Free-Vibration Response of OC3HYwind Spar turbine

I have modeled the OC3-Hywind floating turbine in FAST V.8. My intention is to obtain the free vibration response by subjecting the platform to inital displacement of 1 m in sway direction(or surge direction). All the loading, hydrodynamics, aerodynamic etc are disabled, only loadings are from mooring lines and buoyancy. Only 6 Dof (surge, sway, heave, roll pitch, yaw) out 24 are active (others disabled). i have disabled the additional damping and additional stiffness as well. After getting the response i observe that the response is damping out and reduces with time. I want to know what is the source of damping, because i think the response should be of constant amplitude. Another strange thing is that response only at that DOF is damping out in which i have applied initial displacement. the response at other DOF does not damping. I have shown the response due to initial displacement of 1 m in surge direction. I will appreciate the prompt response. Thanks

Another point which i dont understand is the increase in the response in sway direction due to initial displacement in surge direction. If i consider the mooring line orientation in the OC3-Hywind system, it seems that if there is no load external loading and only initial displacement is applied in surge direction, then there should be no displacement in sway direction (because all the mooring line fairlead forces cancel out in sway direction). But in this case i see that the sway displacement is gradually increasing with time as shown in the attachment, which i dont understand. I would appreciate the valuable comments.
sway.JPG
heave.JPG
Surge.JPG

Dear Zahid,

You say “all the loading, hydrodynamics, aerodynamic, etc. are disabled”; can you clarify how you did this?

The OC3-Hywind model, as provided in Test24 of the FAST CerTest, models the hydrodynamics via a hybrid combination of strip theory and potential flow. Thus, the model includes both hydrodynamic radiation and viscous damping unless you’ve disabled some of these terms.

Regarding Sway, the motion is small, but my guess is there has been a small amount of coupling to the platform-yaw DOF e.g. due to a slight imbalance in the mooring tensions. You say you’ve removed the additional platform stiffness, which was set in yaw to mimic the effect of the delta connection of the moorings to the platform, which are not modeled directly in this model. Thus, the resulting model will have a very small yaw stiffness.

Best regards,

Dear Dr. Jonkman
First of all i am thankful for your kind response on the problem i am facing.

  1. Regarding the increasing sway response and yaw stiffness (as you suggested), i enabled the addition yaw stiffness and after that i analyzed the system and found it helpful. Now the sway response is almost zero.

  2. Regarding the free decaying response in surge direction, i am still facing problems. As you have asked for more clarification on disabling of “all the loads”, I have attached the input file, if you can give a look to it, it will be appreciated. I have highlight the sections where i have disabled the loading. In my opinion there might some problem in the hydrodynamic input file, So please look at that if there is anything i am missing. In brief i can describe that:
    a). I switched the CompAero, Compservo to zero, do disable the aerodynamic loads.
    b). CompHydro is switched to 1 in the initial input file, however than to disable hydrodynamic loading, i switched the “WaveMod =0 (for still water condition)”. Also current mod is switched 0. HasWAMITis switched to False. Additional linear damping is made zeros. (only additional yaw stiffness is enabled).
    c). All the simple hydrodynamic coefficient (Ca, Cp, Cd etc.) are made zeros.

I have attached files for more details. Thanks for your help.
Regards,
Zahid
NRELOffshrBsline5MW_OC3Hywind_HydroDyn.txt (16.9 KB)
OC3 Hywind Input Files.docx (319 KB)

Dear Zahid,

Regarding (2), I see that you still have a nonzero hydrodynamic viscous-drag coefficient (SimplCd = 0.6). It is this term that is still introducing damping into your model.

Just FYI, I see that you’ve eliminated the potential-flow solution in your model, meaning that the hydrodynamic solution will involve only strip theory. When doing this, I recommend that you set appropriate added-mass (Ca) and dynamic pressure (Cp) coefficients (and appropriate viscous-drag coefficients (Cd) when you want to add back damping in your model) to obtain an appropriate response.

Best regards,

Dear Dr. Jonkman

I highly appreciated your help. After making “SimpleCd = 0”, I obtained the exact response i was looking for. I have another query about the hydro-static restoring matrix based on the buoyancy force. In the HyroDyn manual it is mentioned that if WAMIT is flagged to False, and only Strip theory is enabled. The hydro-static stiffness matrix is based on the initial undisplaced position of platform and strip theory does not update the hydro-static stiffness matrix for displaced postion. It is mentioned that it should included manually as “AddClin”. I want to ask few points, your comment on those will be appreciated.

  1. At every time step there will be change in center of buoyancy and buoyancy force and so the hydro-static stiffness matrix needs to be updated, so i don’t understand how can i include/update it manually in the HydroDyn input file, at every time setp.

  2. IF WAMIT is flagged to True, Does the hydro-static stiffness matrix is automatically updated at each time step or it still needs to be computed manually for each displaced position?

Regards,
Zahid Ullah

Dear Zahid,

Regarding (1), I agree that the center of buoyancy and buoyancy force will change (with displacement) every time step, but in a linear sense, this change can be represented as a C*q, where C is the (constant in time) hydrostatic stiffness matrix (AddCLin) and q is the platform displacement.

Regarding (2), AddCLin should be calculated for strip-theory members that are not also associated with the potential-flow solution (PropPot = False). If all strip-theory members are also associated with the potential-flow solution (PropPot = True), then AddCLin may be set to zero.

Best regards,

Dear Dr. Jonkman
I understand your point. I have added the linear hydro-static stiffness coefficient, attached below is snapshot of the input, please verify if this is correct. I obtained the stiffness coefficient from the technical report entitled, “Definition of the Floating system for Phase IV of OC3”. I am confused about the negative valued of roll and pitch stiffness coefficient.

1). It is mention in the report that the negative values of roll and pitch stiffness coefficient are because COB is below SWL. I don’t understand this very well. If i calculate these value from the equation given in HydroDyn theory manual, section 6.8.3, it may be come out as positive value. please help me in understanding this point.

2). Currently the response of system is in inertial reference frame which is placed at SWL. Is it possible to obtain the response of at the center of mass of the system (body reference frame) ?

Thanks

Dear Zahid,

Yes, your specification of AddCLin looks correct for the OC3-Hywind spar.

Regarding (1), the term rhogV_0*z_b from section 6.8.3 of the draft HydroDyn User’s Guide and Theory Manual is the dominant term and is negative-valued because z_b is negative i.e. the center of buoyancy is below SWL.

Regarding (2), it is not possible to directly output the motion of the system center of mass from the list of available HydroDyn outputs. However, It is possible to place the platform reference point in ElastoDyn to be close the system center of mass (at least in the vertical direction) and thus the ElastoDyn platform motion outputs (PtfmSurge, etc.) would represent the motion of the system center of mass. Please keep in mind, though, that FAST treats the floating wind system as a flexible body, so the actual system center of mass location may change e.g. with blade and tower deflection. So, the platform reference point in ElastoDyn can only placed close to some “mean” position of the system center of mass.

Best regards,

Dear Dr. Jonkman

I am very much thankful for your help and i feel sorry for keep asking questions, Its because i am new Fast user so, need to understand a lot, and you have been a lot helpful in that regard. Again i need to ask another point. This time i apply a 1 m displacement only in heave direction to obtain the free vibration response. The rest of result are fine, but i found that there is also some displacement in the surge direction due to initial displacement in heave direction (surge displacement shown below in figure). So, in order to order to find the cause of this surge displacement, i check force equilibrium in surge direction and find out that there is residual force in surge direction due to mooring lines. After summation of fairlead forces, still there is residual force (shown below) in surge direction which cause displacement. I wanted to know how can i fix this so that i the fairelad forces are satisfy equilibrium.

2). Additional i tried to use FEA_Mooring instead to MAP, but i receive some fatal error and the program aborted, i will be thankfull if you can comment on the source of error.
Error.JPG
Surge response.JPG
Surge Force.JPG

Dear Zahid,

How do you know that the residual force in the mooring system is causing the surge displacement, versus being the result of the surge displacement? I would have assumed it is the latter. My guess is that the small amount of surge displacement is coupled to platform-pitch and results from the overhanging weight of the rotor/nacelle, mostly independent of the initial heave displacement.

Where are you getting the NRELOffshrBsline5MW_OC3Hywind_FEAMooring.dat file? This is not a file we’ve supplied in the FAST archive.

Best regards,

Dear Dr. jonkman

I just assumed by looking at the forces in the mooring lines, however after your suggestions, i realized i was wrong, it was actually due to overhanging weight of the rotor/nacelle, as you said. I used the NRELOffshrBsline5MW_OC3Hywind_FEAMooring.dat input file, provided with the FEA Mooring program (attached below). However, i am not sure the input file is in the required format of FAST. I appreciate your help and cooperation, your comments help me a lot in understanding FAST. Thank you very much.
NRELOffshrBsline5MW_OC3Hywind_FEAMooring.txt (5.06 KB)

Dear Zahid,

Your FEAMooring input file appears to have an extra line with white space and paragraph return between each valid line. I’m not sure how the file got reformatted, but that would certainly cause trouble trying to read it in. I’ve attached the version of the FEAMooring input file for the OC3-Hywind spar that comes directly with the FEAMooring archive when I download it.

Best regards,
NRELOffshrBsline5MW_OC3Hywind_FEAMooring.dat.txt (4.96 KB)

Thank you very much Dr. Jonkman

Dear Dr. Jonkman

I have another query regarding OC3HYwind (Test24) and will appreciate your help. I am trying to understand FAST and dynamic response of floating body. I subjected the OC3Hwyind turbine to initial displacement of 2 m in surge direction. I flag the potential flow model “0”, and use only strip theory. Also in strip theory i made all the coefficient to “0”. I requested “HydroFzi”," AddFzi" in the output. There is some force in “HydroFzi”, which i am unable to understand. I have attached the plot of “HydroFzi”, “AddFzi”, “initial buoyancy from hydrodynamic summary”. I know about the source of “AddFzi” and “Buoyancy” but i don’t know about “HydroFzi”. I will be thankful if you can help in understanding “HydroFzi”.

Additionally, If you can send me unofficial “Fast theory manual” and “MoorDyn input file for Test24”. It will help me to understand the FAST more deeply.
i really appreciate your help and cooperation. (my email id: ktk.zahid@gmail.com)

Regards,
Zahid
AddFzi.PNG
Bouyancy.PNG
HydroFzi.PNG

Dear Zahid,

The “odd” square shape you are seeing in the time series is likely due to the small round-off error that results from the fact that FAST is only outputting 4-digits of precision based on the setting of OutFmt = “ES10.3E2” in FAST’s primary input file. My guess is that the shape would look more rounded if you added at least one more digit of precision in the output format. See the following forum topic for more information on OutFmt: http://forums.nrel.gov/t/fast-integrator-step-size-precision/247/1.

I’ve sent you the “Unofficial FAST Theory Manual” by e-mail.

I don’t have copy of the MoorDyn input file for OC3-Hywind model of Test24, but it is probably not too difficult to make one yourself because the MAP++ and MoorDyn input files have many similarities.

Best regards,

Dear Dr. Jason Jonkman
I added one more digit precision as you suggested and now the results are better…more round shape. I also wanted to know about the source of force in “HydroFzi”. I have made all the coefficients zeros but still there is force in “HYdroFzi”. My first assumption was that may be its the the initial buoyancy force in the “HdroFzi”. but the force is smaller than “buoyancy” , so i am confused to track the source of force in “Hydrofzi” vector. Please Help me on the issue. (May be my statement is not very understandable so i attached figures).

About the Moordyn input file i used the input file available with MoorDyn program, but its shows some format error and gives reading error.
Always been thankful for you help.
HydroFzi.PNG
AddFzi.PNG
Initial Buoyancy.PNG

Dear Zahid,

I’m not sure how you are plotting buoyancy, but the value of HyrdoFzi = 8.07E7 N matches the static buoyancy reported in the OC3-Hywind specifications report: nrel.gov/docs/fy10osti/47535.pdf. The small oscillation in HydroFzi is from the small oscillation in AddFzi, which is included in the calculation of HydroFzi.

I can’t comment on the issue with MoorDyn without knowing more about what file you are using an what error you are receiving.

Best regards,

Dear Dr. Jason Jonkman

Upon your suggestion i looked into the buoyancy force in the hydrodynamic summary file and also in the report you referred. There is difference in the buoyancy force in the summary file, however the submerged volume is same in both “the report” and “HD.sum”. I have attached the “HD.sum file” and also a snapshot of it. In the snapshot there is total structure volume and submerged volume, if I multiply the submerged volume with “10259.8066", then i get buoyancy force equal to that given in the report, but if i multiply “total volume” with "10259.8066”, then i obtain buoyancy force which is given in the summary file. Please help me to figure out the problem. (i am using only strip theory and flagged the WAMIT to False)Thank you very for your time.

Regards,
Zahid
HD summary file.txt (205 KB)

Dear Zahid,

Indeed, there is a bug in HydroDyn whereby the external and total buoyancy written to the HydroDyn summary file is incorrect for the version of HydroDyn included in FAST v8.15 and v8.16. This bug has been discussed in the following forum topic and will be fixed in the next release of HydroDyn: http://forums.nrel.gov/t/water-ballasting-buoyancy-loads/1380/4. The time series is not effected.

Best regards,

Dear Dr. jonkman

I am thankful for clarification on buoyancy force. For the MoodrDyn error, i have attached input file i am using, and snapshot of the error. I think the input file is not in the proper format, but i don’t how to fix it.

Regards,
Zahid
NRELOffshrBsline5MW_OC3Hywind_MoorDyn.txt (2.34 KB)
Error.PNG