[buckngnr11]

I have a rather complicated problem and am not sure if the shell conduction method might be able to help me out. I understand the basic premise of the model, but am not sure of all of the details of its implementation. The user guide states it allows for in plane conduction calculations. So the layers you specify I assume are across the thickness of the wall. To allow for in plane conduction, there must be columns of layers right? Since there isn't a physical mesh to visualize this, I'm wondering how this is done? Is there a column of layers for each wall cell and the shell perimeter conforms to the face?

Another question somewhat related. The user defines a thickness of each layer. Is there any way to allow for locally defined thickness? Imagine a plate with thickness varying spatially. I know meshing this would be ideal but my application makes this very difficult. Can I assign a different thickness to each cell with an expression? Or does the shell have to have constant thickness across over the interface?

[vinerm]

There are two options, thin wall and shell conduction. Thin wall models conduction only across the thickness of the wall, i.e., along the local normal of the boundary. A 1-D heat conduction is solved. Shell conduction does the same, except it is done in full 3-D, i.e., thermal energy conservation is solved in all the three directions.

Unfortunately, it is not allowed to have variable thickness over the wall. However, new version has expression option. You may try that, provided, thickness could be described using an expression. But a profile is not allowed.

[buckngnr11]

I figured as much. My issue is I'm using dynamic meshing. When I use this with conjugate, the fluid domain moves, but I don't see any way to move the shadow wall of the adjacent metal mesh. What ends up happening is the fluid mesh deforms and I get a gap between the fluid and solid wall. Because the mesh is conformal, the nodes on the wall and shadow wall have the same global address. So there's no obvious way to tell the node on the solid wall to morph on the same vector as the fluid wall.

[vinerm]

If you are using Dynamic mesh then you have to move both, wall and its shadow. Shadow needs to be moved in passive manner, i.e., based on the movement of the wall itself. I'd expect the shadow to move along with the wall but I have never tried dynamic mesh at fluid-solid interface with CHT, so, it is quite possible that the shadow does not move on its own and requires movement via dynamic mesh.

[buckngnr11]

Originally Posted by vinerm

If you are using Dynamic mesh then you have to move both, wall and its shadow. Shadow needs to be moved in passive manner, i.e., based on the movement of the wall itself. I'd expect the shadow to move along with the wall but I have never tried dynamic mesh at fluid-solid interface with CHT, so, it is quite possible that the shadow does not move on its own and requires movement via dynamic mesh.

Yeah it doesn't move automatically weirdly enough. Here's a look at what happens.

[vinerm]

Then, you have to move it using the UDF. But movement should be based on the displacement calculated for fluid domain boundary.