Dear Sirs
I am trying to develop an Apparent Fixity model for a fixed OWT. I have followed Erica Bush’s and Lance Manuel’s paper “Foundation Models for Offshore Wind Turbines”, and have managed to get the AF length, for a layered soil. However, the next few steps have got me confused.
-
It is required to include the AF length into the FAST file. As i understand, in FAST, ‘tower’ stands for the entire structure supporting the turbine, which includes the transition piece and the pile (upto the mudline). So including the AF length (let us assume 20m, in this case) means increasing the length of the tower. Since the tower base level “TwrRBHt” can’t be less than the default value of ‘0’, does this effectively means increasing the hub height (by 20m, in this case)?
-
since the tower height is normalized (0-1), if i assume the same properties for my pile (below the soil) as the same for the tower above the soil, i wouldn’t have to tinker with the ‘DISTRIBUTED TOWER PROPERTIES’ in the tower file?
-
the mode shapes are required to be altered in the tower file, considering the additional AF length, by using “ModeShapePolyFitting” spreadsheet.
Could you tell me if this is how the AF model is implemented in FAST?
Thanks
Abhinav
Dear Abhinav,
By your description it sounds like you are using FAST v7. I’ll answer your questions below assuming you want to keep using FAST v7:
-
In FAST v7, “yes,” the entire support structure (tower + pile) would be referred to as the “tower”. However, you don’t want to change the tower or hub height, which would impact the aerodynamic calculations. Instead, you should change the tower draft (input TwrDraft) in the FAST platform input file. The total length of the support structure is: TowerHt + TwrDraft - TwrRBHt.
-
Unless the entire support structure has uniform properties, you will need to change the distributed tower data file. This is because increasing the length of the pile will impact the normalized distance where the pile transitions to the tower. That is, you may not need to change the sectional mass or stiffness data, but you’ll need to change the normalized distance in the first column of the table.
-
“Yes”, changing the length and structural properties of the tower will require a recalculation of the tower mode shapes (using BModes and ModeShapePolyFitting.xls).
In FAST v8, the offshore substructure is not modeled as part of the tower within ElastoDyn, but within SubDyn. You may wish to upgrade to FAST v8.
Best regards,
Dear Jason
Thanks for the really quick reply.
I was indeed writing about FAST v7. I have just finished compiling FAST v8. I just had a quick glance at the documentation.
Please let us go back to my initial question regarding the apparent fixity length.
In subdyn, the pile is clamped to the seabed with all 6 DOF’s restrained (-20m in your example). So, is it possible that the apparent fixity length can be simply added to this seabed value? Will this be the correct representation of the model?
Sincerely
Abhinav
Dear Abhinav,
I’m not exactly sure what you are proposing for your approach, but in SubDyn within FAST v8 it would be easy to add an apparent fixity model. Simply:
- Remove the 6 DOF restraints for the joint at the seabed.
- Add a joint below the seabed with all 6 DOFs restrained
- Connect the joints at and below the seabed with a member with the desired cross-sectional properties of the apparent fixity model.
- Depending on how much the apparent fixity model influences the overall stiffness of the substructure, you may need to update the tower mode shapes within ElastoDyn.
Best regards,
Dear Jason
Sorry for the lack of clarity in my previous query. But you certainly managed to clear my doubts.
Thank You
Abhinav
Hi
How do i find the natural frequency of a complete OWT structure (including tower and substructure) in FAST8?
I am getting a value of 30 m for AF depth, probably since the soil i consider is too soft. I have modeled the pile in HydroDyn and SubDyn. Such a long pile can change the overall structural properties, i believe i should re-evaluate the mode shapes of the tower. But how can one be certain when to change the tower properties? Is it based on the difference in natural frequencies between fixed and AF models?
Is it just required to modify the ElastoDyn_Tower file in this regard? This file just contains the mass densities and FA and SS stiffness properties, but BModes requires more parameters. Is there a different version of BModes for FAST8?
Thanks
Abhinav
Dear Abhinav,
We have plans to include linearization functionality in FAST v8 that will enable direct calculations of full-system mode shapes and natural frequencies via eigenanalysis. Until that feature becomes available, one can identify the natural frequencies by the classical method of post-processing time-domain simulations e.g., by computing power-spectral densities (PSDs) of response time series.
You should rederive the mode shapes of the tower in ElastoDyn (changing only the ElastoDyn tower file) whenever the substructure mass or stiffness has been changed in SubDyn. The SubDyn ReadMe file provides some guidance on how to use BModes to derive the mode shapes needed by ElastoDyn when FAST is used with SubDyn.
Best regards,
Dear Jason
Thanks for the reply. I have read the SubDyn Readme file. It is mentioned that one should substitute the hydro_M and hydro_K matrices from obtained from SubDyn, in the BModes file. I haven’t understood a couple of things.
Are these two matrices obtained from SubDyn by running it as a standalone code, with the modified support structure?
It is also said that i should modify the tower properties accordingly, to get eigenmodes compatible with FAST. Is there a way in which i can validate the correctness of the mode shapes i derive for the tower?
Sincerely
Abhinav
Dear Abhinav,
The equivalent substructure stiffness and mass matrices derived by SubDyn are available in the SubDyn summary file created when input parameter SSSub is set to True in the SubDyn input file, both when SubDyn is run in standalone mode and when SubDyn is run coupled to FAST v8.
FAST is not capable of judging whether the specified mode shapes are correct or not. You could verify that the derived mode shapes are correct by verifying the solution against the results of another code with similar capabilities.
Best regards,
Dear Jason
Thanks for the clarification.
Sincerely
Abhinav
Dear Sir
Is wind.nrel.gov/public/jjonkman/BModes/ the latest version of BModes available? Does the OWT tower transition uniformly in diameter and thickness, over the length of 87.6 m?
Sincerely
Abhinav
Dear Abhinav,
Regarding BeamDyn, “yes”, the version located at wind.nrel.gov/public/jjonkman/BModes/ is the newest version. See the following forum topic for an explanation of the differences between the two versions: Run BModes with Hydrodynamic effect..
There are several versions of the tower for the NREL 5-MW turbine, depending on the substructure type (monopile, tripod, jacket, floating platforms), but in most cases the diameter and thickness vary uniformly along the tower length.
Best regards,
Dear Sir
Bmodes require 7 properties at each section of the tower, for calculating the mode shapes. For the FAST8, Monopile supported tower, only the mass densities and the FA and SS stiffness along the tower are specified in the ElastoDyn_Tower file. How do i calculate the other properties?
I tried to compute the SS and FA mass moments of inertia’s using standard formulae, but got wrong values (i first checked it for the bottom section of the CS_monopile_tower_secs.dat file). I used the formula for a hollow cylinder:
I = 1/2 m ( ri^2 + ro^2)
also, i tried modifying the equation you mentioned in the post BModes flp_iner, for a hollow cylinder, but got nowhere near the correct value. Please help me out.
Sincerely
Abhinav
Dear Abhinav,
FAST’s tower model only takes into account fore-aft and side-to-side bending (with no axial, torsional, or shear DOFs) for a straight beam of isotropic material with no mass or elastic offsets. You are using BModes to derive mode shapes for FAST, so, you want the BModes model to be consistent with FAST. I suggest the following values in BModes when deriving tower mode shapes for FAST:
str_tw = 0
tw_iner = 0
flp_iner = very small number (you can’t specify exactly zero for this input in BModes)
edge_iner = very small number
tor_stff = very high number (BModes won’t allow infinite; just set it high enough that it doesn’t impact the BModes results for the modes you care about)
axial_stff = very high number
cg_offst = 0
sc_offst = 0
tc_offst = 0
Best regards,
Dear Jason
Thanks for the reply. Sorry for disturbing you again!
BModes seems to give me no answers. I have changed the tower properties, replaced the hydrodynamic stiffness and mass matrices with those from the SubDyn sum file and changed the boundary condition to ‘free-free’. But i am getting all eigenvalues as 0, when i run BModes. It seems that BModes expects values for ‘flp_iner’ and ‘edge_iner’. I had used ‘0’ for these parameters, as you suggested. Could you please take a look at my files?
Sincerely
Abhinav
Dear Abhinav,
I forgot that BModes doesn’t like flp_iner and edge_iner to be exactly zero; BModes won’t complain, but it won’t deliver the correct results either. I’ve modified my forum post above to clarify this. A similar discussion took place in the following forum topic: Tower Eigenfrequencies of NREL 5MW Turbine.
I also noticed that you have tower draft (draft and ref_msl) set to 20 m in your BModes input file. If its not, you want this value to be equivalent to the platform reference point in FAST v7 (or in ElastoDyn module of FAST v8).
Best regards,
Dear Jason
Thanks for the reply. Yes, my tower draft value was a wrong one. I have replaced it with the platform reference point value (10 m) from FAST8. I have added very small values for flp_iner and edge_iner and BModes is working now. Based on my understanding of the FAST8 files, i have drawn a schematic diagram of the monopile OWT. Please see the attached drawing.
The tower begins at 10 m above the MSL and ends at 87.6 m. So, in effect, the tower height is 77.6 m? Jason, if i am adding a few nodes below the mudline in SubDyn (to model the apparent fixity), should i add the same nodes in HydroDyn too? Or is the first point below mudline enough, in HydroDyn? I have tried both approaches and the answers don’t differ.
Sincerely
Abhinav
Dear Abhinav,
I’m glad you got BModes to work. Yes, a tower that begins at 10 m above MSL and ends at 87.6 m will be 77.6-m long.
HydroDyn needs at least one joint below the seabed to avoid having a static pressure load applied on the bottom of the structure. But there is no need to add members that are fully below the seabed, as HydroDyn won’t compute hydrodynamic loads below the seabed. Nodes need not match between HydroDyn and SubDyn. See Section 6.5 of the HydroDyn User’s Guide and Theory Manual for more information.
Best regards,
Dear Jason
Thanks for the really quick reply. It has helped me a lot.
sincerely
Abhinav.
Dear all,
I am trying to develop an “apparent fixity model” using FAST v.8 and SubDyn. The AF model has the following characteristics:
• AF length = 16.964m
• Cross-section diameter = 6m
• Cross-section thickness = 0.03457m
Following the instructions presented above in this topic, I modified the existing SubDyn dat file of the 5MW Baseline turbine in order to add the apparent fixity model and I made no changes in any other input file. However, FAST didn’t run and I received a notification of a logical error in SubDyn.
The modified dat file is attached below. Any advice would be most appreciated.
NRELOffshrBsline5MW_Monopile_RF_SubDyn.txt (7.06 KB)
Thank you in advance.
Best regards,
Georgia
