Mesh orientation

[lamia]

Hello,

I have convergence issues with my case : complex geometry but very simple physics (an inlet with imposed flow, 2 symmetry, free outlet and a wall with constant heat exchange for the moment)

My calculation diverges and looking more closely at the listing file I find outliers: exploding speeds, negative temperatures ...

I also noticed that my inlet flow is negative and that the outlet is detected as free inlet (and not free outlet). I think the problem comes from the mesh orientation (done under gambit) but I'm not sure and I don't know how to be sure. and if this is the case, would it be sufficient to check the reorient box when loading the mesh in code_saturne gui ?

Thank you

Regards

[Yvan Fournier]

Hello,

Yes, fixing orientation when possible simply requires checking the bax. But even when orientation is not fixed, it is always checked, and if incorrect, usually leads to negative volumes, which stop the code immediately.

So convergence issues are probably due to mesh quality / numerical option combinations.

The outlet being detected as a free inlet and negative inlet flows are more surprising. Do you seen this in the run_solver.log or when visualizing the results. Did you double or triple-check the group names and boundary condition assignment ?

Otherwise, if you have outliers, here are a few recommendations:

- use the least squares gradient reconstruction, with full extended neighborhood
- If the flow is unsteady, add more right-hand side reconstructions (sweeps) to the velocity (5 instead of 1 for example).
- if it makes sense, use an imposed pressure outlet instead of a free outlet
- otherwise, if you can add one layer of cells using code_saturne's mesh extrusion preprocessing option, the increased regularity may make a good difference.
- check also the CFL number. A too large CFL may lead to divergence, but if you have a very small max CFL number, this can cause different issues also.

Best regards,

Yvan

[lamia]

Hello Yvan,

Thanx you for the quick answer.

The outlet being detected as a free inlet and negative inlet flows are more surprising. Do you seen this in the run_solver.log or when visualizing the results. Did you double or triple-check the group names and boundary condition assignment ?

I saw that in the run_solver.log file,

The groups names were imported from the preprocessing log, and I check that in the med file written when preporcessing the mesh.

Here is how I define my boundary conditions:
INLET (nature inlet)
velocity: volumic flow rate = 0.03935 m3/s (normal direction to the inlet)
Turbulence: hydraulic diameter 0.27m
temperature =50°C

OUTLET (nature Free inlet/outlet)
K=0

and this is how it is interpreted by code_saturne
---------------------------------------------------------------
Boundary type Code Nb faces Mass flow
---------------------------------------------------------------
Inlet 2 1419 -0.278283968E+00
Smooth wall 5 388520 -0.360323548E-19
Rough wall 6 0 0.000000000E+00
Symmetry 4 45925 0.000000000E+00
Free outlet 3 0 0.000000000E+00
Free inlet 14 2396 0.278283969E+00
Convective inlet 16 0 0.000000000E+00
Free surface 15 0 0.000000000E+00
Undefined 1 0 0.000000000E+00
---------------------------------------------------------------

[Yvan Fournier]

Hello,

Is this stable across time steps or does it change over time ?

Unless your case is confidential, depending on its size, could you post the setup, or simply a few visualizations, and the other files (logs and setup) listed in this forum's usage recommendations ?

Best regards,

Yvan