Learning Forum

[danilo.llano]

Hi All,

I'm working on a thermal model to explore the temperature distribution in a cubic container with different layers due to an object inside the cube. The object inside (sphere in this case) will generate certain amount of heat and I want to explore what is the thermal distribution across the cubic container walls. The container is a hollow cube with walls made of different materials. The centre of the cube is air, then there is an inner copper wall, a layer of insulation material and an outer copper wall. The air temperature inside the cube and the temperature of the outer copper walls (external faces) are the boundary conditions. I have created the geometry, selected the materials and properties, but once I run the Transient thermal simulation, the simulation result does not seem ok.  I am sure I am missing some configuration or setup to deal with this particular structure. I am looking forward to hearing from you. Any advice would be great.

 

[Rob]

Please can you post some images? A sketch with numbers will also be useful to help the community assist you. 

[peteroznewman]

Hi Danilo,

If you want the temperature distribution, I expect you would use a Steady State Thermal model first. If you later want to add a Transient Thermal, then you can see how long it takes to reach that steady state.

Regards,

Peter

[danilo.llano]

Hi All,

Thank you for your replies. I have attached a couple of figures that might help to explain my question. The dimensions and layers are shown in a 2d view and I have also attached a 3d view from the Geometry designer. Finally, I have included a simulation done by a former colleague years ago (I don't have any ANSYS files for this unfortunately). I am aiming to get something similar as I am using pretty much the same geometry and it is the same physical problem. Further down the road I want to use those results as starting point for my current work. I can share my project file if that is somehow helpful.

A further question, currently I have "5 parts 5 bodies" (circled in the geometry picture), should I transfer them like that to the Thermal analysis block? or should I group them in one single part? Thank you very much for your advice.

[Rob]

Please can you post an image of the mesh cutting through the domain? 

[danilo.llano]

Please find attached some pictures of the mesh

[Rob]

I'll leave community members with Mechanical experience to answer - you've shown that the volumes have been split which was my concern.  However, I suspect that you need more cells in the domain (I cover CFD and we have higher meshing standards!). 

As an aside, when you created the cut plane there is a pale blue tetrahedron icon. Selecting this shows the full cell rather than cutting through it; this makes it easier to see the mesh. 

[peteroznewman]

I would add that for any simulation, Structural, Thermal or CFD, there should always be a Mesh Refinement Study.

Here is an example of a Structural Mesh Refinement.

[danilo.llano]

I have a further question. Based on the same geometry, I am trying to set the temperature of different faces and see the temperature and total heat generation. 

I have been setting the temperature of the inner copper faces (bold orange square in the 2d scheme) to 22.2 C (individual boundary condition for each face) and the temperature of the outer copper faces (bold green square) to 22 C  (individual boundary condition for each face). I don't have any other heat source or boundary condition. Then I use a reaction probe to get the total heat generation. As expected the heat across the 6 inner faces is equal to the heat across the 6 outer faces. 

Now, I want to set one of the inner faces to a temperature slightly different, lets say 22.1 C and see the temperature distribution and heat generation. After simulation, the temperature distribution seems ok and realistic, but the heat calculated in the reaction probes doesn't seem right. If I check the total heat across the outer walls (all the faces at 22 C) it is Ok (similar to the value I got in the previous step) but the heat across the inner walls is much higher and obviously different from the value measured at the outer walls. 

I have to explore the changes in total heat generation when having faces (inner and outer walls) at different temperatures, therefore I need a way to measure the total heat. Does anyone have an advice on how to do this? Or maybe I am missing some additional setup to deal with the different temperature in the faces.