OpenFOAM CFD Simulation of Wind Turbine
Unsteady CFD simulations of wind turbines yield detailed insights into the structure, size and power of trailing vortices. Those down-stream eddies combined with strongly decellerated flow in the direct slipstream effect efficiency of wind turbines far down-stream. Parameters like wind-speed, the shape of the wind-profile, airfoil pitch, landscape peculiarities etc. are included in physically correct models. From the simulation results design criteria like positions and placement patterns of highly efficient wind parks are obtained in the planning phase already.
As an example the simulation of a typical wind-turbine with a nominal power of 2.3MW is presented. The applied wind profile and the simulated geometry are shown in the image on the right.
Parameters of the simulated wind turbine:
Rotor height: 78m, Rotor diameter: 80m
Exponential wind profile, Turbulent intensity: 5%
Wind speed at rotor height: 15.3 m/s
Tip speed ratio: 6
Air foil pitch: 10°
Code-development of parametrised wind profiles as boundary conditions of the simulations
Using the approach explained above, we ran a transient simulation of the turbine and visualized the vorticity caused by the rotation of the blades (see video below). The vorticity is highest where the blades of the turbine move fastest: at their tips. Apart from the turbulence, you can also see that there is a distinct vertical difference in wind speed causing a distortion of the "turbulence helix". Also visible (indicated by the blue color in the rendering) is the downwind loss of fluid velocity over the cross section of the rotor - the rotor harvests impulse from the moving air.
In this video we used a high-fidelity LES approach. In our experience this consistently achieves results with consistently higher quality compared to other, less involved simulation approaches. Additionally, we included generated turbulence at the inlet (incoming "blobs") to demonstrate that this kind of simulation can also be done with gusts.
Gusts can also be generated on top of arbitrary vertical velocity profiles. Especially for onshore turbines we do use a number of precursor simulations to determine this inlet profile. For example: turbines are often located on top of hills / ridges, however in such situations, the wind rarely approaches horizontally, but has a distinct upwards component, that can drastically change the load on blades, tower and transmission. Therefore, using a progression from coarse meteorological simulation as input for a better resolved RANS which in turn is the input for a LES gives a much better impression of the real flow condition around the turbine. Factoring in this information into a planning process, allows tailored engineering of the turbine for the spot it is installed in.
Download video: HD ready (720p) | Full HD (1080p)
Compare this with the simpler approach of a k-epsilon simulation without generated turbulence at the inlet in the video below. While the overall picture is similar to the LES above, there are significant differences in the turbulence around the blades, with the LES showing a picture much closer to reality.
In a large scale PIV (particle image velocimetry) experiment at the University of Minnesota all these phenomena were demonstrated using a industrial scale wind turbine. Details about this can be found in this Paper titled "Natural snowfall reveals large-scale flow structures in the wake of a 2.5-MW wind turbine" published in Nature. A nice video about the experiment showing the vortices can be found on the dedicated homepage about Super-Large-Scale Flow Visualizations of the University of Minnesota. Compare the video of the actual phenomenon to the slow motion animation of our our simulation (see below). Even though the conditions of the experiment and simulation have not been synchronized (we only became aware of the experiment after completion of our simulation), qualitatively, the resulting images look remarkably similar.