I am trying to model steady-state, 2-D (X-Y plane) conjugate heat transfer (CHT) with one fluid two solid domains. It's a simple channel flow type of thing. I am resolving the boundary layer around Y+=0.4 . In CFX you have to use a 3-D geometry to model a 2-D problem. I kept the depth (z-direction) small and only have two nodes in Z direction. For the flow domain I am using symmetry BC but should I use a adiabatic wall for the solid domains' walls in X-Y plane? I'd like to calculate how much of the heat flux goes to the fluid and solid domains. To verify the energy balance in a solid domain, for a steady state HT problem, if I area integrate the wall fluxes around the surfaces, should the result be zero? I am having some problems about this. Interfaces are used between the domains. Thanks for your help.

Dear Barbos I think you should take atleast 3 nodes in z-direction, bocz with 2 nodes each of them will be on the boundary walls which you are specifying as symmetry BC. Regarding solid domain wall BC both BC's are same mathematically. bye Ram Dayal

Hi Ram,

No, there is no need to use 3 nodes in the z direction for a 2D model. Two works just fine as there is no gradients in the z direction to resolve.

Glenn Horrocks

Hi,

On the high Z and low Z planes use symmetry boundaries.

Yes, an area integral of the total heat flux into the solid should give zero (if no heat sources and a steady state simulation). Also have a look at the imbalances in the simulation to check you have accurately balance the heat, often convergence on residuals alone is not sufficient for CHT simulations.

Glenn Horrocks

Hi Glenn Thanks for making it more clear for me. Bye Ram Dayal

Thanks for your response Glenn. I have a heat flux input for the bottom surface. Still heat_in should equal to heat_out since it's a steady state. But I'm getting 0.032W heat flux in vs 0.040W heat flux out. The difference is not acceptable. I kept the iteration up to 1500 but nothing changes.

Problem SOLVED!

I used double precision and it converged.