[denis.sanga@hotmail.it]

GoodMorning, I would ask you some suggestions to apply the residualControl on PimpleFoam solver of OpenFoam 3.0.1.

Until now I added the following strings in the fvSolution of the tutorial PITZDAILY:

PIMPLE
{
nNonOrthogonalCorrectors 0;
nCorrectors 2;

residualControl

{
p_rgh
{
tolerance 1e-2;
relTol 0;
}

p
{
tolerance 1e-2;
relTol 0;
}
U
{
tolerance 1e-3;
relTol 0;
}

}

}

but doesn't stop when it reach convergence condition.
thank you really much
best regards

[chubb87]

Hi Denis,

you should check in the solver output at the beginning, if the dictionary is read correctly:
it should say
"PIMPLE: max iterations = xyz
field p_rgh : relTol 0, tolerance 0.01"
etc.

What did you specify for nOuterCorrectors, I would choose a high value for that.
Another thing to try might be to comment out the control of p and U, so that only p_rgh is checked and see if that works. Then add one variable at a time. I suspect p_rgh is not solved, only p. Maybe start without specifying p_rgh.
Regards

[denis.sanga@hotmail.it]

dear chubb87,
thank you really much for all your suggestions but i didn't reach good news.

I posted here one iterations

Courant Number mean: 0.811001 max: 3.24254
deltaT = 0.000166667
Time = 0.0333333

PIMPLE: iteration 1
smoothSolver:  Solving for Ux, Initial residual = 0.00250871, Final residual = 9.35103e-06, No Iterations 5
smoothSolver:  Solving for Uy, Initial residual = 0.00683136, Final residual = 4.88956e-06, No Iterations 7
GAMG:  Solving for p, Initial residual = 0.0225734, Final residual = 0.000200193, No Iterations 4
time step continuity errors : sum local = 1.77537e-07, global = -7.97343e-10, cumulative = -1.18619e-08
GAMG:  Solving for p, Initial residual = 0.0138574, Final residual = 0.000120176, No Iterations 4
time step continuity errors : sum local = 1.06425e-07, global = -3.60448e-10, cumulative = -1.22223e-08
GAMG:  Solving for p, Initial residual = 0.000814347, Final residual = 6.56738e-06, No Iterations 5
time step continuity errors : sum local = 5.83031e-09, global = -8.92967e-10, cumulative = -1.31153e-08
GAMG:  Solving for p, Initial residual = 0.000206963, Final residual = 1.76587e-06, No Iterations 6
time step continuity errors : sum local = 1.5697e-09, global = -4.39032e-10, cumulative = -1.35543e-08
GAMG:  Solving for p, Initial residual = 4.93477e-05, Final residual = 4.13283e-07, No Iterations 5
time step continuity errors : sum local = 3.67266e-10, global = 9.66343e-11, cumulative = -1.34577e-08
GAMG:  Solving for p, Initial residual = 2.22904e-05, Final residual = 1.49174e-07, No Iterations 5
time step continuity errors : sum local = 1.32559e-10, global = -2.4542e-11, cumulative = -1.34822e-08
GAMG:  Solving for p, Initial residual = 1.32491e-05, Final residual = 7.66981e-08, No Iterations 5
time step continuity errors : sum local = 6.81563e-11, global = 1.37968e-11, cumulative = -1.34685e-08
GAMG:  Solving for p, Initial residual = 8.42648e-06, Final residual = 6.60599e-08, No Iterations 4
time step continuity errors : sum local = 5.87032e-11, global = 7.25156e-12, cumulative = -1.34612e-08
GAMG:  Solving for p, Initial residual = 5.62289e-06, Final residual = 3.73129e-08, No Iterations 4
time step continuity errors : sum local = 3.31577e-11, global = 8.6153e-13, cumulative = -1.34603e-08
GAMG:  Solving for p, Initial residual = 3.86804e-06, Final residual = 7.67479e-08, No Iterations 3
time step continuity errors : sum local = 6.82013e-11, global = -2.41396e-12, cumulative = -1.34628e-08
smoothSolver:  Solving for epsilon, Initial residual = 0.00205922, Final residual = 5.91189e-06, No Iterations 5
smoothSolver:  Solving for k, Initial residual = 0.00285806, Final residual = 8.64015e-06, No Iterations 6
ExecutionTime = 1.84 s  ClockTime = 1 s

and the changes applied are:

PIMPLE
{
nNonOrthogonalCorrectors 0;
nCorrectors 10;

residualControl

{
p
{
tolerance 1e-3;
relTol 0;
}

}


}




I'm sorry for disturb
Thank you really much
Regards

[chubb87]

Well, what you have posted is only 1 pimple iteration, with 10 pressure corrections (because of nCorrectors). After that convergence is reached and so no second pimple iteration is executed. The solver is doing what he is told to do. 10 pressure corrections is quite high, normally 2 or 3 should be enough. But you can specify another parameter nOuterCorrectors, which is the number of pimple iterations and should be higher than nCorrectors in your case.

[denis.sanga@hotmail.it]

Dear chubb87,
I thank you really really much for your help and suggestions. I would ask you another information. Looking to the solution seems that after few time steps the solutions converges and doesn't change so much from one time step to the following one. I would ask you if in PimpleFoam there is a way to stop the pimple loop as in simpleFoam when the residual control is satisfied, becouse now each simulation ends just when the endTime is reached.

Thank you really really much again and I'm sorry for disturb.
Best regards
Denis

[chubb87]

If you want to do a steady state simulation which stops when convergence is reached, why don't you use simpleFoam?
In pimpleFoam, the next time step is calculated after convergence is reached. If you think there are unsteady phenomena in the simulation maybe it makes sense to use pimpleFoam and stop the simulation when the solution does not change anymore (besides the transient phenomena) according to your judgement.

[denis.sanga@hotmail.it]

Dear chubb87,
thank you really much for your answer. So I will see the solution and I will extimate how much time the transient requests, and I will decide the endTime.

All my difficulties starts when I tried to use the twoPhaseEulerFoam for my master thesis. This function uses the PimpleFoam and with my knowledge I couldn't convert the code in order to use the simpleFoam. For this reason I started to stuudy the pimpleFoam solvers.

Thank you really much and I'm sorry for disturb again.
Best regards
Denis