| Japsor |
August 27, 2013 04:47 |
Discrepancy ensight slice and full data for parallel decomposed case
4 Attachment(s)
Dear all,
Yesterday I encountered some unwanted behaviour when analyizing only a slice of data, generated by an addition to system/controldict. This occures for a decomposed case run in parallel.
When comparing results from the full dataset, by opening the .foam file with paraview4.1 the results look fine but when loading an ensight .case file used to generate only a slice, the slice shows sudden high velocity regions where it is both physically unrealistic and also does not match the full data. I reproduced this problem with the simple tut/icofoam/cavity-case.
I work with OpenFoam-2.2.0, viewed the full .foam file with paraview4.1 and generated an additional .foam file for each processor (in this case 3).
This was done to check if the boundaries of the processor-domains match the edges of the anomalies. It turns out they do.
I attached some figures to illustrate the problem. Plotted are the seperated domains, using a slice at the centre, just like the generated .case slice-file. The controldict and decomposePardict are below this message.
This problem was not found in OpenFoam-2.1.0. Has anyone else encountered a problem like this and know how to solvethis? As I can work with OpenFoam2.1 as well this means that it problably will not pose much of a problem for me, but I would at least like to report the bug and hopefully the problem can be solved. I hope you can help. Lastly, I hope this whole post is in an acceptable layout, as it is only my first post.
gr. Jasper
Code:
/*--------------------------------*- C++ -*----------------------------------*\
| ========= | |
| \\ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \\ / O peration | Version: 2.2.0 |
| \\ / A nd | Web: www.OpenFOAM.org |
| \\/ M anipulation | |
\*---------------------------------------------------------------------------*/
FoamFile
{
version 2.0;
format ascii;
class dictionary;
location "system";
object controlDict;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
libs (
"libOpenFOAM.so"
// "libdynamicSmagorinskyModel.so"
"libmyEnsightSampling.so"
);
application PimpleFoam;
startFrom startTime;
startTime 0;
stopAt endTime;
endTime 20;
deltaT 0.001;
writeControl timeStep;
writeInterval 500;
purgeWrite 0;
writeFormat binary;
writePrecision 6;
writeCompression on;
timeFormat general;
timePrecision 6;
runTimeModifiable true;
// ************************************************************************* /
functions
{
sliceCentre
{
type surfaces;
functionObjectLibs
(
"libsampling.so"
);
outputControl outputTime;
outputInterval 1;
// surfaceFormat reducedEnsight;
surfaceFormat ensight;
fields
(
U
p
nuSgsi
// Psi
// J
);
surfaces
(
sliceCentre
{
type plane;
basePoint (0.0 0.4 0.0);
normalVector (0.0 0.0 1.0);
interpolate true;
}
);
interpolationScheme cell;
}
timeAverage
{
type fieldAverage;
enable on;
functionObjectLibs ("libfieldFunctionObjects.so");
cleanRestart false;
resetOnOutput false;
outputControl outputTime;
outputInterval 1;
timeStart 20;
timeEnd $endTime;
fields
(
U
{
mean on;
prime2Mean on;
base time;
}
);
}
}
Code:
/*--------------------------------*- C++ -*----------------------------------*\
| ========= | |
| \\ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \\ / O peration | Version: 2.2.0 |
| \\ / A nd | Web: www.OpenFOAM.org |
| \\/ M anipulation | |
\*---------------------------------------------------------------------------*/
FoamFile
{
version 2.0;
format ascii;
class dictionary;
note "mesh decomposition control dictionary";
object decomposeParDict;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
numberOfSubdomains 3;
method scotch;
scotchCoeffs
{
//processorWeights
//(
// 1
// 1
// 1
// 1
//);
//writeGraph true;
//strategy "b";
}
// ************************************************************************* //
|
