CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   natural convection (

rachid September 18, 2005 14:06

natural convection
Hi I am working on simulating the basic natural convection in 2-D. There is a small rectangle (hot brick at constant temperature) and there is the large rectangle (flow field). I have given the temperatures for the brick and two side walls on the outer rect. In the tutorial n.5 they say we can model this similar flow with turning off the radiation model. Which boundary conditions the top and bottom edges should have?

I tried wall BC's for the outer rect. and no flow happened. After that i tried p out and p in but the flow in this case occured from the pressure difference

What is the correct BCs?

Evan Rosenbaum September 19, 2005 16:42

Re: natural convection
Wall BCs should work. You did specify a temperature-dependent fluid density somehow (i.e., ideal gas, Boussinesq, etc.), right?

rachid September 20, 2005 10:21

Re: natural convection
No i haven't. I suppose you mean the materials panel (Boussinesq) because the radiation model is off.

This makes sense though thank you so much... So how do I do that? I mean specyfying a temperature-dependent fluid density. Is there any literature or fluent tutorial you can advise?

rachid September 20, 2005 11:04

Re: natural convection
I found out that thermal expansion coefficient parameter which is available with the boussinesq type density is the key.

My new question is: Is it right to define a constant thermal expansion coefficient while it is temperature dependent? (3.47*e-3 K-1)

Evan Rosenbaum September 20, 2005 12:25

Re: natural convection
If you use Boussinesq and if temperatures don't vary too much you can use a constant expansion coefficient. Don't forget to set gravity too.

rachid September 20, 2005 14:43

Re: natural convection
Thank you Evan Rosenbaum. It works!!!

All times are GMT -4. The time now is 21:36.