[Sidharthkp]

Hi,
I am doing a natural convection problem with imposed rotation in an annular space using ANSYS CFX as part of my academic work(yes I am new to this). My Temp difference (Ts-Ta) is around 20 degree which is above the boussinesq approximation limit(15deg.).

My problem is with material option.

When i choose ideal gas(in which i persume boussinesq is not used as i specify the reference temp), I am getting results very close to real situation.
When i choose "air at 25deg" option where density is treated as const property and bousinessq is applied by default, temperatiure is overpredicted by more than 10degree.

My doubt are:
1. how can i apply a fluid which is not treated as ideal and also boussinesq is not applies as well. The third option in materials eqn of state is defining it as 'real gas', but i don't know how to obtain all values that i have to specify in it.

2. Is there any other way in which I can choose a library fluid (air) by not choosing boussinesq approximation?

[ghorrocks]

1) Have a look in the CFX tutorial examples on real gas. You can get them from the ANSYS Customer site.. Not sure if the ANSYS Student Portal has them, it is a bit advanced for a student application.

2) Boussinesq approximation is the only constant density approach to bouyancy in CFX. The alternative is to define a density function versus temperature and then that will define the buoyant effects. A simple function would be ideal gas, but real gas and other EOS are also applicable.

[Sidharthkp]

Thank you for the reply.
I would try your suggestions and come back if anything troubles me