I try to use AUSMDV fluxing splitting scheme in hypersonic flow. But i am confused of it. In detail, the AUSMDV scheme combines AUSMD and AUSMV in normal monentum computaton. (supposed in xi direction),and no detail about the other two direction (eta direction, zeta direction), AUSMD or AUSMV which shoud be uesd? the same preplex also occur in enegy equation. who can give me some detailed advices on this. in addition, can MUSCL + AUSMDV scheme obtain high resolution in hypersonic flow? if so , can it get on well with two equation turbulent model? what about its ability of convergence? thanks in advance

Hi

I have coded and tested both AUSM+ and AUSMDV couple of years back for 2D compressible flows and a one-D shock tube. You can combine MUSCL and AUSM schemes and it shouldn't give you a problem.

To get 'u' normal to the face, say in 2D ( i am asuming it is a structured grid) u = U = u*xi_x + v*xi_y u = V = u*eta_x + v*eta_y

I hope this explains your question.

Ravi

Thank for your advice. which scheme is more robust in hypersonic .AUSM.AUSM+,AUSMDV, AUSMPW.AUSMPW. AND In ausmdv the modification only is made in normal flux, isn't it

Hi AUSM+ and AUSMDV behaved the same way. I think AUSM+ is little robust than AUSMDV. I don't know much about AUSMPW. I think it also may be some variant of these schemes.

In AUSMDV, the modification is made in normal flux.. What they mean by this is, only the normal flux component is evaluated for each face in xi ( or x) and eta (or y). If you are writing a unstructured code, you won't have xi and eta, in that case your normal velocity is U = u*nx + v*ny, where u and v are velocities in x and y directions and nx and ny are face normals. Now in structured grid, we do a transforamtion to computational plane ( from xy to xi-eta) and evaluate along each direction.

U_xi = u*xi_x + v*xi_y U_eta = u*eta_x + v*eta_y

here u and v are again velocity components in x and y directions xi_x and xi_y are face normals in xi direction and eta_y and eta_y are face normals in eta direction. We do the AUSM treatment to these velocities.

Ravi.

thanks for your interpereting in details, i wiil code it and then get help from you

Now i meet some difficulty in coding AUSMDV. in 3d structured grid, U= u*xi_x+v*xi_y+w*xi_z V = u*eta_x + v*eta_y + w*eta_z W = u*zeta_x + v*zeta_y + w*zeta_z

in my code , in xi direction, splitting U velocity get AUSM interface mass flux.but in AUSMDV. in referrence .it tells to modify normal flux with A combination of AUSMD with AUSMV. that is to say.only in F_xi ,the modification is used.and the other two G_eta,H_zeta keep the original scheme.as AUSMD. is this the correct usage?