FAST theory manual

[Juan.LopezMuro]

Dear Jason,

Thank you so much for your answers.

1. I did receive the “Unofficial FAST Theory Manual”. Thank you very much. Anyway, as a new beginner of OpenFAST, I want to follow your advice and review your last modeling workshop. Unfortunately, the link in your Sep 15, 2014 post in the forum topic is broken.
   1.a. Could you please provide me with a new link?
   1.b. Also, is there a website/repository where I could find the user’s guides and theory manuals?

2. Ok, thanks. Now I am using both (as you will notice in the next point), but I will try to use Ubuntu in the future.

3. I am showing the OpenFAST output from an Ubuntu virtual machine, a guest in a Windows host (I am using Oracle VirtualBox). I created this Ubuntu virtual machine to try the installation process before making it in my “real” Ubuntu system. Since I was running into the error mentioned above, I also downloaded the precompiled binaries from the GitHub repo to Windows. So, I have the same setup, and I was finally running into the same error in both OS. I hope this makes sense.
   Now, I followed your instructions, and I did add the DISCON.dll file you provided to the path indicated in the “NRELOffshrBsline5MW_Onshore_ServoDyn.dat” file. That solved the problem in Windows, but the problem persists in Ubuntu.
   Which file should I add to my Ubuntu machine to fix the problem and make it work?

Many thanks,
Juan

[FuHang.Gan]

Dear Jason and Juan,

I think I met the same problem as Juan. I am sure the .dll file is in the appropriate path. Look at sreenshot 1.
![Y7YJQO}@(UDMAZH9PQTFV6.png|1069x1433
And I found that the .fst file without using bladed-style DLL can be execute by openfast on Linux. Maybe, the .dll file can not be execute on Linux and the needed type is .so file. Look at screenshot 2.![K1]2[NSJP1$J21Y436A_BK.png|1713x58
However I am also a beginner in Linux. So, I want to find some help from professors here.

I thanks any help and suggestion from here!
Fuhang.

[Jason.Jonkman]

Nikolas – I’ve now sent you the “Unofficial FAST Theory Manual”.

Best regards,

[Jason.Jonkman]

Dear Juan,

Here are my answers to your questions.

1a) As mentioned on our forum here: Other News, the old NWTC Information Portal has been decommissioned; the most important files from there have been moved to my Google Drive. The modeling workshop materials are now located here: drive.google.com/drive/folders/ … sp=sharing.

1b) The User’s Guides and Theory Manuals previously published on the NWTC Information Portal are now available here: github.com/OpenFAST/openfast/pull/805.

3. On Linux, I recommend the Conda Installation approach documented here: openfast.readthedocs.io/en/main … index.html.

Best regards,

[Jason.Jonkman]

Dear Futtang,

A dynamic library has a *.dll extension on Windows and a *.so extension on Mac and Linux. Again, on Linux, I recommend the Conda Installation approach documented here: openfast.readthedocs.io/en/main … index.html.

Best regards,

[Hauke.Juergensen]

Dear Jason,

I am looking for more Information about the Theory behind the ElastoDyn Module. Could you send me a copy of the unofficial Theory Manual to my Mail at hauke.juergensen@campus.tu-berlin.de?

Best regards,
Hauke

[Roberto.Echeverria]

Dear Jason,

would it be possible to receive a copy of the unofficial Theory Manual?

My email address is roberto.echeverriadelgado@2benergy.com

Thanks in advance and kind regards

Roberto

[FuHang.Gan]

@Jason.Jonkman

Dear Futtang,

A dynamic library has a *.dll extension on Windows and a *.so extension on Mac and Linux. Again, on Linux, I recommend the Conda Installation approach documented here: openfast.readthedocs.io/en/main … index.html.

Best regards,

Hey, Jason
Maybe it is because I don’t know the custom in USA. I think you could be more careful. Using wrong name is not polite . My name is Fuhang! If you think it is not convenient, you can call me Gan.

I have done the installation by conda before my try. But there is no information about producing *.so extension in the document you mentioned and it is only to teach how to install in Linux, compile from souce code, and test. Maybe this is a very low level question, I will continue to find the way to solve it.

Anywhere I wrong, point out directly.

Thanks.Best wish.

[Jason.Jonkman]

Dear Hauke – I’ve sent you the “Unofficial FAST Theory Manual”.

Best regards,

[Jason.Jonkman]

Dear FuHang,

Sorry for typing your name wrong; I simply misread it.

OK, I’m not too familiar with the installing OpenFAST on Linux (I use Windows myself) and I thought that the Conda install did the trick. I’ll ask Rafael Mudafort to respond.

Best regards,

[FuHang.Gan]

@Jason.Jonkman

Dear FuHang,

Sorry for typing your name wrong; I simply misread it.

OK, I’m not too familiar with the installing OpenFAST on Linux (I use Windows myself) and I thought that the Conda install did the trick. I’ll ask Rafael Mudafort to respond.

Best regards,

That is all right!
Solving problem matters much more.
Just make the software more usable and convenient in wind energy research.

Best regards,
Fuhang

[Rafael.Mudafort]

Hi Fuhang,

The controller libraries are not currently distributed through Conda. However, on linux you can easily compile it with free and open source tools.

Are you running your test case within the regression test suite? If so, see the documentation for that on how to get the controller libraries into the correct location: openfast.readthedocs.io/en/main … _test.html.

Thanks
Rafael

[Rafael.Mudafort]

@Juan.LopezMuro

Dear Jason,

2. I have access to Windows and Linux/Ubuntu. In the near future, I aim to use the Simulink interface to use the ROSCO controller and to test different custom controller designs. Do you suggest using one operating system over the other (e.g. Linux over Windows)? Will I need to compile the source code in order to test custom controller designs?

Hi Juan,

You’ve probably seen this already, but just in case I wanted to point out that the DISCON libraries for ROSCO are slightly different than the ones we have within the testing suite for OpenFAST. The ROSCO controller libraries are available on that project’s GitHub Releases page: github.com/nrel/rosco/releases.

Thanks
Rafael

[Juan.LopezMuro]

You’ve probably seen this already(…)

Hi Rafael,
Thank you so much for jumping in.

Honestly, I am still struggling to make it work.

I did the conda install in Ubuntu. Next, I wanted to go through the testing of OpenFAST (r-test and pFUnit), but I could not find straightforward documentation on how to do that.

So I started by cloning the whole repository (git clone --recursive openfast.readthedocs.io/en/main … executable )

After, reading this GitHub “NOTE: External ServoDyn controllers for 5MW_Baseline cases” ( github.com/openfast/r-test#note … line-cases ) I understood that I had to compile the controllers, so I did create build directories at
_____r-test/glue-codes/openfast/5MW_Baseline/ServoData/DISCON/build
_____r-test/glue-codes/openfast/5MW_Baseline/ServoData/DISCON_ITI/build
_____r-test/glue-codes/openfast/5MW_Baseline/ServoData/DISCON_OC3/build
_____and run cmake … in each one

Next, I did run make in each build directory and obtained a DISCON.dll file witch I cut and pasted to the ServoData directory.

Finally, I went back to the “3.2.2.4.3. Python driver with a given executable” document, where I understood that to run the tests; I needed to run
python manualRegressionTest.py ~/miniconda3/envs/openfast_env/openfast Linux GNU 1e-5

As you can see in the following Figure, I tried running a specific case with apparent success.

Honestly, I am a bit surprised that the case is running, given that Jason pointed out that

A dynamic library has a *.dll extension on Windows and a *.so extension on Mac and Linux.

and in this case, I am using a *.dll extension on Linux.

Furthermore, I tried to run the r-test, but all of them FAILED, as you can see in the Figure below.

I believe that I am NOT doing it correctly.

Could you please give me some directions on how to do it right?

I genuinely thank you in advance,
Juan

[Rafael.Mudafort]

Hi Juan,

Good work in getting to this point. There are some things I want to point out that should add some clarity.

First, the tests and accompanying documentation are most relevant to people who have compiled OpenFAST themselves. More specifically, they are useful when someone is making changes to the code and compiling because they allow developers to (somewhat) quickly know whether something has been changed unintentionally. In the case of installing OpenFAST with conda, you can be sure that your version is fine as it is well tested before its distributed with conda.

And on the “dll” vs “so”, you’re right that “.dll” are for Windows and “.so” are for Linux. However, those extensions are part of the name of the file and do not determine the contents of the file. For the regression test suite, we simply change the name of the DISCON.so to DISCON.dll after it is compiled so that we can use the same input files (specifically the ServoDyn files) for all systems. Here’s how its done in CMake: github.com/OpenFAST/r-test/blob … ts.txt#L90.

Now about the manual regression test script, try adding the verbose flag to that command, “-v”. It’s possible that you have not installed the dependencies for Python that are required to run the tests. And just to reiterate, you can trust your conda installation so getting these tests to pass is not a requirement. In fact, the results for these tests are stored in double precision and the conda installation is compiled in single precision, so a lot of these tests will not pass when the results are compared even if the do execute on your system.

FWIW, there is a clear gap in the testing documentation about all of this, so thanks for your persistence and questions. I’ll work on updating the documentation.

[James.Dilworth]

Hello Jason,

Please could you send a copy of the
“Unofficial FAST Theory Manual” to james.dilworth@gmail.com

Thanks very much,
James

[Jason.Jonkman]

Sent.

[Cagatay.Topal]

Hi Jason,

Please could you send a copy of the
“Unofficial FAST Theory Manual” to cagatay@cgtytpl.com

Thanks
Cagatay
[url]https://www.goksanmakina.com/urunler/disk-yaylar[/url]
https://www.kilitlipulx.com
https://www.tieig.com

[Jason.Jonkman]

Sent.

[Warit.Chanprasert]

Dear Jason
If it is posible, could you please send me the unofficial Elastodyn theory manual?
Email: wcha533@aucklanduni.ac.nz.

Thank you very much
Warit