[Peralta]

Hi! I am currently writing a 2D structured cell-centered Euler code using the central spatial discretization scheme with scalar artificial dissipation (JST) and a simple explicit 5-stage R-K solver. Lately, I am stuck debugging the code (it blows up) and I am a bit frustrated so I wanted to ask you for any idea regarding how to fix my problem.

I have implemented the supersonic inlet/outlet farfield boundary conditions (for start) and I am using an O-mesh. I observe that even though I have implemented the wall boundary conditions using ghost cells, the condition for zero-normal velocity on wall is not satisfied. I am getting a "blowing" velocity starting from leading edge and this produces a "cushion" that surrounds the airfoil (see attached screenshots).

I am updating the boundary conditions as in the following pseudocode

// Desired values on BC wall face
rBC  = r[wallcell];
uBC = u[wallcell] - V[haloface]*nx[haloface];
vBC = v[wallcell] - V[haloface]*ny[haloface];
pBC = p[wallcell];

(V is the contravariant velocity and haloface is the same as the wallface with reversed (negative) face vectors)

//Then the first layer of halo cells is
r[halo1] = rBC;
u[halo1] = 2*uBC - _u[wallcell];
v[halo1] = 2*vBC - _v[wallcell];
p[halo1] = pBC;

//and the second
r[halo2] = rBC;
u[halo2] = u[halo1];
v[halo2] = v[halo1];
p[halo2] = pBC;

Any idea on what I am missing will be appreciated! Thanks!

Screenshots -> https://drive.google.com/open?id=16V...PaQUqi7BtAGR1Q