Coordinating fuzzy ART neural networks to improve transmission line fault detection and classification

This paper demonstrates several uses of adaptive resonance theory (ART) based neural network (NN) algorithm combined with fuzzy K-NN decision rule for fault detection and classification on transmission lines. To deal with the large input data set covering system-wide fault scenarios and improve the overall accuracy, three fuzzy ART neural networks are proposed and coordinated for different tasks. The performance of improved scheme is compared with the previous development based on the simulation using a typical power system model. The speed and accuracy of detecting continuous signals during the fault is also evaluated. Simulation results confirm the improvement benefits when compared with the previous implementation.