[bentkj]

Hi everyone,

I am new to openfoam (4.1) and currently I'm working on simulations in the field of aerodynamic interactions.

I have been using reconstructPar to reconstruct my decomposed case which now takes up a lot of time with a simulation that involves about a thousand timesteps.

I have just discovered that I can postprocess my solutions without reconstructing the case by use of the command

paraFoam -builtin

and then selecting decomposed as the case type and then clicking 'Apply'.

At this stage I run into an output error message:

ERROR: In /home/openfoam/OpenFOAM/ParaView-5.0.1/VTK/IO/Geometry/vtkOpenFOAMReader.cxx, line 8500
vtkOpenFOAMReader (0x2e6e200): Can't open /home/ben/OpenFOAM/ben-4.1/run/pisoFoam_URANS_LES/sample_cylinder/processor0/constant/

ERROR: In /home/openfoam/OpenFOAM/ParaView-5.0.1/VTK/Common/ExecutionModel/vtkExecutive.cxx, line 784
vtkPVCompositeDataPipeline (0x38a6330): Algorithm vtkPOpenFOAMReader(0x38bb260) returned failure for request: vtkInformation (0x18ead30)
Debug: Off
Modified Time: 200793
Reference Count: 1
Registered Events: (none)
Request: REQUEST_INFORMATION
FROM_OUTPUT_PORT: 0
ALGORITHM_AFTER_FORWARD: 1
FORWARD_DIRECTION: 0

From the error messages, it seems to me that the the constant folder is supposed to be decomposed into the processor* folders as well? *processor number

There must be something that I am doing wrong with my workflow.

After setting up the case files (0, constant, system),
typically I would run...

blockMesh && checkMesh

decomposePar && mpirun -np 36 snappyHexMesh -parallel && reconstructParMesh

I would then copy the files within 0 (U, p, nut, nuTilda etc.) into the folder with the latest time step containing my mesh that was generated with snappyHexMesh, rename it to 0 while deleting all the other timesteps so now I'm left with (0, constant, system) folders again, with 0 containing my polyMesh file with (U, p, nut, etc.)

decomposePar && mpirun -np 36 pisoFoam -parallel && reconstructPar

paraFoam

...to view my results.

My processor* folders do not contain a constant folder.

I have been using this workflow for my RANS (simpeFoam) and now URANS (pisoFoam) which is where I have to deal with a large amount of saved timesteps making reconstructPar extremely time consuming hence I stumbled upon the 'paraFoam -builtin' command to work with a decomposed case.

Please help! the following are the Dict files which I think may be need to help me solve my problem.

Thank you very much!

EDIT: also, any feedback regarding my workflow is greatly appreciated!

[bentkj]

I seem to have solved the problem. It seems like my workflow was not exactly correct and how I managed the overwriting of files during the snappyHexMesh process.

I followed the workflow in the tutorial:

https://www.youtube.com/watch?v=ObsFQUiVi1U

and my problem is fixed. Thank you!

[Antimony]

Hi,

Just to add on a bit. You don't need the paraFoam command.

In your case folder create a dummy file called, say, case.foam. And open the file with Paraview, and in the Case Type, choose Decomposed.

Cheers,
Antimony

[bentkj]

Originally Posted by Antimony

Hi,

Just to add on a bit. You don't need the paraFoam command.

In your case folder create a dummy file called, say, case.foam. And open the file with Paraview, and in the Case Type, choose Decomposed.

Cheers,
Antimony

Thank you, Antimony.

Just for understanding, is there a difference between the .foam and .openfoam formats? In what way do they differ? I realize this difference in formats since I've adjusted my workflow.

Thank you,
Ben

[Antimony]

Hi,

Sorry, haven't used much of .OpenFOAM to really comment on it. Typically use .foam and it has always been convenient. If I am not mistaken the built-in OpenFOAMReader needs a .foam extension, but I could be wrong on this.

Cheers,
Antimony