This is problem of natural convection in a vertical duct with proper insulation.Air at room temperature is sucked in at the inlet with uniform velocity. The duct is heated on the wall with a constant surface heat flux.The flow is in the laminar region,incompressible and we are considering the steady state. I am solving it in the PHOENICS CFD Software.I am facing problems with condition of gravity.Without gravity i am getting a parabolic velocity profile in the pipe as expected.But with gravity,the results are ambiguous and unexpected as i am getting maximum velocity at the grid nearest to the wall and minimum velocity at the centerline..violating the no-slip condition.

Can anyone help me out with this!! I am final year student in Mechanical Engineering.

You will get a high velocity near the wall, as that's where the heated fluid is! If the grid near to the wall is fine enough (in the normal direction), you will capture the wall boundary layer. If the grid is not fine enough, you won't. I suspect that your near-wall grid is not fine enough to see the boundary layer.