Hi everybody,
I was wondering if I can adapt the Tensioned Leg system to one of the main cylinders of the semisub structure, while leaving the other two cylinders without any mooring lines. This would create a kind of a pivot point in one of the ylinders while the other two would be free to rotate around it. I’ve done so, simply by replacing the semisub moordyn file by the tlp moordyn file. For that I just changed the localization of the center of buoyancy of the semisub platform putting the centroid of the cylinder in which I want to use Tl in (0,0).
Unfortunately, when I run the simulation, it crashes after 2.5 seconds displaying the following message:
FAST_Solution:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutput:BEMT_CalcOutput(node 17,
blade 1):Compute_UA_AirfoilCoefs:UA_CalcOutput:Mach number exceeds 1.0. Equations cannot be
evaluated.
Is there any way to fix this issue? Is it even possible to do what I’m trying to?
Best regards,
Tomás
Dear Tomás,
Yes, FAST / OpenFAST can be used to model a semisubmersible-shaped floater with a taut-line or tension-leg station-keeping system.
I’m not sure I understand fully how you’ve set up your FAST model, but clearly the model is going unstable as a result of some physical property of the model or a numerical problem. It is difficult for me to guess as to what the problem is. I would probably start by looking at the basic hydrostatic stability of the system i.e. disable aerodynamics and wave excitation and see if the system remains in hydrostatic equilibrium.
Best regards,
Dear Jason,
Thaks for your suggestion. I’ve done that, turning off hydrodyn and aerodyn. The simulation run till the end but when I looked in the output file, my turbine was 4000 meters deep…
Then I turned hydrodyn on putting WavMode=0. It crashed before reaching 3 seconds. Again, I took a look at this new output file and the pltform was sinking… Also, the fairlead tension just jumped from 10^5 to around 10^10 in one time step…
So the problem should be here… Any further hints?
Best regards,
Tomás
Dear Tomas,
You can’t model a floating platform without HydroDyn because there will be no buoyancy to keep the turbine afloat.
Yes, enabling HydroDyn with WaveMod = 0 is what I meant by running a case without wave excitation. But from your description it sounds like the model is still unstable. From the blow-up in tension that you describe, it sounds like there is a numerical problem that reducing the time step would perhaps resolve. But there could also be a problem with the hydrostatic stability of the structure as well. For the latter, I would disable structural DOFs until you can isolate which DOF (heave, pitch, roll) is unstable.
Best regards,
Dear Jason,
I’ve concluded that the problematic DOF is pitch. If its flag is turned to false, no problems arise. In this case, what would you suggest me to do? How can I fix this issue?
Best regards,
Tomás
Another issue: When I change the volume of water used for ballast, in the hhydrodyn file, and consequently change the overall system’s moments of inertia (roll, pitch, yaw), do I need to change the platform’s moments of inertia in the elastodyn file? Or the moments of inertia considered in the elastodyn file are only concerned with the “metallic” platform substructure not taking into account the water used as ballast?
Best regards,
Tomás
Dear Tomás,
Again, I’m not sure I fully understand how you’ve set up your FAST model, but from the symptoms you’ve described, I would guess that your model is not hydrostatically stable in pitch (a physical instability) or the platform-pitch natural frequency is so high that a smaller time step is required (a numerical instability).
If you are modeling the water ballast in HydroDyn, then the platform mass, center of mass, and inertias in the ElastoDyn input file only pertain to the “metallic” platform structure.
Best regards,
Dear Jason,
I’ll try to be more clear regarding the model setup.
1st:in hydrodyn, I’ve set PltfCOBxt=28,87m (which is the distance from the semi-sumersible vertex to the center of the triangle defined by the semisubersible structure). I’ve this so that I’ve one vertex on the center of coordinates which eases the use of the TLP system in this vertex.
2nd: in elastodyn I’ve set PtfmCMxt=40,101m, which is the distance from the center of coordinates to my tower localization coordinates (the 28,87m plus the distance from the center of the triangle defined by the semi-submersible structure to the middle point of the line defined by the two other vertexes)
3rd:in hydrodyn, I’ve adapted the ballast by:
keeping the same total water mass (what I’ve added in one of the cylinders is what I took from the other two)
moment balance (The displacement of the tower position and the change it causes in the internal momentum equilibrium must be nullified by the momentum caused by the addition of water in one cylinder and the extraction of water in the other two. The momentum equation is done around yy axis. Around xx, nothing changes, due to the symmetry of the system )
4th: I’ve copied the moordyn file from the 5MW_TLP_DLL_WTurb_WavesIrr_WavesMulti into my semisubersible file.
Is there any chance that with this information you can provide any hint on the problem going on?
Best regards,
Tomás
Dear Tomas,
I have a hard time picturing everything you are describing, but it sounds like you want to model the tower atop one of the three offset columns of the semisubmersible; is that correct? The reference point in the ElastoDyn platform and in HydroDyn is associated with the intersection of the undisplaced tower centerline and still water level, so, all matrices (the WAMIT data and additional platform loads) and geometry in HydroDyn must be specified about this point.
It sounds like you are changing platform center of buoyancy (PtfmCOBxt) in (1) to accommodate the change in reference point, but are you changing the associated WAMIT matrices, additional platform loads, and strip-theory member geometry accordingly?
I’m not sure I understand why your platform center of mass offset in xt in (2) would be different than your center of buoyancy offset in xt in (1).
When, changing the water ballast in (3), are you changing the additional linear stiffness matrix (AddCLin) accordingly)? Also, please be aware of the bug regarding the filled member option in HydroDyn that we’ve reported on OpenFAST issues: github.com/OpenFAST/openfast/issues/245. I’m not sure this bug applies to your case, but something you should consider.
I hope that helps.
Best regards,
Dear Jason,
Thanks again for the help and sorry for the confusion. I’ve attached one image of my model so that it gets clearer, I hope… Would you pllease take a look at it?
Thanks in advance,
Tomás
Dear Jason,
As far as the bug you’ve mentioned, that’s only an issue when one has high yaw angles, right? When that doesn’t happen, there should be no problems? What would be more or less the threshold, concerning the yaw limit? Do you think it would make it impossible to test/simulate the ability of the overall structure to weathervane? Just as X1wind proposes: x1wind.com/x1wind-technology/
From your message, I take that the center of the system of coordinates (CSC) is ALWAYS defined by the centerline intersection with MSL. In this case, my “MY MODEL” drawing doesn’t makes sense, as I’ve defined the CSC not connected to the tower localization… Is this correct?
Regarding the setup of the model displayed:
Should I define the offsets by making PtfmCMxt=PtfmCOBxt=-11,231m (which is the distance from the center of the triangle to the wind turbine)?
With this setup, do I need to make further modification to the file? Apart from the offsets and changing the ballast to balance the momentum caused by the new tower position relative to the platform? If yes, which ones and how?
Best regards,
Tomás
Dear Tomás,
Yes, the drawing helps. The platform reference point in HydroDyn is always along the centerline of the undisplaced tower at MSL. So, all geometry and matrices specified in HydroDyn must be specified relative to this point. From your description, it does not sound like you’ve set up your model correctly. You must make many more changes then setting PtfmCMxt and PtfmCOBxt. You must change all of the strip-theory geometry (location of joints), you must change the reference location of all of the WAMIT data (*.hst, *.1, *.3 files, plus the second-order files if you are using those), and you must change any matrices specified for the additional platform loads in HydroDyn.
The issue regarding the ballast in HydroDyn is always present, but I’ve only seen it cause problems with the solution for sizeable platform-yaw angles e.g. beyond 10 degrees. I would not recommend enabling ballast in HydroDyn for weathervane-related applications until this bug is fixed.
Best regards,
