Dear Fluent users,

I have recently started using Fluent. I am trying to model supersonic jet through convergent-divergent nozzle into a vacuum chamber at 1E-06 atmosphere. The inlet pressure is 10 bar of helium and I have tried both pressure outlet and pressure-far-field boundary conditions. The model converges fine when the outlet either boundary conditions are at 1 atmosphere, but the flow diverges as soon as the pressure in the oulet boundary is reduced. The model sover utilised is coupled, implicit, axisymmetric, and the viscous model uses k-epsilon RNG option.

*My questions is, What I am attempting is right?

*if any one has attempted similar problem of this kind with any success, and if so please your help will be most appreciated.

*If no, is fluent capable of handling this extrem pressure difference. i.e. is the algorithm valid in vacuum environment.

Thank you for your time and effort.

I think 1E-06 atmosphere is too low for NS equations to work.

wxl is probably right!

You have to work out the Knudsen number for your flow. If it is in the order of magnitude of 1, NS is not valid anymore. It must be a very small number...

Kn = lambda/L

lambda=mean free path of the molecules L=characteristic length of geometry

Has anyone an idea of the mean free path for molecules in air at 1E-6 atm ? Someone with good knowledge of kinetic gas theory?

I'm curious to know what the pressure-limit is to use NS for normal air at normal temperatures.

Laika, still orbiting

laika, I think you are right.

For Navier-Stokes algorithm to work the Knudsen number has to be less than or equal to 1e-03. For helium at 1e-6 atmospher I calculate the Knudsen number to be 0.2615, which is well beyond the scope.

Thank you for your help

Ali Khan, where did you read the criterior of Kn<0.001 for NS to work? If you use the slip wall correction, i guess you can push Kn to larger number.

laika, the mean free path for air at 1e-6 atm is about 66 mm at room temperature.