alpha.water field zero for parallel run with particular number of processors
 

Hello everyone,

I'm working with a two-phase problem (droplet on a surface) and a solver based on interFoam with an user-defined alpha contact angle BC.

When I'm running a case in serial mode it works. When I'm running it in parallel mode on 5 processors, it also works. But for the same geometry, if I'm running the calculation on 6 processors, the alpha.water values get everywhere zero or nearly zero (which is not the case in serial mode or on 5 processors) and the code is consequently doing the wrong thing without giving an error.

Though alpha.water field is actually set and the droplet is existing. This is a code part from the user-defined alpha contact angle BC:

if( this->db().foundObject<volScalarField>("alpha.water") )
{
volScalarField& alpha1 = const_cast<volScalarField&>
(
this->db().lookupObject<volScalarField>("alpha.water" ) );
... }

Furthermore, if I don't change the geometry but only the number of cells in my simulation model, it's still always 6 processors and more that cause the problem. If I increase the geometry dimensions, it starts working for 6 processors, but the same problem appears with 10 and more processors.

Important: it's only the partition of the wall with the user-defined alpha contact angle BC that is causing the problem. But it doesn't matter if the droplet is set completely on one processors or if the droplet is partitioned.

I'm quite new to openfoam. So I hope I could describe the problem relatively clearly and somebody could help me or just give a hint. I would like to post some code parts, but I don't know at all what the problem is. So if you have further questions I will try to answer.

THANK YOU in advance!

Best regards
VKr

Hi VKr! I have same question. Do you solve it ?