Evaluation of Coherency-based Aggregation Methods

This project applies coherency methods for dynamic analysis of power systems. It has been observed that in case of disturbances machines swing together, forming groups of coherent generators. In [1], a dynamic network reduction is described based on a slow-coherency method, aggregating coherent generators into one equivalent machine. In [2] it is demonstrated that such methods can also be used for large disturbances, in order to identify groups of coherent generators and apply controlled islanding. The objective of this project was to implement slow coherency methods in a medium and a large scale power system and test their effectiveness. We implemented algorithms that would group the machines offline based on coherency between them independent of the disturbance and then evaluated their performance in comparison to actual simulations performed in PSS R © E. In this way, conclusions were drawn regarding the effectiveness in representing the dynamic properties of the full network by an equivalent reduced one and the possible applicability of slow coherency in controlled islanding algorithms.