A on-line coherence identifying method based on independent component analysis
Jing Wang, Ke Guo · 2014
For the large interconnected power system, dynamic equivalence could significantly reduce the computing load and manifest the dominant characteristics. Coherent method is one approach of dynamic equivalence and its core is the automatic identification of coherency. Based on the angle perturbed trajectory measured by WAMS, a new coherency recognition method based on Independent Component Analysis is proposed in this paper, which can take advantage of different fault types and the dynamic behavior of nonlinear systems varying characteristics into account. First of all, the power angle collected by WAMS were centralized and whiten and then the angle perturbed trajectory feature was obtained by the FastICA algorithm and the coherent generator groups were identified consequently. This method avoids building of the system model, determining of the system parameters and the model of system operation. The simulations of New England 39-bus system prove the validity of the proposed method. It could give the clustering results accurately.