Abstract
Accurate characterization of atmospheric turbulence is essential for wind energy applications, influencing turbine design, load assessment, and wake modeling. Floating lidar systems offer a cost-effective alternative to offshore meteorological masts, but their turbulence measurements are affected by several factors, including spatial averaging, sequential scanning, platform motion and lidar configuration. This study introduces an N-beam method that exploits platform motion to derive the full Reynolds stress tensor from a commercially available five-beam pulsed lidar. The approach constructs a large set of effective beam orientations within each averaging period using deterministically motion compensated line-of-sight data, and estimates the second-order turbulence statistics through an overdetermined least-squares system. The method is evaluated against a sonic-anemometer reference and a standard velocity-azimuth-display (VAD) retrieval as benchmark. Compared to the VAD benchmark, the N-beam method significantly improves the estimation of velocity variances and therefore turbulence intensity (TI), reducing inter-beam contamination and achieving close agreement with the sonic reference. Velocity covariance retrievals show limited improvement, reflecting fundamental constraints of sequential lidar scanning. Future work should include stability-dependent error analyses and comparisons with measurements derived from a fixed six-beam lidar. Such measurements can also be used to estimate second-order turbulence statistics through an overdetermined least-squares system, helping to isolate the effects of platform motion. Moreover, the flexibility of the N-beam framework could be leveraged to selectively limit contributing beams based on geometry, enabling the construction of virtual six-beam or other optimized configurations.

This publication has 10 references indexed in Scilit: