How effective is the Mann turbulence model in predicting horizontal turbulence from vertical data?

Abstract

Offshore wind turbines are approaching 300 m.a.s.l., where the dynamics of the atmospheric turbulence remain unclear. Present understanding is limited, with direct measurements above 100 m from the sea surface being scarce.

The thesis assesses the feasibility of deducing the along-wind spectrum from measurements of the vertical velocity combined with numerical modeling, using the uniform shear model (Mann, 1994). The vertical component experiences less influence from mesoscale motions, enhancing stationarity and potentially simplifying instrumentation and data collection. The primary objective is to determine if the along-wind standard deviation the governing parameter in modeling wind loads on wind turbines (Wiley et al., 2023), can be accurately estimated from the deduced along-wind spectrum.

The method was tested on sonic anemometer data from the FINO1 platform at 81.5 m collected during 2007 and 2008 for two different cases: 1) all velocity components were known, and 2) only the vertical component was known. Fitting to all components showed good agreement between the estimated along-wind standard deviation and the target values. Encountering local minima was a challenge when fitting to only the vertical component. This was partially avoided by using a narrower wavenumber interval and lower iteration tolerances in the fitting. The method performed quite well for mean wind speeds up to 17 m/s. For higher mean wind speeds, the standard deviation was significantly underestimated, possibly due to underestimation of the anisotropy parameter .

Additionally, the method was briefly tested on vertical velocity data from a Leosphere WindCube 100S, with a novel attempt to correct for spatial averaging. Lack of simultaneous point-measurements prevented verification of the correction method and the deduced along-wind spectrum obtained from fitting the uniform shear model to the original and corrected spectrum.