[Joachim]

Hey everyone!

When using the LU-SGS scheme on a structured grid (let's assume 2D for now), one needs to know the change in conserved variables in the ghost cells. How can I deal with that issue? Can I simply assume Dirichlet BC, so that deltaU = 0?

For some reason, my implicit scheme is highly unstable. It works for a uniform rectangular grid (freestream capture), but if I add a very small angle of attack (slightest perturbation), the residuals start increasing until the simulation diverge (oscillations appear as stripes in the flow field...).

Would anyone have an idea regarding the source of this issue?

Thank you for your help,

Joachim