Rahman Dashti

The inrush currents of transformers are among the main causes which compromise the performance of the protection relays in practice. Therefore, the discrimination between the inrush currents and faults is very important. A new technique is proposed in this paper to distinguish the internal fault of transformers and other normal transient conditions (such as inrush current). Accordingly, a new operational characteristic is defined for feature extraction and classification of the instantaneous current behavior based on the modal transform. The internal fault current is accurately and rapidly discriminated from the other transient current conditions (such as load switching, and inrush currents) by tracking the differential current curve. Furthermore, the real transformer is simulated to validate the proposed method by the PSCAD-EMTDC and MATLAB software. Furthermore, the real data of the actual fault in the transformer of the Iranian power network is studied. The simulation results show the stability, reliability, simplicity, and high speed of the proposed method in the discrimination of the fault current from inrush current. The method is compared with its other intelligent counterparts and the results confirm a clear improvement.