FSI Questions: Slamming of wedge

Andreas_21549

Hi!

I'm trying to learn more about setting up FSI analyses in HyperWorks by coupling Radioss and AcuSolve. I first want to start with a simple case, which is a 2D wedge impacting water, and I've tried to model this as best I could. There are a number of issues I stumble upon, but I lack too much of the knowledge of the software to even understand why there is a problem there. I have attached the solverdeck files to this thread. If any of you feel like being a very good samartian, now is a golden opportunity to highlight my mistakes and misunderstandings. It would be highly appreciated!

I can perhaps start off with a few questions:

I find boundary condition treatment quite different from what I'm used to with FVM based codes. I would ideally model only half of the wedge and domain, and add a symmetry plane to mirror the solution, but I'm not sure of how to do this. I therefore modelled the full domain, and have now made a part called 'Boundary' which includes all the surfaces except the wedge. I made a surface mesh, and made a LAW51 material with '3: Outlet boundaries to Multi-Material Law' and ALE CFD Formulation set to Eulerian. Is this sensible? Or do I need to make a solid volume mesh instead of a surface mesh? The reason I ask is that I get a huge number of errors when I try to run the file in the HyperWorks solver. Nearly all of them are connected to the elements, for example:

ERROR ID :    62
** ERROR IN MATERIAL/ELEMENT DEFINITION (LAW)
DESCRIPTION :  
   -- MATERIAL ID :    2
   -- MATERIAL TITLE : Boundary
   INVALID MATERIAL LAW 51
   FOR SHELL ELEMENT ID=963

It doesn't make much sense (to my mind) to make another volume mesh which also includes the fluid domain, just in order to get the surface boundary conditions right - but is that the right thing to do?

 

Maybe we can start there, and I can follow up with more questions as we go along.

Best regards,
Andreas.

Mathis

Hi Andreas,

The error you encountered is due to the fact LAW51 is only compatible with solid, hexa elements.

 

So, in order to model your Non Reflecting Frontier (or "boundary"), it is recommanded to use LAW51 with iform=6. And you should mesh like this:

 

It is advised that each outlet element share one face with one element from the ALE domain(s), so forner the corner, you should do it like this:

 

If you want just to model a symetry plane, then only define /BCS and block the directions according to the symetry plane you want to model.

 

Regards,

Mathis.

Andreas_21549

Thank you! That helps. Where did you find this info? I can't seem to find this in the Radioss 2020 user guide.

Mathis

Hi Andreas,

It can be found on the Radioss reference guide. In law51, iform=6:

regards,

Mathis.

Rogerio Nakano_21179

As a side note, you can run the FSI coupling with Optistruct and Acusolve. But not possible RAdioss Explicit and Acusolve as for now.

Regards