Wind-field characterization using synthetic lidar measurements and proper orthogonal decomposition

This paper presents a simple least-square method combined with Proper Orthogonal Decomposition to reconstruct full-rotor flow, using synthetic measurements from a pulsed lidar mounted on the turbine hub. The proposed lidar effectively overcomes blade blockage effects, enhancing data availability. Conducted at a wind speed of 11.4 m/s with 10% turbulence intensity, the study assesses wind-field reconstruction accuracy with the proposed method by examining the influence of mode count and measurement range selection. Comparisons with a baseline, derived from averaging line-of-sight across the rotor plane, reveal that including more modes generally improves reconstruction performance, achieving up to 57% error reduction in the wind-field reconstruction over the baseline. However, this benefit is constrained by the availability of measurements at each time step; limited data coupled with an increased number of modes can lead to overfitting, escalating errors. The method demonstrated here offers advantages in characterizing turbine responses, particularly in capturing low-frequency content in the wind-flow. Yet, channels like tower base moment necessitate a substantially higher number of modes for accurate characterization. Overall, this approach shows potential for real-time wind-flow estimation in lidar-assisted control applications.