[hjasak]

It runs fine here, the last change was

commit 6b50e03f4417291ba2c9184244d37e2630ef2dae
Author: Henrik Rusche <h.rusche@wikki-gmbh.de>
Date: Thu Dec 12 00:41:18 2013 +0100

Below is the fvSolution file:

/*--------------------------------*- C++ -*----------------------------------*\
| ========= | |
| \\ / F ield | foam-extend: Open Source CFD |
| \\ / O peration | Version: 3.0 |
| \\ / A nd | Web: http://www.extend-project.de |
| \\/ M anipulation | |
\*---------------------------------------------------------------------------*/
FoamFile
{
version 2.0;
format ascii;
class dictionary;
object fvSolution;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

solvers
{

p
{
solver PCG;
preconditioner Cholesky;

tolerance 1e-07;
relTol 0;
minIter 0;
maxIter 800;
}

U
{
solver BiCGStab;
preconditioner ILU0;

tolerance 1e-6;
relTol 0;
minIter 0;
maxIter 1000;
}

tau
{

solver BiCGStab;
preconditioner ILU0;

tolerance 1e-6;
relTol 0;
minIter 0;
maxIter 1000;

};

}

PISO
{
nCorrectors 2;
nNonOrthogonalCorrectors 0;
pRefCell 0;
pRefValue 0;
}

relaxationFactors
{
p 0.3;
U 0.5;
tau 0.3;
}

// ************************************************** *********************** //

[sita]

Thanks for the information. How can I check what my last change was?

[sita]

OK, I give up: last Thursday I installed foam-extend-3.0 from source, tested it on the Oldroyd-B tutorial, AND IT WORKED!!! No problems with finding the Cholesky preconditioner or anything.
So today I tried to run a case of my own, but now suddenly foam seems to have forgotten everything it knew previously and it gives the same error I got earlier, about Cholesky being an unknown preconditioner. WHY?? I didn't change anything, as far as I know. Anyone?

[yxiong]

Originally Posted by sita

OK, I give up: last Thursday I installed foam-extend-3.0 from source, tested it on the Oldroyd-B tutorial, AND IT WORKED!!! No problems with finding the Cholesky preconditioner or anything.
So today I tried to run a case of my own, but now suddenly foam seems to have forgotten everything it knew previously and it gives the same error I got earlier, about Cholesky being an unknown preconditioner. WHY?? I didn't change anything, as far as I know. Anyone?

i have the same problem,when i try smenon's example in the extend-3.0:
--> FOAM FATAL IO ERROR:
Unknown asymmetric matrix preconditioner Cholesky

Valid asymmetric matrix preconditioners are :

7
(
AMG
none
GAMG
diagonal
SymGaussSeidel
ILU0
DILU
)


file: /home/yxiong/FileOfYxiong/projectile/system/fvSolution::solvers::U from line 82 to line 90.

From function lduPreconditioner::New
(
const lduMatrix& matrix,
const FieldField<Field, scalar>& coupleBouCoeffs,
const FieldField<Field, scalar>& coupleIntCoeffs,
const lduInterfaceFieldPtrsList& interfaces,
const dictionary& dict
)
in file matrices/lduMatrix/lduMatrix/lduMatrixPreconditioner.C at line 141.

FOAM exiting

sita,did you have solved your problem?tell me,thank you!

[hjasak]

Just change the name from Cholesky to ILU0.

Hrvoje

[yxiong]

Originally Posted by hjasak

Just change the name from Cholesky to ILU0.

Hrvoje

Thank you very much,Mr Hrvoje Jasak,it seems very nice now!

[sita]

Sorry, no, I haven't found a solution yet, maybe the Cholesky preconditioner just disappeared? Although some time ago I saw a post by Hvorje Jasak, who didn't have problems with Cholesky. I switched to using PAMG instead of Cholesky, that works.

[Raymond.Leoi]

I suppose the divergence of (tau_p + tau_s) should be returned for the momentum equation. However in the viscoelastic foam, it is
fvc::div(tau_/rho_, "div(tau)")
- fvc::laplacian(etaPEff/rho_, U, "laplacian(etaPEff,U)")
+ fvm::laplacian( (etaPEff + etaS_)/rho_, U, "laplacian(etaPEff+etaS,U)")

It is not identical to div(tau_/rho_ + tau_s), isn't it?

[hjasak]

I am sorry: did you read the paper describing the methodology?

[Raymond.Leoi]

what's the paper?

Originally Posted by hjasak

I am sorry: did you read the paper describing the methodology?

[hjasak]

Apologies: paper.

[Raymond.Leoi]

I've gone through some papers for the theory. What paper do you mean, Hjasak?

Originally Posted by hjasak

Apologies: paper.

[hjasak]

Usually when I work on something in FOAM, there are papers published with the methodology, numerics and validation results.

Try googling Favero Jasak viscoelastic, or

http://scholar.google.co.uk/scholar?...ed=0CCUQgQMwAA

Does this help?

Hrv

[Raymond.Leoi]

Cheers, Hrv. No luck. Anyway, I suppose that divTau is the key for the viscoleastic solver no matter what model it is. Anyone dig this part?

Originally Posted by hjasak

Usually when I work on something in FOAM, there are papers published with the methodology, numerics and validation results.

Try googling Favero Jasak viscoelastic, or

http://scholar.google.co.uk/scholar?...ed=0CCUQgQMwAA

Does this help?

Hrv

[heliana60]

Hi Hrvoje,

I am using a model similar to the White Metzner-Larson model in the visco solver, I want to study shear thinning.

I am studying star-up flow in a 2D channel.

The model looks like this:

fvSymmTensorMatrix tauEqn
(
fvm::ddt(tau_)
+ fvm::div(phi(), tau_)
==
etaP_/lambda_*twoD
+ twoSymm(C)
- fvm::Sp(1/lambdaM, tau_)
);

where:

lambdaM = (lambda_*gammaC_/(gammaC_+lambda_*sqrt(0.5)*mag(twoD)))

In few words, the relaxation time is in function of the shear rate (my gammaC_ is 0.1)

When I do this, the centerline velocity acts like if it was a almost a newtonian fluid, and it doesn't reach the steady state (analytical results).

I checked also the nonNewtonian and Newtonian stress on time at a quarter of the channel and it is like \tau_ doesn't "change" anymore with time, though the newtonianStress does.

I am attaching here the plot of the nonNewtonian and Newtonian shear stress at quarter of the channel vs time, so you can see how it actually goes.

If I actually remove the convective term (I was just trying to see what was affecting the costitutive equation), my results reach steady state.

How can this convective term affect my model, so when I use it it doesnt reach steady state like expected?


Thanks,

Heliana

[sita]

Dear Jovani,

In your FENE-P and FENE-CR models, the values of etaS and etaP should be specified. What do these parameters mean physically? My first guess would be that etaS + etaP = eta_eff (i.e. effective viscosity of the solution = solvent viscosity + polymer contribution), but in the FENE-CR code, you use an eta_eff that's defined as eta_eff = etaP. So, is etaP then the viscosity of the polymer solution?

Thanks in advance for helping me out,
Sita

[ron_OFuser]

How to calculate drag coefficient on a cylinder for viscoelastic fluid?

[sita]

Originally Posted by ron_OFuser

How to calculate drag coefficient on a cylinder for viscoelastic fluid?

Hi Ron,

Maybe this helps?

Viscoelastic flow past a cylinder: drag coefficient. Nhan Phan-Thien & Hua-Shu Dou (1999). Computer Methods in Applied Mechanics and Engineering, 180 (3-4), pp. 243-266.

Good luck,
Sita

[ron_OFuser]

Thank you Sita for the same.

Do you know how would I implement the drag coefficient equation for these fluids in OpenFOAM?

[ron_OFuser]

Anyone?
Any idea for the above post on implementation of drag coefficients for viscoelastic fluids?