Elmer Discussion Forum

Hi Peter,
Yes I will see a little more how SaveScalar works.
Say just that, as I told it in the former reply, when solving a multiphysics model (like in our case an electromagnetic processing problem with electromagnetics, heat transfer, moving mesh, fluid mechanics coupled physics), you have often to make some calculations like:
- local volume Joule losses or/and integration in a body,
- integration of the current density in the inductor section to get the total current,
- heat flow in a body or in a boundary (local and/or integrated),
- fluid flow rate in a boundary, etc......
All these indirect quantities are function of the basic dependant variables (vector potential, scalar electric potential, temperature, velocity, etc....) and of the physical material properties.
This is why it would be very convenient to be able to calculate all these quantities (which you need to know gradually when advancing in the results analysis) either by staying in Elmer, but rather in Paraview where you are already for the results graphical analysis, you see? This is why it would be very interesting to be able to export all the basic variables (this is the case, ok) and the physical material properties from Elmer to Paraview.
But say also that even if this is not possible, it is very exciting and interesting to discover all the data analysis possibilities of Paraview (like the different "extract selection", "extract block", "extract surface", etc...., which I am currently investigating)!
We keep us informed.
Roland

Hi Peter again,
I just tried what you explained about the GeometryIds for the bodies and the boundaries on your "BlockDomainsVtuParts" Github model.
1/ for the bodies GeometryIds threeshold Min/Max = 2/4 (here attached) it shows nicely the 3 Bodies 2 to 4 (as defined in the sif file)
2/ For the Boundaries GeometryIds threeshold Min/Max 101/104 (here attached), which should show boundaries 1 to 4 (according to your explanations), I see nothing in the Renderview....Probably I make a mistake somewhere.
Could you please explain again how to see these boundaries?
Thanks in advance for your help!
Roland

...Oups, I forgot the 2 attached png Files!
They are here attached

Hi,
In order to test the possibility of selecting in Paraview the bodies/boundaries (as you explained it Peter) I solved a simple heat transfer "Heating_block" model (temperature_result plot and zipped GUI project here attached). The model works correctly without following commands in the Model/Setup/Simulation free text field:
vtu: Output Directory = File "par"
vtu: Vtu Part Collection = Logical True
But when adding these 2 command lines, the solver stops and shows an error message given in the here attached Solver_log file.
Could you please give some help (Peter ?) for fixing that?
Thanks in advance!
Roland

Hi

What if you remove the line or create the directory by hand.

Things may not always work exactly the same way in Windows. So maybe the internal make directory failed?

-Peter

Hi Peter,
At first many thanks for your efforts in helping.
Unfortunately it does not work if I remove the command line "vtu: Output Directory = File "par", neither if I create the "par" directory by hand.
What is very strange is that it works nicely in the 2D Phase_Change model (here attached ) while it does not work in the 3D "Heating_block" model sent in my former post(??).
How could this be fixed?
Roland

Hi,

Sorry, I cannot reproduce this. Also your binary is ~10 months old so there may have been changes. Can somebody else reproduce the issue?

Btw, if you use the "Geometry Ids" to pick pieces of the mesh then you don't really need "vtu part collection". These are alternative ways to choose entities.

-Peter

Ok Peter, I will try all this and we keep us informed.
Many thanks
Roland

It works Ok for me on my system.

Hi,
It works now also on my 3D "heating_block" example after rebuilding it from zero.
Thanks to you all!
Roland