Import Mesh from FLUENT/CFX or similar

[Martin]

Hello,

 

I just want to know if it's possible to import mesh structures that I generated with a different simulation progamm like FLUENT or OpenFoam into a FDS File. In the PyroSim-Forum I heard that it is not possible to do that. But maybe you have other experiences. If it is not possible, how do you handle problems connected to these limitations?

 

Best Regards

Martin Rudert

[F-Sim.de]

if it's possible to import mesh structures that I generated with a different simulation progamm like FLUENT or OpenFoam into a FDS File.

No chance.

 In the PyroSim-Forum I heard that it is not possible to do that.

PyroSim-Forum is right.

If it is not possible, how do you handle problems connected to these limitations?

For example what???

[Randy McDermott]

Boris,

Is this because Fluent can't generate a DXF from their geometry file?  I thought if you had a DXF, then you could use Dave Sheppard's DXF2FDS or BlenderFDS or something like this to create the OBST lines, no?  Sometimes people have the DXF even before they create the Fluent geometry.  So it may not really be a question of needing to go directly from Fluent to FDS.

Anyway, regarding the question in the original post about how we handle unstructured geometry, at the moment FDS treats all geometry as Lego blocks aligned with the Cartesian grid.  So even if you manage to convert the DXF to FDS input, you lose fidelity.  We are developing ways to treat higher order surfaces directly, but that it planned for v7, perhaps a couple years away.

Randy

[Mohamed ASSAL]

HI Guys,

I wanna add some clarifications here.

with Gambit you can > import *.igs file mesh it and export it as *.msh file.

*.igs files can be created by CAD softwares like rhyno, Solid works
Pro engineer and Catia.

other way we can creat directly a geometry on Gambit GUI, mesh it and
export it to *.msh. to

the solver Fluent.

2- About Pyrosim, this GUI for FDS CFD code use DXF and STL file.
Randy is right ( FDS

treats all geometry as Lego blocks aligned with the Cartesian grid). a
same for

inclined and curved walls. finaly Pyrosim is GUI Graphical user
interface for FDS.

the Mesh line is created by giving number of cells on X, Y and Z

3-Originally Fds is script of data, so Martin you are not able to
import Mesh generated from

other CFD codes. you need to know the number of cells on X Y Z and
where you should

refine.

Cheers Mohamed.





2012/12/8, Randy McDermott <randy.m...@gmail.com>:

> --
> You received this message because you are subscribed to the Google Groups
> "FDS and Smokeview Discussions" group.
> To post to this group, send email to fds...@googlegroups.com.
> To unsubscribe from this group, send email to
> fds-smv+u...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/fds-smv/-/TLwhWIRwFDAJ.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
>

[F-Sim.de]

Randy,

I talked to Martin before and I think I know what he is aiming at. He tries to simulate very big space and is interested in things like mesh refinement.
I think he is not interested in importing geometry but mesh structures like tetraeders and variable cellsizes. Importing geometry should be no problem.


One question because of your hint to IBM in FDS 7 "perhaps a couple years away". Is AMR also planned for v7 or might this become true in FDS 6.x?

Boris

[Martin]

Yes Boris, you are right, my wish is to refine the mesh locally. But I guess that this is not possible with FDS. That makes it harder to simulate big buildings with a fine mesh :-(

[Mohamed ASSAL]

Hello Martin I don't understand your question and what do you want to
do, refine Mesh locally!!.

do you want something like this?

&HEAD CHID='Mesh' , TITLE='Local Mesh.' /
&MESH IJK=100,100,100, XB=0,100,0,100,0.0,100 /
&MESH IJK=100,100,100, XB=10,50,10,50,0,100 /
&TIME TWFIN=0.0 /
&TAIL /

The default Grid for fds is cartesian. please give more clarifications.

Cheers Mohamed.




2012/12/10, Martin <martin...@gmail.com>:

> --
> You received this message because you are subscribed to the Google Groups
> "FDS and Smokeview Discussions" group.
> To post to this group, send email to fds...@googlegroups.com.
> To unsubscribe from this group, send email to
> fds-smv+u...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msg/fds-smv/-/1IuJ36SpqB4J.

[Jamie]

Martin,

It is possible to do some refinement in FDS using transforms, or with creative use of multiple meshes.

Keep in mind, FDS uses a rectilinear grid.  No tet meshes, etc.

While that can at first seem restrictive after using Fluent, you will find that it saves you loads of time not having to run through Gambit or some other mesh generator.  You'll also find it to be a much faster code for a lot of fire applications, so you might find you can run more cells with FDS than with Fluent.

[F-Sim.de]

It is possible to do some refinement in FDS using transforms,

Yes, but with some restrictions. As far as I know just in 2 of 3 dimensions...

 

or with creative use of multiple meshes.

 
Embedded meshes not with a full interaction and connected meshes with changes to VELOCITY_TOLERANCE and the number of iterations.

AMR and IBM will be something like a final solution for the restrictions remaining here.

[Mohamed ASSAL]

Martin

I suggest you to add a sample input.

Tell us about the refinement region.

Mohamed.

2012/12/10, F-Sim.de <In...@f-sim.de>:

> --
> You received this message because you are subscribed to the Google Groups
> "FDS and Smokeview Discussions" group.
> To post to this group, send email to fds...@googlegroups.com.
> To unsubscribe from this group, send email to
> fds-smv+u...@googlegroups.com.
> To view this discussion on the web visit

> https://groups.google.com/d/msg/fds-smv/-/CAS8pACRfmAJ.

[Randy McDermott]

Boris,

I foresee AMR development proceeding in two phases.  The first will be adding the capability to handle two-way coupling with embedded (nested) meshes.  This could conceivably be worked into a 6.x release.  For better or worse, I have more pressure right now to get IBM into a workable state.  Once that is done I can refocus on embedded meshes.  The second phase of development for AMR will be true automatic refinement.  Let's just say this is v7, at least.  Could even be v8 :-).

Cheers,

Randy

[F-Sim.de]

Randy,

thanks for these informations.
From the viewpoint of an user: both, AMR and IBM, are real killer features and will solve the restrictions we have today. Some years ago I visited a symposium on CFD at Technical University of Dresden. They have their own chair for fluid dynamics and numerical simulation. At the very end of the symposium, Prof. Fröhlich looked into the far future of CFD and what he saw was IBM and LES. It is amazing that FDS users can use this feature in the not-so-far future.
AMR, no matter if full automated or manual adopted, is something like the holy grail of LES. A quantum leap in qualityfor all users...