Coherency identification using growing self organizing feature maps [power system stability]

Stable operation of a power system following a disturbance is very important from the point of view of reliability. For this purpose, online assessment is needed to evaluate the impacted system components in a short time. Fast evaluation of a disturbance impact requires the formulation of dynamic equivalence of external systems. On the other hand, preventive measures for stability enhancement requires a priori knowledge of the components that will be affected by the disturbance. This paper presents the identification of coherent generators in power systems using an unsupervised learning neural network called a "growing self-organizing feature map" which dynamically generates the network architecture. The data for the neural network has been obtained from the simulation of a 1000 bus, 62 generator system.