[Cryo1991]

I'm doing 2D turbulent wall jet simulation using fluent recently.
I got a good convergence in residual but the results were not right.

The decay maximum velocity is much faster than experiment results in reference.



I tried different sets of grid and I'm sure it is fine enough (△x/b=1*e-6, △y/b=1*e-5 ) near inlet.
I also tried different turbulent model including STD RNG Realizable k-e with enhanced wall treatment and low-Re k-e as well.
The steady and transition solver were applied, different discretization methods were used, however the results are same.

inlet: velocity inlet 1m/s jet height: b=0.1m
outlet: outflow top: symmetry bottom: wall H=50b L=150b
fluid: air (fluent default)

Really don't know where the problems are. I think it is quite a simply case
Please help~~

[LuckyTran]

You have a 2D slot jet right? Not a round jet? Yeah something does not look right.

I think it is your outflow boundary. Try changing it to a slip wall. And second check would be to run a laminar case with on turbulence, I think you will still find that the decay of maximum velocity for the laminar case is also too fast. Can you confirm if the experiment is a free jet or confined jet?

[Cryo1991]

Originally Posted by LuckyTran

You have a 2D slot jet right? Not a round jet? Yeah something does not look right.

I think it is your outflow boundary. Try changing it to a slip wall. And second check would be to run a laminar case with on turbulence, I think you will still find that the decay of maximum velocity for the laminar case is also too fast. Can you confirm if the experiment is a free jet or confined jet?

thanks for reply
I tried laminar case and yes, as you said, decay rate wasstill too fast
Data from reference are all for turbulent wall jet without buoyance
I can't understand why outlet boundary should be slip wall. if so there will be no outlet.
I use outflow condition for outlet because such condition was used in reference and first derivative is 0 seems OK.
Or you mean the upper boundary should be moving wall not symmetry? i tried stationary wall, still not right.

[LuckyTran]

I meant the top BC. Btw what is your outlet BC on the right?

The reason I suspect it, is because it looks like your your flow is just going out the top when there should be some entrainment inwards. But it is hard to tell based on just x-velocity. Check a 3D velocity vector plot of the same region. Hence switching to a hard wall will be a solid check.

Switching to laminar is to show it's not some deep model problem but just something silly going on.

[Cryo1991]

Originally Posted by LuckyTran

I meant the top BC. Btw what is your outlet BC on the right?

The reason I suspect it, is because it looks like your your flow is just going out the top when there should be some entrainment inwards. But it is hard to tell based on just x-velocity. Check a 3D velocity vector plot of the same region. Hence switching to a hard wall will be a solid check.

Switching to laminar is to show it's not some deep model problem but just something silly going on.

well, actually the top BC is far from the inlet. H=50b
entire u contour

entire stream function


I just changed the top BC to wal. I have tried wall for top before because I
thought it was more similar to the experiment condition. But was a coarse mesh at that time. Hope it good this time.

[LuckyTran]

Sorry I'm not too familiar with the nomenclature and I don't really get what "2D turbulent wall jet" means. But I think it's just a matter of boundary conditions and making sure your mesh is fine enough to resolve the shear layer. If you have really coarse cells then numerical diffusion dominants and the jet mixes out really fast like you have here. But you want to make sure you have the right BC's.

The outflow on the right makes sense, the question is the top BC and the BC on the left. In a free jet with no nearby walls, the jet entrains the external flow so that you need some sort of freestream BC at the dump plane and top, walls are no good. A confined jet on the other hand, obviously you need the same walls as in the experiment.

[Cryo1991]

Originally Posted by LuckyTran

Sorry I'm not too familiar with the nomenclature and I don't really get what "2D turbulent wall jet" means. But I think it's just a matter of boundary conditions and making sure your mesh is fine enough to resolve the shear layer. If you have really coarse cells then numerical diffusion dominants and the jet mixes out really fast like you have here. But you want to make sure you have the right BC's.

The outflow on the right makes sense, the question is the top BC and the BC on the left. In a free jet with no nearby walls, the jet entrains the external flow so that you need some sort of freestream BC at the dump plane and top, walls are no good. A confined jet on the other hand, obviously you need the same walls as in the experiment.

Thank you very much. I'm thinking refining mesh. I checked y+ is less than 5 at wall. Maybe need finer near inlet for jet flow.

[LuckyTran]

The mesh resolution across the shear layer is the most important, and this you cannot detect with any y+. For example you may as well not have any walls and a uniform velocity inlet.

[fizsics]

Originally Posted by Cryo1991

Thank you very much. I'm thinking refining mesh. I checked y+ is less than 5 at wall. Maybe need finer near inlet for jet flow.

Hi Cryo,

Were you able to resolve the problem? I encountered a similar problem when trying to simulate a jet flame. I have tried using different grids and refined the mesh near the jet exit, but still the problem persists.

[LoGaL]

If it is a circular jet (axially symmetric), all RANS models have problems.