Influences of Lidar Scanning Parameters on Wind Turbine Wake Retrievals in Complex Terrain

Scanning lidars provide the ability to collect spatially-resolved measurements of wind fields. Lidar measurements may be used to estimate wake magnitude, extent, and trajectory. The characterization, however, may be subject to distortions due to the observational system: resolution of the measurement points across the wake, the projection of the winds onto the beam, averaging along the beam probe volume, and intervening evolution of the flow over the scan duration. We use a large-eddy simulation (LES) and our virtual instrument model (Robey and Lundquist 2022) to assess how the parameters of scanning lidar systems might influence the properties of the retrieved wake using a case study from the Perdigão campaign.

Here, simulated lidar measurements of the wake are compared against reference simulations with WRF-LES to provide insights into the observation system behavior. We consider three lidar performing range-height indicator (RHI) sweeps in complex terrain based on the deployments of the DTU (Danish Technical University) and DLR (German Aerospace Center) lidars in Perdigão. The unwaked flow, measured by the DTU lidar, is generally well-captured by the lidar, even without combining data into a multi-lidar retrieval. The two DLR lidars measure a waked transect from different downwind vantage points. In the region of the wake, the observation systems interact with the more spatially and temporally variable winds and smaller spatial scales similar to that of the lidar probe length, allowing more significant discrepancies to arise. While the measurements do largely capture the wake structure for much of its ~4D extent (Figure 1), the reduced resolution of measurement points and volume averaging lead to a quicker loss of the two-lobed distinction of the near wake, displacement of the wake center (frequently 5-10m too high), smearing of the vertical bounds of the wake (< 20 m), and dampening by up to a third of the peak normalized velocity deficit (ΔU=1-U_wake/U_unwaked).

Figure 1: For a five-minute period, (a) vertical profiles of horizontal wind deficit for the LES wake and those measured using each of the DLR lidars along with the detected wake center line and height. Comparison of the (b) normalized maximum wind deficit magnitude (c) wake center relative to the terrain-following turbine hub height (gray line in panel a) and (d) the vertical extent of the wake. Dashed vertical lines show where the fit transitions from a double to single Gaussian, i.e. near to far wake.