[Iaroslav]

Hello,

I am having trouble with modeling free supersonic air jet flow in Fluent 2.3.26. the issue arises from setting boundary conditions. This is how I set up my problem:

I am solving the flow axisymmetricaly. Because of this, my domain is 27D in the radial direction and 41D in the axial direction of the flow (where D is the diameter of the inlet). The inlet is located on the bottom of the left boundary(0.5D high), where the rest of the boundary is a wall. I set the lower boundary to "axis" and the upper boundary to "pressure far-field" where the pressure and temperature are ambient with mach number 0. The inlet is set to "pressure inlet" with appropriate values. Also, I am using the standard k-e model to start with.

The issue arises from not knowing what to do with the right-end boundary. Ideally I was to set the conditions to undefined, however I cannot find this option in Fluent or Gambit. Setting it to farfield would be inapropriate because that would imply a constant mach number which is not the case here. Does anyone have any suggestions?

The main questions is: How do I set an undefined boundary condition?

Thanks,
Iaroslav

[martingariepy]

If your outlet boundary condition his supersonic, you can set a pressure outlet of whatever you want cause the resolution will be only on one side (what I mean is that the cell before your outlet boundary is entirely defined by the cell upwind, so no calculation will involve your outlet boundary).

If your outlet boundary will not be supersonic, then you will surely have a shock and you will need to guess the outlet pressure with the Rankine-Hugoniot relations (if I understand your simulation clearly...)

[Iaroslav]

Hi Martin,

Thanks so much for the help. I was able to get a converged solution that made sense. One more question, would the value I set for back flow total temperature have an impact on my final solution? As I have read online, this value is only a guess to what the total temperature would be if the flow is backwards at that boundary and if anything, is only used as initial guess. Let me know if I am wrong.

Thanks,
Iaroslav

[martingariepy]

Your rigth. Backflow temperature is only in the case that you have a backflow. Usually, you should create your geometry to avoid that kind of situation at an outlet...

[Iaroslav]

I appear to be having problems with my result for the simulation I stated in my original post. Here are my exit conditions:
M=2.5, T=129.72, k=49, e=181930, u=570.75, density=2.721, p=ambient
My medium is air (I am using the compressible ideal-gas model). Initial values are all 0 with temperature at 300K.

The issue I am having is that my jet "core" is too long. From experimental results it should be about 10D, where I am having over 30D. I believe the issue is with my right-end boundary condition (which I have set to pressure outlet). Since I don't have any back flow I left back flow temperature at its default (300K) and gauge pressure to 0. Later on I tried varying those numbers slightly to see if they affect my final solution. What I have noticed is that in some cases my solution even diverged (for example, after I have changed the gauge pressure to 10000 Pa).

Can anyone help me with this? Am I using the wrong boundary condition? I also attempted to use "outflow", but it is not valid for compressible flows.

Thanks,
Iaroslav