Dear Lorenzo,
I haven’t checked all of the off-diagonal inertia terms, but at least the diagonal entries of your inertia matrix, as well as the mass and center of mass, make sense to me.
Best regards,
Dear Jason,
The diagonal terms return to me too, while I have doubts about the off-diagonal terms of inertias. In fact, in my model I get accelerations in x, y, z right while rx, ry, rz do not coincide very well.
Can you please check them?
Thanks for your help,
Best regards,
Lorenzo.
Hi Lorenzo,
Yes, I agree with your numbers. Here are my calculations:
[code]>> M_swl = [ [ 8.0660481545E+006 0 0 0 -78.0013014648.0660481545E+006 -1.1132995542E-0028.0660481545E+006 ];
[ 0 8.0660481545E+006 0 78.0013014648.0660481545E+006 0 -1.3918125391E-0028.0660481545E+006 ];
[ 0 0 8.0660481545E+006 1.1132995542E-0028.0660481545E+006 1.3918125391E-0028.0660481545E+006 0 ];
[ 0 78.0013014648.0660481545E+006 1.1132995542E-0028.0660481545E+006 6.8025977471E+010 9.9135213851E+006 -8.6878157798E+006 ];
[ -78.0013014648.0660481545E+006 0 1.3918125391E-0028.0660481545E+006 9.9135213851E+006 6.8022535284E+010 6.965876724E+006 ];
[ -1.1132995542E-0028.0660481545E+006 -1.3918125391E-0028.0660481545E+006 0 -8.6878157798E+006 6.965876724E+006 1.9157300092E+008 ] ]
M_swl =
1.0e+10 *
0.000806604815450 0 0 0 -0.062916225372230 -0.000008979927815
0 0.000806604815450 0 0.062916225372230 0 -0.000011226426962
0 0 0.000806604815450 0.000008979927815 0.000011226426962 0
0 0.062916225372230 0.000008979927815 6.802597747100000 0.000991352138510 -0.000868781577980
-0.062916225372230 0 0.000011226426962 0.000991352138510 6.802253528400000 0.000696587672400
-0.000008979927815 -0.000011226426962 0 -0.000868781577980 0.000696587672400 0.019157300092000
TransMat = [ [ 1 0 0 0 78.001301464 1.1132995542E-002 ];
[ 0 1 0 -78.001301464 0 1.3918125391E-002 ];
[ 0 0 1 -1.1132995542E-002 -1.3918125391E-002 0 ];
[ 0 0 0 1 0 0 ];
[ 0 0 0 0 1 0 ];
[ 0 0 0 0 0 1 ] ]
TransMat =
1.000000000000000 0 0 0 78.001301463999994 0.011132995542000
0 1.000000000000000 0 -78.001301463999994 0 0.013918125391000
0 0 1.000000000000000 -0.011132995542000 -0.013918125391000 0
0 0 0 1.000000000000000 0 0
0 0 0 0 1.000000000000000 0
0 0 0 0 0 1.000000000000000
M_cg = TransMat’M_swlTransMat
M_cg =
1.0e+10 *
0.000806604815450 0 0 0 0 0
0 0.000806604815450 0 -0.000000000000000 0 0
0 0 0.000806604815450 0 0 0
0 0 0 1.895050184890263 0.000991227154749 0.000006894335879
0 0 0 0.000991227154749 1.894705909912941 -0.000003858384188
0 0 0 0.000006894335879 -0.000003858384188 0.019157043867686[/code]
Best regards,
Thanks a lot for your prompt reply.
Best regards,
Lorenzo.
Dear Jason,
I report here my problem because despite using the right mass matrix my model does not work.
I noticed that in the case of offshore turbines the two predominant forcings are the loads at the base of the turbine tower and the moorings (which must compensate for the tower loads).
In my model the loads at the base and the moorings add up (with those relating to hydrodynamics, restorings, etc. which are smaller) and, divided by the mass matrix, give the accelerations, from which I get the positions to feedback and obtain the loads of the mooring (via Map ++).
The problem is that Fx and My of the tower start with a step, which if too large leads the moorings to diverge rather than converge and compensate for these loads. To demonstrate this, I tried to manually enter the average value of the forces at the base of the turbine tower, and as can be seen in Figure 1, beyond a certain value of Fx the mooring diverges (at 2 * 10 ^ 5 N converges, while at 3 * 10 ^ 5 N diverges).
I asked you about the off-diagonal moments because I noticed that it is not so much the x position that makes Fx_moor diverge, but rather ry which is much higher than the value obtained with FAST.
By setting an increasing trend instead of a step (Figure 2) it is possible to reach higher values of Fx (up to 10 ^ 6, a value higher than the Fx_tower values of the tests carried out), however this would mean making up the first instants of Fx at the base of the turbine.
How does FAST prevent this problem?
Thanks for the reply, best regards.
Lorenzo.
Dear Lorenzo,
I’m not sure I can really comment on the results from your model, which I’m not familiar with.
What does FAST / OpenFAST do for this same load case?
Please note that FAST / OpenFAST include many nonliearities that may not be present in your model. A related discussion in the following forum topic may provide some insight: Hydrodynamic implementation.
Best regards.
Hi everyone
I’m a Ph.D. student studying concrete floating offshore wind turbines. (using WAMITv7 and OpenFAST.)
I quite don’t understand the 2nd process.
(2)Redefine the added mass and damping as zero in the file named “spar.1”
As the spar.1 is the output from WAMIT, does this mean the zero values of added mass and damping will be input at WAMIT *.frc files then do the analysis again, or it will be input at PLATFORM ADDITIONAL STIFFNESS AND DAMPING of OpenFAST HyDroDyn?
Thank you in advance for your kindly answer.
Wichuda
Dear Wichuda,
The steps outlined by Minxi where to obtain the rigid-body mass matrix of the system, not including hydrodynamic added mass. In OpenFAST, it is possible to zero out the added mass completely, but this was not possible in FAST v7. So, the workaround was to specify WAMIT output files with all coefficients set to zero. These are not obtained by running WAMIT; instead, the files were created by hand to look like they were WAMIT-generated. I’ve attached the files to my post dated Oct 30, 2012 in the following forum topic: Questions about the HydroDyn Module - #4 by Jason.Jonkman.
Best regards,
Dear Jonkman
Thank you very much. Now I got it.
Regards,
Wichuda
Dear @Jason.Jonkman ,
Regarding the hydrostatic-restoring matrix of OC4 floating platform, I have a question.
In your paper, it said:
The only non-zero components of C_ij^Hydrostatic are (3,3), (4,4), (5,5), (3,5) and (5,3) when the body-fixed xz-plane of the submerged portion of the support platform is a plane of symmetry
and
If the body-fixed yz-plane of the submerged portion of the support platform is also a plane of symmetry, the (3,5) and (5,3) components of C_ij^Hydrostatic are also zero.
For OC4 floating platform, it is only x-z symmetric, the (3,5) and (5,3) components of hydrostatic-restoring matrix should be non-zero.
But in “Definition of the Semisubmersible Floating System for Phase II of OC4”, it provides the hydrostatic-restoring matrix as:
The (3,5) and (5,3) components of hydrostatic-restoring matrix are zeroed.
In contrast, I use HAMS to simulate OC4 floating platform, and get a hydrostatic-restoring matrix with non-zero (3,5) and (5,3) components:
0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00
0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00
0.00000E+00 0.00000E+00 0.37221E+07 -0.99995E-09 -0.10649E+02 0.00000E+00
0.00000E+00 0.00000E+00 -0.99995E-09 -0.37610E+09 0.28113E-06 0.27304E+07
0.00000E+00 0.00000E+00 -0.10649E+02 0.28113E-06 -0.37610E+09 0.00000E+00
0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00
In HAMS, I locate the center of gravity at (0,0,0), so my hydrostatic-restoring matrix has no contribution from gravity, and it only results from buoyancy. In addition, my hydrostatic-restoring matrix has not been non-dimensionalized. And the (3, 4) (4, 3) (4, 5) (5, 4) components are close to zero, so I think they are negligible.
Now I am confused which hydrostatic-restoring matrix is correct.
Regards,
Ran Tu
Dear @Ran.Tu,
I agree with your comments.
The (3,5) and (5,3) elements of the hydrostatic stiffness matrix should be:
-WtrDens * Gravity * INTEGRAL( x, dA ) over the waterplane area. This integral is definitely zero for a structure that is symmetric about the yz plane. While the OC4-DeepCwind semi is not symmetric about the yz plane, this integral should also be zero because of its equilateral triangular shape. I also see in your result that you have a nonzero (4,6) element, which should be -WtrDens * Gravity * PtfmVol0 * x_b, where x_b is the x offset of the center of buoyancy. This value should be zero for the OC4-DeepCwind semi, again because of its equilateral triangular shape. I’m guessing the geometry is not specified exactly correctly in your HAMS model.
Best regards,
Oh, I have not noticed (4, 6). After checking the buoyancy center, I confirm that previous HAMS model definitely is inaccurate as it has a offset in x axis. I will model them again.
By the way, do you know any free/open meshing software which can construct mesh for WAMIT or HAMS?
Regards,
Ran Tu
I’m not familiar with free/open source meshing software compatible with WAMIT or HAMS, but hopefully someone else can comment.
Best regards,
Dear all,
I am trying to reproduce RAOs for OC3 system by using nemoh. The reference papers are: “Definition of the floating system for phase IV of OC3” (Jonkman, 2010) and “Investigation of Response Amplitude Operators for Floating Offshore Wind Turbines” (Ramachandran, Robertson, Jonkman, and Masciola, 2013).
First, in nemoh, I assum the center of gravity position is at the intersection of the still water level(0,0,0) and compute the hydrodynamic quantities(added mass,damping and wave excitation forces), later I compared the hydrodynamic quantities from nemoh with that from the paper “Definition of the Floating System for Phase IV of OC3” and agreed very well.
Then, I added additional linear damping matrix matrices, mooring stiffness matrix, additional yaw stiffness matrix according to the paper “Definition of the Floating System for Phase IV of OC3”. For convenience, I combine the mooring stiffness matrix with the additional yaw stiffness matrix. The MassMatrix(M_swl) refer to the matrix provided by jason, these matrices are shown below:
Fig 1
I use the following equation to solve RAO, find the inverse matrix first, and then multiply the wave force matrix F_swl
Fig 2
Fig 3
AMMatrix, DAMPMatrix and FMMatrix are added mass,damping and wave excitation forces from nemoh. Here are my RAO results:
Fig 4
As it can be seen, heave RAO is OK both in frequency and amplitude, while pitch RAO is completely wrong and surge RAO don’t have the second peak. Here are the results from the reference:
Fig 5
I have checked several times to confirm that the matrix used is consistent with the matrix provided in the first reference, and I also followed the advice on the forum and used transfor matrix between the SWL and CG, but it doesn’t work. and I don’t know which step went wrong
I would really appreciate it if anyone could offer me some suggestions!
Dear @Lin.Yang2,
Overall, your process sounds OK. Just a couple comments:
- You mention assuming that the center of gravity (CoG) at (0,0,0), which is not the correct CoG location for the OC3-Hywind spar system. I assume you mean this is the reference point for the hydrodynamic variables. I do see that your mass matrix includes the CoG offset.
- You equation seems to be missing the contribution of full-system weight to the restoring (stiffness) in the pitch and roll directions (that is mass * Gravity * CoG. This explains your miscalculation of the pitch RAO and incorrect coupling to surge.
Best regards,
Dear Jason,
Thank you for your reply, I did miss the contribution of full-system weight to pitch and roll restoring in the stiffness matrix, now I have recalculated the stiffness matrix, and for C44 and C55, it should be -4999180000-80660489.878=-11164867090 Nm/rad.
I recalculated the RAO using the new stiffness matrix, however the results are still wrong, here are my RAO results:
Fig 1
In nemoh, the position of the center of gravity of the full-system does not affect my hydrodynamic results, because I only need the frequency-domain added-mass and damping matrices and the frequency-domain wave-excitation force vector. All other terms in RAO calculation are specified by me. In order to match the results in the reference paper, I also set the reference coordinate point to the still water surface (0,0,0) instead of the center of gravity.
For the paper “Definition of the Floating System for Phase IV of OC3”, I currently only use the additional linear damping matrix matrices, mooring stiffness matrix, additional yaw stiffness matrix in the paper, for linear hydrostatics and linear mooring system, I did not consider the previous item(Fig 2 and Fig 3), because I found that considering these two forces will affect the result of heave RAO. Besides, I think, for all stiffness and damping matrices and the wamit results of added-mass, damping and wave-excitation force provided in this paper, the reference point is still water level and don’t need to get transformed between SWL and CG, am I thinking right?
Best regards,
Lin.Yang
Dear @Lin.Yang2,
I’m having a hard time following everything you are saying. Can you share your full set of mass, stiffness, and damping matrices you are using (not including the added mass matrix, radiation damping matrix, and wave-excitation force from NemoH, which it sounds like you have already confirmed are correct)? I’m guessing the problem is in one of these matrices.
Best regards,