Hello, I´m working on a project to ensamble Nastran and fluent to get F&S interaction. The way i do is taken my structure with his forces from fluent, to get the deformations by nastran, and then get the new mesh deforming the old one like hundreds of springs using, conjugate gradient method to get the new mesh. if we close this into a loop we get a form of working with f&s interaction.

All of this has a dificult in the mesh numeration, one of each program works diferent on a way. the gambit-fluent one, the nastran one, and mine of the conjugate gradient. i use a pair of programs that translate the meshes for any step of the loop. but the one i have more problems is to extracting from fluent. When i export my surface that form the structure, the numeration of the nodes change, and the surface is renumerated again. Thats is when i use the export way. The question i have is there is another way of export the surface forces and node position whithout renumeration takes place, and if nastran or marc will admit a non consecutive numeration and nonstarting on zero. Maybe i could do a userfunction to do that?.

Thank you

Javier

The easiest thing, of course, would be to do both sides within the same software. I am aware of quite a lot of solid-fluid interaction work in FOAM (http://www.nabla.co.uk) where both solid and fluid were done in a closely coupled manner and using autimatic mesh motion.

If this is not an option, there were several projects in the past specifically for the prupose of coupling popular commercial fluids and stress analysis codes (Ansys was definitely one of them). Unfortunately, I cannot remember the name, but it sounds just like what you need. somebody else on this forum may remember or may have worked on the actual project...

Good luck,

Hrv

CFDRC or ESI Software, can also do it.

http://www.cfdrc.com

http://www.scai.fraunhofer.de/index.php?id=222&L=1

That's the baby! Thanks Jorn.

Hello, I work with Godunov's method for both Fluids and for Structures. In this case Fluid Structure interaction is rather simple. on the boundary FLUIDorGAS-STRUCTURE I use Riemann solver (it is itterative, structure is elastic) in Lagrang to calculate the movement of the boundary line. If you construct Euler mesh on the boundary after its Lagrang movement as on the line (2D) or surface (3D) and then Euler in your FLUID or STRUCTURE it is necessary to take care about appropriate fluxes. regards abouziar