I have been running a fully ax
 

I have been running a fully axi-symmetric case with axi-symmetric geometry (grid) and axi-symmetric flow data. This case refers to the flow developed between two co-axial cylinders with no blades between them; the outer cylinder is still and the inner one rotates.

The problem we are facing is that there is no peripheral symmetry of the results ( p, U). The flow is not peripherally identical. This is visible from the velocity profiles (even the axial!) plotted at the same axial position over different meridional planes.

At a first glance, it seems that the source of the problem might be that the interpolation of velocities at faces is based on Cartesian velocities components rather than cylindrical components. Can anyone confirm this?