Posters

Abstract

LiDAR technology plays an essential role in wind resource assessments and energy yield prediction. LiDAR measurements rely on aerosol backscatter of the emitted laser beams to calculate the wind speed up to several hundred meter above ground. Physical factors such as beam blockage from vegetation, terrain features, man-made structures, or low aerosol content caused by overly clean air or a general decrease with height can reduce data availability. A decrease in data quality leads to an increase in measurement uncertainty and can contribute to project financing costs. This study validates the capabilities of two previously developed availability-improving algorithms by comparing a total of 528 days of ensemble data from 14 LiDAR devices to a 200m IEC-compliant meteorological mast in northern Germany. We observe a significant increase in availability for both individual and combined algorithms (average increase of 4% at a height of 200m) and we anticipate that the availability increase would be even higher in areas with low aerosol concentration. The algorithms demonstrate reliable detection and compensation for beam blockage caused by obstacles such as wires. Our analyses show excellent data quality and a strong correlation of 0.99 to reference measurements for the additionally gained data points, providing a compelling case for the application of these algorithms.  The deep learning quality control (DLQC) algorithm applies semantic segmentation to images of line-of-sight wind speed time series to identify ambiguities, outliers, hard targets and valid wind data. The dynamic selection of wind field reconstruction algorithm (D3LOS) detects beam obstruction and dynamically calculates the wind speed based on the remaining unobstructed beams. D3LOS uses a Moore-Penrose inversion with 3-, 4- or 5-beam.  These new algorithms represent a significant improvement over traditional filters and can be applied to any WindCube LiDAR including LiDAR profiler, nacelle LiDAR, and scanning LiDAR. They enable the continuous generation of high frequency, high quality wind measurements under conditions that would typically be filtered out by the conventional -23dB cutoff. They enhance the reliability and flexibility of LiDAR, especially in areas with low aerosol concentration, as a replacement for met masts and support the transition to renewable energy.