[JohnPuentes]

Thanks. I am working on the residuals output right now. I hope to identify the problematic region soon.

Meanwhile, I am playing with different boundary conditions for the Energy equation. I think this is also causing convergence problems in the model. I want to add a heat flux source on a wall (boundary source, which is more realistic in my case). The problem is that I have to define 3 parameters:

on wall-details: I need to choose the heat transfer mechanism.
on wall-solid: I also need to define the heat transfer mechanism
and there is wall-source, where I can also add a heat flux.

Am I overdefining the system if I select the 3 of them as heat flux? How should I define the heat transfer for the boundary details and the boundary solid values, perhaps adiabatic? See attached pictures

Thanks for any hint.

[ghorrocks]

The choice of boundary condition should be made based on what you are physically modelling.

So how is the heat applied int eh real system? Is the die heated? Is the die constant temperature? Or constant heat input? What other paths does the heat in the die have?

If it is not possible to define a sensible heat boundary condition then you have to move the boundary out a layer - which probably means modelling the die in this case, then the resin/fibre to die boundary is internal (so you do not need to specify all this stuff) and you define the heat conditions on the heating coils or whatever is driving the heat. And that should be much easier to specify acuurately.

[JohnPuentes]

That makes sense now. My die has a variable heat input. Since a chemical reaction (due to the epoxy nature) generates heat inside the system, the die will decrease the heat input to keep the temperature constant. I will add another solid and specify this heat source in the new solid and see how it goes.

[ghorrocks]

So it looks like you should model the chemical reaction as a heat source, and maybe the cooling coils in the die could be set at constant temperature. Then the resin/fibre to die interface is internal and you do not need to specify it. The heat flux is going to depend on the thermal resistance of the entire heat circuit - and that circuit includes the die. So it is starting to look like you need to model the die.

This is assuming that the cooling coils in the die are at a known temperature. Then they will be ideal for a boundary condition.

[JohnPuentes]

Thanks for all the great ideas. I was able to identify some problematic regions in my geometry with high residuals for the energy equations. I also identified the problem aof my heating source, which can be solved by adding the die as a solid (as Glenn suggested).

I am working on improving the geometry, boundary conditions and mesh for a better convergence. Meanwhile, I had some rough estimations but I have problems visualizing the results.

My problem now is that after obtaining a rough solution (poor convergence (10^-3) for temperature and (10^-2) additional variable added as a source in the energy equation, When I go to CFX Post, the default view lengend 1 is not available. I wonder is this happens becuase of the poor convergence or if it is a software bug. It is the first time that I am not able to see/activate the default legend view.

Thanks

[ghorrocks]

Then reset the view in CFD-Post.