A New Method for Discriminating Between Internal Faults and Inrush Current Conditions in Power Transformers Based on Neuro Fuzzy

in this paper, a new algorithm based on Neuro Fuzzy for differential protection of the power transformer is proposed. This algorithm consists of considering the ratio and the difference phase angle of the second harmonic to the fundamental component of differential currents under various conditions. These two protection functions are computed and the protective system operates in less than one cycle after the occurrence of disturbance. Another advantage of this algorithm is that fault detection doesn´t depend on the selection of thresholds and one of the best advantages of this algorithm is that all kinds of internal fault can be recognized from magnetizing inrush current even those which are mixed with inrush currents. The proposed Nerue Fuzzy is trained by obtained data from simulation of power system under different faults and switching conditions. The suitable operation of this algorithm by simulation of faults and various switching conditions on a power transformer is shown.