6th International Conference Energy & Meteorology: Abstract Submission

WRF-LES evaluation of turbulence measures using observations from a 250-m tower (665)

Alfredo Peña 1 , Rogier Floors 1 , Andrea Hahmann 1
  1. Technical University of Denmark, Roskilde, GREAT COPENHAGEN, Denmark

The model-chain evaluation of wind atlases (MEWA) project aims, among others, at evaluating meso-micro-scale coupling techniques for wind resource and site assessment. One of those techniques downscales a mesoscale flow solution dynamically via a large-eddy simulation (LES).

Here we investigate the ability of the Weather Research and Forecasting (WRF) model to perform LESs. This is achieved through comparison of WRF-LESs with measurements from a 250-m lightning mast deployed at the Østerild test center for large wind turbines, which is located in northern Jutland, Denmark. Østerild is on a flat but rough site and for the predominant wind directions the flow might be close to homogenous, which allows us, in principle, to compare the measurements with LES of canonical flows. The lightning mast is equipped with, among others, Metek USA-1 sonic anemometers at five levels above ground (7, 37, 103, 175, 241 m). The sonic anemometer measurements provide details on the turbulence structure of the atmosphere within the range of heights where modern large wind turbines operate.

The canonical flows we are aiming at simulate represent the atmosphere under the three common atmospheric stability regimes: unstable, neutral and stable. From preliminary analysis of the sonic anemometer measurements at Østerild, we observe a clear dependence with atmospheric stability and height above the ground of the three parameters we use to model the three-dimensional turbulence structure of the atmosphere. The flow becomes less anisotropic the higher we observe and the more unstable the atmosphere is. We also observe an increase of turbulence dissipation the closer to the ground, particularly for neutral conditions. Finally, we observe an increase of the turbulence length scales with height, being the length scales larger under unstable conditions. The idea is to be able to reproduce these findings also from LES and to show how close or far our simulations are from the measurements.