Aha, you must be working on a common class?? project as others 
You might find some help in my post to this topic: To capture the mesh better around sharp edges - #8 by DaleKramer
EDIT: If I were you, the first thing I would investigate is whether your expected Cd value of 0.0138 comes from a reliable and trusted (peer reviewed etc) source. A lot of NASA NACA real world studies have been found to be not reliable. If it is a reliable number to verify against with CFD, you still need to need to know all of the conditions and assumptions that were made in determining 0.0138. Only then can you determine which CFD ‘best practices’ to implement in your CDF study to even expect to get close CFD results. Otherwise you will just be forcing your CFD results to match a given result which is sorta useless in my book