[Chris T]

Hi everyone,

I have a chtMultiregionFoam case that I run in parallel. After Introducing a view factor radiation model I noticed that the results of

faceAgglomerate -region air -dict system/air/viewFactorsDict
viewFactorsGen -region air

where not present in the processor directories after decomposition with

runApplication decomposePar  -allRegions

To get around that I used

mpirun -np $1 faceAgglomerate -region air -parallel -dict system/air/viewFactorsDict
mpirun -np $1 viewFactorsGen -region air -parallel

.
This was ok for running the case but in order to use functionObjects for post-prosessing after run time was getting an error because finalAgglom files where missing from the reconstructed case.
Now to get around this problem I do both ways of calculating view factors. But I noticed that the results I get in the log file for each method are different.


Serial:

/*---------------------------------------------------------------------------*\
  =========                 |
  \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox
   \\    /   O peration     | Website:  https://openfoam.org
    \\  /    A nd           | Version:  9
     \\/     M anipulation  |
\*---------------------------------------------------------------------------*/
Build  : 9-b456138dc4bc
Exec   : viewFactorsGen -region air
Date   : Mar 30 2023
Time   : 14:25:32
Host   : "ct-System-Product-Name"
PID    : 178199
I/O    : uncollated
Case   : /home/ct/OpenFOAM/ct-9/run/iSTORE/iSTORE_conical_radiation_2
nProcs : 1
sigFpe : Enabling floating point exception trapping (FOAM_SIGFPE).
fileModificationChecking : Monitoring run-time modified files using timeStampMaster (fileModificationSkew 10)
allowSystemOperations : Allowing user-supplied system call operations

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create mesh air for time = 0


Creating single cell mesh...

Created single cell mesh...

Total number of coarse faces: 104

View factor patches included in the calculation : 4(0 1 2 3)

Calculating view factors...
Writing view factor matrix...
F00: 0.53782006
F01: 0.10066795
F02: 0.31692687
F03: 0.062410839
F10: 0.49382079
F11: 0.4450708
F12: 0
F13: 0.20058446
F20: 1.0042635
F21: 0
F22: 0
F23: 0
F30: 0.31134188
F31: 0.20398426
F32: 0
F33: 0.52597451
End

Parallel

/*---------------------------------------------------------------------------*\
  =========                 |
  \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox
   \\    /   O peration     | Website:  https://openfoam.org
    \\  /    A nd           | Version:  9
     \\/     M anipulation  |
\*---------------------------------------------------------------------------*/
Build  : 9-b456138dc4bc
Exec   : viewFactorsGen -region air -parallel
Date   : Mar 30 2023
Time   : 14:26:22
Host   : "ct-System-Product-Name"
PID    : 178341
I/O    : uncollated
Case   : /home/ct/OpenFOAM/ct-9/run/iSTORE/iSTORE_conical_radiation_2
nProcs : 8
Slaves : 
7
(
"ct-System-Product-Name.178342"
"ct-System-Product-Name.178343"
"ct-System-Product-Name.178344"
"ct-System-Product-Name.178345"
"ct-System-Product-Name.178346"
"ct-System-Product-Name.178347"
"ct-System-Product-Name.178348"
)

Pstream initialised with:
    floatTransfer      : 0
    nProcsSimpleSum    : 0
    commsType          : nonBlocking
    polling iterations : 0
sigFpe : Enabling floating point exception trapping (FOAM_SIGFPE).
fileModificationChecking : Monitoring run-time modified files using timeStampMaster (fileModificationSkew 10)
allowSystemOperations : Allowing user-supplied system call operations

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create mesh air for time = 0

[3] 
Creating single cell mesh...
[0] 
Creating single cell mesh...
[6] 
Creating single cell mesh...
[4] 
Creating single cell mesh...
[7] 
Creating single cell mesh...
[1] 
Creating single cell mesh...
[5] 
Creating single cell mesh...
[2] 
Creating single cell mesh...
[7] 
Created single cell mesh...
[5] 
Created single cell mesh...
[3] 
Created single cell mesh...
[6] 
Created single cell mesh...
[1] 
Created single cell mesh...
[0] 
Created single cell mesh...

Total number of coarse faces: 123
[0] 
View factor patches included in the calculation : 2(1 3)
[4] 
Created single cell mesh...
[2] 
Created single cell mesh...

Calculating view factors...
Writing view factor matrix...
F11: 0.42525914
F13: 0.18779075
F31: 0.18889182
F33: 0.53471009
End

Notice how there are fewer surfaces and view factors even though both processes read from the same dictionary viewFactorsDict during agglomeration.
That dictionary is the folloing

/*--------------------------------*- C++ -*----------------------------------*\
| =========                 |                                                 |
| \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox           |
|  \\    /   O peration     | Version:  9.0                                   |
|   \\  /    A nd           | Web:      www.OpenFOAM.org                      |
|    \\/     M anipulation  |                                                 |
\*---------------------------------------------------------------------------*/
FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    object      viewFactorsDict;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

writeViewFactorMatrix     true;
writeFacesAgglomeration   true;
writePatchViewFactors     true;

maxDynListLength 999999;

debug        1;
//dumpRays     false; //true;

free_air
{
    nFacesInCoarsestLevel     20;
    featureAngle              5;
}
air_to_receiver_hot
{
    nFacesInCoarsestLevel     20;
    featureAngle              5;
}
air_to_receiver_mid
{
    nFacesInCoarsestLevel     20;
    featureAngle              5;
}
air_to_receiver_out
{
    nFacesInCoarsestLevel     20;
    featureAngle              9;
}


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

1.) What might be causing the discrepancy? Could it be a bug? Now I don't trust as much the validity of my simulation.
2.) What is the most appropriate way of combining the viewFactor model with the decomposePar, reconstructPar and post processing with functionObjects?

[muanyagdino]

Hi Chris,

have you found an answer to your questions? I have now the very same experience in a similar case and it's making a headache, whether my parallel generated faceAgglomeration and viewFactors are correct.

Regards

[Chris T]

Hi muanyagdino,


No I didn't find anything to satisfactory solve my problem. I'm still using the process described above. I haven't been working on this for a while now though. I'm only recently getting back to it.


I'm wondering if the problem persist the same in OFv10 and OFv11 I'm using still version 9 from openfoam.org and I didn't get to try my problem on the newer versions.

If I may ask what version are you working with.



Best,
Chris