[haiteng]

I am simulating the natural convection of a heated cylinder immersed in a very large water tank, with a submersion depth quite large so the influence of buoyant plume on the free surface of water should be neglibigle. The free surface of water tank should be largely stagnant. The problem is, how should I specify the boundary condition on the free surface?

I am considering symmetry boundary condition on free surface because there is no gradient of velocity, temperature on this interface, or zero-shear wall boundary condition with a constant room temperature. However, fluid medium is different across the interface (water on the bottom and air on the top). Will this make any difference on the selection of boundary conditions?

[vinerm]

If the container's submergence depth is rather larger as compared to the container itself, why go into multiphase at all. Just apply a free-slip wall or pressure outlet. Both are equally good.