Workbench - electic boundary conditions

[SRC90]

Hi,
I'm trying to simulate temperature of my conductor if a certain current is passing through it. It's a long flat rectangular conductor.
I'm interested how to correctly put boundary conditions, especially electric ones. I tried to put 0V and 4V at the ends but then i get enormous currents, if i try to define current that is entering at 4V end i get an error, which is understandable because ohm law. Should i just put 0V at one end and specify current that is entering in conductor or 4V and current that is exiting the conductor or something else? Any explanation would be useful.
I'm using thermal-electric option in workbench.
Thanks

 

[lamboram]

Hi,

this might sound odd. But why don't you simulate using Ansys-Thermal? I simulate electrical systems for thermal investigations too. In your case I know the current value, I have the internal resistance of the body and I have the 3D model of the component. I can calculate power loss as I²R and divide it by the active volume to get w/m³. This is my BC that I will define on my heat generating body. also with necessary thermal BCs (if any). This will be much simpler, in my opinion.

Cheers,
Ram

 

[SRC90]

Thank you for your reply,

I tried that but i get is still way off from my measurements, i got about 39°C when i did the test and in the simulation i get 65°C with thermal electric and about 200°C when with just thermal when i put internal heat generation. As thermal BC i put radiation emissivity of 0.04 and convection with film coefficient of 4 W/m^2 K

I think i have a problem with determining correct boundary conditions.

Its a small strip of copper about 140x7x0.1 mm with about 8A and voltage drop of 0.2V going through it.

Regards.

 

[lamboram]

the physics should be the same. regardless if you simulate with thermo-electri/thermal. theoretically you should get the same temperature if the BCs are right. Try looking for possible heat paths through which your generated heat might flow away from the heat source. Now your simulation is hotter than the experiment (this is a solvable situation). Your simulation model has lesser sinks (or bad thermal paths leading to sinks) in comparison to reality. I would check the possible heat sinks (through ambient air, direct contact with boards or plates below in which case conduction). Is your convection coefficient is a measured value?

It will be interesting if you can fit it with thermal, because your task is relatively simpler. You just need to find a viable thermal path and also possible sinks.

Cheers