Dear @Jason.Jonkman,
Thank you for your response.
I totally agree that RAOs need wave input since they are transfer function normalized with wave height. (they are typically used in naval engineering).
But, in OpenFAST, one can add a wave elevation field as a white noise spectrum and at the same time add a wind speed field denoted V. Therefore, the RAO computed considers the effect of wind and wave.
So, the FOWT will become a black box where i could know the amplitude of tower top fore-aft displacement for any combination of wave height with V (under the assumption that the system is linear). For me, it is equivalent to a FRF.
I dont know if i answer to your inquiry regarding the input.
Best Regards,
Riad
1 Like
Dear Dr @Jason.Jonkman ,
Many thanks for your previous comments, which have been of great help to me. I hope you don’t mind me troubling you again with a question I have.
I am currently trying to implement soil interaction in my models of the NREL 5MW OC3 Monopile Offshore Wind Turbine. I aim to analyse the behaviour of different Tuned Mass Dampers in three operational modes: Idling Rotor, Parked Rotor, and Normal Operation.
However, when I include the SSI files in the model, I encounter the attached error. It is worth noting that this error does not occur when I use the apparent fixity model. The values for the soil stiffness matrix were obtained and entered into the SSI file as per your previous recommendations (I have also attached the matrix and a snapshot of the SSI file).
In light of this error, I would like to ask whether the issue may be due to a misinterpretation of the data—resulting in an incorrect stiffness matrix—or whether it stems from a problem with the geometry of my models. How would you suggest I go about resolving this issue so I can continue with my simulation batches?
As always, I am grateful for your support.
Kind regards,
Emanuel Rergis
Dear @Emanuel.Rergis,
It looks like you are specifying SSI stiffness in the x, y, roll, and pitch directions at the seabed. Presumably you have disabled motion in the z and yaw directions at the seabed because you have no stiffness there, by setting RctTDZss = RctRDZss = 1; can you confirm?
Your simulation errors out after 1.855-s into the simulation. Can you clarify what is happening to the ElastoDyn platform motions and/or motions of SubDyn nodes over this time?
Best regards,
Many thanks, Dr @Jason.Jonkman , for your prompt response.
Just to clarify, RctTDZss = RctRDZss = 0. In fact, please find attached a screenshot of the full SubDyn configuration I am currently using. I have also attached an image showing what is happening across the four basic platform motions.
I am not sure how to obtain the displacements of the nodes — is this an output from ElastoDyn?
Additionally, I would like to confirm whether I am using the correct soil stiffness matrix for the OC3 Monopile model. The relevant images are also attached.
I look forward to your response.
Kind regards!
Emanuel Rergis
Dear @Emanuel.Rergis,
I agree with your coupled springs stiffness matrix for the OC3-monopile.
You have not shown the platform heave or yaw motion, but because you have no stiffness and no constraint in these directions, I would guess that is the problem you are facing: the monopile is simply falling due to gravity. Setting RctTDZss = RctRDZss = 1 in SubDyn should solve the problem.
Best regards,
Dear Dr. @Jason.Jonkman:
Many thanks once again for your response. I followed your advice, and the simulations are now running without any issues.
At this stage, I once again need to use BModes for my model. I created my own .bmi file, but I’m not entirely sure if it will run smoothly. After reading the manual a bit and following my own logic, this is what I managed to write (please see the attached file).
My intention is to obtain the modes (at least the first three) of the NREL 5 MW OC3 Monopile, including soil interaction through a 6x6 stiffness matrix.
Am I on the right track? Am I missing anything? Do you think it will work?
I look forward to your comments and, as always, I greatly appreciate your help.
Best regards!
Emanuel Rergis
Dear @Emanuel.Rergis,
I’m not fully sure what you are sharing, given that what you share is not in the format of BModes. Also, the tower for the NREL 5-MW baseline turbine atop the OC3-monopile starts at 10 m above MSL.
A BModes model for the full support structure of this offshore wind turbine configuration is available on my Google drive: BModes - Google Drive. However, this model was developed for FAST v7, before SubDyn was introduced. But you could update this model by starting the tower at 10-m above MSL, eliminating the monopile properties from the tower section properties table, and replacing hydro_M and mooring_K with the 6x6 Guyan mass and stiffness matrices that are calculated by SubDyn and reported in its summary file.
Best regards,
Dear Dr @Jason.Jonkman :
Many thanks for your responses and recommendations. I have not yet had the time to familiarise myself with BModes, but I will do so as soon as possible. In the meantime, I produced a quick PSD plot using the displacements in the fore-aft direction for two soil models: the Apparent Fixity Model and the cross-coupling model obtained from the soil matrix I previously shared with you.
This has raised a few questions which I hope you might be able to clarify:
- Is it normal for the natural frequency of the tower in its first mode in the cross-coupling model to appear further to the left than that of the Apparent Fixity Model?
- I injected white noise from HydroDyn but was unable to observe the natural frequency of the second mode. Can I calculate this from another type of analysis, or would it be preferable to use BModes?
I have attached the plots with the relevant information. Many thanks again, and kind regards.
Emanuel Rergis
Dear @Emanuel.Rergis,
Regarding your first question; normally I would expect the apparent fixity model to be set-up to match the full-system natural frequencies of the more complicated model, which in your case sounds like a coupled springs model. Can you clarify how you defined your apparent fixity model?
Regarding your second question; did you inject white noise from HydroDyn up to a frequency that encompasses the second support structure mode of interest? That said; you could always use an OpenFAST linearization analysis followed by Eigenanalysis to compute full-system natural frequencies of OpenFAST models (including for the second bending modes of the support structure), rather than relying on white noise excitation.
Best regards,
Dear Dr @Jason.Jonkman ,
I assumed that, by default, OpenFAST uses the Apparent Fixity model if nothing is specified in the SubDyn input file. Is this assumption correct?
Is the information on how I defined the Apparent Fixity model the one that appears in the SubDyn file? Or is there another submodule from which you would need additional information?
I look forward to your comments. Many thanks.
Emanuel
Dear @Emanuel.Rergis,
An apparent fixity model is defined in SubDyn by specify the base reaction joint below the seabed and placing a beam element between a joint at the seabed and this base reaction joint. This beam element (length, stiffness properties) would be set up so that the lowest modes of the support structure match the natural frequencies of a more complicated model with soil-structure interaction.
Best regards,
Dear Dr. @Jason.Jonkman:
Please find attached the SubDyn input file that defines my apparent fixity model. Is this set-up the expected one for an NREL 5MW OC3 Monopile? Let me know your thoughts, please. Thanks once again.
Emanuel M. Rergis
Dear @Emanuel.Rergis,
This file does not have an apparent fixity representation of the foundation; rather this file uses a rigid foundation at the seabed.
Best regards,
Dear @Emanuel.Rergis,
If you are interested in the apparent fixity method, I would recommend you looking at section 3.3.2 here: https://docs.nrel.gov/docs/fy21osti/79938.pdf
Specifically, since you are using Timoshenko beams in SubDyn, you will have to apply the approach proposed in section 3.3.2.2.
The eigenfrequencies obtained from the apparent fixity method and the coupled springs method (i.e., stiffness matrix at the seabed) should be equivalent since both methods result in the same compliance at the seabed location.
I hope that helps!
1 Like
Dear @Roger.Bergua,
First of all, thank you very much for taking the time to read my message and for sharing this valuable information. I truly appreciate your guidance and the reference you provided—it’s incredibly helpful.
The clarification regarding the use of the apparent fixity method with Timoshenko beams in SubDyn is particularly useful. I’ll make sure to review section 3.3.2.2 in detail, as you suggested. It’s also reassuring to know that both approaches—apparent fixity and coupled springs—should yield consistent seabed compliance and eigenfrequencies.
Thanks again for your kind support!
Warm regards,
Emanuel
1 Like
Thank you for your guidance, Dr @Jason.Jonkman . I would like to let you know that I have already started working with BModes. Unfortunately, despite having read the user manual and several forum posts, my inexperience with the subject has led me to the following questions:
- Should
KBBt replace hydro_M and MBBt replace mooring_K? What about the CBBt value? Should I include it in the matrices as well? If so, where exactly?
- I cannot find the value of the parameter
mass_pform in the summary file (.yaml).
- From where should I take the value for the 3×3 inertia matrix (
i_matrix_pform)?
- What exactly should I modify in the
tower_sec_props.dat file (please find attached a snapshot of said file)? I cannot identify the properties of the monopile. Are these values still useful nonetheless? Should they be modified?
- Finally, where do I obtain the value for the
cm_pform parameter?
Thank you very much for your time and kind support.
Best regards,
Emanuel Rergis
Dear @Emanuel.Rergis,
Presumably you are using BModes to compute tower mode shapes for ElastoDyn for a support structure in OpenFAST where the monopile is modeled in SubDyn; is that correct?
I’ll answer your questions assuming your answer to my question is “yes”:
- You have these switched, but
hydro_M in BModes should equal MBBt from SubDyn and hydro_K (or mooring_K, with the other matrix zeros) in BModes should equal KBBt from SubDyn. BModes does not make use of damping, so CBBt from SubDyn is not needed.
- There is no lumped platform mass in this case, so,
cm_pform, mass_pform, and i_matrix_pform in BModes can all equal zero.
- See 2.
- The distributed tower properties in BModes should only include the distributed mass and stiffness of the tower, not including the monopile (as specified in ElastoDyn).
- See 2.
Best regards,
Thank you once again, Dr @Jason.Jonkman . I will take your comments into consideration. I still have a few doubts, as the variable names in the ElastoDyn summary file and in the tower_sec_props.dat file differ significantly. If I understood your message correctly, are tor_stff and axial_stff the only values I need to input into the tower_sec_props.dat file? What about the other parameters — should they be set to zero? Should I disregard the blade values and also set them to zero? Is sec_loc important in this case?
Additionally, I would like to ask whether I should run ElastoDyn and SubDyn separately. Are the values in the SubDyn and ElastoDyn summary files still useful if they were obtained from a baseline-type simulation that also included HydroDyn, AeroDyn, Inflowind and ServoDyn?
I’ve attached a snapshot of the ElastoDyn summary file to ask one final question: from this file, should I only extract TMassDen, FAStiff, and SSStiff? Should I ignore Node, TwFract, HNodes, and DHNodes?
By the way, the answer to your previous question is yes, I’m using BModes to compute tower mode shapes for ElastoDyn for a support structure in OpenFAST where the monopile is modelled in SubDyn
Kind regards!
Emanuel Rergis
Dear @Emanuel.Rergis,
Regarding the relationship between the distributed mass and stiffness specified in BModes and ElastoDyn, see my response dated Jun 1, 2012 in the following forum topic: Tower Eigenfrequencies of NREL 5MW Turbine.
Best regards,
Dear Dr @Jason.Jonkman ,
Many thanks for your help. I would like to inform you that I have reviewed the discussion thread in the forum to which you kindly provided the link.
I successfully obtained the first three modes in both the fore-aft and side-to-side directions, and I was pleased to see that at least the first modes coincided with those I had previously identified through a quick PSD analysis (please find the corresponding screenshot attached and note that even if it says apparent fixity model, it represents rigid soil).
However, upon plotting these first three modes, I noticed that the third mode in the side-to-side direction appeared to be inconsistent. I mention this because, according to the plot (also attached), this particular mode shape seems to show greater displacement closer to the base of the tower.
I should note that I did not use the Polyfitt Excel file, but instead plotted the raw mode shapes directly. My aim is to identify the height at which the excitation is greatest, in order to tune Tuned Mass Dampers (TMDs) at that level and thereby reduce the modal dynamic response.
That said, I am unsure whether this apparent inconsistency is due to normalisation of the height or a misinterpretation on my part. I therefore have the following questions:
- Shouldn’t the values near the base of the tower be close to zero?
- I am also unclear whether this zero corresponds to 10 metres above the seabed, or if it represents the Mean Sea Level (MSL).
- Additionally, I am unsure whether the
.bmi or .dat files are correctly or appropriately updated. I have also attached screenshots of these files for your reference.
Thank you once again for your kind support.
Warm regards,
Emanuel Rergis
