Detection of Short Circuit Faults in Power Transformer by the Measurement of Its Windings Voltages and Currents Using a Neuro-Fuzzy System

Insulation failure between winding turns is one of the main causes of incipient winding fault in a transformer. During the operation of a transformer, strong electric fields are applied to the dielectric material of its windings. Dielectric deterioration and aging is one of the major cause of short circuit faults in transformer windings. Due to the probable occurrence of this type of defect and its extension, its early detection is a very important task in power systems. In this paper, it is shown that by measuring the phase difference between winding voltage and winding current of a loaded transformer, the existence of internal winding fault can be detected. For online fault detection, an intelligent system (neural-fuzzy system) has also been proposed.Both simulation results and laboratory tests confirm the ability of the proposed method for the detection of internal winding faults especially at light loads. With this method, there is no need to de-energize the power transformer.