Modelling Thermal Interface between two parts

[peglor]

I'm trying to model the thermal bahaviour of an assembly of solid parts. A heat flux is applied to one surface and there is a convection condition from surfaces on another part in contact with it.

The problem is in reality there isn't complete thermal contact between the two parts. The parts I'm trying to model using FEA have been made, so I can test the validity of my model experimentally, but in order to make a valid FE model I'll need to add a thermal resistance at the interface between the mating parts. Is there a standard technique which can be used for this process?

I don't want to have to modify the shapes of the parts to allow a thin layer of elements to be put between the mating parts as this will drastically increase the element count as well as making the geometry less valid. Shell elements don't seem to support a temperature gradient through their thin direction either, but this may just be the package I'm using.

I'm using ProMechanica, which is more than adequate for solving the thermal problem but doesn't seem to have any settings to create an interface resistance between parts.

 

[corus]

In Abaqus you can use gap conductance where the gap conductivity can depend on the gap distance or on the pressure between the surfaces. As for ProMechanica, I don't know, but if it's possible you could have forced convection from one surface to another where the heat transfer coefficient is equal to k/x where x is that distance apart, k the conductivity of the gap.
One problem that may occur is what do you assume for the thermal resistance of the gap?, unless you intend to modify that value to suit the results from your experiment.
In practice there is always some thermal resistance between surfaces that are in contact but in general it is assumed that perfect thermal conductance is achieved. Unless you suspect that your results will be grossly incorrect I would assume perfect contact and expect some errors in your model that could be justified.

 

[peglor]

Thanks for the advice corus. Unfortunately, ProMechanica doesn't have a gap conductance option for modelling thermal interfaces. All it will model is convection (Using a single Convection coefficient) and conduction through solids. My plan was to vary the resistance of the interface until I got results in line with what I needed. Looks like I need to upgrade my FEA package to build the kind of model I was hoping for.

I'll run it with full contact between parts and see what happens. Thanks for your help.