Discussion Forum

I'm relatively new to Comsol, so I may have made some rookie mistakes, but I am trying to validate a model's result.

I'm trying to model a steady-state problem which finds a temperature profile between competing convection and conduction in axi-symmetric 3D. A compressible, non-isothermal fluid passes through a restriction and then cools a pipe which is in contact with a hot surface at the outlet. What's been causing me problems is the heat transfer between the fluid and the pipe.

Bulk fluid velocities range from 2-30 m/s, so I applied the k-epsilon turbulence model. This looks quite nice and is nicely in the range empirical results. Using the convection/conduction models under heat transfer, the conduction also works quite well, and I've validated this with hand calculations in simplified models. Radiation effects have been added as manual terms in the boundary equations.

What I'm still having trouble with is the forced convection. The simulation converges, but typically shows hardly any heat transfer into the gas, no matter the velocity. I'm using the default thermal wall function as suggested in the model files. I've tried to follow the example "Forced Convection Cooling of a Hot Flat Plate."

At first I thought it was an entrance effect, but when I modeled flow down a long pipe (L/D = 60), the heat transfer was 1-2 orders of magnitude less than predicted by empirical equations such as Dittus-Boelter.

Are there any tips or tricks to setting this up correctly? Is it wrong to couple the k-episilon model with the heat transfer model? If someone's interested in taking a look I can send what I've done for the pipe model.

Version: 3.4.0

Thanks all.