Posters

Abstract

Methodology Ground based lidar has been considered as a promising wind measurement technology in wind resource assessment and wind turbine power performance test. Thanks to doppler effect, lidar devices reconstruct wind flow filed by assuming that the flow is homogenous. However, this assumption is challenged in complex terrain due to inhomogeneity of the wind flow and especially flow curvature. In consequence, the measurement uncertainty is increased in complex terrain, and it has become one of the main constraints for using lidar more widely. With the topographical data of a specific site, microscale CFD modeling is used to reproduce the inhomogeneity of the wind flow and use it to correct lidar measurement. The correction considers site-specific conditions and the specific wind flow reconstruction method used by the lidar device which could differ from one provider to another. Then the correction technique for a given location is implemented to do batch processing on all the locations in an efficient way, for varying incident wind directions and at different heights. Thus, the spatial distribution of lidar error can be determined precisely. Contribution On one hand, the study improves the understanding of lidar measurement error in complex terrain. On the other hand, the method can be generalized and used in any site with complex topography, to predict the measurement error at potential locations prior to the installation of lidar, and thus helps to optimize lidar placement from the perspective of quantifying and reducing uncertainty. Result The spatial distribution of lidar measurement error varies with incident wind direction and height. Following characteristics have been identified: * Lidar trends to underestimate wind speed on the ridge. The most severe underestimation occurred when the wind direction is perpendicular to the ridge. * The spatial disparency of error is the most notable when the wind direction is perpendicular to the ridge, and less notable when the wind direction is parallel to the ridge. * The variation of lidar error at the heights 100m/150m/200m shows decreasing disparency as well as decreasing extreme value of error with increasing height. However, a finer study shows that the extreme value of error may occur at a particular height. This phenomenon is related to the geometry of lidar scanning cone, combined with the inhomogeneity of wind flow at different height (which is furtherly impacted by the topographical condition around the specific location). Thus, an accurate flow model is fundamental to have a good assessment of the error. * The main influencing factor on lidar measurement error is the flow curvature, causing difference of inflow angle between upstream and downstream measurement points. Conclusion Lidar measurement error has been evaluated from wind flow parameters obtained with CFD-based numerical simulation. By using this technique, it has become possible to carry out a full analysis on the spatial distribution characteristic of lidar error over a specific site. It also provides the possibility to run an optimization of lidar placement in a specific site.