Marker interface
 

I have been trying to run a simulation using SU^2 to analyze a rocket nozzle flow in a supersonic exhaust diffuser. I have been encountering some issues with the problem that I believe is due to information not being communicated across the nozzle exit and the diffuser inlet. For this situation, I believe I should be using MARKER_INTERFACE to define the interface between the rocket nozzle exit and the diffuser inlet. The reason I say this is that even with the rocket nozzle and diffuser running with the design conditions, the flow goes from supersonic at the nozzle exit to instantly subsonic at the diffuser inlet. I have verified by increasing and decreasing the total pressure and diffuser exit pressure, respectively. Is this true? If so could someone provide me an example on how to specify MARKER INTERFACES in SU^2's configuration file?

More information
 

Here is some more information about my file. I am using Euler as my solver, and the setup is 2D, axissymmetric. For the initial solution, I do not use a turbulence model. I hold my CFL number constant at 1.0.

I have attached a screenshot of my geometry. The rocket nozzle is a Mach 4.4, contoured nozzle. The diffuser geometry is a second-throat exhaust diffuser with a subsonic diffuser at the end. The total pressure and temperature at the nozzle inlet is 500 psia and 5600 deg R, respectively. The ratio of specific heats (gamma) is 1.22. The diffuser exit pressure is ~ 14.174 psia.

My objective of this simulation is to determine the diffuser's starting pressure ratio (total pressure / diffuser exit pressure). According to my calculations, the diffuser should be started for a pressure ratio greater than 21. I have ran this case with another CFD solver (ISAAC) to help solve my problem, and I was able to run successfully with ISAAC. Increasing the number of iterations did not help with this problem; it actually made it blowup (NaN).

Mesh?
 

That thumbnail shows an outline of your nozzle, but doesn't show your mesh domain at all...would you post an image of your mesh?

Another approach
 

After talking to Dr Power, I do not believe the marker interfaces are the issue. I will try re-running with a lower test cell pressure for marker inflow, and also try using the SA turbulent model.

It's working!!!
 

Finally, I got it working. I think the issue may be in my nozzle design. The total pressure had to be increased to 5000 psia, 10x's more than the calculated value. For amusement, enjoy the video:

http://www.youtube.com/watch?v=AXwGVXD7qEQ

Great! Anyway, in this kind of problem the right boundary conditions are MARKER_INLET and MARKER_OUTLET, and keep in mind that the initial values are computed using MACH_NUMBER, AoA, and FREESTREAM_PRESSURE.

Best,
Francisco