Wind speed reconstruction from mono-static wind lidar eliminating the effect of turbulence

Abstract

Light detection and ranging (lidar) is an attractive technique to measure wind speed via the Doppler effect, thanks to its flexibility. Although stand-alone deployment is growing, current standards for wind energy applications often require wind lidars to be verified against an in situ met mast with cup anemometers, which deliver point-like measurements. Analyzing 38 lidar-to-cup verification campaigns at eight accredited sites, we investigate the difference between these techniques during turbulence influenced atmospheric conditions. We demonstrate that strong turbulence results in the lidar 10-min scalar average wind speed overestimate the cup value, whereas the lidar vector average results in an underestimate. Although scalar and vector averaging provide acceptable measurements, we show that a linear combination, with appropriate weighting, reduces the lidar turbulence sensitivity by more than one order of magnitude. The resulting overall lidar-to-cup bias is within the one-standard-deviation comparison uncertainty at all sites. The resulting ensemble mean bias is (- 0.14 ± 0.42) %, where the uncertainty stems from variations between sites. We explain the experimental observations by a model derived from first principles. The developed hybrid averaging is implemented in the Leosphere WindCube v2.1 wind lidar.