An efficient approach to the evaluation of mid-term dynamic processes in power systems

This paper presents some improvements in the methodology for analysing mid-term dynamic processes in power systems. These improvements are: an efficient application of the hierarchial clustering algorithm to adaptive identification of coherent generator groups and a significant reduction of the mathematical model, on the basis of monitoring the state of only one generator in one of the established coherent groups. This enables a flexible, simple and fast transformation from the full to the reduced model and vice versa, a significant acceleration of the simulation while keeping the desired accuracy and the automatic use in continual dynamic analysis. Verification of the above mentioned contributions was performed on examples of the dynamic analysis of New England and Yugoslav power systems.