Curl wake not working as it should

[tomblanks1]

Hi. I am working with the OC4 semi-sub model in fast farm and am looking how the wakes change and develop down the field. As you can see my curl is recovering slower than my polar which i find strange and there is no kidney shape wake as seen in papers you have published. Are these correct? If this isn't what you expect what may be causing this. My input is a 11.4avg Turbsim at ref height 90 and all i'm doing is changing mod wake from 1 to 2 and not changing anything else, is that the wrong method? best wishes and thanks

[jjonkman]

Dear @tomblanks1.

Can you clarify what you are plotting? Are these time-averaged flow fields normalized by the time-averaged wind speed at a given reference height?

Regarding the "kidney shape", this will show up in the curled wake model when there is skewed flow. Can you clarify what skew you have? Do you have mean yaw misalignment between the nacelle orientation and wind direction, or is a skew coming from the shaft tilt and platform-pitch?

The wake recovery in the curled wake model will be determined by FAST.Farm inputs k_VortexDecay and k_vCurl. I assume you are simply using the DEFAULT values for these inputs, but you could update these if want the polar and curled wake solutions to give more consistent results for a given case.

Best regards,

[tomblanks1]

thanks @jjonkman

Yes I was trying to plot the time-average flow wind speeds around the hub height. I believe i currently have these skew settings ShftTilt -5 NacYaw 15 PtfmPitch 1.9 Are these what you were talking about? If not could you point me in the correct direction please

As for those curl and vortex delay, yes i was running the simulations with default settings, again could you please suggest updated values for my more basic floating case?

Thanks

[jjonkman]

Dear @tomblanks1,

With background turbulence and resulting wake meandering, it may not be possible to see "kidney shape" of the curled wake in the time-averaged fixed frame of reference. This would be easier to see without inflow turbulence or in the meandering frame of reference, and then, only if there is enough skew to form the kidney shape. It sounds like you have 15-degrees of yaw and 6.9 degrees of tilt (on average), which I would expect to be sufficient to see the kidney shape, but due to the combination of yaw ant tilt, it will be skewed to one corner (not left/right or top/bottom).

The curled wake settings we applied when modeling the NREL 5-MW baseline turbine atop the OC4-DeepCwind semisumbersible are explained in this paper: https://wes.copernicus.org/articles/9/1827/2024/wes-9-1827-2024.pdf.

Best regards,

[tomblanks1]

Thanks @jjonkman I adjusted the cut off frequency as you describe in the paper and removed the turbulence, and can now see the kidney shape and surroundings better as expected. I was wondering why the curl still has a lower recovery compared to polar as from theory i believe this is not expected. Are there any other parameters that may be causing that lack of wake regeneration in the curl model ?

[jjonkman]

Dear @tomblanks1,

I'm glad that you can now see the "kidney shape" in the curled wake solution.

Wake diffusion can differ a bit between the polar and curled wake models because the underlying formulation is different, including finite-difference (polar) versus Euler (curl) solution of the x-momentum equations and the lack of a continuity solve in the curled wake model. FAST.Farm input k_vCurl was added as a tuning parameter so that the diffusion of the curled wake model can be adjusted to match that of the polar model without changing other eddy viscosity inputs. This is discussed a bit in our paper describing the introduction of the curled wake model in FAST.Farm: https://onlinelibrary.wiley.com/doi/full/10.1002/we.2785.

Best regards,