[bannari]

In Fluent code, there is a possibility to recover the number of a zone of the domain (Zone_ID) with (THREAD_ID(t)=number of zone).
I will Know if ther's a function which makes the same thing in OPENFOAM?
if No, How can I do this?
thanks a lot

[bannari]

it is possible to have plus than one zone in the domain with different properties?
and how can I recover the pointer addressed to each zone ?

[mattijs]

No there is no standard way of having different zones in a mesh. You'll have to program it yourself.

[hjasak]

... the idea being that each "zone" is in fact a mesh itself and you then deal with mesh to mesh coupling instead of chopping up a perfectly good mesh you started from.

If you just want to have different properties in various part of your domain but still solve the same equations everywhere, you can specify a spatially varying properties without using zones.

Enjoy,

Hrv

[bannari]

thanks for you,
but I dont solve the same equations in the different regions.
So I will tried to program this...

[bannari]

thanks for you,
but I dont solve the same equations in the different regions.
So I will try to program this...

[panara]

look in the tread 'Heat transfer with solid elements (conduction)'

you may find some ideas

Daniele

[fra76]

At first, sorry for the very long post, but I'd like to explain my problem clearly.

I've developed a model for porous media within OpenFOAM 1.3, paying attention to respect the programming conventions used in the code. It's based on Brinkman's equation, with the inverse porosity tensor defined as tensor::zero outside the porous media, and with a value defined by the user using a new dictionary file.
I can specify an arbitrary number of porous "boxes" within the mesh, each one with its own properties. And that's very fine!

The equation I solve is derived from the standard simpleFoam:


tmp<fvvectormatrix> UEqn
(
fvm::div(phi, U)
+ turbulence->divR(U)
+ (G & U)*nu
);

UEqn().relax();

solve(UEqn() == -fvc::grad(p));


This works quite well, but I have to choose very low underrelaxation factors to reach convergence, when the values of G are high (The eigenvalues of G, derived measuring the pressure lost between the inlet and the outlet of the porous media and comparing it with experimental data, are about 1.e+8 m^-2, for a velocity of about 10 m/s).
Keeping underrelaxation factor so low (even 0.001 or less) requires too much computational time, of course.
I guess the instability problem is produced by the discontinuity of the [G] field between the porous region and the outside.

So, I'd like to solve this equation only within the porous region and the standard Navier-Stokes for the flow outside it, coupling somehow (explicitly or implicitly) the regions.

In the thread named "Heat transfer with solid elements (conduction)" Henry Weller gave some hints on the way to do something like this, like transfering boundary condition explicitly between the two zones.
Mattijs Jansen posted a twoMeshes version of simpleFoam, which I'm studying to learn how to threat two different meshes.

The questions are:
How can I define two ore more different meshes within one case? I mean, while creating the mesh and importing it in OpenFOAM.

How can I couple the regions within the SIMPLE loop?

Thanks a lot for your patience!
Francesco.

[melanie]

Hi Francesco,
I wouldn't be able to help you improve your solution, but I would be very interested if you share your code or explain a little bit more how you implemented it... I'd like to develop a model for heat exchangers, and I'm quite sure that your porous media is close to what I'd need !
thanks
melanie