Skip to Main Content
This study aims to analyse different linear quadratic Gaussian (LQG) controllers' performances in terms of reducing the fatigue load of wind turbines' (WT) most costly components caused by the spatial turbulence of wind speed. Five LQGs with increasing control model complexity and a greater number of objectives are designed, the first four with collective pitch control (CPC), and the fifth with individual pitch control (IPC). In the design of the controllers, firstly a linear control model is obtained in the operating point corresponding to a wind speed of 18 m/s. Then, the Kalman filter (KF) and the rest of the controller are tuned with simulations in order to obtain the lowest possible fatigue loads while respecting certain generator power and speed variation limits. Finally, the five controllers are tested with processor-in-the-loop (PIL). Fatigue loads are evaluated by rainflow counting algorithm and then applying the Palmgren'Miner rule. Tests results show that drive-train loads are significantly reduced from LQG1_CPC, that the complexity of the controllers does not have a significant influence on the reduction of tower loads, and that LQG3_IPC allows fatigue loads on blades to be alleviated considerably.