The general procedure is correct. You mention coordinate systems. The Ansys input and results are probably in global cartesian unless you have transformed them. Are the analytical results perhaps in a cylindrical coordinate system? You didn't give the Ansys results. How "incorrect" were they?