Distributed Control of Wind Farm

Abstract: The growing use of wind energy as major energy source, has led to the construction of large wind farms. The control of such wind farms is typically separated into control at wind turbine level and at wind farm level. The wind farm controller assures that the wind farm produces the demanded amount of power, by providing set-points for the turbines in the farm. The local turbine control assures that the local set-point is tracked. Current wind farm controllers conduct the power distribution in a static manner. In this work it is explored, how fatigue reduction in the wind farm can be achieved, by letting the controller distribute the power demand dynamically. In the first part of this work, the problem of reducing the fatigue on the turbines by dynamic power distribution is presented. A controller is designed based on this problem formulation, taking both tower and shaft fatigue into account. The controller is designed, such that it is both modular and scalable. This has the advantage of the controller being identical on all turbines in the wind farm. Also, the modularity allows turbines to be added or removed from the wind farm, without changing the controllers on the remaining turbines. Evaluation of the controller in a realistic simulation environment verifies the functionality of the controller, and shows that fatigue reductions of the tower and shaft in the magnitude of 10 % and 50 % respectively, can be expected. The final part of this work examines how the concept of fatigue reduction through dynamic wind farm control can be implemented on the offshore wind farm Thanet. Simulations indicate that the limitations of the server system at Thanet does not render this type of control possible.