Wind Farm Sliding Mode Control and Energy Optimization with Fatigue Constraints

Wind energy is the fastest growing market in the energy production, and it is estimated that by 2030, over the 20% of global energy production will be done with this technology. However, with the increase of dimensions of both the single turbines and of the clusters of wind turbines, generally known as wind farms, the engineering challenges increase too. In this thesis, a model for a wind farm composed of four wind turbines and the relative control algorithms are presented. A simplified model for an Horizontal Axis Wind Turbine (HAWT) has been developed considering both the aerodynamics and the electrical generator; this last one has been developed considering a Doubly-Fed Induction Machine (DFIM) controlled and connected to the grid by a back-to-back converter, composed of two bi-directional voltage source inverters (VSI). For defining the aerodynamic interaction between the four turbine, assumed to be installed in a diamond shape layout, a wake model, based on the Jensen model has been developed; this model shows how the overlapping effect between a wake and another turbine, reduces the wind input speed, reducing the available power to extract. In order to control the single wind turbines, a vector control based on the super-twisting sliding mode control (STW-SMC) has been developed for both the VSIs; the results shows a performance tracking of a desired rotational speed for the DFIMs, while the DC-bus voltage regulation for the grid side converter still presents too high ripples for considering it acceptable. Finally, a supervisory control for the wind farm has been developed, where instead of tracking the maximum power point of energy extraction, an optimization algorithm compute the best rotational speed value on order to have a trade-off between power production, mechanical damages and turbulence generation.