Alexander Raul Meyer Forsting

Modelling Wind Turbine Inflow: The Induction Zone

A wind turbine decelerates the wind in front of its rotor by extracting kinetic energy. The wind speed reduction is maximal at the rotor and negligible more than five rotor radii upfront. By measuring wind speed this far from the rotor, the turbine’s performance is determined without any rotor bias. However, the measured wind speed decorrelates from the one interacting with the rotor especially in wind farms and mountainous terrain. This is exacerbated by the ever growing rotors, as the physical distance to the measurement location grows equally. Decorrelation is mitigated by measuring closer to the rotor, but requires exact knowledge of the flow deceleration to estimate the available, undisturbed kinetic energy. Thus this thesis explores, mostly numerically, any wind turbine or environmental dependencies of this deceleration. The computational fluid dynamics model (CFD) employed is validated with velocity measurements from lidars upstream of an operational turbine. A new stochastic validation methodology in combination with extensive uncertainty quantification and propagation allows validating the CFD model under these realistic conditions for an area covering the majority of the decelerating flow upstream. This is the first validation of its kind and it demonstrates the advantage of including uncertainties in the process. The flow behaviour upstream of a single rotor is largely insensitive to specific rotor designs and operating conditions. In fact the rotor thrust coefficient is the single most significant parameter. Exploiting this singular dependency, a fast semi-empirical model is devised that accurately predicts the velocity deficit upstream of a single turbine. Near-rotor measurements in combination with this model are able to retrieve the kinetic energy available to the turbine in flat terrain. Complex terrain and multiple turbines are more demanding, though, as they enhance non-linear interactions.