Dear @Jason.Jonkman
I hope this message finds you well.
I am currently studying aeroelasticity and using OpenFAST for research purposes. I would like to confirm my understanding of how aeroelastic coupling is handled within OpenFAST.
From the documentation and related literature, it appears that OpenFAST employs an explicit, loosely coupled (non-iterative) time-marching scheme between the aerodynamic and structural modules, rather than a fully strongly coupled (iterative) approach where loads and displacements are iterated to convergence within each time step.
Could you please confirm whether this interpretation is correct?
Also, if you dont mind, could you please suggest references about aeroelasticity (i.e. NREL technical reports/papers) ?
Best Regards,
Riad
Dear @Riad.Elhamoud,
OpenFAST v4 and earlier uses an implicit loose coupling scheme, which is iterative (using a predictor-corrector scheme) and involves implicit coupling wherever there is force-acceleration interaction (e.g., for structure-structure coupling or hydrodynamic added mass), which requires that a constraint equation be solved. The term “loose” implies that each module has its own time-integrator/solver. When AeroDyn is enabled for a wind turbine, explicit loose coupling is used, which is still iterative, but does not have specific force-acceleration interaction where a constraint equation must be solved. This algorithm is numerical stable and robustly converges to the correct solution, but often requires smaller than desired time steps when BeamDyn or SubDyn are enabled.
In the upcoming release of OpenFAST v5, we have added a tight coupling option to avoid the small time-step requirements when BeamDyn or SubDyn are enabled. By “tight”, we mean that in this version, the structural modules (BeamDyn, ElastoDyn, SubDyn) all share the same time-integrator that is now at the glue-code level rather than within each module. This results in a 10-100x speed up when BeamDyn or SubDyn are enabled. Other modules are coupled similar to how there were before.
We have drafted a paper that explains the original loose and new tight coupling algorithm and plan to submit it to a journal when OpenFAST v5 is first released, hopefully in the next month or so.
In the meantime, you can review a couple of our older papers and implementation plans available here: 4.1. General — OpenFAST v4.1.2 documentation.
Best regards,
1 Like