[munishsharma]

Hi,

I am working on ANSYS Icepak (which uses fluent solver) for a natural convection problem. There is a heated plate mounted on a vertical wall. Rest 5 sides, I am applying open boundary conditions at a significant distance (at least 10X of plate dimensions in each direction) from the heated plate. I am applying atmospheric pressure on these openings and no slip on the plate. My solution is converging but no matter what I do, I am getting flow vectors in the direction of gravity . I have tried various combinations of under relaxation values, boussinesq approximation, ideal gas approach, very fine mesh, increased iterations, transient solution and many other options. I guess, there is some issue in figuring out boundary conditions on the open sides.
@Experts, please suggest something.

Thanks,
Munish