A wind tunnel study on cyclic yaw control in a wind farm model

Dynamic control strategies, such as dynamic induction control [1], were reported as efficient and beneficial in improving wind farm power production. A recent work based on computational fluid dynamics simulation [2] shows that by controlling wake meandering through cyclic yaw angle variation, a faster wake recovery and higher available power in the wake of a wind turbine can be achieved. To investigate the potential of this cyclic yaw control (CYC) strategy in wind farm power improvement, we conducted wind tunnel experiments on a model wind farm consisting of three miniature wind turbines (arranged to be aligned with the inflow direction with a streamwise spacing of five rotor diameters). The yaw angle of the most upwind turbine was controlled to vary sinusoidally. The two control parameters, including the yaw angle amplitude and the yaw Strouhal number (i.e., the normalized frequency), were adjusted to optimize the power performance of the wind farm. Based on both power and wake measurements, we found that cyclic yawing can enhance the lateral flow entrainment and thus increase the power production of the wind farm. The power performance of the wind farm is found to be dependent on the control parameters. A maximum power gain of 15.2% is achieved in our study. We found that the controlled wake meandering dynamics are highly periodic. The phase-averaged wake center trajectory also highly resembles the sine wave, making it possible to predict the instantaneous wake deflection using the wave equation. Furthermore, it is found that the amplitude of periodic wake meandering first increases and then decreases with the increase of the downstream distance from the turbine location. The critical downstream distance (where the amplitude attenuation starts) is found to be around one wavelength. At the growth stage, the amplitude can be well predicted with the yawed wake model [3] (at a yaw angle equal to the yaw angle amplitude of CYC), while the physics of amplitude decay still needs to be better understood to predict the decrease stage. The amplitude decrease can be related to the damping effects (i.e., energy dissipation) due to turbulent wake mixing, which will be considered in future work on analytical model development.