Coherency identification based on Maximum Spanning Tree Partitioning

This paper presents a novel method for finding coherent groups of generators in large power systems, based on graph partitioning and clustering methods commonly used in other areas like image processing and gene expression analysis. First, the power system is modeled as a graph in order to obtain a maximum spanning tree that represents the strongest connections between generators in the system. Then, inconsistent edges are removed from the tree according to proposed criteria. This leads to the identification of the coherent groups of generators. Two algorithms are proposed for the elimination of the edges in the maximum spanning tree, both based on Fukuyama-Sugeno validity index. The method is tested on 68-bus test system and 678-bus Italian equivalent system. Results are compared with classical approaches.