Age calculation problem

[bernardombraga]

I've tested the new features of Baram 25.1 in Windows 10. I've gotten the age variable after a steady-state solution was obtained for an internal flow problem. However, the result was bad:

I'm not sure, but I suspect that the problem arises from the use of a non-bounded advection scheme ("Gauss linearUpwind momentumReconGrad"). Since scalars also use the same scheme, their solution may have the same problem.

[kdg5424]

Hello @bernardombraga. This is Donggyu Kim.

Could you check your flow result and mesh?
Of course, numerical scheme can affect the Age calculation results.
But I guess in this problem it's not important. So could you check your mesh and flow results?

Best Regards

Donggyu Kim

[bernardombraga]

Hello Donggyu Kim,

The mesh was created for an old project using Ansys Meshing. I exported it as a .msh file and imported it into Baram:

The output of the checkMesh command was recorded in the file below:
checkMesh.txt

The mesh seems ok.

The flow solution converged:

I also monitored the inlet pressure and the mean velocity to check convergence:

The values of the monitors are in good agreement with the result that I obtained in the past using Ansys CFX.

The distributions of velocity and eddy velocity (nut) also seem reasonable:

That's why I suspected of a numerical issue.

Best regards,
Bernardo.

[kdg5424]

Hello @bernardombraga.
It seems very weird. Could you let me know how you made the settings about age?
In our test, using Gauss linearUpwind momentumReconGrad made the calculation converged.
Why don't you change the dictionary and try it on terminal?
You can test it on terminal and also can check the residuals

Best Regards

Donggyu Kim

[bernardombraga]

Hello @kdg5424,

In my first post, I ran age calculations using the BaramFlow GUI (Results -> Reports -> Collateral Fields), so I used the default Baram settings. Since the calculation result was unsatisfactory, I looked for the divScheme used in the fvSchemes dictionary.

I don't know how to run NextFOAM on the terminal from Baram installation in Windows. I suppose this requires setting some environment variables and/or folder paths to PATH. Could you instruct me?

Then, I calculated age in BaramFlow from the Numerical Calculation tab (Advanced button). This way, the convergence history is shown in the Console:

So, I identified the cause of the problem. The solution of the age equation was not converged and the Reports tab doesn't inform that for the user.

Finally, I edited the fvSchemes dictionary to use the upwind scheme for age:
"div(phi,age) Gauss upwind;"
And I calculated age from the Reports tab. The result was reasonable:

I can't test the upwind scheme from the Numerical Calculation tab since BaramFlow modifies the fvSchemes dictionary in this case to use the "Gauss linearUpwind momentumReconGrad" scheme.

Best regards,
Bernardo.

[bernardombraga]

In another test, I modify the fvSchemes dictionary:
div(phi,age) bounded Gauss linearUpwind momentumReconGrad;

And the fvSolution dictionary:
relaxationFactors
{
...
equations
{
...
age 0.7;
ageFinal 1;
}
}

The result was:

[kdg5424]

Hello @bernardombraga.
I forgot to instruct you how to use BARAM solver on terminal. Sorry for that.
You can use it on windows cmd with this command.

set WPD=BARAM INSTALLATION PATH\solvers\openfoam
set MGW=BARAM INSTALLATION PATH\solvers\mingw64
set WM_PROJECT_DIR=%WPD%
path %MGW%\bin;%MGW%\lib;%WPD%\bin;%WPD%\lib;%WPD%\lib\msmpi;%Path%

Well I guess the numerical condition might be the reason.
I'll discuss it with other engineers.

Best regards

[bernardombraga]

Hello @kdg5424 ,
Thank you for the instructions. It worked fine.

[kdg5424]

Hello @bernardombraga.
We decided to change the numerical scheme of age.
As you told, it will be changed to bounded Gauss linearUpwind momentumReconGrad;

I'm very thank you for your suggestion

Best Regards

Donggyu Kim