Smith prediction approach for the design of a robust wide-area measurements based hierarchical controller

This paper presents a robust and delay insensitive version of the wide-area measurements based two-level hierarchical controller introduced in F. Okou et al. (2004). The hierarchical structure consists of a central controller at the secondary level dedicated to compensate inter generators interactions and conventional controllers (AVR-PSS/SPEED GOVERNOR) at the primary one to dampen local oscillations. A Smith prediction approach is used to preserve the controller performance in presence of remote measurements delays and communication delays between the central and the local controllers. The controller robustness and insensitivity to the nonlinear dynamical behavior of various power system loads is improved by an online load estimation algorithm which continuously updates both the central controller and the predictor parameters. The proposed structure is tested in a realistic situation using a large power system with an induction motor as dynamic load together with voltage dependent load. Simulation results reveal that delay compensation and load estimation characteristics are effectively required to enhance the hierarchical structure performance. Furthermore, local controllers´ performances are considerable increased by the secondary controller action and the system stability is improved in presence of severe contingencies.