Doppler LiDAR observations of wake meandering at a utility scale wind turbine
Wake meandering is a low-frequency oscillation of the entire wind turbine wake with respect to the time-averaged centerline. These oscillations of the instantaneous wake position contribute to the turbulence received by a downstream turbine in a wind farm leading to power and load fluctuations. One hypothesis on the origin of wake meandering assumes that large scale turbulent eddies in the atmospheric boundary layer cause a lateral and vertical displacement of the wake as it is transported downstream (Larsen et al., 2008). For modelling purposes, a common assumption is that the wake behaves like a passive scalar during this advection process. Here, we investigate (i) the downstream propagation velocity of the wake meandering, (ii) the quality of a linear relationship between large scale turbulence and the instantaneous wake center position, and (iii) the errors of a passive advection based prediction of the instantaneous wake center position. Field measurements of wake meandering from two Doppler LiDARs mounted on the nacelle of a utility scale wind turbine (79 m hub height, 96 m rotor diameter) are used for the investigation. One Doppler LiDAR was scanning the longitudinal velocity field of the wake at hub height, from which the instantaneous wake center position was estimated. The other Doppler LiDAR used a stare setup to provide the lateral inflow velocity component. The dataset covers a wide range of wind speeds (5 - 11 m s-1) and turbulence intensities (1% - 9%). First, the wake measurements were used to investigate the downstream propagation velocity of wake meandering. The downstream propagation was estimated with a cross-correlation approach between the instantaneous wake center positions at downstream distances separated by three rotor diameters. The results yielded downstream propagation velocities between 0.7uhub and 0.9uhub (with uhub being the inflow wind speed measured at the nacelle) and that it increased slightly with downstream distance (Fig. 1). The found downstream propagation velocities were lower than the wind speed outside of the wake, but higher than the velocity at the wake center. Second, we investigated the relationship between the lateral velocity and instantaneous wake center position. In agreement with wind tunnel studies (e.g. Bastankhah and Porté-Agel, 2017), we found that the wake meandering strength (or amplitude) increases with the lateral turbulence intensity and with downstream distance. Further, we found that the correlation between the lateral velocity and the instantaneous wake center position shows a dependency to the ratio of the integral time scale of the lateral velocity to the time delay due to downstream advection (Fig. 2). This finding could imply that wake meandering predictions based on a passive advection might exhibit increasing errors for large downstream distances or fast evolving turbulent velocity fields. Lastly, we made predictions of the instantaneous wake center position using the lateral velocity and the downstream propagation velocity. Preliminary results show that these predictions underestimate the strength of wake meandering compared to the measurements, and that the root-mean-square error between predictions and measurements increases with the aforementioned ratio of integral time scale to downstream propagation time.
2021-05-27
15
Event name | Event place | Event date |
[online] Hannover, Germany [online] Hannover, Germany | 25 - 28 May 2021 | |