Real-Time Identification of Coherent Generator Groups

This paper presents a two-step approach for defining coherent generators in disturbed power systems based on the similarity among their inter-area oscillations and swing curves. In the first step, an online visualization of the generator swing curves is performed by extracting time domain solutions of the swing equation using PMU measurements. In the second step, the similarity between each pair of the generator swing curves is examined based on an intraclass correlation analysis. The pairwise similarity coefficients obtained are served as edge's weights to a fully connected graph. The grouping of the coherent generators is completed through a graph minimization algorithm. The proposed methodology is applied to the 39-Bus New England System and coherent generators are obtained for different operating points to provide a more accurate and realistic grouping.