LIDAR-assisted exact output regulation for multi-megawatt wind turbines

© 2019 Amin MahdizadehWind energy is free and accessible almost everywhere on earth. Nevertheless, the utilization of wind energy is associated with the costs of the material, manufacturing, installation, maintenance, and operation. These factors define the Levelized Cost Of Energy and its competency against other energy sources. In order to improve the competitiveness of wind energy in the energy market, wind turbines operational performance and efficiency should be further optimized. The wind turbines performance is improved by maximizing the energy harvest whilst minimizing the mechanical extreme and fatigue loads on the tower structure, blades and the rotor. Over the past few years, various technologies and methods have been implemented to improve the performance of the wind turbine. Among these technologies, the Light Detection And Ranging devices (LIDARs) have been a point of interest for researchers. The LIDAR’s ability to provide information about the upcoming wind has originated a new generation of wind turbine control techniques. However, effective augmentation of the wind preview information into the wind turbine control systems has proven to be a challenging problem to-date. This thesis explores the application of a classical control method, namely Exact Output Regulation, to improve the control performance of utility-scale wind turbines by utilizing the LIDAR measured wind information in the control loop. Here, the Exact Output Regulator is designed to reject the known input disturbances, i.e., wind, whilst ensuring that the system output tracks a desired reference signal. The proposed Exact Output Regulator-based wind turbine controller is comprised of a state feedback controller together with a feed-forward gain matrix. The LIDAR wind preview information can be used to obtain a low-order exosystem that models the wind dynamics. This exosystem wind model is used to obtain the feed-forward gain matrix that enables the Exact Output Regulator to effectively reject the impact of the input wind disturbance on the rotor speed and achieve the desired reference tracking. To perform realistic simulations and analysis and obtain reliable results, different software, such as high fidelity aero-elastic simulators (e.g., FAST), 3-D turbulent wind field generators (e.g., TurbSim), and performance and fatigue evaluation code are required. To this end, a toolbox is developed alongside this thesis which can synthesize the Exact Output Regulator in real-time and apply it to the FAST model of the wind turbine. This toolbox is called Turbine Output Regulator, used for simulations throughout this thesis. Various simulations are performed on the Exact Output Regulated wind turbine in both partial and full load operating regions using different wind fields of different mean wind speeds. The full nonlinear aero-elastic model of the National Renewable Energy Laboratory’s 5-MW reference wind turbine is used as the wind turbine model. A Continuous Wave LIDAR simulator is used in order to replicate the LIDAR measurements of the wind fields constructed by TurbSim. For performance comparisons, we implemented a feedback Baseline controller and a feed-forward Disturbance Accommodating Controller. It is also shown that the Disturbance Accommodating Controller may be represented as a special case of Exact Output Regulator, if some important parts of the Exact Output Regulator architecture are omitted. The simulation and analysis results on the turbulent wind scenarios show that, in comparison with a Baseline and Disturbance Accommodating Controller, the Exact Output Regulator can provide a substantial reduction of extreme and fatigue loads. The Exact Output Regulator is shown to be much more reliable in preserving the wind turbine states within the permissible range