Development of a model predictive controller for floating offshore wind turbines

In this work, an Economic Model Predictive Controller for a floating offshore wind turbine is presented. The classical Model Predictive Control for floating offshore wind turbines provides promising results. In addition, research on onshore wind turbines revealed the potential of the economic control method, which can improve the closed-loop behavior and simplify the control design in comparison to the classical version of this control method. The aim of this work is, to develop a novel Economic Model Predictive Controller for a floating offshore wind turbine based on these two research results. A simplified low order model of a floating offshore wind turbine serves as a basis for the controller design. Including the disturbance preview and constraints, the controller computes optimal trajectories for the blade pitch and the generator torque. To apply the control technique to a floating offshore wind turbine two things have to be done: Firstly, the cost function is designed, to fulfill the main objectives of, maximizing the generated power and alleviating the structural fatigues. Secondly, the constraints are integrated into the control problem. After selecting a suitable solver, the controller is discretized and scaled, thus a proper implementation and smooth operation is possible. Afterwards, the successful functioning of the algorithm, a multi-objective optimization is done, to find appropriate weights to adjust the cost function for the required objectives. Finally, the developed controller is tested with realistic wind and wave disturbances. A significant reduction of the standard deviation of the generated power can be shown, while maintaining real time capability. Furthermore, the structural fatigues of the tower and the platform are decreased