Analysis of Voltage Stability Margin and Prediction of Asynchronous Machine using ANN

This paper presents an ANN based model for predicting stability margin for an asynchronous machine power system prone to voltage instability. Such a model may be employed either for direct prediction of the stability margin based on the existing loading conditions or for forecasting the loading conditions for a future time period and then providing an estimate of the stability margin. The neural networks employed are the multi layer perceptron (MLP) with a second order learning rule and the radial basis function (RBF) network and feed forward neural network. The simulation results for a sample 5-bus system indicate that the ANN models provide a fairly accurate and fast prediction of the stability margin making them, suitable for application in an on-line energy management system.