Colloquia

Summary

Wind turbines constantly react and adjust their blades to changes in the incoming wind which are detected through monitoring the measured blade loading and rotor speed. Accurate and real-time predictions of the flow allow for more pro-active behavior resulting in improved efficiency and lowered blade loading. In previous research a simplified wind model has been constructed aiming to predict the short-term incoming wind field. The current research focuses on validating the most important underlying assumption: Taylor's Frozen Turbulence hypothesis. Using Large Eddy Simulation data, several cases of varying atmospheric stability and mean wind speed were analyzed. The analysis mostly focuses on the longitudinal cross-correlation at the delay where the turbulent perturbations would arrive if they were advected with the mean flow. The analysis is applied to the entire temporal velocity signal and on velocity signals where a specific frequency band of interest is isolated through a coherence analysis. Estimates for the validity distance (the distance up to what turbulent eddies can be considerd advected without significant changes to their properties) are provided based on several criteria: a cut-off cross-correlation, an auto-correlation cut-off and the integral length scale.