Hello!
I have a question regarding the loads at the top of the wind tower. I tried the search function of the forum, but I came to different answers in different time frames for different versions of openFAST so I was hoping to clarify this vital point.
I’m modelling the OC4 reference jacket paired with a 5MW NREL wind turbine. I had to simplify my model but the wind loads I extract directly from openFAST and apply at the top of the tower (I use a lumped mass approximation to the Rotor-Nacelle-Assembly with weight and Inertias). My question is, what would be the most correct output channel for loads and moments to be applied at the very top of the tower in x, y and z that would not be affected by some Inertia openFAST may consider?
Dear @Filipe.Rocha,
A similar question was recently asked and answered in the OpenFAST GitHub repository: Calculation of wind turbine aerodynamic forces · Issue #2809 · OpenFAST/openfast · GitHub. Does this discussion answer your question?
Best regards,
Dear Jason, hi! Sorry for my delayed reply but I think I understood which loads I should use. Interestingly there’s a spike in the force value in the direction the wind is dominant (in my case, x direction) at the beginning of the simulation. Do you have some recommendation as to discard these initial values because of some mechanics the rotor may have? I’m using the values of RtAeroFxi.
Dear @Filipe.Rocha,
Perhaps you want to clarify your simulation set up and the type of spike you are seeing. Regardless, we generally recommend ignoring the output of the simulation until the simulation start-up transient has passed before post-processing for response and load metrics.
Best regards,
Dear Jason, sorry for not getting back to you, for some reason my NREL notification e-mails went straight to the spam folder. At the moment I’m developing a code for my PhD thesis considering the OC4 reference jacket coupled with the 5MW NREL wind turbine for an optimization scenario. At the moment I’m modeling the setup as a FEM and applying the RNA loads and inertia at the top of the tower. I’m using OpenFAST as a way of benchmarking the code. That being said, I was trying to match the displacements I get after applying openFAST wind loads to my FEM model. I got to a good convergence as you can see by the attached picture (this is the wind tower top). Now my next challenge is to check the loads produced by OpenFAST on the OC4, apply them to my own modeled OC4 code and check if the displacements also have a good agreement. I found the option in the subdyn file (“OutAll”) to output all the members files, but then I got confused by the documentation saying this:
"For a given element node, the loads are the 6-vector with index 1-6 or 7-12 for the first or second node respectively. By convention, the 6-vector is multiplied by -1 for the first node and +1 for the second node of the element:
(4.171)
Fn16=−Fe12(1:6),Fn26=+Fe12(7:12)
The above applies for the inertial and stiffness loads."
Does this mean that for a given member, let’s say member 1 (connected to nodes 1 and 2 respectively) the forces M1J1FKxe, M1J1FKye, M1J1FKze, M1J1MKxe, M1J1MKye, M1J1MKze, M1J1FMxe, M1J1FMye, M1J1FMze, M1J1MMxe, M1J1MMye and M1J1MMze need to be multiplied by -1 before being applied and the the same forces for the second node are already in correct convention (meaning don’t need to me multiplied by anything).
Dear @Filipe.Rocha,
I would say the reaction load outputs from SubDyn do not need to be multiplied by -1, unless that is the convention you want.
FYI: There is nice discussion about the sign convention used by the SubDyn member-level outputs through a simple example in the follow issue on OpenFAST GitHub: Node forces and moments from beams in SubDyn · Issue #2781 · OpenFAST/openfast · GitHub.
Best regards,
Dear @Jason.Jonkman, thank you very much for the answer and pointing me in the right direction. One last thing, inverting the reaction would make sense to understand how OpenFAST produces the displacements though, correct? Because if I don’t, I get the inverted signal for my displacements compared to what OpenFAST produces (please see attached pictures).
Dear @Filipe.Rocha,
I don’t know what the output convention of your FEM model is to comment as to why the sign of the OpenFAST/SubDyn output needs to be flipped to match.
Best regards,
