Add Thermal Contact Resistance to Existing Transient Analysis in APDL

[allintime]

Hi, I've taken over someone else's work and need some advice on how to add some additional complexity to the model. The problem is a 2D axisymmetric, transient, non-linear, heat transfer analysis on an assembly of components with convection on the external surface. The analysis was written in APDL so it can be run in batch mode on a cluster.

The analysis was created by simplifying the 3D geometry into a 2D IGES file using SolidWorks, importing the IGES into ANSYS, defining each of the component areas, and then glueing all of the areas together. Each of the areas are further simplified so that a mapped structured mesh can be created across the entire assembly by defining the number of divisions of each line. The resulting mesh is highly accurate but is very cumbersome to refine or edit. Materials and BC's are then applied and the run is solved.

One feature we would like to add to the model is thermal contact resistance to certain component interfaces. The goal is to run several cases: no resistance, perfectly insulated, and others in between. I am looking for a method to quickly add this feature to my current model without having to rebuild the geometry or mesh.

QUESTIONS: Is there a method of adding thermal contact resistance to a line? I don't think I can use the CONTA171 and TARGE169 because the components share the same line. Is it possible to 'unglue' a line after meshing?

I'll be working on similar analyses in the future, so any advice on a better method of setting up the analysis is appreciated!

- Greg

 

[scheah]

Hi Greg,

Interesting problem you have there. To 'unglue' a line between the meshes on both side after meshing, you could try copying the elements on one side of the line (using EGEN) then repeat it for the next side. That way you have two meshes that do not share the same nodes on the line. The original disused mesh can be deleted.


Kind regards,
Jason

 

[allintime]

Thanks for the reply Jason. If I go that route, would I need to do that for every component interface? There are a few dozen components so that would be fairly difficult. It would probably be easier to just rebuild the mesh.

So there isn't a way to just add contact resistance to a line? So would the correct way to set up the problem be to mesh each component independently and then apply the contacts? I'm relatively new to this type of analysis so I apologize if this is a basic question.

 

[scheah]

Sorry Greg, I've not done heat transfer through contacts before so please keep that in mind when evaluating my suggestions.

If I go that route, would I need to do that for every component interface?

Yes, you would have to do that for every component so that the CONTA171 can be defined at the interface nodes. I agree it would be easier to remesh than to go down this route.

So there isn't a way to just add contact resistance to a line?

No, I don't think you can add contact resistance to a line as the adjacent elements share the nodes thus would have the same temperature value.

So would the correct way to set up the problem be to mesh each component independently and then apply the contacts?

Yes, I believe that is true.


Kind regards,
Jason

 

[allintime]

Ah, yes that does make sense. I guess another option would be to add a very thin component at the interface by offsetting the line, and then give it material properties of an insulating material. This would require updating a small portion of the mesh but not the entire thing. I was hoping to avoid this.

Thanks for the help.