Determining voltage unstable area in power systems using kohonen neural network

This paper presents a new method to determine voltage unstable area in power systems using Kohonen neural network (KNN) from dynamic voltage stability viewpoint. Using KNN, the buses in a power system are classified as critical and non critical buses based on the power transfer stability index values. The critical buses are then clustered to form the voltage unstable area in a power system. The proposed method was implemented on the IEEE 39-bus test system in which for dynamic simulation of voltage collapse, two contingencies such as load increase and line outage were considered. The dynamic voltage collapse simulation results were used for generating training and testing data sets of the KNN. The results on the determination of voltage unstable area by using the KNN were also compared with the learning vector quantization technique. The results showed that the proposed method using KNN is more accurate than the linear vector quantization technique in forming the voltage unstable area in power systems.