Augmented LQG controller for enhancement of online dynamic performance for WTG system

Operation of variable speed wind turbine generator (WTG) in the above-rated region characterized by high turbulence intensities demands a trade-off between two performance metrics: maximization of energy harvested from the wind and minimization of damage caused by mechanical fatigue. This paper presents a learning adaptive controller for output power leveling and decrementing cyclic loads on the drive train. The proposed controller incorporates a linear quadratic Gaussian (LQG) augmented by a neurocontroller (NC) and regulates rotational speed by specifying the demanded generator torque. Pitch control ensures rated power output. A second-order model and a stochastic wind field model are used in the analysis. The LQG is used as a basis upon which the performance of the proposed paradigm in the trade-off studies is assessed. Simulation results indicate the proposed control scheme effectively harmonizes the relation between rotor speed and the highly turbulent wind speed thereby regulating shaft moments and maintaining rated power.