[1988]

hello everyone
1-It has been claimed that Pimple is combination of Piso and Simple but how does how it work?
2-when I set " nOuterCorrectors 1"; it works as piso but when I change it to 2 or 3 it starts as pimple however really time consuming, so what does it do in the pimple algorithm ?
3- nCellsInCoarsestLevel 100; I have read it sets a coarse mesh to start a solution and merge levels is also related to this so if I am right how it can help to getting better answers?Is it for just first time step or not!!?

 solver          GAMG;
        tolerance       1e-06;
        relTol          0.01;
        smoother        GaussSeidel;
        cacheAgglomeration true;
        nCellsInCoarsestLevel 100;
        agglomerator    faceAreaPair;
        mergeLevels     1;
    }

    pFinal
    {
        solver          GAMG;
        tolerance       1e-06;
        relTol          0;
        smoother        GaussSeidel;
        cacheAgglomeration true;
        nCellsInCoarsestLevel 100;
        agglomerator    faceAreaPair;
        mergeLevels     1;
    }

    "(U|k|epsilon)"
    {
        solver          PBiCG;
        preconditioner  DILU;
        tolerance       1e-05;
        relTol          0.1;
    }

    "(U|k|epsilon)Final"
    {
        $U;
        tolerance       1e-05;
        relTol          0;
    }
}

PIMPLE
{
    nOuterCorrectors 1;
    nCorrectors     2;
    nNonOrthogonalCorrectors 0;
    pRefCell        0;
    pRefValue       0;
}

relaxationFactors
{
    fields
    {
    }
    equations
    {
        "U.*"           1;
        "k.*"           1;
        "epsilon.*"     1;
    }
}

4-I also have a question about relTol (relative tolerance) and I have heard that it multiples by initial residual to give final residuals but as you can see here it is not true !!! how can I be sure that the solution is converged by take a look at lines in the terminal ?so what is the role o f tolerance if the solution divergence is related to reltol?

Courant Number mean: 0.0886374 max: 0.523535
deltaT = 7.93317e-06
Time = 0.000127352

DILUPBiCG:  Solving for Ux, Initial residual = 0.00804195, Final residual = 3.54141e-06, No Iterations 1
DILUPBiCG:  Solving for Uy, Initial residual = 0.00137271, Final residual = 6.18957e-07, No Iterations 1
DILUPBiCG:  Solving for Uz, Initial residual = 0.0245264, Final residual = 6.07321e-07, No Iterations 2
GAMG:  Solving for p, Initial residual = 0.100308, Final residual = 0.000989593, No Iterations 20
time step continuity errors : sum local = 2.45355e-07, global = -4.49789e-08, cumulative = 8.13124e-06

Is there any file or tutorial that explains more about pimple and these things?
thanks for your attention

[1988]

no one is interested in this subject?

[Bernhard]

Many people are interested, but did you actually search the forum? The PIMPLE and PISO algorithms are very extensively discussed, with many useful references.

GAMG is the matrix solver. It does not do anything with your mesh or equations, it is just a clever way to solve the system of equations.

With respect to your question 4, you can see that it is exactly true. Why do you think it doesn't?

[1988]

Hi Bernhard and thanks for protect
I have tried to find more information abot pimple , maybe I did not make a good search but I think I have read most of posts about pimple so if it is possible please give the adress of one refrence.
Is soething which was mentioned in number3 is correct?
About question number 4 I should say that reltol was 0.01 and initial residual was 0.008 but the final is 3.54e-6!!!so this is confusing for me.
really thank you

[Bernhard]

Reference: Search for Hrvoje Jasak's thesis, easy to find.
http://openfoamwiki.net/index.php/Op...hm_in_OpenFOAM
I don't remember specific topics on this message board.

3. The solvers generally do not get you better answers, as they solve a given matrix equation. Your solution depends rather on the matrix itself (i.e. discretization, equations and mesh). If your set-up is faulty, your matrix-solver will solve the equations, but the results are bogus. The reason to choose GAMG over other is rather a choice of the convergence towards the solution of your matrix equations. nCellsInCoarsest level is a parameter of the GAMG, which controls the algorithm in some way. Look up references for GAMG if you want to now what the exact meaning is, I am not sure about it. It is unrelated to your mesh.

4. It reached convergence in one iteration, so stops immediately. Check e.g. for your pEqn, which needs 20 iterations, and thus the residual decreases slower.