Jacket analysis using FAST and Abaqus

Dear all,

I am doing an undergraduate project in which I have to analyze offshore wind turbine jacket support structure. I am using the reference jacket from FAST Test21. I have completed FAST simulations with my environmental conditions and now I need to do further analysis of the jacket using FE software Abaqus. I have read a few papers where similar jackets are analyzed taking tower top or tower bottom reactions and calculating/applying hydrodynamic loads on the jacket separately. However, my goal is not to compare the results obtained using SubDyn and FE software, but to represent the results from FAST and do the further analysis using Abaqus. My question is if it is possible to represent FAST results in FE software using only FAST outputs? Which outputs should be used?

I have tried to use reactions at the top 4 nodes of the jacket and reactions at the bottom 4 nodes, but I also had to fix the nodes at the bottom and the response of the jacket was different.

Thank you for any thoughts.

Kind Regards,
Dainius

Dear Dainius,

Perhaps I’m not understanding your question. What do you mean when you say that you want to represent the results from FAST within ABAQUS?

Best regards,

Dear Jason

I am sorry for unclear questions. Ideally, I need to apply identical loads to the jacket in ABAQUS as they were applied in FAST simulations (aerodynamic and hydrodynamic loads). As I understand by applying reaction forces at the top of the jacket I capture all the wind and gravity loads from everything that is above. However, it seems that in the recent FAST version I cannot get the hydrodynamic loads at each node of the jacket from HydroDyn. Should I calculate them manually using Morison’s equation or there is another way to capture these loads using any outputs from FAST?

Kind Regards
Dainius

Dear Dainius,

While FAST/HydroDyn is not currently set up to allow you to output the hydrodynamic loads at all nodes throughout the structure, these loads are certainly calculated internally and you can modify the source code to output what you need.

However, you must realize that the hydrodynamic loads, as well as the reaction loads at the top and bottom of the jacket, depend on structural motions (deflections) of the jacket, so, if you apply these loads to your ABAQUS model, the loads will not be correct if the ABAQUS motions are not identical to those of FAST.

What further analysis are you needing to do in ABAQUS?

Best regards,

Dear Jason,

Thank you for your response.

In ABAQUS I will have to analyse the deflections of the jacket, highest stress locations on the joints and members. Furthermore, I may try changing jacket dimensions or adding extra members and observe the change in jacket response. The official project objectives are as follows: “To learn FAST and ABAQUS to characterise the uncoupled load effects on the support structures. To study the performance of these structures based on the analysis and to highlight the current limitations.” Also, I am analysing only Ultimate Limit State response, so no fatigue calculations are needed.

How could I avoid or decrease the error due to differences in jacket motions? Is it a better idea to design transitional piece and tower in ABAQUS and use only the forces at tower top from FAST?

Kind Regards
Dainius

Dear Dainius,

Simply taking the tower-base or tower-top loads from FAST as input to ABAQUS for the jacket design/analysis is a one-way coupled approach that is not common in the offshore wind industry.

The most common approach in use by the offshore wind industry is the sequentially coupled approach whereby the jacket is designed and analyzed in an offshore design tool e.g. ABAQUS, which also exports a superelement for use in an aero-elastic code e.g. FAST for coupled aero-hydro-elastic simulation. The loading at the interface (e.g. tower base) is then passed back from the aero-elastic code to the offshore design tool for stress recovery and code checks.

Another approach that is being applied is the fully coupled approach, whereby the coupled aero-hydro-servo-elastic simulations are performed in one tool e.g. FAST. and only the structural reaction loads are output for import into an FEA tool for stress recovery and code checks. This is discussed e.g. in the following forum topic: http://forums.nrel.gov/t/stress-analysis-of-blades-and-a-tower/489/13. However, if you are changing the design of the jacket in the FEA solution then this would likely require updating and rerunning the coupled aero-hydro-servo-elastic FAST analysis apriori.

I hope that helps.

Best regards,

Dear Jason,

Thank you, your explanation made everything much clearer.

I have a few questions regarding the second approach you mentioned:

  1. Just to clarify: the structural reaction loads mentioned in your response are the 3 forces and 3 moments from SubDyn at both ends of each member of the jacket?
  2. In the forum topic from the link, you mentioned: “These outputs are then used to feed FE-based models of individual components for calculations such as stress, strain, and buckling.” What is an individual component in my case? Full jacket or individual members of the jacket?
  3. I understand how these loads could be used to analyze the tower or single member of the jacket, but for full jacket analysis, it looks a bit more complicated. So I assume I would need to write a code as it would be quite unrealistic to do all the calculations by hand. Do you agree?

Thank you very much, again.

Kind Regards
Dainius

Dear Dainius,

Here are my answers to your questions, but I cannot answer in detail as I haven’t done this myself:

  1. Yes, or wherever you wish to calculate stress or code checks.
  2. Individual members of the jacket.
  3. I’m not sure I understand your question, but you can use the OutAll option in SubDyn to output the three forces and three moments at the end node of each member connected to all joints.

Best regards,

Thank you for your help!

Dear jason
According to the message, how to import a superelement approach to tower base in FAST 8?

Besides, how to customized the code to add distributed springs to represent the soil conditions?
I have reviewed the SubDyn theory basis from section 6 of the draft SubDyn User’s Guide and Theory Manual.
Does the method resemble to model soil stiffness/damping in FAST v7 offshore turbines through the user-written UserPtfmLd() routine?
And how to change this boundary condition in SubDyn before adding distributed springs?
Please give me some suggestions.
Thanks for your help again.

Best Regards
jason.Lai

Dear Jason,

Since the release of FAST v8.16, we’ve added the option for a new user-platform module (ExtPtfm) that will allow you to implement user-specified 6x6 mass, damping, and stiffness matrices, and a 6x1 load vector to be applied to the platform in ElastoDyn (or a more complicated user-specified platform load). This feature will allow you to import a superelement (expressed in terms of mass, damping, stiffness and load) for use within FAST. The new release of FAST–with this and other features–is imminent.

Regarding soil-structure interaction, this has been discussed a few times on this forum e.g.: http://forums.nrel.gov/t/foundation-stiffness-and-damping-in-fast-v8/1059/1.

Best regards,

Dear all,

I am doing a similar project to Dainius and am wanting more information on the one way coupled approach. I am new to FAST and am wondering if this approach simply involves applying the tower base moments from the FAST output file to the ABAQUS model to capture the environmental loading?

Many thanks,
Rory

Dear Rory,

If you simply take the time series loads output from FAST and apply them to the top of the substructure in ABAQUS, I would be afraid of a couple innacuracies e.g. (1) the natural frequencies of the support structure (and consequent influence on the loading) would not be correct and (2) the damping of the support structure (especially the aerodynamic damping, and consequent influence on the loading) would not be correct. Again, the one-way coupled approach is not common in the offshore wind industry.

Best regards,

Dear Jason,
I have used “outAll” option in subDyn in order to result three forces and moments at all joints in the jacket which include 112 members for OC4 jacket and 224 joints consequently.(A 10-min simulation was implemented.)Now there is something vague for me and that is why the static forces are changed within each time step since the static force is calculated using F=[k]*u which is time independent.Might there be some considerations which have been out of my sight?!
Best regards,
Arsalan Shahmohammadi

Dear Arsalan,

The nodal displacements (u in your equation) are time varying based on the instantaneous deflection of the substructure. See section 6 of the draft SubDyn User’s Guide and Theory Manual for more information: wind.nrel.gov/nwtc/docs/SubDyn_Manual.pdf.

Best regards,