thermal analysis of a vacuum chamber

Numerical methods and mathematical models of Elmer
slav
Posts: 15
Joined: 10 Oct 2013, 00:20
Antispam: Yes

thermal analysis of a vacuum chamber

Post by slav »

Hi, I'm new to FEA and Elmer. I need to perform a thermal analysis of a vacuum chamber and wil be very grateful for hepl. My test chamber is in a shape of a stainless steel sphere (with a thickness of 2 cm or so) with 2 viewports (glass). There is high vacuum inside and air outside. Let say that one viewport is heated. I'm interested in a steady-state temperatude distribution on the chamber on the vacuum side (that means inside).

I've prepared my test model in FreeCad, opened it in Salome (I work on Windows) and meshed it (with the "Multiple bodies from Salome to Elmer" tutorial). I have 4 bodies: chamber, 2 viewports and vacuum inside - I see them when I want to assign materials or so but I'm not able to click on a body separately (there is only one, outside, surface). Then I perform all the steps given in ElmerTutorials:

1. Model -> Equation
- I activate Heat Equation and should I assign it to all bodies? Or only to chamber and viewports?
- I don't see any option like "ambient air" (air outside). I think I need to take convection into account but how to do this?
- How Elmer will guess how the heat should pass from one body to another?
- Is there any way to assign names to bodies? In my final project I will have many bodies and will got lost for sure with names "Body 1" etc.

2. Model -> BodyForce
What exactly "Heat Source" is?

3. Model -> InitialCondition
Should I assign an initial temperature for all bodies except for the heated one?

4. Model -> BoundaryCondition
- I have only one surfce boundary to assign my conditions, I suspect it is not enough.
- In the "Heat Equation" window I found "Heat Flux conditions" and "Radiation Settings", is there any tutorial how to use these options?

What is the resultant file format? Is this something like temperature + coordiantes of a given surface element?
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: thermal analysis of a vacuum chamber

Post by mzenker »

Hi,

very short and incomplete answer (no time right now, sorry): first look into the documentation, especially the GUI and Models manual. You will find some answer there.
Maybe someone else will also give you some hints...

HTH,

Matthias
slav
Posts: 15
Joined: 10 Oct 2013, 00:20
Antispam: Yes

Re: thermal analysis of a vacuum chamber

Post by slav »

I didn't notice Models manual so I will check it, thanks. In GUI manual I didn't find the information needed.
slav
Posts: 15
Joined: 10 Oct 2013, 00:20
Antispam: Yes

Re: thermal analysis of a vacuum chamber

Post by slav »

Ok, I've read some theory and know now that:

a) "The Elmer heat equation module is capable of simulation heat transfer by conduction, convection, and diffuse gray radiation." - so it is perfect for me.
b) I can "have either temperature or heat flux" in boundary conditions (1.2.5 in Models manual). So I think I should use heat flux (of air outside and the one consisting of radiation for vacuum inside my chamber), am I right?
c) Ok, I know what the heat source is.
d) I learnt that I can save my results in other formats, like VTK.

So I still don't know:
- How Elmer will guess how the heat should pass from one body to another? -> Ok, there is an equation of heat flux but I'm not quite sure how to use it.
- Is there any way to assign names to bodies?
- Should I assign an initial temperature for all bodies except for the heated one?
- I have only one surfce boundary to assign my conditions, I suspect it is not enough. -> I don't know how to obtain more surfaces.
- And additional question - can I see something (eq. the volume mesh) inside a body, that means can I make a crosssection of my object to see if the mesh is correct.
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: thermal analysis of a vacuum chamber

Post by mzenker »

Hi,
slav wrote: b) I can "have either temperature or heat flux" in boundary conditions (1.2.5 in Models manual). So I think I should use heat flux (of air outside and the one consisting of radiation for vacuum inside my chamber), am I right?
As I understand your setting, I think you should use radiation inside and convection outside.
So I still don't know:
- How Elmer will guess how the heat should pass from one body to another? -> Ok, there is an equation of heat flux but I'm not quite sure how to use it.
It will just pass, you don't need to do anything special. Since you used the Salome/Elmer tutorial, I assume that your mesh is conformal between two bodies.
- Is there any way to assign names to bodies?
Principally yes, check "View->Volume mesh" and "Model->Set body properties" and doubleclick on a body. You can then assign a name. This will not always work, however.
- Should I assign an initial temperature for all bodies except for the heated one?
For a transient simulation, all bodies need an initial temperature, even the heated one. For a steady-state simulation, you don't need an initial temperature.
- I have only one surfce boundary to assign my conditions, I suspect it is not enough. -> I don't know how to obtain more surfaces.
You have to go back to Salome or even FreeCAD to do that. I don't know any of those, so I cannot help you here.
- And additional question - can I see something (eq. the volume mesh) inside a body, that means can I make a crosssection of my object to see if the mesh is correct.
You can do that in principle with ElmerGUI with "View->Volume mesh", however I would advise that you check your mesh within Salome and just verify that ElmerGUI has imported it correctly.

HTH,

Matthias
slav
Posts: 15
Joined: 10 Oct 2013, 00:20
Antispam: Yes

Re: thermal analysis of a vacuum chamber

Post by slav »

Hi, thank you very much for your help. Things are getting more and more clear for me but in general I still see darkness ;)
As I understand your setting, I think you should use radiation inside and convection outside.
You have to go back to Salome or even FreeCAD to do that. I don't know any of those, so I cannot help you here.
I agree with the convection and radiation.

But still don't know how to obtain more surfaces to assign more boundary condition. My idea is to remove "vacuum" body and have "nothing" inside my chamber. I guess that the inside surface will then appear. Maybe then, in BoundaryCondition setting box I can fill Heat Flux and Heat Transfer Coeff. with 0 (as I think it should be for vacuum), I don't know what to put for External Temperature. And I will set the radiation (idealized). The problem is that I have to set the emissivity and this surface (one "element", like the outside one) is of steel and galss. Nevertheless, I will try to fight with this in Salome but I'm open for suggestions :)
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: thermal analysis of a vacuum chamber

Post by mzenker »

Hi,
But still don't know how to obtain more surfaces to assign more boundary condition. My idea is to remove "vacuum" body and have "nothing" inside my chamber. I guess that the inside surface will then appear. Maybe then, in BoundaryCondition setting box I can fill Heat Flux and Heat Transfer Coeff. with 0 (as I think it should be for vacuum), I don't know what to put for External Temperature.
Yes, you can try that. For vacuum, there is no temperature defined in the strict sense. But I would put some finite value to avoid possible numeric problems.
If you still have problems defining your surfaces, you could post your STEP file (if it is less than 1 MB) so someone can have a look.
And I will set the radiation (idealized). The problem is that I have to set the emissivity and this surface (one "element", like the outside one) is of steel and galss. Nevertheless, I will try to fight with this in Salome but I'm open for suggestions :)
Since steel and glass have different properties, you should draw them as separate bodies, most likely in FreeCAD already. And for radiation, "idealized" means radiation exchange with the outside world which is a black body at "external temperature". If I get it right, you have one side of the viewport radiating inside the chamber, which is made of (polished?) steel and glass. So you might need to think about emissivity and reflectivity of your surfaces, and of view factors. Elmer can calculate view factors, but I have never done that. Yuu would have to consult the Models manual , and there are also some threads on the subject in this forum.

HTH,

Matthias
slav
Posts: 15
Joined: 10 Oct 2013, 00:20
Antispam: Yes

Re: thermal analysis of a vacuum chamber

Post by slav »

Hi,

there is some progress in my simulation. I managed to create more boundary surfaces. In Salome, besides creating volume groups for geometry and mesh, I created also surface/face groups for the outside and inside of my model (now I have only 3 bodies - chamber and 2 viewports). I exported the mesh to .unv and opened it in Elmer. I have, as expected, 3 bodies and 7 surfaces (here I expected 6: outside and inside parts of 1 chamber and 2 viewports, but I have also surprisingly "round" sides of both my viewports as one surface - I don't know why as they are not present in the mesh faces groups in Salome). The problem is to guess which surface is where. I enable the Numbering feature and the surfaces are numbered but still there is a problem with inside surfaces - I switched off the outside ones but their numbers are still visible. But as I understand there is no smart way to recognise surfaces and bodies :(

I set all the properities as I told in former posts (this more complicated radiation model I will consider later). When I set (in Equation) Convection to Computed I get many warnings - "WARNING:: HeatSolver: Convection model specified but no associated flow field present? " How and where to specify this flow field? When I set no convection I have no warnings and my simulation is "successful" but probably many others things are wrong and there is no temperature distribution.

Another problem is the Heat Source. In reality, I will have (somwhere on the chamber) 2 coils in mounts and with cooling. I know some parameters of this "heat source" but not all. The easiest quantity to measure is temperature so my question is - can I set something like a "thermal load" just by setting a temperature of the heating object? There are some examples of treatment of this kind for SolidWorks Simulations but I don't have this software to check how it works.
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: thermal analysis of a vacuum chamber

Post by mzenker »

Hi,
slav wrote: The problem is to guess which surface is where. I enable the Numbering feature and the surfaces are numbered but still there is a problem with inside surfaces - I switched off the outside ones but their numbers are still visible. But as I understand there is no smart way to recognise surfaces and bodies :(
Yes there is, but not straightforward for you as you use Salome: convert the Elmer mesh in gmsh format (elmergrid 2 4 meshdir where "meshdir" is the directory where your Elmer mesh files reside) and open it with gmsh. There the handling of the 3D object is much better, you can discover the numbering of bodies and boundaries and assign your boundary conditions in Elmer. If you ask kindly and someone of the Elmer team has some time, they might implement a similar re-export to Salome, who knows. Or you do it yourself if you have time and some C programming skills.
I set all the properities as I told in former posts (this more complicated radiation model I will consider later). When I set (in Equation) Convection to Computed I get many warnings - "WARNING:: HeatSolver: Convection model specified but no associated flow field present? " How and where to specify this flow field? When I set no convection I have no warnings and my simulation is "successful" but probably many others things are wrong and there is no temperature distribution.
You can assume a constant heat transfer coefficient by convection. Then you have to set the ambient temperature and the heat transfer coefficient under "heat flux conditions" in the boundary conditions.
Another problem is the Heat Source. In reality, I will have (somwhere on the chamber) 2 coils in mounts and with cooling. I know some parameters of this "heat source" but not all. The easiest quantity to measure is temperature so my question is - can I set something like a "thermal load" just by setting a temperature of the heating object? There are some examples of treatment of this kind for SolidWorks Simulations but I don't have this software to check how it works.
If the heating object has outer surfaces you can just set them to a constant temperature using a boundary condition.


HTH,

Matthias
slav
Posts: 15
Joined: 10 Oct 2013, 00:20
Antispam: Yes

Re: thermal analysis of a vacuum chamber

Post by slav »

Thanks Matthias, something works! I mean I have some results and not only errors ;)
If the heating object has outer surfaces you can just set them to a constant temperature using a boundary condition.
It is what I needed! I tried to do this in BodyForce but in BoundaryConditions it finally works.
You can assume a constant heat transfer coefficient by convection. Then you have to set the ambient temperature and the heat transfer coefficient under "heat flux conditions" in the boundary conditions.
I assumed constant convection and set a convection coefficient (Heat Transfer Coeff.) and External Temp. of air in Boundary Condition. I set no Convection Velocity in Equation module.

I will try to use gmesh to deal with the multitude of bodies and surfaces :)

Please tell me where are the results of my simulation hidden? Is this case.ep? There is written in the SolverManual that I can have results also in ascii format, I tried "Binary output = true" but don't get any additional files.
Post Reply