FAST.Farm

Wind & Water General
61

Dear Jiaping,

Here are my answers to your questions:

  1. The sample FAST.Farm model provided by Kelsey includes an OpenFAST model formatted for OpenFAST v2.3. The precompiled pre-release of FAST.Farm was compiled with OpenFAST v1.0.

  2. The current version of FAST.Farm in github.com/jjonkman/OpenFAST/tree/f/fast-farm is linked to OpenFAST v2.3. So, you should be able to use Kelsey’s sample FAST.Farm model directly after compiling this version.

  3. The format of the FAST.Farm input file has not changed between the pre-release and the current version in GitHub.

Best regards,

62

Dear Jason,

I have been working on the pre-release version of FAST.Farm that you sent me via e-mail. Last week I managed to use, with this version, the turbsim example that Kelsey shared in this forum. After that I was traying to generate my own wind grid but, since I have the turbsim’s version 1.4, I haven’t been able to build a large enough wind grid needed for FAST.farm (memory allocating errors). Is it possible to get an updated version of Turbsim? In that case, where should apply for it?

Furthermore, my team and I have been producing, in the last years, many precursors with different wind characteristics with our CFD, that we use to simulate wind farms (LES-ALM). I wish to compare our CFD results with FAST.farm, but we only have information from these precursors stored in 2D planes, since we give this planes as input to our wind farm LES simulations. So, I would try to give these 2D LES-generated wind grids to FAST.farm as input, instead of the whole (3D) wind field information. What do you think ? Have you already tried this?
I guess that, in order to acomplish this, I should set the WindType to 6 (User defined) in the “InflowWind.dat” file. But I’m not sure how to define in FAST.farm the structure of my wind data. Could you instruct me?

Many thanks!
Kind Regards,

Bruno López

63

Dear Bruno,

TurbSim v2.0 is available on my Google Drive: drive.google.com/drive/folders/ … sp=sharing. And there is a slightly newer version of Turbim, based on v2.0, included in the OpenFAST repository in github.

If you already have LES-generated planes of 2D+time ambient wind data that you wish to use in FAST.Farm, you could try to read the data into InflowWind using WindType = 6. This will require that you write own logic for processing your wind data within the IfW_UserWind.f90 source file and recompiling FAST.Farm. Alternatively, and perhaps preferably, you could convert your 2D+time wind data into a wind format that InflowWind already makes use of, e.g. TurbSim format.

Best regards,

64

Dear Jason,

Many thanks for your quick reply. I could successfully download and test TurbSim v2.0.

Regarding the 2D LES wind data, I agree that it’s more convenient to convert it into a wind format that it’s already being read. I had already thought about it, but since all these formats are written as binary files I was not sure how to do the conversion. I could see that in the folder you shared via Google Drive there is a matlab script to read bts files. I think I could use this script along with the information provided in Annex D of TurbSim user’s guide to write a conversion script from my data into bts format. In case you have any suggestion, please let me know.

Thanks again!
Best regards,

Bruno

65

Dear Bruno,

Yes, I agree.

I’m not aware of a MATLAB-based writer of TurbSim binary (*.bts) formatted wind data files, but there is a MATLAB-based reader, readfile_BTS.m, in the MATLAB Toolbox: github.com/OpenFAST/matlab-tool … /Utilities. I’m sure with a bit of work, this script could be updated to become a writer. Of course, there is a Fortran-based writer inherent within the TurbSim source code.

Searching this forum, I also recall a user posted a MATLAB-based writer for Bladed binary wind (*.wnd) formatted wind data files–see: binary .wnd specification - #7 by Simon.Fobe.

Best regards,

66

Dear Jason,

Thanks a lot for your help. You’re right, I can also look the fortran source code of turbsim for bts writing. In case I come up with a Matlab script for bts writing I can also share the code in this forum.

Best regards,
Bruno

67

Dear Jason,

Would you please share the pre-release version of the FAST.Farm software package?

Many thanks in advance. The attention you pay to the forum is very helpful!

Best regards,

Nicolás

68

Sent.

69

Dear Jason,

With your help I can run the FAST.Farm inflow=2 case successfully, but in this case, the turbine model is NREL-5MW. When I switch the model to the IEA-15MW, it appears some errors, I have try to modify some parameters in the primary input .fstf file for many times but always break down. So I want to search a help.The situation as followings:

F:\Turbsim_test\FAST.Farm_Turbsim_test_15MW>bin\FAST.Farm_x64.exe F:\Turbsim_test\FAST.Farm_Turbsim_test_15MW\TSinflow.fstf

FAST.Farm

Copyright (C) 2017 National Renewable Energy Laboratory
Copyright (C) 2017 Envision Energy USA LTD

This program is licensed under Apache License Version 2.0 and comes with ABSOLUTELY NO WARRANTY.
See the “LICENSE” file distributed with this software for details.

FAST.Farm-Unversioned from aad78b53445694288e03eb2f7bb8274a73a2eb3c
Compile Info:

  • Architecture: 64 bit
  • Precision: single
  • Date: Oct 08 2020
  • Time: 23:05:36
    Execution Info:
  • Date: 10/17/2020
  • Time: 13:59:25+0800

Heading of the FAST.Farm input file:
Sample FAST.Farm input file
Running AWAE.
Running InflowWind.
Running WakeDynamics.
Running FASTWrapper.
Running OpenFAST.
OpenFAST input file heading:
IEA 15 MW offshore reference model monopile configuration

Running ElastoDyn.
Running AeroDyn.
Running BEM.
Running UnsteadyAero.
Warning: Turning off Unsteady Aerodynamics because C_nalpha is 0. BladeNode = 1, Blade = 1
Warning: Turning off Unsteady Aerodynamics because C_nalpha is 0. BladeNode = 2, Blade = 1
Warning: Turning off Unsteady Aerodynamics because C_nalpha is 0. BladeNode = 1, Blade = 2
Warning: Turning off Unsteady Aerodynamics because C_nalpha is 0. BladeNode = 2, Blade = 2
Warning: Turning off Unsteady Aerodynamics because C_nalpha is 0. BladeNode = 1, Blade = 3
Warning: Turning off Unsteady Aerodynamics because C_nalpha is 0. BladeNode = 2, Blade = 3
Running InflowWind.
Running ServoDyn.
Running ServoDyn Interface for Bladed Controllers.
Running HydroDyn.
Generating incident wave kinematics and current time history.
Time: 0 of 400 seconds.

FARM_InitialCO:AWAE_CalcOutput:LowResGridCalcOutput:The center of a wake plane for turbine #1 has
passed the lowest-most Z boundary of the low-resolution domain.
T1:FARM_InitialCO:FWrap_t0:FAST_Solution0:CalcOutputs_And_SolveForInputs:SolveOption2:SolveOption2
c_Inp2AD_SrvD:InflowWind_CalcOutput:CalcOutput:IfW_4Dext_CalcOutput [position=(-26.436, -90.875,
98.864) in wind-file coordinates]:Interp4D:Outside the grid bounds.
SolveOption2:SrvD_CalcOutput:

Running a controller implemented through NREL’s ROSCO Toolbox
A wind turbine controller framework for public use in the scientific field
Developed in collaboration: National Renewable Energy Laboratory
Delft University of Technology, The Netherlands
Primary development by (listed alphabetically): Nikhar J. Abbas

Aborting FAST.Farm.

And in the primary input .fstf file, the parameters of OutList haven’t a standard introduce or explanation, so where can I find a explanation for these paratemers?

Else, I want to know wheather having a example that including the super controller flag is true and the Mod_AmbWind=2, if having ,can you send one to me ?

Best regards,

Jiaping.Cui

70

Dear Jason,

I have read the comparison of FAST.Farm simulations against LES. However, there are no quantitative results presented in the comparison. The conclusion is that FAST.Farm predicts “reasonably well” generator power and torque.
Do you know if I can access quantitative results from the comparison? Or if there are similar validations with numerical results?
I have searched for publications, but have not found anything. I am interested in power prediction.

I have also read “Validation of FAST.Farm considering structural loads at alpha ventus”, but no numerical results are presented either.

Many thanks in advance,

Best regards,

Nicolás

71

Dear Nicolás,

Can you clarify what you mean by “quantitative results”? What exactly do you want to see?

Best regards,

72

Dear Jason,

When running FAST.Farm, I used the IEA-15MW wind turbine model([url]https://github.com/IEAWindTask37/IEA-15-240-RWT/tree/openfast_dev[/url]), instead of 5-MW model. Because of the changing of the size of the two kinds of wind turbine model, I modify the size of low-resolution wind domain and high-resolution wind domains. But the running process will have some errors, so I want to konw if having some notes when modifying the parameters of low and high-resolution wind domains.

And I don’t know the significance of the parameters in the primary input .fstf file. Can you tell me where I could find the introduce about these parameters ? The parameters of Outlist as followings:

“RtAxsXT1 , RtAxsYT1 , RtAxsZT1”
“RtPosXT1 , RtPosYT1 , RtPosZT1”
“YawErrT1”
“TIAmbT1”
“CtT1N01 , CtT1N02 , CtT1N03 , CtT1N04 , CtT1N05 , CtT1N06 , CtT1N07 , CtT1N08 , CtT1N09 , CtT1N10 , CtT1N11 , CtT1N12 , CtT1N13 , CtT1N14 , CtT1N15 , CtT1N16 , CtT1N17 , CtT1N18 , CtT1N19 , CtT1N20”
“WkAxsXT1D1 , WkAxsXT1D2 , WkAxsXT1D3 , WkAxsXT1D4 , WkAxsXT1D5 , WkAxsXT1D6 , WkAxsXT1D7”
“WkAxsYT1D1 , WkAxsYT1D2 , WkAxsYT1D3 , WkAxsYT1D4 , WkAxsYT1D5 , WkAxsYT1D6 , WkAxsYT1D7”
“WkAxsZT1D1 , WkAxsZT1D2 , WkAxsZT1D3 , WkAxsZT1D4 , WkAxsZT1D5 , WkAxsZT1D6 , WkAxsZT1D7”
“WkPosXT1D1 , WkPosXT1D2 , WkPosXT1D3 , WkPosXT1D4 , WkPosXT1D5 , WkPosXT1D6 , WkPosXT1D7”
“WkPosYT1D1 , WkPosYT1D2 , WkPosYT1D3 , WkPosYT1D4 , WkPosYT1D5 , WkPosYT1D6 , WkPosYT1D7”
“WkPosZT1D1 , WkPosZT1D2 , WkPosZT1D3 , WkPosZT1D4 , WkPosZT1D5 , WkPosZT1D6 , WkPosZT1D7”
…

Best regards,

Jiaping.Cui

73

Dear Jason,

Sorry, I need to add one problem. There is one inflowfile in FAST.Farm and there is one inflowfile in OpenFAST, so I want to know wheather the two kinds of inflowfile will have a conflict when I run the FAST.Farm Executable. If they have a conflict, what I need to do to solve this conflict？

Best regards,

Jiaping.Cui

74

Dear Jiaping.Cui,

Just a few comments:

  • The error you are receiving when running FAST.Farm is the second message written–see quoted section below. From the error, it sounds like your high-resolution domain is not big enough to fit the IEA Wind 15-MW reference wind turbine you are simulating.
  • We have been developing modeling guidance for running FAST.Farm, which includes guidance for creating low- and high-resolution domains of appropriate size and spatial-temporal discretization. Which version of the draft FAST.Farm User’s Guide and Theory Manual are you using? I’ll send you the newest draft via e-mail, but we are still working on finalizing this document before official publication.
  • The OutList parameters of FAST.Farm are described in Appendix C of the draft FAST.Farm User’s Guide and Theory Manual.
  • We are still finalizing the implementation of the super controller, and so, we do not have have a sample model to share.
  • As mentioned in the draft FAST.Farm User’s Guide and Theory Manual, the InflowWind module input parameters WindType and PropogationDir are ignored in the InflowWind input file set within the OpenFAST model used by FAST.Farm.

I hope that helps.

Best regards,

75

I refer to some measure of the differences between time-averaged power outputs between simulations in FAST.Farm and SOWFA (for example, Root Mean Squared Relative Error).

Best regards,

76

Dear Deza,

A paper we wrote titled, “FAST.Farm Development and Validation of Structural Load Prediction Against Large Eddy Simulations” was recently accepted for publication in the Wind Energy journal. Included in that paper are quantitative comparisons between FAST.Farm and high-fidelity coupled SOWFA-OpenFAST simulations in terms of percent error of mean and standard deviation values for the structural outputs (deflections and loads) of several wind turbines. I gather this is what you are looking for. The publishing process is underway and should be available soon.

Best regards,

77

Dear Jason,

Could you please send to me the pre-release of the fast.farm software together with its documentation? Do you have an approximate date for the official release?

Thanks a lot,

Leire

78

Sent.

We are working on the official release of FAST.Farm and expect it to be soon, but I don’t want to state a date because I’ve been wrong so many times before.

Best regards,

79

Dear Jason,

Thank you for your explanation patiently, that solves many my doubts.

I scan the FAST.Farm User’s Guide and Theory Manual, and I found that the primary input .fstf file have a few changes, so if the pre-version FAST.Farm have a update ? If have, I hope I can get the lastest pre-version and better also the lastest sample and Executable on windows.

In the FAST.Farm primary input .fstf file, I found the column of SUPER CONTROLLER only have the location of the dynamic library{ .dll [Windows] or .so [Linux] }, and it haven’t a user-defined way. So can I think that there hanen’t a interface to other softwarea like Matlab ? And do you have some material that using FLORIS to develop the super controller?

Best regards,

Jiaping.Cui

80

Dear Jiaping,

The primary FAST.Farm input (*.fstf) file has not changed since the version of the precompiled Windows executable and the newer version on github, other than the addition of the Mod_AmbWind = 3 option. The Mod_AmbWind = 3 option does not change the formatting of the file (it does not add or remove lines), but it does allow the 3rd option of Mod_AmbWind, which changes some of the line comments. NREL does not intend to release other pre-compiled versions of the FAST.Farm Windows executable until the official release of FAST.Farm.

The super controller implementation being finalized now only supports a super controller, together with individual controllers, defined in dynamic library format (that is, the DISCON.dll or DISCON.so for the individual turbine controllers). This was done so as to ensure that the super controller implementation in FAST.Farm is consistent with the super controller interface available in the C++ API for interfacing OpenFAST with CFD (e.g., SOWFA). NREL does not currently have funding to develop a Simulink interface for the super controller.

FLORIS is a steady state tool, and so, can be used to derive optimal mean nacelle-yaw angles (to optimize power capture via wake steering) for a given steady wind condition. But FLORIS has no dynamics, and so, cannot be used to develop dynamic (time-varying) implementations of controllers. More information is available in the many FLORIS publications.

Best regards,