Discrimination of Magnetic Inrush and Phase to Phase Fault in Transformer using Probabilistic Neural Network

In this paper, the probabilistic neural network (PNN) is proposed as the classifier to discriminate between the magnetizing inrush and the phase to phase fault of a power transformer. The probabilistic neural network (PNN), is investigated for the protection of power transformer. Current waveforms due to different types of events, such as phase to phase faults and circuit breaker switching were generated using power transformer simulated in MATLAB simulation software. Simulated data were used to train the classifiers. To evaluate the developed algorithm for various operating condition of the transformer, including internal faults, are obtained by modeling the transformer in MATLAB Simulink. The data from the simulation are trained and tested by using Probabilistic Neural Network for discrimination of magnetic inrush and phase to phase fault.