[Kinnikuman]

Hello everyone. I'm new to CFD and I want to run k-epsilon model for initialization of LES. However, I got some errors like

Reversed flow on 391 faces (32.2% area) of pressure-outlet 26.
2 1.0000e+00 7.8929e-04 8.4984e-04 7.4759e-04 4.6880e-04 2.5813e-01 1:02:22 498

I think this is because of bad mesh as previous similar threads pointed out because I got this error after I changed my mesh. But problem is that the error happens after I make the mesh fine using inflation and element size. Would you tell me what should I change to remove the error and calculate correctly? And Would you tell me how to evaluate the mesh?

The model is 500 in pipe with 4.5 in diameter. There is small branch (diameter 0.4in, length 8 in) at 450 in as inlet outlet.png shows. I setup inflation and make the mesh fine (Cross Section.png). I also attached BC at inlet.
Thank you.

[LoGaL]

it's not a superfine mesh actually. I doubt it is the problem. Did the flow develop? Can you plot a velocity contour in a plane cutting the pipe into two? Do you see all red (3 m/s)?

[Kinnikuman]

Actually I changed a little bit (Cross section.png, BC inlet.png, overview.png).
The geometry is same (500 in pipe with 4.5 in diameter. There is small branch with 0.4 in diameter and 8 in length at 450 in from end of the main pipe. The fluid exit from the end of the branch.).

The calculation did not converge and flow did not developed (concole.png and velocity.png).If possible would you point out what I am making mistake? And if we conduct experiment in this condition, does reverse flow happens at outlet?
Cross Section.png

Overview.png

BC inlet.png

Console.png

Velocity.png

[LoGaL]

So there's no main flow in the large pipe? Only the one coming from the small pipe?

[Kinnikuman]

The flow is injected form the large pipe and exit from the small pipe. The end of the large pipe is closed.