LQG / LTR controller for speed governing of hydro-turbine

This paper attempts to investigate the approach of linear quadratic Gaussian methodology with loop transfer recovery (LQG/LTR) for hydro turbine control. The Kalman filter estimates the state of the plant and an optimal control law is used to minimize the performance index represented as quadratic objective function. First the model is simulated with LQR (Linear Quadratic Regulator) considering that all the states can be measured. Next, the results of the pure LQ controller are compared with a LQG controller. With LTR (loop transfer recovery) a satisfactory return ratio is recovered for the Kalman filter in LQG compensator design. It is shown that turbine control based on LQG/LTR is nearly as good as the turbine control based on LQR.