FAST.Farm: Time Step, Wake Model & Wind-Wave DLC Setup

[6Mr-Y]

Hello everyone,
I recently read discussions about TurbSim time step settings in the community and have several questions regarding my FAST.Farm configuration.
My computational domain dimensions (where D stands for rotor diameter):
High-resolution domain: X = 720 m (3D), Y & Z = 360 m (1.5D)
Low-resolution domain: X = 4800 m (20D), Y = 1200 m (5D), Z = 600 m (2.5D)
In the .fstf file, the high-resolution time step is 0.1 s and the low-resolution time step is 1 s, with Mod_AmbWind = 2. The TurbSim time step is set to 0.05 s, which is the general recommended value from the user manual.
I have been using this configuration for all my simulations. I believe the settings on the FAST.Farm side provide decent accuracy, but I am concerned about time step mismatch between the wind field and FAST.Farm. Could you please advise whether this combination of time steps will significantly affect the simulation results?
Besides, I noticed that Mod_AmbWind = 3 requires a dedicated high-resolution wind field for each turbine plus a global low-resolution wind field. This setup seems inconsistent with the characteristics of real natural wind fields. I would also like to know the proper application scenarios for this model.

My second question is about the wake model. I use Mod_Wake = 2 (Curl wake model) in .fstf. Except for RotorDiamRef, dr, NumRadii and NumPlanes, all other parameters under Wake Dynamics and Curled-Wake Parameters are kept as DEFAULT. After checking the user manual, I did not find any mandatory parameters to adjust. Could you confirm if the default parameters are sufficient to meet the accuracy requirements for general simulations?

My last question relates to Design Load Cases (DLC) and wind-wave conditions. I cannot find any built-in scripts for automatic DLC generation in the OpenFAST Toolbox. I need to perform combined wind and wave load simulations for wind farms. Given a target wind speed and wave height, I am not sure how to define:
Meteorological boundary conditions and grid settings in the TurbSim input file;
Wave parameters in the SeaState input file.
Is there any available code or tool to automatically generate these wind and wave condition parameters? If not, could you recommend relevant standards or technical references for defining operating conditions in wind farm simulations?
Thank you very much for your help!

[jjonkman]

Dear @6Mr-Y

A TurbSim TimeStep of 0.05 s is not needed for the IEA Wind 15-MW RWT (which is what I presume you are simulating) and certainly does not need to be smaller than DT_High in FAST.Farm; I recommend increasing TurbSim's TimeStep to match DT_High, or 0.1 s in your case.

When generating ambient wind inflow with TurbSim for use in FAST.Farm, we generally recommend using Mod_AmbWind = 3. When you follow the FAST.Farm modeling guidance, any mismatch between the ambient wind in the low- and high-resolution domains is minimized by constraining the high-resolution domain to match the time-series of the low-resolution domain at the hub height location.

Just a few comments on your spatial dimensions:

• The high-resolution domain must encompass the turbine and its structural displacements, but an X-range of 3D sounds larger than it needs to be.
• To support wake meandering, we recommend having the Y and Z boundaries of the low-resolution domain to extend to 3D to the left, right, and above the hub, so, your Y-range and Z-range do not seem large enough.

If you are using the scripts in the OpenFAST toolbox to define the spatial-temporal discretizations of FAST.Farm and TurbSim, this process will be straightforward and all FAST.Farm modeling guidance will be followed.

The DEFAULT wake parameters are meant to be a good starting point, but if you have wake results from a higher-fidelity model (such as Large-Eddy Simulation), you'll likely obtain a better solution from FAST.Farm if you deviate from the DEFAULT values by calibrating against the higher-fidelity solution. For the IEA Wind 15-MW RWT, the curled wake parameters were calibrated against LES under neutral atmospheric boundary layer conditions for a range of yaw errors in the following paper: https://wes.copernicus.org/articles/10/1033/2025/.

For running large number of FAST.Farm simulations across a range of wind and wave conditions, we've typically customized the OpenFAST Toolbox for our specific application.

Best regards,

[6Mr-Y]

Thank you very much for your patient guidance @jjonkman . I have come across formulas for turbulence intensity in papers, but I cannot find the relevant standards or theoretical references for them. I intend to set different turbulence intensity values for varying wind speeds and would appreciate it if you could point me to supporting documents.

Besides, I noticed that the wind-wave combinations adopted in some publications do not match the conditions listed in Table 12. I am confused about how to reasonably define such wind and wave operating conditions.

I also have a question about FAST.Farm: does it support simulations with two different turbine models in one wind farm, for example, NREL 5MW and IEA 15MW turbines?Thank you again for all your support. Your guidance has been extremely helpful to my work.

[jjonkman]

Dear @6Mr-Y,

The IEC wind turbine classes are defined in the IEC 61400-1 design standard. Your description sounds like a paraphrase of the IEC. For category "B", I_ref is 0.14, rather than 0.12 for category "C".

Regarding your description of wave data, again wave data is typically chosen site specifically, where each site differs. So, I wouldn't be surprised to find different wave data in different publications.

You can consider different wind turbines in the same OpenFAST model, but we do plan at some point to improve FAST.Farm for this application, e.g., changing RotorDiamRef to be turbine-specific. Until then, I would set RotorDiamRef in FAST.Farm equal to the largest diameter of the wind turbines you have, which does compromise a few calculations for the smaller turbines.

Best regards,

[6Mr-Y]

Thank you for your guidance @jjonkman . So it would be more reasonable to define wind and wave conditions based on a specific site, right? Could you tell me where I can obtain wind and wave data for typical sites?Regarding the wake settings in FAST.Farm: If my wind farm consists of both NREL 5MW and IEA 15MW turbines, and I configure all parameters under WAKE DYNAMICS based on the IEA 15MW turbine, will this lead to inaccurate results for the NREL 5MW turbine? Is my understanding correct?I am currently running simulations with mixed turbine models. If this approach greatly compromises the accuracy, are there alternative tools suitable for wind farm simulations with different turbine types? For instance, does SOWFA support such cases?Thank you very much again for your help!

[jjonkman]

Dear @6Mr-Y,

The source of good wind and wave data likely depends on the specific site you have in mind. I gathered from your earlier comments that you don't have a good site in mind. For U.S.-based metocean conditions, the following paper we published a few years ago is a likely a good reference: https://onlinelibrary.wiley.com/doi/full/10.1002/we.1881.

I believe FAST.Farm should still be accurate for wind farms with turbines of different sizes. RotorDiamRef is only used in a few calculations, e.g., in sanity checks and to ensure wakes propagate the desired distance downstream, which is why I would use the larger of the diameters.

Best regards,

[rthedin]

I still have a few points I’d like to clarify with you: When using Mod_AmbWind = 3, how should I configure the settings to ensure that the time series at hub height in the high-resolution domain matches that in the low-resolution domain? You mentioned a script in the OpenFAST Toolbox for defining spatial-temporal discretization parameters for FAST.Farm and TurbSim. Could you confirm whether this refers to the following two files in openfast_toolbox/fastfarm/: Ex1_FASTFarm_discretization.py Ex2a_FASTFarm_TurbSim_driven.py I have reviewed all example scripts, but I am not sure if these are the ones you meant. If not, could you please specify which one to use?

The openfast_toolbox for FAST.Farm has all of these steps automated. You should run the low-resolution turbulence box, and the toolbox will automatically open the low-res low-resolution data, get the time series at the correct location, shift it in time depending on where in the domain the turbine is, write to an input file, and write the TurbSim input file for the high-resolution box(es) that will take the time-series input file into account. This procedure is illustrated on the Ex2a_FASTFarm_TurbSim_driven.py example.