hi,

the problem is similar to that of a shock tube. it is initialized as follow: inside the circle, rhoI, pI, u=v=0; outside the circle, rhoO, pO, u=v=0 (pI > pO, the other way around is also interesting). any reference/suggestion is appreciated. thanks.

linhbao

Hi there, I am not quite sure what the problem is, since you don't give much details, neither is clear waht you meant to ask. I would suggest to write the equations in polar coordinates (r, theta), and if the flow is not rotating and is axisymmetric, then just solve for the radial equation like for the shock tube test. The only difference should be terms due to the geometry. So if you have a solver for the shock tube test (one dimensional and cartesian) you should only add terms for the distortion of the geometry in each equation (and maybe a factor 1/r or so). My guess is that a shock propagating inwards will amplified while a shock propagating outwards will decrease in amplitude (inwards=to smaller radii, outwards=to larger radii). If the problem is really 3D then you need spherical coordinates, etc... and I would expect the same behaviour. P.