dolfin team mailing list archive

["Johan Hoffman" <jhoffman@xxxxxxxxxx>]

> Dear everyone,
>
> I am trying to simulate a fully inflated parachute fixed with cables and
> the
> air flow around it.
>
> I am new to gmsh, dolfin and unicorn. In my understanding by learning this
> week, in order to simulate an extremely large deformation fluid-structure
> interaction problem, I need to use gmsh generate parachute and surrounding
> volume mesh separately. Then, use python-convert convert the *.msh file to
> *.xml so dolfin can read it. Then, I can use the package unicorn developed
> by TACO group to do the FSI couple work. I did not go deep into each part
> to
> understand them better yet.
> I think I need your help to make sure I am on the right track.
>
> Thanks a lot,
> Zhanjun

Dear Zhanjun,

Yes, you are on the right track! Some things to consider along the way:

* The fsi solver implemented in the current version of Unicorn is based on
an incompressible elastic solid model together with an incompressible
fluid model. In particular, you'll need to give your parachute a thickness
to treat it as a solid (not a shell/membrane). The same goes for the
cables, unless you model them by forces.

* The fsi solver uses a global Unified continuum formulation, so that the
full fsi problem is solved on one single mesh (parachute + air), so you'll
only need one mesh for the full fsi-domain, but with markers indicating
what is solid (parachute) and what is fluid (air).

* I am not sure what you mean by "extremely large deformations": but the
solver should be able to handle fairly large deformations so long as there
are no topology changes.

* It sounds like you will need a huge number of degrees of freedom for
accurate modeling of this problem. In an upcoming release of Unicorn (in
1-2 months) there will be support for fully distributed parallel
computing, but today there is only a serial version. You may use the
recently developed shared memory parallel tools in DOLFIN to speed up the
computation, but memory may still be an issue.

Good luck!

/Johan

> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev
>