question about implementing boundary conditions

saygin

I'm confused about implementing wall and symmetry boundary conditions, does anybody have comments about the questions below?

I'm using an Euler (invisid-compressible) 2-D cell-centered FV (general geometry-not cartesian) approach with 3 dummy cells at each boundary. First question is:

Should I find extra solutions on the cell boundary for wall and symmetry b.c.? That means do I have to make the normal component of the velocity zero? Because I'm using dummy nodes and using a mirror image of the conservative variables, except the corresponding momentum has a reverse sign. I think this reverse momentum at the dummy nodes would give me the zero flux and zero normal velocity automatically. I used this technique for cartesian grid case (without any special computations at the boundary cell face) and it worked, in addition to that I'm using a dimension by dimension computation.

Second question is:

Some books say that for symmetry b.c. use the mirror image technique, with making the corresponding momentum negative at the dummy cells. So what is the difference between wall b.c. and symmetry b.c. for this invisid case?

Regards