[eugene]

Your run is converged. The cumulative error is just that, the sum of all errors over previous iterations. For a steady state run, you don't have to worry about it. As long as the local and global errors are small, you will be fine.

[eduardasilva]

Hello Eugene,

Thanks for your reply. ok I understand that I shouldn't consider the cumulative error. But is there anyway to automatically ignore this sum and only consider that the convergence is dependent on the residuals for U and p? I would like that the run would precisely stop at the moment in which the residual are bellow the limit so I could get information about the number of iterations needed, and ClockTime.

Thanks again,
Eduarda

[eduardasilva]

Good morning,

Can someone give me a suggestion on how to overcome the problem posted previously by me?

Thank you very much for your time.

Eduarda

[tian]

Hi,

there was a good example for convergence dependet on the residuals for U, p and T in the forum but you need to change the main code of simpleFoam.

Search for "BoussinesqBuoyantSimpleFoam":

http://openfoamwiki.net/index.php/Co...yantSimpleFoam

The excample used the file "fvSolution" for convergence:

SIMPLE
{
nNonOrthogonalCorrectors 0;
pRefCell 0;
pRefValue 0;
UConvergenceCriterion 1e-3;
pConvergenceCriterion 1e-2;
TConvergenceCriterion 1e-3;
}

You need to include 2 files in the main code: "convergenceCheck.H" and "initConvergenceCheck.H".

I do not know how to attach the files here. So if you need help give me your email...

Bye
Thomas

[lhcamilo]

Dear thomas,


I am also interest in those files, could you perhaps send them to me at lhcamilo@gmail.com .

At your discretion you might want to consider posting them on a file hosting website such as rapidshare.

anyway thanks in a advance

leo

[eduardasilva]

Thanks Thomas. My e-mail:
anaeduardasilva@gmail.com

Can you please tell me what changes do I need to make in the main code of simpleFoam?

Eduarda

[santos]

Hi,

The standard simpleFoam solver already includes the files mentioned by Tian (lines 57 and 75 of simpleFoam.C).

Regards,
Jose Santos

[tian]

Hi,

I upload my solver:

http://rapidshare.com/files/20536685...eFoam.tar.html

in fvSolution you can use:

SIMPLE
{
nNonOrthogonalCorrectors 0;
UConvergenceCriterion 1e-2;
pConvergenceCriterion 1e-1;
}

Jose, that is true but I made some modifications.

Bye
Thomas

[nandiganavishal]

Hi Tian,

I was going through the code in this site
http://openfoamwiki.net/index.php/Co...yantSimpleFoam

I would like to know what is the exact difference between

TEqn().relax();

and T.relax();

I know former is Implicit relaxing the matrix.. But how exactly this is being done in OpenFOAM and the latter is explicit form relaxing the solution using previous iteration value...

I would like to know which of the two we should usually choose ...what is the advantages of the respective relaxation techniques...

Kindly throw some light on this...

Thanks

Regards

Vishal

[eduardasilva]

Hello again,

Thank you very much Thomas. I was trying to compile the solver you sent using wmake nut I got an error saying.
make: *** No rule to make target `/home/USER/OpenFOAM/OpenFOAM-1.5/src/OpenFOAM/lnInclude/dimensionedDiagTensor.H ', needed by `simpleConvergenceFoam.dep'.

Any suggestion?

Thanks again.
Eduarda

[kbr]

Hi Eduarda,

first use 'wclean' and then try again compiling the solver by 'wmake'.

Maybe that helps.

Kerstin

[eduardasilva]

Thanks,

Now I have a new error coming.

-lincompressibleRASModels -lincompressibleTransportModels -lfiniteVolume -lmeshTools -llduSolvers -lOpenFOAM -ldl -lm -o /home/eduarda/OpenFOAM/USER-1.5/applications/bin/linuxGccDPOpt/simpleConvergence Foam
/usr/bin/ld: cannot find -llduSolvers
collect2: ld returned 1 exit status
make: *** [/home/eduarda/OpenFOAM/USER/applications/bin/linuxGccDPOpt/simpleConvergenceFoa m] Error 1

Any help?

[kbr]

Hi Eduarda,

I think it tells you that your lduSolver-library is missing or can't be found. Another possibility might be that in 'Make/options' there is a wrong source given, i.e. have a look at your file 'options' under 'EXE_INC'. There you should find '-I$(LIB_SRC)/lduSolvers'. Is it there? And under 'EXE_LIBS' you should find '-llduSolvers'.

Kerstin

[eduardasilva]

Thank you so much Kerstin.

The file "options" looks like:

EXE_INC = \
-I$(LIB_SRC)/finiteVolume/lnInclude \
-I$(LIB_SRC)/turbulenceModels/RAS \
-I$(LIB_SRC)/transportModels

EXE_LIBS = \
-lincompressibleRASModels \
-lincompressibleTransportModels \
-lfiniteVolume \
-lmeshTools \
-llduSolvers \
/* $(LIB_WM_OPTIONS_DIR)/libfbsdmalloc.o */

[eduardasilva]

Just to add that the options file is exactly the same as in simpleFoam.

Thanks.

[kbr]

Hi Eduarda,

is in your '/home/eduarda/OpenFOAM/USER-1.5/lib/linux**GccDPOpt/' directory a file available with name 'liblduSolvers.so'? If not it's missing, if yes, I have no other idea which problem it could be.

Kerstin

[eduardasilva]

hello again Kerstion,

And thanks.

The directory '~/OpenFOAM/USER-1.5/lib/linuxGccDPOpt$' is empty.

The directory '~/OpenFOAM/OpenFOAM-1.5/lib/linuxGccDPOpt' doesn't include that file either.

How can I get the file?

Eduarda

[kbr]

Hi Eduarda,

I think I understand now the problem. I'm working under OF 1.4.1 and there this library exists. In OF 1.5 this library doesn't exist anymore. The solver you've got is possibly written in another version. I think that's the reason why you can't compile it without some modifications.

I'm sorry I can't help you really.

Kerstin

Kerstin

[tian]

Hi Eduarda,

can you compile the original simpleFoam solver without trouble? If yes, you can replace four files from the simpleConvergenceFoam to the original folder (make a copy before):

initConvergenceCheck.H
convergenceCheck.H
pEqn.H
UEqn.H

Bye
Thomas

[eduardasilva]

Hello Kerstin and Thomas,

Thank you very much for your help! It is working perfectly now.

Eduarda