Reply To: Different residuals by changing boundary conditions

[Rob]

OK. The best choice of boundary here depends on how the flow is when it reaches the surface. A suction velocity will assume a uniform value so may give convergence issues if the flow is very disturbed. You've also picked up an asymmetry in the flow, which is probably moving slightly with each iteration. It's probably not far off what would happen in reality as perfectly symmetrical flows almost never exist.
Check the flux balance and see how good that is.