Robust power system controller design based on measured models

Abstract

This paper presents combined power system identification and controller design methods to dampen low-frequency oscillations in multimachine power systems. An iterative closed-loop identification method is used to find a linear model for the power system. Linear quadratic Gaussian controller design with loop transfer recovery (LQG/LTR), based on a generalized technique for the nonminimum phase (NMP) power system model, is used to design the controllers. Simulation results are presented to demonstrate the robustness of these controllers based on closed-loop identified plant models and the amount of loop transfer recovery that is possible for NMP plant models.