Application of the ARTMAP neural network to power system stability studies

The authors discuss the application of a novel variation of the adaptive resonance theory (ART) neural network called fuzzy ARTMAP to the determination of the steady-state stability of a synchronous generator. The model of the generator includes the voltage regulator, the excitor, and the power system stabilizer. The results obtained with the fuzzy ARTMAP network are compared with those obtained with a backpropagation network. For online training, the fuzzy ARTMAP network was found to be a better choice because of its faster convergence, but in some cases, the fuzzy ARTMAP network did not perform as well as the BP network. >