LQG decentralized controls for FACTS based on structure of modeling by subsystems

It is well known that the main disadvantage of centralized controls systems in power systems is the time delay that is produced by communication systems. This paper presents a design technique of decentralized LQG controllers, which conserves the optimality characteristic of an LQR regulator. The design of these controllers is based on a proposed structure of subsystems of low order, which can be calculated using either order reduction techniques or identification techniques. This structure is noncoupled in state variables, that allows the computation of independent LQR regulators. The paper presents the calculation of equivalent estimators for each subsystem taking into account the other optimal controllers. The proposed method is used to coordinate two SVC controllers. Results of simulations of the controlled system are shown.