Dynamic coherency identification of generators using self-organising feature maps

Abstract

An important step in constructing dynamic equivalents of large power systems is the coherency identification and grouping of generators. Self-organising feature maps can do this task, if the attribute vectors, which characterise the features of the generator dynamics inside the network are well chosen. It is shown in the paper that the principal components of the correlation matrix of the simulated time responses of the generators after faults are especially suitable to form the attribute vectors. The results are compared with the use of right eigenvectors and participation factors of the linearised system matrix as attribute vectors.