shell conduction @ thin walls

Jiri Beran

Dear Fluent Users,

I am simulating concentric pipe flow with two pipes (8cm and 15cm in diameter) each 12m long as part a of my problem. The liquid of my interest is prosaicly water . Pipes are made of carbon-steel.

My calculation involves heat transfer and since my focus is elsewhere in larger domain, I didn't really mesh fluid-zone in pipes / annulus very well.

Additionally, I am using "carbon-steel" coupled 5mm thin wall (viz. Fluent user guide 7.13.1) with SHELL CONDUCTION.

The end result for this conjugate heat transfer problem I am getting looks very peculiar.

My "ambient" operating temperature is 325K, the laminar yet buoyant flow in both pipes is along the gravity vector at 310K. The "spooky" shadow-wall zone was initialized at 310K and solution converges well.

The moment I shut the flow in the annulus (walls B.C. swapped for outlet&inlet B.C. and solution in annulus set trivial) Colorful Fluid Dynamics begins.

After only few seconds, the thin wall gets "shell conducted" up to 325K, while both flows are still at approx. 310K due to its large thermal capacity.

I certainly believe that this is non-sense, since there is no material on Earth, which would make 5mm thin pipe shell conducting over 12m like hell and haven't exchanged any heat to the flow around it in such a short time.

When I switch off "shell conduction" Fluent starts behaving again. Had anybody ever noticed similar trouble?

As I said I am not interested in the pipe else I would fine-mesh it and run a numerical solution instead of fudging in some semi-analytical fix. However, the aspect ratio 12m x 0.005m is a good reason not to do so. I wished Ansys INC. ever disclosed how they do "shell conduct" it in their code.

Best regards Jiri