[haze_1986]

Hi All,

I'm not sure if anyone is experiencing this in interFoam, that reducing the maxCo and alphaMaxCo from the default 1 will change the result quite a lot. Does this simply mean that a smaller Courant number is required? From some quick runs the results seems to not change much when the Courant number is 0.1.

[hchen]

Hi,

InterFoam uses a explicit time scheme, so I suppose the time step will affect the accuracy of the result. I always use Co < 0.35 - 0.5. However this is case sensitive.

Best regards
Hao

Originally Posted by haze_1986

Hi All,

I'm not sure if anyone is experiencing this in interFoam, that reducing the maxCo and alphaMaxCo from the default 1 will change the result quite a lot. Does this simply mean that a smaller Courant number is required? From some quick runs the results seems to not change much when the Courant number is 0.1.

[haze_1986]

Originally Posted by hchen

Hi,

InterFoam uses a explicit time scheme, so I suppose the time step will affect the accuracy of the result. I always use Co < 0.35 - 0.5. However this is case sensitive.

Best regards
Hao

Thanks Hao, I assume that's just for the alpha equations? Is there a way to run them using an implicit time scheme?

[santiagomarquezd]

Hi, you can go up in Co increasing nAlphaSubCycles accordingly.

Regards.

[haze_1986]

Just an update on this, in version 2.4.0 alpha is now solved semi-implicitly, and I no longer have this issue. Solution converged better with larger Courant numbers.

[JF_tx]

Originally Posted by haze_1986

Just an update on this, in version 2.4.0 alpha is now solved semi-implicitly, and I no longer have this issue. Solution converged better with larger Courant numbers.

Hi all,
I have been researching in the maxCo number in interFOAM. I checked that in the tutorial (OpenFoam 3.0) inteFoam/ras/waterChannel, maxCo and maxAlpha are given a value of 6. Same configuration is working for the problem that I'm studding (flow over a sloped channel with a control structure (a wall with a small notch on it)), obtaining a good result when comparing to the experimental results (turbulence and velocity profiles).
However, when I try to justify the value and when I check for Co numbers, all references seem to agree that Co must be less than 1, and other tutorial and post seems to confirm this, with reasonable justifications.
Can anyone provide me some light into this?
Now, I’m simulating the same scenario with a maxCo and a maxAlpha equal to 1 to see the possible consequences of decreasing the values and, in turn, the max time step.
best

[JF_tx]

Hi again,
So, the results obtained using maxCo=1 are similar to the ones obtained with maxCo=6 for the case that I'm studding, of course, the time steps are smaller. Thus, seems a good alternative the use high maxCo first until you reach to equilibrium and after simulate with smaller maxCos. A good explanation of why is possible to use higher maxCo can be found in:

http://www.openfoam.org/version2.3.0/multiphase.php

best