[FMDenaro]

You have mainly forced convection from nozzle thus, provided that the temperature difference is not high, in a first approach you could disregard buoyancy effect. However, far from the nozzle the buoyancy can be relevant.

[FMDenaro]

Originally Posted by medaouarwalid

i noticed experimentaly that velocitie at let saying 2 meters in axial direction is close to zero , So to reduced the dimensions of the room i set pressure outlet at 2 meter instead of putting the real dimension of the room and set it as wall. I will change the geometri by setting the real dimensions and real boundary conditions . After that I will share with you the result.
A question, could the temperature of the jet in my case have an influence in compressibility ?

Sent from my F1f using CFD Online Forum mobile app

How can you set an outlet as wall??? the mass entering will cumulate in time without exiting and this case would be fully compressible! I see you are not aware of the basis of fluid dynamics....

[medaouarwalid]

Originally Posted by FMDenaro

How can you set an outlet as wall??? the mass entering will cumulate in time without exiting and this case would be fully compressible! I see you are not aware of the basis of fluid dynamics....

My english is not too good, maybe i didnt explain it well. the sides wall of the room are 2,5 meter from the outlet of the nozzle. The velocity at 1 meter was les than 0.1 m/s , so instead of drawin a room with real dimensions and setting the sides as wall (at 2,5 m), i reduced the domain and i put pressure outlet at 1 meter.

Sent from my F1f using CFD Online Forum mobile app

[medaouarwalid]

this is the real domain.

[Altomare]

Try to converge the temporal mean of those fields.

Sent from my Moto G (4) using CFD Online Forum mobile app

[FMDenaro]

Originally Posted by medaouarwalid

this is the real domain.

If you want to try to simulate the whole domain, you must take into account for some outlet region otherwise the incompressible solver cannot work.

[medaouarwalid]

Originally Posted by FMDenaro

If you want to try to simulate the whole domain, you must take into account for some outlet region otherwise the incompressible solver cannot work.

my mesurement was not for all the room but only for 20D ( 0.8 m) in the axial direction and 20D also radially . i'am not interssesting to simulate all the room but only in the region were the jets interact. that's why i reduced the domain and i put the sides and the down as pressure outlet where the velocities were almoste stable and close to zeo

[medaouarwalid]

Originally Posted by FMDenaro

Pressure outlet must allow the mass flow rate entering from the nozzles to exit. If that does not happen, the continuity equation is not satisfied, the mass increases and the error in the divergence-free constraint acts as a source driving to divergence.

hi , i did an other mesh just to verify , i dropped the number of elements to almost the half. in initialization, the warning of "convergence tolerance of 1.000000e-06 not reached during Hybrid Initialization."has disappear; the warning '' reversed flow ...'' has disappear also, i am in the 25th time step now and therer is no divergence detected . maybe the problem was in the mesh and not in the BC ! what do you think Proffesor ?

[FMDenaro]

Let the run going on ... it could be something in the mesh generation or, likely, a numerical instability on the finest mesh.