Discrimination between Winding Mechanical Defects, Internal and External Electrical Faults and Inrush Currents in Transformers using Hybrid Method

In this paper, an efficient method is suggested for distinguishing winding mechanical defects from transient fault currents and inrush current. In this method, a signal processing and artificial intelligence tool are simultaneously utilized. Firstly, the transformers winding mechanical defects are investigated on the real model of a 1.6-MVA transformer winding. Then, the parameters of the detailed model of the transformer winding are estimated in MATLAB software by Genetic Algorithm and compared with experimental results for validation. Thereafter, the winding mechanical defects, internal and external electrical faults and the inrush current is stimulated using ATP/EMPT software in order to obtain the differential currents. Afterwards, distinctive features are extracted using the wavelet transform. Finally, these features are used to train an ANN classifier and the disturbances are distinguished. The proposed method is able to distinguish among winding mechanical defects, internal and external electrical faults and inrush current in transformers with a good accuracy.