[niklas]

whats there to discretize???

I do not understand what the problem is.

[maritozzo]

Hi guys thank you everybody for the help you gave me, Thank you Bernhard first, you always answered my questions even if my questions were quite silly.
I wrote the solver and it seems work, i have the loss of pressure in correspondance of the space occupied by the porous media... Here is the final form of the velocity equation:

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

I did not changed anything in the rest of the solver leaving it as in the original form of SimpleFoam. may someone tell me if i should change something in the rest of the solver because of the adding of the new term in the velocity equation or I can leave as it is?

[vvqf]

Ì think you can leave as it is.
But I don't understand "fluid through a porous media OUTSIDE it", does it make sense?
Could you explain a little bit?

Did get different G in each time directory?

[liugx212]

Hi Muzio Grilli:

Could you please paste your solver to here, And let us to study that how to solve porous question.

Thanks.
Guoxiang

[nico]

Hi everybody,

Bernhard Gschaider wrote:

By Bernhard Gschaider on Thursday, October 20, 2005 - 10:22 am: Edit Post

Sorry. With "out-of-the-box" I meant "a solver available in the OF-distribution". You'll have to write a solver yourself.

One approach would be to simply extend an existing solver by adding Darcy as a source-term. In the Darcy-term there is a permeability/resistivity (whatever formulation you prefer). By using a field for that and specifying appropriate values for certain regions you can define porous/non-porous-zones.

My Question: How can I define different appropriatet values for different regions? What boundary conditions do I have to choose for the interface between the porous and non-porouszone?

bye nico

[nico]

Muozo Grilli, please post created.H, createG.H, createNu.H -files?
Thanks

[maritozzo]

My createG utility is very slow..you'd better use the setFields utility which is very simple to use, you only have to create the setFieldsDict file which you can find in thedamBreak tutorial and specify the opposite vertices of the box or the center and radius of the sphere in which you want to define the G tensor is really simple

[nico]

The problem is, that it only works, when you use blockmesh. I create my meshes in gambit and then this way doesn't work. It would be a great help for me, if you send me your createG utility.

Thanks
Nico

email : nico.petry@gmx.de

[nico]

Hi!
I have directed permeabilities and therefore I define G as a volTensorField. My problem is, that fvm::SuSp(nu*G,U) doesn't work for volTensorFields and nu*(G & U) does not give the right solution.
I am not quite sure why.

Hope somebody can help me.

Thanks

[maritozzo]

Hi nico the second form nu*(G & U) is the right one. You are using simpleFoam so you must choose in the right way the relaxation factors in the fvSolution file because otherwise the process will diverge and the solution will not be right.
Try diminuishing the relaxation factors.
You have to monitor the residuals to see that convergence is achieved.

[nico]

Thanks.

Is it possible to calculate the relaxion factors, that will bring the right solution? Or is the only possibility trying out?

bye Nico

[nico]

sorry, i mean relaxation factor

[maritozzo]

The default values are 0.3 on p an 0.7 on all the other variables. First you should tell me which is the shape of the G tensor you defined and which are the values you inserted.
Anyway try starting with 0.1 on p and 0.5 on the others.
Then you should also look at the relative tolerances, try starting with 0.01 on p and 0.1 on the others then after a certain number of iteration switch to zero on all the variables

[nico]

Hi Muzio:

[30000 0 0
G= 0 15000 0
0 0 30000]

Why shall I switch the relativ tolerances to zero. I don't understand the sence of it.

Thanks for taking time for my problems.

Nico

[liugx212]

Hi, friends,

Could you please sand the creatG.h and creatNu.h to me to learn?

Thank you very much.

email: gliu@mix.wvu.edu

[guido_adriaensen]

Hello Muzio or someone else,

Sorry to ask for the same files as well. :-) But could you be so kind to share the created.H, createG.H, createNu.H -files? Thanks.
If you can email it to guido.adriaensen@modesi.nl, I would be very gratefull, thanks again

Guido

[michele]

Dear OpenFoamers,
has anyone of you tried to simulate the unconfined seepage problem?
The governing equation is

grad(p)=-rho*nu*U/[K] + rho * [g]

where [K] is the hydraulic conductivity (variable tensor) and [g] the gravity acceleration vector.

Things are complicated by the fact that the phreatic level is not a-priori known, but is part of the solution (equilibrium condition at atmospheric pressure).
For example, the goundflow in the picture below


How the problem can be addressed? Is the customisation of interFoam the unique possibility?

Thanks
Michele.