1

I'm new to Comsol and I'm trying to simulate the effect of a metal heater resistance (gold) on a laminar cooling flow (water). I would like to get the stationary temperature distribution in the fluid due to the heating effect of the gold resistance. I'm probably missing some boundary condition since I cannot get my solution to converge.

What I have so far:

  • 2D axisymmetric geometry
  • Non-Isothermal Flow, which is a coupling between a laminar flow and heat transfer in fluids.
  • inlet boundary condition (flow rate)
  • outlet boundary condition (pressure)
  • fluid temperature initial condition : room temperature (298.15 K)
  • boundary temperature condition for gold resistance (310 K)
  • convective heat flux defined with heat transfer coefficient

I attached a screen-shot of the model builder and of my geometry. Does anyone know which boundary conditions may be missing or otherwise why my solution isn't converging?

Thanks!

Geometry

Screenshot of Model Builder

General Grievance
  • 4,555
  • 31
  • 31
  • 45
thylakoid12
  • 137
  • 1
  • 9
  • 1
    Can you describe your geometry please. –  Apr 21 '17 at 19:07
  • @totyped Thanks, yes of course. The geometry is a tube (2mm diameter) filled with water. I'm actually only modeling the "water interior" of the tube. The tube has a gold resistance embedded into its wall. As the resistance heats up it transfers some heat to the fluid. I would like to know the temperature distribution in the fluid due to the heat of the resistance. The geometry is 2D axisymmetric so the tube translates into a rectangle with 1mm radius and the gold resistance is just a smaller rectangle stuck on its wall. – thylakoid12 Apr 22 '17 at 06:37

1 Answers1

0

It looks like you have the correct interfaces, assuming your heat flux is constant or definable, but I would have to look at your entire model in order to identify the problem. Not enough information here.

Since you have resistive heating, that can be included as another interface. In this case you would have to include the geometry of the cylindrical wall as well. This would replace your heat flux interface.

Steven Conrad
  • 51
  • 2
  • Thanks a lot for your answer! Indeed I think I will go for a simpler set-up and directly model the heat transfer to the fluid by setting part of the fluid at a higher temperature (rather than modeling the resistor). – thylakoid12 Apr 24 '17 at 15:03