A Maximum Power Point Tracking Approach for Wind Farm Control

This paper presents a new methodology for controlling wind farms in a decentralized framework, where the wind turbines exchange information with their neighbours, and update their own control parameters in order to optimize the total power produced by the wind farm in a time-efficient manner, taking into account effects of wake interaction. The method can be characterized as being model-free since it does not require a model of the aerodynamic interaction between the turbines. The total power optimization is performed using gradient-based optimization. The gradients are approximated based on information of the past control actions, the power response of the turbine itself, and the power response of neighbouring turbines. To achieve time-efficiency, only the influence on the nearest neighbouring turbine is taken into account. The method is tested on a simulation of the Princess Amalia Wind Park, a 60 turbine offshore wind farm in The Netherlands. Because of the improved time-efficiency of the control update scheme, the new method yields a faster convergence of the power optimization when compared to an existing model-free game-theoretic wind farm power optimization method.