Artificial neural networks for transient stability assessment of power system

The paper explores the suitability of using two artificial neural network models (ANN) as tools for power system transient security assessment (TSA). Firstly, a TSA problem of a local power net is changed into a pattern recognition problem suitable for an ANN, and sample data are preprocessed. Then BPN and KNN models are used respectively for this TSA problem. The suitability and advantages of the two models are discussed and compared on mapping capability of the problem, estimation of certainty factor of interpolating results and ANN size. The results show that from the point of view of TSA application, the KNN model is better than BPN.<>