Hello Hung Pham,

I want to ask you something because I don't feel clear, There are two different nozzles?. If that is the case, because it seems to be, I recommend you to check the mesh and the choked flow condition.

Regards.

Hi Jose,

Thanks for your answer. Two picture there only illustration this problem when i solve model 2D and 3D of C-D Nozzles. As you say, i haven't check the choked flow condition yet. I misunderstand i solve no. of interactions not enough, but  i increase no. of interactions so result velocity of air don't increase until the exit end of nozzle.

Picture below show as the result velocity of 3D nozzle i solve.

Regards.

Your question is not clear to me ... 2D and 3D flow problems are always quite different. Why do you plot only half domain, are you using symmetry? Do you use an inviscid model? How do you compare the flow rate for both case? The presence of a shock can also be a correct solution for that case

Hi Filippo,

This is 3D model, and these is plane result i export from 3D model to see easy. I use an inviscid model to solve it. You say "The presence of a shock can also be a correct solution for that case", but with compressive flow theory , velocity spectrum increase along length of nozzle.

Regards.

It depends on the pressure configuration, see Zucrow, Vol 1, sec. 4.7

Thank you, Filippo.

In the 3D case your solution displays flow separation and the formation of a free jet.

The flow separation in a supersonic flow will be the result of the formation of a normal shock. Do you observe this normal shock?

The solution in such cases does depend quite strongly on the boundary condition used at the end of the nozzle (p=constant??? or (grad p)_x=0???). Furthermore a correct description of the flow separation involves an accurate description of the viscous boundary at the side walls. Is you mesh sufficiently layers refined to resolve the boundary layer. Do you use a no-slip boundary condition on the side walls???

Yours sincerely, Mico

Hi Avraham,

Thank for your answer. I solve this 3D nozzle follow model nozzle of G.S Settle  in Fluent and ANSYS AIM, and it have the same results, inlet condition is compressible air 0.79MPa, 300 K, the result velocity of G.S.Settle is  ~530 m/s, but the result simulating is 480 m/s. As you say, i set inflation condition with 10 layer, with thickness 0.001 m and it is inviscid model.

Yours sincerely.

As you know, outlet boundary condition has a great influence on the simulation results inside the nozzle, especially, supersonic part, thus, care must be taken when you set outlet BC. Check it again and set a suitable outlet BC besides good mesh distribution in the throat region.

Regards.

You should also check wether the answer you obtain is grid independent. To check this you can make the numerical grid coarser or finer. The numerical result should converge to a well defined limit for a sufficiently fine grid. Regards, Mico