Problems Setting up with Elmer Gui

[wjpowers]

I have been using the ElmerGui Manual. I find that my version of Elmer Gui does not respond in the way that the manual indicates.

For example, in adding an equation, I do Model -> Equation -> Add.
I add Heat Equation, by typing in the name, clicking on Add, and then OK.
Now, Heat Equation is an option under Model -> Equation
If I now go to Model -> Equation -> Heat Equation, and hit the Edit Solver Settings, Elmer bombs.
The GUI manual says that if I hold down on the shift key and double click on a surface, a pop-up menu will
indicate the possible of Equations that can be attached to that surface. This does not work.
It does not work if I have no Equation attached to that surface or if I have Heat Equation attached.
This same feature also doesn't work for adding materials.

The logger says that I have Version: 2.1 Mesa 7.7.1.

Any suggestions?

bill

[raback]

Hi

These should indeed work. Have you tried runnig through some Tutorial cases? There may perhaps be problems with the mesh or some missing steps in the workflow. The tutorials are pretty extensively tested and if problems persist these would suggest some problem with the code.

-Peter

[wjpowers]

I set up the mesh with Gmsh. It is a 2D problem.
What should I look for to see if there is a problem with the mesh.
When it plots with Elmer Gui, I have to rotate it to see that it is 2D.
All the surfaces, edges, and boundaries appear to be there.

What else?

thanks,

bill

[raback]

Hi

If you need to rotate it the mesh might not be x-y-plane. Usually this is the case so I don't know whether problems arise if you are working in x-z or y-z coordinates.

Additionally, if you use physical entities their numering may be unappropriate. To this end choose in Gmsh "Save As" -> "Gmsh 2.0" + "Save all (ignore physical entities)". The geometric entities will be automatically renumbered at import.

-Peter

[wjpowers]

I've created a test.msh that is a 2D figure in xy plane, with two planar surfaces.
My input file for Elmer consists of
$MeshFormat
2.1 0 8
$EndMeshFormat
$Nodes
217
[followed by 217 nodes]
$EndNodes
$Elements
450
[followed by 450 elements]
$EndElements

That's it. Should there be more or less for Elmer to work with?

This mesh has the same problems that my slightly more complicated mesh has.
Except that since this one is in xy plane, I don't have to rotate it when I open Elmer.

thanks,

bill

[raback]

Hi, Perhaps you could attach your case here (in geo & msh formats).

[wjpowers]

You ought to see attached a test.msh

test.msh

(18.26 KiB) Downloaded 334 times

generated with Gmsh. The software would not allow be to upload the test.geo file. It seems that the test.msh file contains all and some that the geo file contains.

I don't know whether this will be useful, but you ought to at least be able to test the file with your version of elmer gui.

thanks,

bill

[raback]

Hi, the mesh works without any problems for me (rev. 4312) so it's difficult to judge where the problem could lie. -Peter

[wjpowers]

Peter:

Maybe it's what I'm doing with the mesh that wrong.
Suppose we treat this mesh as a steady state heat flow problem.
Let one material be air, the other say aluminum.
I choose these because they are part of the materials library.
Let the air boundary be at 273 K and the aluminum at 300 K.
I know this makes the temperature discontinuous at the air-aluminum boundary.
I'm not really certain how to set the problem up.
I could try to make the aluminum vary in a linear fashion until it matches the air boundary termperature.

What steps would you follow to set up and run this problem?

The problem I am having is how I am suppose to set up initial or boundary conditions.
When I try to do it, I bomb.

BTW, when you take the mesh I have given you and you add heat equation to both bodies,
can you edit solver settings and not bomb? My Gui bombs whenever I hit this button.

Thanks,

bill

[raback]

Hi Bill

I really can't reproduce any of your problems. Adding convection should not be a reason for the problems either. Of course it adds a number of things that could go wrong. I plan to write some GUI based tutorials for multi-domain/multi-solver problems. This would hopefully help many people struggling with the same questions.

Now regarding the discontinuity, it should not be a problem. You could of course use a functional expression, linear dependence or user defined subroutine to give the profile. Look at the start of SolverManual or search this forum for "matc", on how to do that. This can hardly be the reason to user "bomp" though.

-Peter