Hi,

I has used unstructured FVM public code NSC2KE to simulate a traveling shock of Ms=1.5 (actually I used 2D grid to simulate 1D shock ),the outflow B.C. used NSC2KE's Steger-Warming method . So this problem is a subsonic problem and post-shock mach number < 1 . Before the shock went out of outflow B.C. the result is very good , but when the shock went out my result is very poor because the B.C. reflecting was so big that there was a density and pressure down.

Then I used Thompson's non-reflecting B.C. , but the result is also poor .Could anyone help me to deal with the outgoing boundary conditions ?

Thanks.

Hi,

May I suggest you to read the discussion Oleg and I had yesterday on 'BC problem'. It treats boundary conditions through the method of Characteristics. Let me know if you need more details and/or references to the litterature.

Cheers, PG.

Hi,

Thanks for your help !

The problem has been solved by Thompson's non-reflecting B.C. (J.C.P 68,1-24,1987) using a characteristic method . I made a foolish mistake so the result is poor because I didn't convert the primitive form to conservative form . Now everything is good for subsonic and supersonic , although there is a little reflection about the most one persent and I should put this B.C. to un-regular geometry .