Uporaba umetne inteligence v distribucijskem elektroenergetskem sistemu.

Recent developments indicate that the Artificial Neural Networks (ANNs) may be useful assisting the power system dispatchers at their work. The paper presents an approach to detection a faulty feeder during a phase to earth fault (also a high impedance fault) in a power distribution system using a feed-forward ANN. In the proposed ANN algorithm, the standard backpropagation technique with a sigmoid activation function is used. ANN is trained with real data obtained with a disturbance recorder during an actual fault. The study results show that the proposed algorithm performs excellently in detecting the phase to earth faults and locating the faulty feeder. [ABSTRACT FROM AUTHOR]