[DGIorio]

Hello, I would like to know how the supersonic outlet boundary condition works on SU2?
Physically it's not necessary a boundary condition (BC) on supersonic outlet, since the flow properties just depends on the upstream properties, but numerically it's necessary some BC.
Usually it's extrapolated, using a zero gradient (∇⋅u = 0) or just ∂u/∂x = 0.

I took a look at the function BC_Supersonic_Outlet (file CEulerSolver.cpp) and there is a comment saying that the flow variables are interpolated (probably meaning extrapolation). There is also a few lines like this:
V_outlet[0] = V_domain[0];

So, is it right to say that what SU2 does is define the flow properties in the outlet as the same value as the upstream?

Thanks in advance!

[pcg]

That is correct, the supersonic outlet consists of just evaluating the convective flux (as done for the interior of the domain) since there are no incoming characteristics.

[DGIorio]

Originally Posted by pcg

That is correct, the supersonic outlet consists of just evaluating the convective flux (as done for the interior of the domain) since there are no incoming characteristics.

Ok, thanks for the answer.

Just to clarify, it means that it is used ∇⋅F=0 instead of just ∇⋅U=0, where F is the convective flux and U={ρ,ρv,ρE} (using SU2 notation presented in governing equations page).

[pcg]

What is extrapolated are the primitive variables (not multiplied by density).
In the absence of source terms (spacial or temporal) the divergence of the flux is zero at every node.