FSI issue in abaqus co-simulation: misaligned nodes

[rrieger]

Hi everyone,

I am performing FSI simulation under abaqus.
The simulation consists in a porous solid (elastic isotrope and homogeneous under Abaqus/explicit) filled by a non newtonian (viscous fluid under Abaqus/CFD). The whole is contained in a cubic shell in order to apply proper boundary condition. The cube is pinned on one face and undergoes a compression displacement on the opposite face. I have defined an interaction condition between the solid surface and the fluid model surface so Abaqus can exchange information during Explicit and CFD simulation.
The simulation runs well until I diminish the element's size, then the simulation doesn't start and give me the following error message:

***WARNING: When performing co-simulation mapping from the partner code (the
source) to /Explicit (the target): search failed for the following
target elements:
These nodes are written to the output database (odb) in nodeset
COSIMULATION_ENGINE_SEARCH_MISALIGNED_NODES

I looked into the Abaqus documentation which gives the advise to put a skin on the solid geomtry in order to paste the mesh on the 2nd model geometry. This works only for a co-simulation between Abaqus/Standard and Explicit, but the skin option does't seem to exist for the fluid model.
I also tried to define exactly the same size of element in both solid and fluid model, but I still have the same error message. Also the explicit simulation performed alone works fine, so seems to indicate that the problems remains in the interface between the solid and the fluid.

Have you guys encountered the same kind of problem or could you give me some advice in order to solve this problem? I would very much appreciate.
Thanks in advance.

Romain

 

[PedroCG]

Hi Romain,

Were you able to solve your problem? I'm getting the same error, I've structured the Input file based on the tutorial that comes with STAR-CCM+ (which runs fine).


Regard's,

Pedro.

 

[rrieger]

Hi Pedro,

Yes actually you have to be careful not to select the surface shared by solid parts. Make sure your FSI surface is only shared between the solid et the fluid.

Sounds logical but you have to think about it.

Regards,

Romain

 

[PedroCG]

I only have one solid part, I tried splitting the surface but the error persists and the elements that don't get mapped are the same, I then tried to refine the solid mesh (making it the same size of the fluid mesh) but got the same error.

Do you know if the fluid region needs to surround completely the solid? My problem is 2D but I'm using a one cell thick mesh and it's the only significant difference to the tutorial case.

Thanks for your help,

Pedro.

 

[PedroCG]

Well turns out my guess was right... one cell thick meshes don't work. Now it's just a question of how little cells can one put across the thickness. It's curious to see that all vertexes got mapped properly, I'd expect for the outermost ones not to be and I was ready to attribute that to the shared boundary condition (symmetry)... All that matters now is that it's working!