Posters

Abstract

Measurement uncertainties from remote sensors (RSD) are quantified following the IEC 61400-50-2. Each new RSD model must undergo a rigorous evaluation of its environmental sensitivities, known as Classification. Environmental sensitivities are a component of an RSDs overall uncertainty budget during measurement campaigns. In this research, we present a new technique to compensate for RSD environmental sensitivities in a transparent, traceable way, allowing for lower overall wind measurement uncertainty. The two primary RSDs used in wind energy applications are sodar and lidar. Both technologies use focused beams of energy to probe the atmosphere, measuring the Doppler shift of the backscattered energy to estimate wind velocity. These energy probes are spread out in space, and the RSD line-of-sight (LOS) measurements are a volume average of the wind velocities within the probe. Even for a perfect, idealized RSD, this volume averaging effect leads to a wind shear sensitivity due to the typical, non-linear form of the wind profile power law, and the shape of the probe’s range-weighting function (RWF). In this research, the wind shear sensitivity observed in an official device Classification is successfully replicated in a simulated environment. A virtual lidar is configured using device-specific data gathered during the lidar’s manufacturing process. The simulated atmosphere is configured to match the wind conditions during the Classification measurement campaign. The simulated sensitivity slopes at all altitudes are within ±1 percent-per-unit-shear of the sensitivity slopes derived from the Classification campaign. Using the simulated lidar behavior, a correction is applied to the lidar measurements from the Classification campaign. After correction, the wind shear sensitivity slope is reduced below the threshold of significance, eliminating the lidar’s wind shear sensitivity and reducing the direct Classification uncertainty to zero. There remains a residual uncertainty due to the uncertainties of the correction inputs and the simulation. This residual uncertainty is quantified for each altitude, with average values of roughly 0.2%. This residual correction uncertainty is entered into in the lidar measurement campaigns’ uncertainty budget in place of the Classification uncertainty, significantly reducing the overall uncertainty of the lidar. This white-box approach allows for full traceability of the measurements, and application to specific site conditions beyond those captured during the Classification campaign.