[Coris]

I am having trouble in defining the correct boundary conditions for an external natural convection problem. In this context, external is that the domain is relatively large and open at several boundaries. At these open boundaries, I don't know which pressure BC I have to apply.

The example case is in the attachment, along with a drawing of the setup. The case can be run using sh buildMesh.sh and chtMultiRegionFoam.

I am looking for the boundary conditions for p_rgh at the Inlet and Outlet. I have tried various combinations of totalPressure, prghPressure and prghTotalPressure, but none seem to work (mostly non-converging cases).

The only solution I see is the new prghTotalHydrostaticPressure BC added to fireFoam: https://github.com/OpenFOAM/OpenFOAM...0c819854ed27aa
but it is not yet available for chtMRF (http://bugs.openfoam.org/view.php?id=2132)

Any thoughts ?

[Bloerb]

I do not think this is a boundary problem. I did the following:

Running this with a courant number of 5 is crazy for this kind of simulation. A value above 1 should only be chosen if your boss tells you he wants pretty pictures by yesterday. I lowered that to 0.1. Also lowered your initital dT. In your schemes I changed div(phi,K) to upwind for good measures. I changed your pressure outlet to zeroGradient activated your heat source and it seems to work on first glance.

Also:
What are these for? Are you using a modified solver or am I missing something? Those shouldn't do anything with this solver.

    hydrostaticInitialization yes;
    nHydrostaticCorrectors 5;

[Coris]

Originally Posted by Bloerb

I do not think this is a boundary problem. I did the following:

Running this with a courant number of 5 is crazy for this kind of simulation. A value above 1 should only be chosen if your boss tells you he wants pretty pictures by yesterday. I lowered that to 0.1. Also lowered your initital dT.

Thanks a lot! That probably explains why a lot of my simulations did not converge at all.

Originally Posted by Bloerb

I changed your pressure outlet to zeroGradient activated your heat source and it seems to work on first glance.

So I have 2 zeroGradient patches, but then I don't have a reference for the pressure?
I tested this, and it diverges..
I am getting really high temperatures and the simulations breaks off.

Originally Posted by Bloerb

Also:
What are these for? Are you using a modified solver or am I missing something? Those shouldn't do anything with this solver.

    hydrostaticInitialization yes;
    nHydrostaticCorrectors 5;

At one point, I made the effort to incorporate the enhancements from fireFoam to chtMRF. These are the settings needed for that.

[Bloerb]

I also lowered the value for your heat source. 100W on such a small area seemed a bit high. I put that down to 1W. Most likely not necessary. I have run it for 10 seconds and the results were plausible. Velocity rising up on the heated wall, as well as rising temperature. You could also set a fixedValue for p_rgh on your inlet. It seems far enough away.

[Coris]

I did exactly the same and I am having these issues at ~ 1.05 s..

Could you post the case with your changes ?

[Mondal131211]

Originally Posted by Bloerb

I do not think this is a boundary problem. I did the following:

Running this with a courant number of 5 is crazy for this kind of simulation. A value above 1 should only be chosen if your boss tells you he wants pretty pictures by yesterday. I lowered that to 0.1. Also lowered your initital dT. In your schemes I changed div(phi,K) to upwind for good measures. I changed your pressure outlet to zeroGradient activated your heat source and it seems to work on first glance.

Also:
What are these for? Are you using a modified solver or am I missing something? Those shouldn't do anything with this solver.

    hydrostaticInitialization yes;
    nHydrostaticCorrectors 5;

Hi Bloerb,

Hope you are doing fine. Though it is a relatively old post, currently I am trying to solve a problem pretty much similar to that discussed in this thread. I found you suggested using zero gradient P_rgh boundary condition for the outlet boundary. After having a lot of hectic tries, I followed your steps and found a pretty good solution with this BC. I validated my model with the experimental result as well. However, the reason for writing this message to you is, is it physical to use zero gradient pressure at the outlet boundary? It might be possible, but it beyond my knowledge. I am very interested to know the reason of using a zero gradient pressure at the outlet as I want to stick with this pressure BC combination. This is what I was looking for. Is it possible to give me a brief explanation of using zero gradient pressure at the outlet? I would really very grateful to you.

Or if anybody in this group can answer my question, would appreciate it.

Regards,
Razon