ADAPTIVE PROTECTION STRATEGIES FOR DETECTING POWER-SYSTEM OUT-OF-STEPCONDITIONS USING NEURAL NETWORKS

This paper presents new strategies for adaptive out-of-step (OS) prote ction of synchronous generators based on neural networks. The neural n etworks architecture adopted, as well as the selection of input featur es for training the neural networks, is described. A feed forward mode l of the neural network based on the stochastic backpropagation traini ng algorithm is used to predict the OS condition. Two adaptive OS prot ection strategies are suggested, The first approach depends firstly on detecting the case of the system through case detection neural networ ks by some prefault local measurements at the machine to be protected, and then calculating the new OS condition through an adaptive routine . The second approach is based on creating a large neural network to b e trained using different outage cases of the power system, The capabi lities of the developed adaptive OS prediction algorithms are tested t hrough computer simulation for a typical case study. The results demon strate the adaptability of the proposed strategies.