Fuzzy Equivalence Relation Clustering-Based Algorithm for Coherency Identification among Generators

Abstract

Coherency identification among generators plays an important role in controlled islanding and wide-area monitoring of a power system with renewable energy sources. So far, wide-area measurement systems (WAMS) have been deployed widely in power plants and key substations, so the trajectories measured by phasor measurement units (PMUs) and WAMS could be utilized to identify coherent generators (CGs). In this study, an algorithm for coherency identification among generators based on the fuzzy equivalence relation (FER) clustering method and synthesized weight is proposed. First, 10 indices for measuring the similarity of trajectories are presented. Then, the technique for order preference by similarity to ideal solution (TOPSIS) and a method based on entropy and multicorrelation coefficients is presented to determine the weight of each index for decision making of trajectory similarities. Next, the FER clustering method is proposed to cluster the trajectories and the F-statistics, which could measure the rationality of clustering results, is presented to determine the optimal cluster number of coherent groups among generators. Finally, the actual Guangdong power system in China, a revised 16-unit 68-bus power system, and the simplified China South Power Grid (CSPG) are used to demonstrate the effectiveness of the proposed methodology. Comparisons with some existing methods are performed and the impacts of integrated renewable energy generation sources on power system oscillations are also investigated.