Hi @Jason.Jonkman:
I am using OpenFAST v2.5.0, I used it because when I carried out linearization I got pitch and roll natural frequencies that compared well with NREL’s. When I used v3.4.1 my pitch and roll natural frequencies were too high.
Regards,
AOAW
Dear @Andre.White,
So, with OpenFAST v2.5.0, are you using the NREL-provided r-test model of the NREL 5-MW baseline wind turbine atop the OC4-DeepCwind semisubmersible? And for OpenFAST v3.4.1, are you using the NREL-provided r-test model of the NREL 5-MW baseline wind turbine atop the OC4-DeepCwind semisubmersible? Or did you convert one of models from one version to the other yourself? I would expect both versions to give similar pitch and roll natural frequencies that match the paper.
Best regards,
Hi @Jason.Jonkman:
Yes for both models I am using the NREL-provided r-test model of the NREL 5-MW baseline wind turbine atop the OC4-DeepCwind semisubmersible. Correction to what I said earlier, the results I sent was for OpenFAST v3.4.1. Unfortunately, I do no get same pitch and roll frequencies for both versions.
I also take it that you have concluded that the previous plots did not compare well with NREL’s.
Regards,
AOAW
Hi @Jason.Jonkman:
I just did the surge free decay test with v2.5.0. The plots are the same as v3.4.1.
Regards,
AOAW
Hi @Andre.White,
As I mentioned in my posts dated Mar 2 and 7, 2023 above, I would expect some differences in results between the different versions, but changes to natural frequencies (other than minor) I would not expect.
Can you share you full set of free-decay results for the NREL 5-MW baseline wind turbine atop the OC4-DeepCwind semisubmersible from OpenFAST v2.5.0, OpenFAST v3.4.1, and original published results from NREL (using FAST v8)?
Best regards,
Hi @Jason.Jonkman is there a way to share the file as a txt file. The only option that I see for uploading data is an image format.
Regards,
AOAW
Dear @Andre.White,
This new forum is limited in what file types can be attached. But you can always link to a file-sharing site such as Google Docs or github.
Best regards,
Hi @Jason.Jonkman thank you.
I believe you will find the files here https://github.com/andreowhite/andresfiles.
The files are:
Please check and let me know if you see anything wrong with results as I still trying to finalize this.
Regards,
AOAW
Dear @Andre.White,
The surge results look quite reasonable to me–the responses are nearly identical between OpenFAST v2.5 and v3.4.1. Do you have the results for other free decay responses?
Best regards,
Hi @Jason.Jonkman:
I do not have the heave free decay for v2.5.0, when I saw that the surge results were not good I held off on doing heave free decay to see if I could solve the problem. However, I think for the heave freedecay test the heave motion is good and for the surge freedecay test the surge motion also shows good agreement with NREL. It is the other plots, for example for heave, the coupled surge and pitch plots do no show good agreement. Similar for surge, the coupled heave and pitch plots do not show good agreement with NREL data. You can look again at the plots I sent to you earlier.
I would have normally only compared the surge free decay results with the surge free decay and similar for the heave, compare heave free decay with heave but I see that for the freedecay test under consideration other motions were also considered in the paper. These are the ones that are out of sync. Hope I am clear. I would like to know if there is a way that I could resolve this?
Regards,
AOAW
Dear @Andre.White,
As I said, I would expect some differences in results between the different versions. I would not be too concerned with these small differences in the coupled motions. I would be more concerned if there are sizeable differences in the primary modes of motion in a given free-decay simulation.
Best regarsd,
Hi @Jason.Jonkman:
I have added the heave free decay for v2.5.0. That is all I have done based on paper by Robertson et al (2014). The heave and surge free decay compare well with NREL for both v2.5.0 and v3.4.1. It is the coupled pitch and surge/heave free decay motions that do not compare well with NREL’s. Let me know what you think.
Regards,
AOAW
Ok. Thanks @Jason.Jonkman . I did not see this before I made the last post.
Regards,
AOAW
Dear @Andre.White,
I would also say that the heave free-decay results compare quite well, with only minor differences in the coupled effects that I would not be concerned with. How does the pitch free-decay look?
Best regards,
Hi @Jason.Jonkman:
Thank you. I did not do the pitch free decay test because it was not shown in the paper by Robertson et al. (2014). They only did the surge and heave free decay test…
Regards,
AOAW
Dear @Andre.White,
Ah, OK. I had thought you had said earlier that the pitch natural frequency had changed. I would suggest comparing the pitch free-decay results between OpenFAST v2.5 and v3.4.1 to verify that the natural frequency has not changed.
Best regards,
Hi @Jason.Jonkman:
Noted. Instead of free decay test I had linearized and used fx_mbc3 function and campbell_diagram_data function to output the natural frequencies in an excel file. In which case, I did not give perturbation to any the platform DOFs. It is this computation that showed the pitch natural frequencies for v2.5 and v3.4.1 to be different. Let me know your thoughts.
Regards,
AOAW
Dear @Andre.White,
Linearization results are often sensitive, e.g., to output precision. I would suggest verifying that you see the same frequency change in the time domain via free-decay tests.
Best regards,
@Jason.Jonkman noted. Thank you.
On another note, I am trying to find someone who could help me to write MATLAB code that would allow me to run a lot of cases in OpenFAST. That is, say I would like to run like 12 seeds and change certain variables etc. I would be able to do it. Trying to expediate the process. Do you know anyone who could assist. Please let me know.
Thanks.
Regards,
AOAW
Dear @Andre.White,
I can’t recommend someone, but such scripts have been discussed before, e.g., here: Automatic DLC input file generation · Issue #658 · OpenFAST/openfast · GitHub. Also, tools such as WEIS can already do what you propose (though WEIS is not MATLAB based).
Best regards,