Ranjbar - ELeclectric Power Research (EPRC) CEnter(EPRC)
Marjanmehr - ELeclectric Power Research (EPRC) CEnter(EPRC)
Yazdani - ELeclectric Power Research (EPRC) CEnter(EPRC)
Mohammad Sadegh - ELeclectric Power Research (EPRC) CEnter(EPRC)

Modeling the characteristic of operational relay curve is inavitable for power system relays coordination software. The methods that have been used so far were based on mathematical models. Four alternative equations for representing relay curves at specified time dial setting have been proposed. These equations have been further used to form composite models of the relay characteristics including the time dial setting. In applying mathematical models a polynominal equation is considered to obtain the operational time of a relay. This paper describes a new method by applying Neural Networks. Perceptron and Hopf ield nets were the components of Neural Network that have been used. For each time-dial setting (T. D. S.) a perceptron net is considered. The applied Hopfield net, chooses one of these perceptron nets according to its inputs (T. D. S). Comparing the results of this new way with mathematical models, shows that Neural Net i s more proper, as it provides high accuracy.