Coherency Based Design of Decentralized Optimal Controllers for Stabilization of Large Power Systems

Abstract

This paper gives a new procedure for the design of decentralized optimal controllers for very large power systems. The group of machines, comprising one decentralized area is identified by a coherency recognition algorithm. The coupling between the elements within the group is strong and the coupling between groups are weak. This principle is used to identify the groups. A pole assignment technique, which requires calculation of one eigenvector only, for shifting one eigenvalue is used. This method requires only less computer storage and less computing time, yet achieving results very much comparable with the design of full controllers. The procedure is illustrated with an example. This has very good applications in the design of optimal controllers for large power systems and in their dynamic analysis.