[landsber]

Hello, respected community.

I am trying to simulate with SU2 the compressible flow in a S-Duct (test case of S-Duct of Wellborn - see paper by Harloff et al. in https://ntrs.nasa.gov/archive/nasa/c...9920024502.pdf ).

mesh_wellborn.PNG

I assign boundary conditions as follows:

1) inlet (entry to S-Duct):

INLET_TYPE= TOTAL_CONDITIONS

MARKER_INLET= ( inlet marker, total temp, total pressure,
flow_direction_x, flow_direction_y, flow_direction_z )

2) outlet (exit from S-Duct)

MARKER_OUTLET= ( outlet marker, back pressure (static) )

3) Inner wall of S-Duct - Navier-Stokes adiabatic wall boundary

MARKER_HEATFLUX= ( inner_wall_marker, 0.0 )

Numerical method - ROE scheme, limiter of Venkatakrishnan = 0.5, CFL = 0.5.

As a result, the calculation does not converge and
the condition of mass flow conservation is not satisfied between the entry and exit planes.

In addition, at some iteration number, the solver gives a warning:
"There are XXX non-physical points in the solution".

In case of use Euler wall boundary, the calculation converges and
the condition of mass flow conservation is achieved.

Maybe someone faced a similar situation - Does anybody have some ideas why the RANS Spalart-Allmaras SU2 diverges for such a standard tes-case for internal flow ?

Any help will be kindly appreciate since currently we are unable to resolve that issue.

Sincerely yours,

Alexander

[nitish_anand]

Hey Alexander,

I usually ramp the back-pressure for supersonic turbomachinery cases, maybe you can try that out.

Thanks,
Nitish