Can we solve 2-d CFD problems by using 3-d CFD codes with periodical boundary condition in third direction?

As my experience(use a 3-d code to envaluate the lift of 2-d aerofoil), I let the length of computaion domain in the third direction to be much biger than the other two directions, then I averaged the third direction. It is expensive but maybe much reliable.

You should use symmetry conditions (zero normal derivatives) in the third direction and not the periodic bounday conditions. The third velocity component should be set to zero (Dirichlet condition).

regards

DML

If there is a real 2D configuration you can solve the problem with a 3D-code by setting symmetry boundary conditions at the surfaces in the third direction. Periodic boundary conditions are not really appropriate because this kind of BC does not imply zero mass flux over these surfaces.