Learning Forum

[koustavbandyopadhyay]

Hi,

I am trying to look at the mean heat transfer rates from a thin isothermal wall (i.e. 0 thickness) under natural convection. Looking at online videos and forums, I have created a baffle surface on a plane in the middle of the control volume in spaceclaim. I merged the surface to the body in spaceclaim to preserve the meshing at the contact region. However, upon importing the geometry, there exists a baffle (wall B.C.) and baffle shadow.

Decoupling the wall and shadow and setting identical thermal boundary conditions (constant temperature) gives me different wall adjacent heat transfer coefficients on both sides. Coupling them results in no net heat transfer as a specific temperature could not be set.

How should I proceed? Is there any way to alter the geometry or the problem setup?

Many thanks!

Capture.PNG

[koustavbandyopadhyay]

[Rob]

OK, two problems here.
Decoupling means you're not passing any information between the two (now separate) walls. I'd expect the two halves of the model to produce similar results, but given it's natural convection, and therefore possibly transient, they're unlikely to be the same.
With a coupled wall heat can pass through the boundary. Depending on the temperature on each side there may be no heat transfer: that needs a temperature difference for that.

[koustavbandyopadhyay]

Hi Rob, thanks for your extremely quick response. This is what my question is precisely. Since decoupling is not working, I understand that I would have to put coupling as thermal B.C. on both of them (as coupling cannot be set on only one). However, when I do that, I cannot set a constant temperature on the wall as shown-
How do I make it a coupled isothermal wall?

[Rob]

It can't be? The wall will reach a temperature based on the local flow if it's coupled. You can add heat via the Heat Generation Rate. Why would the temperature be fixed?

[aitor.amatriain]

As says, if you set the temperature then the wall is not coupled. Where would be the coupling in that case?