Publications Details

Novel, Nacelle-Mounted Spire for Accelerated Wind Turbine Wake Decay

Houck, Daniel; deVelder, Nathaniel d.

Wind turbine wakes are characterized by helical trailing tip vortices that are highly stable initially and act as a shield against mixing with the ambient flow and thereby delay wake recovery until destructive mutual interference of the vortices begins. Delayed wake recovery in turn reduces the power production of downstream turbines that are positioned in the wakes of upstream turbines. The long natural decay length forces wind farms to have large distances between turbines to yield sufficient wake recovery. Herein, we tested a new concept aimed at accelerating the breakdown of wind turbine tip vortices by causing the vortices to interact with one another almost immediately behind the rotor. By adding a spire behind the rotor, essentially a blockage to perturb the paths of the tip vortices, we hypothesized that the altered paths of the tip vortices would cause their destructive interference process to begin sooner. The concept of a nacelle-mounted spire was tested in high-fidelity large-eddy simulations using Nalu-Wind. Four different spires were modeled with wall-resolved meshes behind the rotor of a wind turbine with another turbine five diameters downstream. We compared power and wake data against baseline results to determine whether the spires accelerated wake recovery of the upstream turbine and thereby increased the power of the downstream turbine. The results showed no change in the total power of the two turbines for any spire compared to its respective baseline. These results were further explored by testing at higher spatial resolution and without turbulence in the inflow. The increased spatial resolution increased the apparent stability of the tip vortices while the lack of turbulence did not. We conclude that the spires’ geometry and size were inadequate to alter the helical paths of the trailing tip vortices and that modeling of the formation and decay of tip vortices may be highly sensitive to model parameters.