HeatSolve w/radiation

Numerical methods and mathematical models of Elmer
Post Reply
hellbourne
Posts: 4
Joined: 24 Jan 2013, 18:42
Antispam: Yes

HeatSolve w/radiation

Post by hellbourne »

Hi,
I am trying to use HeatSolve to simulate the heating of a plate by a metal heater.
In the attached .msh file, there are two bodies. The smaller one is the metal (gold) heater and the bigger one is the plate (fused silica).
Attached is also the .sif file. I tried to force heat flux onto the heater and assume radiation from the glass to the environment.

For some reason the solver does not even converge.

Your help will be greatly appreciated. I should point out that I'm a newbie with Elmer.

Amit.
Attachments
quartz_heater.msh
(33.63 KiB) Downloaded 216 times
case.sif
(2.39 KiB) Downloaded 195 times
hellbourne
Posts: 4
Joined: 24 Jan 2013, 18:42
Antispam: Yes

Re: HeatSolve w/radiation

Post by hellbourne »

OK, I simplified the problem a bit (see newly attached .msh file, which I plotted with gmsh).
What I would like is to impose a heat flux entering one face of the plate and let radiation and thermal conductivity do the rest.
Attachments
new.msh
(170.16 KiB) Downloaded 199 times
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: HeatSolve w/radiation

Post by mzenker »

Hi,

without looking at your sif file, I have noticed that your mesh is a surface mesh only. You need a volume mesh (the "3D" button in gmsh) in order to solve your problem. Or did you make a volume mesh, and gmsh didn't save it? In this case, you probably have declared your surfaces as physical, but not the volumes. You either have to declare physical volumes, or check "save all (ignore physical groups)" in the dialog box appearing when you save.

EDIT: I have also noticed that you have two overlapping meshes at the interface between the two blocks. You have to change your geometry so that there is only one surface between the two bodies, otherwise no heat will flow between them.

HTH,

Matthias
hellbourne
Posts: 4
Joined: 24 Jan 2013, 18:42
Antispam: Yes

Re: HeatSolve w/radiation

Post by hellbourne »

I simplified the mesh to a one plate box (see previous post with new.msh). How can you see that there are no volume mesh? I tried saving as you suggested, but where do I check the indication that there is a volume mesh now?

To create the mesh, by the way, I just plotted the points defining the box, connected the points with lines, connected the lines with surfaces, and defined the volume from all six surfaces.

Amit
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: HeatSolve w/radiation

Post by mzenker »

Hi,

when you open the mesh file with gmsh, you see that there are no volume elements. Also in "Tools->Visibility", it shows no volumes, just surfaces (and lines for your "simplified" mesh).
You have to do "File->Save as" in order to see the dialog box where you can check "save all". It doesn't seem to appear when you do File->Save Mesh.
To create the mesh, by the way, I just plotted the points defining the box, connected the points with lines, connected the lines with surfaces, and defined the volume from all six surfaces.
That is not the way to do it. Draw your box with corner points, their connecting lines as edges, and define surfaces and the volume. Then change to mesh mode and press the 3D button. Save the result as .msh file. When you re-open it with gmsh, it should show a surface mesh and a volume mesh.

Matthias
hellbourne
Posts: 4
Joined: 24 Jan 2013, 18:42
Antispam: Yes

Re: HeatSolve w/radiation

Post by hellbourne »

OK, so apparently I have made some progress. I now have both the surface mesh and the volume mesh (attached). I will now try to use ElmerSolver with HeatSolve and update on the post.

Thank you, Matthias.

Amit
Attachments
new.msh
(5.7 KiB) Downloaded 187 times
Post Reply