I use Fluent 6.2 version. I am simulate the Laval nozzile. In the convergent part, it is subsonic. In the divergent flow, the flow is supersonic, when the pressure at exit is very low. Now I increase the back pressure at exit, this pressure would propagate upstream to genreate a shock wave which would be between the throat and the exit, according to the text book. However, the simulate result showed that this back pressure could not be exerted on the flow. The result of the contour showed that the the flow always supersonic flow, the shock wave is not generated. Firstly, I use the converged superonic flow in the divergent as initial condition, the pressure at exit is 7725 pa. Then, I given the "pressure-outlet" as 505325 Pa, then computing. When I plot, I find that the pressure at exit is still 7725 Pa.

Could any body comment this? How to make the back pressure on role to the flow at the exit?

Thanks in advance!

You might be able to calculate the mass flow at the exit relative to your desired pressure, and then use this as a "target mass flow" input into your boundary condition at the exit plane.

Thanks for your response.

However, there is no "target mass flow" at the boundary condition at the exit plane.

If you are simulating just the nozzle itself, with no exterior, then the answer to your problem is simple: the exit of the flow domain is fully supersonic, so whatever condition you try to impose on the exit boundary will be totally ignored by the solver. This is completely physical: in a supersonic region, any perturbation cannot propagate upstream, because the velocity which the perturbation traveles with (sound speed) is always smaller then the flow velocity (supersonic speed = higher then sound speed).

To be able to overcome this problem, you have two options:

- first, if you do not want to modify the grid, then you will have to restart the calculation using a small upstream pressure (let's say, 1.5 x exit pressure), and slowly increase it; at a certain pressure ratio, depending on the nozzle geometry, the throat will become choked, and after a slight increase of the pressure ratio, a shock system will be generated in the divergent part of the nozzle; if increasing the P.R. even more, then the shock system will travel from throat toward exit, and at the design P.R. will reach the exit and vanish.

- secondly, if you are willing to modify the grid, then go back to Gambit, and add a generous amount of exterior domain (at least 25 nozzle diam. in lenght, and 10 nozzle diam. in height); then recalculate the flow with your present conditions; now, if you modify the exit pressure, at 25 nozzle diam. away, the exit being subsonic, the perturbation will propagate upstream and you will be able to obtain the shock system inside the nozzle.

Personally, I prefer the first way, because it requires a smaller grid and it is much more realistic.

All the best,

Razvan

[Zoli]

Thank you !