[OBOLLI]

Good day to all.
I ask you for help, since I myself can’t find a logical solution to my problem.

The principle of operation of the installation under study:
Through two horizontal pipes, air enters a metal container, inside of which there is a metal funnel (Picture A).
The pink part of the funnel will later be made of perforated sheet with a large number of holes. But at this stage of design, it is taken as a continuous sheet of metal, i.e. air from the container should flow through the bottom of the funnel into a thick pipe, and then out.

I built MESH in ICEM 15.0 (Picture A. I can only in this version of the program) with a MESH size of 0.017 (it was this size that ensured the highest MESH quality). The quality of MESH for various indicators is satisfactory (Picture B, "Angle", of course, is doubtful).

There are no errors when opening the Meshfile in Fluent 15.0, and when you check with "CHECK", no errors also arise (Picture C).

I set the air velocity at the inlet 10 m/s, Pressure-Outlet and K-epsilon Viscous Model (Picture C).

Immediately after the start of the solution, errors occur with "turbulent viscosity limited to viscosity ratio of 1.0000e05 in XXX cells" and "reverced flow in XXX faces on pressure-outlet" (Picture D),

these errors continue until the end of the solution (Picture D).

Picture 7 shows one example lasting 500 Iterations. There were other attempts that lasted more than 650,000 Iterations and led to the same result. I open the resulting result in "cfx5POST" and get the following result (Picture E).

1) It can be seen on it that the air flow (Speed 10 m / s), leaving horizontal pipes, simply flows down like a liquid and further vortices are simply illogical.
2) There should also be a symmetry between the two inputs (since 2 pipes).
3) And as a result, the air does not exit at all through a vertical pipe, as if the bottom of the funnel was brewed. But this is not so !!! (Picture E)There are definitely "fluid" blocks.
4) Also, the air velocity over time is constantly increasing inside the container, which is physically impossible

For comparison: the same setup, but without a funnel with a flow velocity of 10 m / s, looks like this (Picture E).

Perhaps the matter is precisely in the poor quality of MESH according to the "ANGLE" indicator.

I dont know. Please help me find a solution to my problem.

[sufjanst]

Do you have the volume elements in different parts in ICEM? If yes, you should set both to fluid in fluent, sometimes this does not happen automatically. Alternatively, you can put all volume elements into one part and you should be fine.


Next thing is, is your scaling correct? Your y-coordinate goes up to 2,4 meters.

[OBOLLI]

Hello, sufjanst
* 1) Yes, they are installed as Fluid in the fluent. I can’t put all the volume elements in one part, since the middle part of the funnel will be a separate perforated sheet, for further research.

2) Yes, that's right. The height of the laboratory stand is 2.4 meters.