To anyone who may be able to help:

I'm having some trouble with a steady state CFX simulation of a supersonic exhaust plume into ambient conditions. I think my problem lies in my boundary conditions, but I've yet to find a viable solution. Here's the deal:

I have a square control volume around a nozzle geometry. The entire volume is only one element thick, with surface mesh extruded through and symmetry conditions on the sides and through the center of the nozzle flow. The control volume of ambient fluid is, in my opinion, satisfactorily larger than the nozzle geometry. The entrance of the nozzle is subsonic, with total temperature and total pressure boundary conditions. The entrance of the control volume is low speed, subsonic flow. I've applied a velocity and static temp boundary condition at that inlet. The exit of the entire control volume (which would be the exit for the subsonic ambient air and the supersonic exhaust gasses) has a subsonic outlet boundary condition, with static pressure assigned.

My problem (on top of the fact that I can't get good convergence) is that the flow along the entire exit appears to be supersonic. The flow begins sub-sonic at the ambient inlet and at the nozzle inlet. The flow chokes at the throat of the nozzle as expected, but by the time the flow reaches the outlet boundary condition, the entire field appears to be supersonic. Also, instead of getting slightly higher pressures at the ambient entrance, the pressure is significantly greater than the volume exit (which does appear to reach the low ambient pressure that I've assigned it)... around 5x or 6x.

Now, I know that there is no 'mixed' supersonic and subsonic outlet in CFX, but I'm sure there's a way to at least get satisfactory convergence of the simulation and a relatively accurate prediciton of the exhaust plume, even if the conditions at the exit are not what they should be. I've even considered taking a guess at what the location of the supersonic and subsonic regions of the outlet are and seperating the geometry there to seperate the oulet boundary conditions into supersonic and subsonic regions.

I've increased the length of the control volume to try and let the flow energy bleed out before the exit, but the result is similar. I've tried turning the entrance of the ambient air into an opening... tried turning the free slip wall at the top of the ambient volume into an opening, an outlet, an inlet... no joy.

Any thoughts would be helpful. I'd like to at least get the boundary conditions correct and the flow in some state of behaving normally before I start messing with adaptive meshing etc. So any comments on the boundary conditions or problem set up would be appreciated. If you think all is well with those, then I'm also ready to listen to other thoughts ideas... anything that might help.

Thanks in advance,

newj

[purushothge]

HI ,
My case is also a similiar to your 's .But I am trying for vacuum condition. And I am using fluent for simulation .And i have obtained the apppropriate result while i am trying to simulate the same for atmospheric condition.
Purushothaman.N

[JMern]

One suggestion might be to switch from a velocity inlet to a pressure or mass flow inlet, velocity inlets and compressible flows don't always play nice.

Also what type of solver are you using, pressure or density based? Your problem may lie there if pressure based.