[macfly]

Hi all,

I can't find how to properly set up basic heat transfer from a fluid to a solid. Here's what I've done so far:

In Gambit:
- create 2 adjacent faces
- connect the 2 adjacent edges into only 1
- specify continuum types to solid face and fluid face
- specify boundary conditions, the merged edge between solid and fluid is specified as wall

In Fluent:
- Fluent automatically created wall and wall-shadow for the fluid-solid boundary and the thermal boundary condition is set to coupled.
- No error messages, check case, everything seems ok.

Looked at attached jpg, the cold solid never heats up, instead, it lowers the temperature of the fluid along the fluid-solid boundary.

Quick help or redirection towards a tutorial that I don't know the existence of would be very appreciated. Thanks.

[krisshoe]

Try adjusting the range of temperature contours that you are plotting to something like 300-330K to capture the contours in the solid.

[macfly]

Hi krisshoe, in my first post/first figure, the solid is entirely at 300 K, no matter how I refine the temperature contours.

Here's another result : the bottom boundary is set to symmetry (or wall with 0 heat flux) instead of fixing its temperature at 300 K. Temperature of solid is initialized at 300 K. The solid heats up, this is good news, but the results are still incoherent : it's a steady-state simulation --> the solid should end up at 1000 K.

[Kwiaci]

Is this steady state or transient simulation? From what material is your wall? Do your results vary when you change number of iterations?

[macfly]

Hi Kwiaci,

1. Steady-state simulation
2. Wall material is copper from Fluent material database
2. I set 500 iterations but the solution is converged after 33 iterations.

Physically, I think that the solid should end up at 1000 K in a steady-state simulation. But I'm a little confused about how to verify by calculations on paper. Also, I don't understand why we have to initialize a steady-state simulation?

[Kwiaci]

To verify your simulation on the paper extract heat transfer coefficient from inner wall so you can asses inner wall temperature and then use equations from for example: http://www.wlv.com/products/databook/ch1_1.pdf

Can you write what are dimensions of your pipe (it looks like a very thin pipe).

[macfly]

Pipe has a radius of 0.1 m.

I'll see what calculations I perform, I think it should be pretty simple actually. Thanks for the nice basic résumé.

[Kwiaci]

One thing more: I would use axis instead of symmetry because now you simulate flow in a rectangular infinite channel. Using axis you will get logarithmic temperature distribution like in pipe - now you have linear distribution.