3D external periodic multi-zone flow with RANS-based Turbulence model
 

Hello

Dear Developers,

Thank you for new release along with educational version which is a great idea.

Recently I am studying wind turbine performance in wind farm. Consequently for numerical study of the concept I decided to use SU2 for some reasons. Although, for better explaining the concepts, a pathline picture have been attached from a simulation of mine albeit with other solver. Following this further, I am defining stationary zone and rotational zone periodically in my grid. First point that I am wondering about is it possible to have stationary and rotational zone in one simulation using SU2's rotational frame options in the last version? anticipating your answer, then should we define interface boundary condition between two zones?
The second point is that numerical study for such a concepts confront with heavy computational cost which enforce me to simplify the simulation by using conformal and non-conformal periodic boundary condition. Consequently, I need to use parallel processing with these periodic boundaries, stationary and rotational zones along with RANS equations. Already, I found out that periodic boundaries could be used in parallel in version 2.0.2 and rotational frame could be used with Euler equation in serial in version 2, but is it possible to use all the aforementioned capability (periodic boundaries, stationary and rotational zones, RANS equation and parallel) in later versions like 2.0.5 or 2.0.8 or even 3.0.0?
If no, how can one develop these capabilities in the code?

Thanks ever so much for your consideration. I really do appreciate your hint.

Bests,
Payam

Dear Payam,

Thanks for your interest in working with SU2: your problem is related to some of the topics that we have been working on over the past year or so (multi-zone, rotating frame calculations, wind turbine applications, etc.).

At this time, we have some of the infrastructure that you will need; however, in order to have a fully functioning framework, some new development will be needed. In particular, a method for combining the two zones will be required, either through a mixing approach or sliding interface. We have worked a little on sliding interfaces in the past, but we have frozen the development on that feature for now. This would be a nice contribution, if you are interested in getting involved in code development.

Another interesting approach is to use actuator disks for modeling this type of interaction without worrying about physically moving or rotating multiple zones in the mesh. This is another area in need of some development (we have some folks working on this a little internally).

Lastly, you might be interested in some of the other unsteady/moving wall type problems that were discussed in our last major workshop. If so, please check out the notes found here: http://su2.stanford.edu/documents/SU...edAnalysis.pdf.

All the best,
Tom

Thank you for your leading advice. I am interested in the code development for moving wall boundaries.
Thank you for the file, I am remembering your last major workshop.

I believe that it would be easier for me to start with rotational reference frame with sliding interface approach for combining two zone. However, do you have plans for developing dynamic mesh for cases that cannot simplify by using rotational frame or sliding zone like cavity pump with non-symmetrical rotor?
Although, It needs to be re-meshed every time step which means another module for re-meshing, It will support other applications with moving boundaries.

As I mentioned, we had a preliminary implementation of a sliding mesh capability, but we paused development on that feature until we have a particular project that requires/supports it. There aren't any new capabilities in this area that will be available in the immediate future.

In the meantime, please let us know if you are developing the code and have any questions.