[lycoming]

I'm attempting to model an aluminum engine cylinder to determine the cooling effect of airflow over the cylinder. The model consisted of a hollow cylinder, capped at one end, with a series of cooling fins along its length. I first ran a CHT with no airflow. The heat transfer coefficient of 200 W m^-2 K^-1 was applied to the outside surfaces. A wall was established on the inside diameter with 300 deg. F applied.

A second model was constructed with the same geometry but with air flowing over the cylinder. The domain interface between the solid aluminum and air (fluid) was set to Conservative Interface Flux. An inflation layer was included when meshing the fluid. Inlet velocity was set to 5 knots and an SST turbulence model was employed.

When comparing the models, the analysis with air flow is reporting hotter temperatures than the one without air flow. What am I missing?

[flotus1]

You are not necessarily missing anything.
Since you have a fixed temperature boundary condition to add heat in both cases, the maximum temperature should be the same.

But since you have a fixed heat transfer coefficient to remove the heat in the first case, whereas in the second case the heat transfer coefficient is a result of the fluid flow around the cylinder, the temperature distributions will be different.

[ghorrocks]

And also be careful about convergence on CHT models. Your solid only model should be fine, but you need to be extra careful with convergence on CHT models. The best way to ensure convergence in CHT models is to include imbalances in your convergence critereon.

[lycoming]

Thanks for the replies.
What do you mean by "include imbalances in your convergence critereon"? How do I do that? (Sorry - I'm new to CFD )

[ghorrocks]

In CFX-Pre: Solver Control/Convergence Criteria, Turn "Conservation Target" on. The tolerance is 1% by default, do a sensitivity analysis on this number to get the accuracy you need.

[lycoming]

Thanks for your help!