6th International Conference on Optimization of Electrical and Electronic Equipments, OPTIM 1998, Brasov, Romanya, 14 - 15 Mayıs 1998, cilt.1, ss.241-244
Overcurrent relays can be used to protect practically any power system element. If the operating time of a relay depends on the current magnitude being smallest when the current is the greatest. Such a characteristic is known as an inverse time characteristic and the relay is known as an inverse time overcurrent relay. It is not possible to show the inverse time-overcurrent relay operating characteristic with a determined function. In this paper the characteristic of inverse time-overcurrenl relay (t=f(I)J is obtained by means of artificial neural networks (ANN). ANN is successfully used in a lot of areas such as fault detection' control and signal processing. One of the major paradigms of ANN is feed-forward network. For this paper' it is used feed-forward network and error backpropagation algorithm.