A New Wavelet-Based Approach for Internal Fault Current Identification in Power Transformers

This paper demonstrates a novel approach for the differential protection of power transformers. this method uses the wavelet transform (wt) and the adaptive network-based fuzzy inference system (anfis) to detect a fault current from an inrush current. the proposed method has been designed, based on the differences between the amplitudes of wavelet transform coefficients in a special band of frequency that is caused by faults and inrush currents. the performance of this algorithm has been simulated and tested under different conditions of the switching on of power transformers, using the pscad/emtdc environment software.