[caitoc]

Hi all,
Q1: Is it possible to model natural convection in a 2D horizontal model?

I am trying to simulating heat transfer of 8 cylindrical batteries by using a horizontal 2D fluent model. I modeld half of the cells.

When there is no forced air flow around the cells, natural convection should be dominant in the model. But since the model is 2D and horizontal, is it still possible to simulate the flow of the air (Bossinesq model is not available because gravity in z direction can not be activated)?

Attached pic shows temperature distribution in one certain time step when I simply use the boundary condition "inlet velocity=0m/s". It is not right, because:
(1) the velocity of air is 0m/s everywhere.
(2) In a correct result, there should be dramatic temperature gradient near the cell wall and then the adjacent air temperature becomes ambient temperature.

If it is not possible to use a 2D horizontal model to simulate natural convection,
(3) Q2: does it means that I have to use a 3D model in the case when air flow velocity=0m/s?
(4) Q3: can I use a very low inlet veolcity(0.05m/s), to simulate the case using a k-e(RNG) model? (I had some experience in modeling very low Re #, and I found laminar model always gives me inaccurate results, while RNG model is better.)

Thank you.