using large data sets.