Fast Assessment of Transient Stability Margins by a Neural Network Approach

This paper reports the use of Artificial Neural Networks (ANN) for fast and accurate evaluation of the transient stability degree for each contingency in a multimachine power system, using only real time monitorable system values. The output of an ANN provides an emulation of the transient stability energy margin. Preventive control measures can be suggested through generation load redispatch, by getting the sensitivities of the energy margin directly from a trained ANN. The approach was tested with several contingencies in the CIGRE test system, giving better results than other methods so far reported in the literature.