Hydrodynamic analysis of OC3-Hywind spar buoy

Dear Dr. Jonkman,

and is it possible to obtain the center of buoyancy position calculated by Hydrodyn with the Strip-theory solution?

Thank you.

Best regards,
Gabriele

Dear Gabriele,

I misspoke in my earlier post. I meant to say that the strip-theory solution is indirectly calculated by distributing the buoyancy loads across the cylindrical members. The center of buoyancy of the strip-theory solution is not directly calculated or output to a file.

That said, the total buoyancy load (3 forces, 3 moments) is calculated and reported in the HydroDyn summary file (with the moments reported about (0,0,0)), and when the buoyancy force is purely directed vertically (with only the Z-component of the force nonzero), you could derive the X and Y offset of the center of buoyancy based on the reported buoyancy moments about Y and X, respectively (but you won’t be able to derive the vertical/Z offset of the center of buoyancy this way).

Best regards,

Dear Dr. Jonkman,

thank you for your reply.
Apologize for my question, but why I won’t be able to derive the vertical Z offset of the center of buoyancy?
I would also check the Ptfm inertias input in Elastodyn but I am not sure how to do calculate them. I tried by making it by hand but the results don’t match.
I read a topic in which you explain that you calculated it by mean of MSC.ADAM considering the total system (platform + ballast + tower + nacelle + rotor). I can understand this reasoning for the mass calculation but I cannot understand for the inertia.

Thank you.

Best regards,
Gabriele

Dear Gabriele,

You can’t derive the vertical offset of the center of buoyancy from FZ, MX, and MY because with these quantities there is only enough information to calculate the X and Y offsets.

The platform inertias should be derived from the mass distribution across the platform. These are likely difficult to calculate by hand, but can be derived numerically e.g. from CAD package.

Best regards,

Dear Jason

  I want to study the second order effects on the motion of OC4-DeepCwind. Could you provide me  the 2nd-order WAMIT input files for the OC4-DeepCwind platform?

 Best regards

Dear Yu.Lei,

Please find the associated WAMIT input files that we used to generate the 2nd-order WAMIT solution for the OC4-DeepCwind semisubmersible attached.

Best regards,

WAMITInputFiles_OC4Semi_2nd.zip (8.86 KB)

Dear Jason

Thank you! It really helps me a lot!

Best gegards

Dear Jason

In your attachments, when you calculated the second-order hydrodynamic coefficients by WAMIT, the second alternative foam of the FRC file was used.

merin_semi_frc.JPG
But in your pervious answer, when you calculated the first order hydrodynamic coefficients, you always set the center of mass, and radii of gyration to zero in WAMIT( the first alternative foam of the FRC file was used) [url]About WAMIT model for 5MW OC3 Hywind - #5 by Jason.Jonkman]. So if I want to get the first and second order hydrodynamic coefficients from WAMIT that will be used in FAST, do I have to run two times of WAMIT with different FRC file?

Best regards

Dear Yu Lei,

Yes, your understanding is correct. The hydrostatic restoring matrix derived from the first-order WAMIT solution should not have contributions from body mass or moorings. However, the second-order quadratic transfer functions depend on the first-order motion response, and so, should use a proper representation of the body mass and moorings. As a result, two separate WAMIT solutions are required for the first- and second-order problems. More guidance is provided in section 6.8.4 of the draft HydroDyn User’s Guide and Theory Manual: wind.nrel.gov/nwtc/docs/HydroDyn_Manual.pdf.

Best regards,

Dear Jason

 I built my own second-order model of  OC4-DeepCwind in WAMIT 6.1S. The high-order method was used and the gemotery model was built in MultiSurf.  And second alternative form of the FRC file was adopted in which I used the same body mass and stiffness matrices as yours. 

I compared my second-order results (.12s, .12d)  with the marin_semi.12d and marin_semi.12s which were downloaded from NREL and found that there was a big difference between them. I have added my WAMIT input files as the  attachments, could you help me to find out the reason? What's more, I find  that the second-order calculation takes a long time, even if I  calculate two frequencies, it will take more than one hour. Is this a normal phenomenon?  What measures should be taken to speed up the calculation?

Best regards
marinsemi_leiyu.rar (5.17 KB)

Dear Yu.Lei,

I took a brief look at your files and the only real difference other than the surface geometry (which I cannot see because I don’t have MultiSurf) is that you use NPF=-1 whereas our model used NPF=0 in the Free Surface Data File (*.fdf). I do recall the second-order WAMIT solution taking a long time to solve…I recall it taking several weeks to generate the full QTFs for the OC4-DeepCwind semisumbersible for the full frequency range we chose.

That said, I’m not an expert on the second-order solution in WAMIT and can’t really offer guidance on how to improve the solution beyond what is already written in the draft HydroDyn User’s Guide and Theory Manual. For further guidance on the WAMIT second-order solution, I suggest reaching out to Chang-Ho Lee of WAMIT, Inc.

Best regards,

Dear Jason

Thanks for your reply. I have checked the setting of NPF according to WAMIT user manual. In WAMIT user manual, there are two method to discretize the free surface: 1) generating the free surface automatically inside the program (form 1 FDF); 2) specifying the horizontal coordinates of the panel vertices (form 2 FDF). Both of us use the first method to generate the free surface (form 1 FDF). But according to WAMIT user manual, when alternative form 1 FDF is used, the NPF must be an negative integer. I couldn’t understand why you used NPF=0 and how could you make WAMIT run because when I run your model errors occured. The attachments is the WAMIT user manual, you can find the setting of NPF in page 7.

Best regards

Chapter 11. VERSION 6.1S.pdf (82.2 KB)

Dear Yu.Lei,

The option NPF=0 is equivalent to alternative form 2 of the fdf with no panels in the inner circle. Essentially, the free-surface forcing terms have not been solved. See section 11.4.5 in your attachment.

Best regards,

Dear Dr. Jonkman,

I found the WAMIT input files for the OC3_Hywind and OC4_Semisubmersible turbines in the former posts. Could you also share me with the MIT-TLP WAMIT input files. Many thanks.

Best regards,
Jian

Dear Jian,

Please find WAMIT input files that we used for MIT/NREL TLP attached.
TLPMIT.zip (33.5 KB)
Best regards,

Dear Dr. Jonkman,

Thank you for your quick reply. If possible, you might also share me with the ITI_Barge WAMIT input files. I planned to refer to Appendix D in your thesis “Dynamics modeling and loads analysis of an offshore floating wind turbine” and construct the files. It seems that you deleted some mesh information in the .gdf for brevity. Thus, you may transfer it to me when free-many thanks.

Best regards,
Jian

Dear Jian,

Please find the WAMIT input files that we used for ITI Energy barge attached.
Barge.zip (321 KB)
Best regards,

Dear Dr. Jonkman,

I am new to WAMIT. I created MultiSurf model and generated the .frc file from MultliSurf; however, it is said as below:

MultiSurf does not write a complete .FRC file –
only the NFIELD and XFIELD portions.

I go and check the WAMIT manual section 3.3 and 3.4 and found that there are some other parameters which not included in my MultiSurf generated .frc file. I understood that I can manually add those missing lines (to specify missing parameters) by myself.
Do I understand correctly? Could you kindly confirm please?

Best Regards,
Subanapong

Dear @Danupon.Subanapong,

I’ve not used MultiSurf myself, and so, I can’t confirm its capabilities to generate *.frc files for WAMIT. But editing *.frc files manually in a text editor is common.

Best regards,

Thank you so much, Dr. Jonkman. I manually edit it and it works.