CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ParaView (https://www.cfd-online.com/Forums/paraview/)
-   -   [OpenFOAM] [Paraviev] problems with small domains (https://www.cfd-online.com/Forums/paraview/109914-paraviev-problems-small-domains.html)

flotus1 November 29, 2012 12:38

[Paraviev] problems with small domains
 
I have a problem with Paraviev when trying to post-process very small (~10^-6 m) domains.
Whenever I try to create streamlines or "plot over line", the created object doesnt contain any data.

If I run a similar simulation with "normal" domain extent, everything works fine.

Is this a known bug or does anyone know how to fix it?
Scaling my geometry is not an option because there are too many dimensionless parameters to be matched.

PV version is 3.10.1 64 bit

wyldckat December 2, 2012 09:35

1 Attachment(s)
Greetings Alex,

Indeed it looks like this is a limitation in the Streamline filter algorithm.
Attached is another example of this. It's a 1e-05 times smaller example of OpenFOAM's icoFoam "cavity" tutorial.

I've used ParaView 3.12.0 for this test. One of the solutions for now seemed to be to apply the Transform filter to the geometry in ParaView. It will only affect the geometry scale, not the content fields scales. Problem is that the vorticity calculated is affected by the rescaling of the geometry, so you'll have to apply a counter calculation of those fields for balancing out the geometrical transformation.

In the attached VTK file, it was enough to scale it up to 3 times the size, but a multiple of 10 always makes it easier to scale things back, if necessary.

For more on this, I suggest two things:
  1. Test ParaView 3.98 RC3 that has been released a few days ago. Maybe this has been fixed in that version.
  2. Use the ParaView mailing list for asking this question as well, because the main developers should be able to help you out further with this: http://www.paraview.org/paraview/help/mailing.html
edit: "Plot over line" also seems to not work in such a small domain...

Best regards,
Bruno

flotus1 December 2, 2012 17:21

Thanks a lot for your reply and the heads up about the vorticity... that would have cost me another few hours.

I will try the latest PV version tomorrow.


Edit:
Unfortunatly, the problem persists in version 3.98 RC3
Applying a transformation in PV to scale the geometry is a feasible workaround.


All times are GMT -4. The time now is 01:55.