Originally Posted by Sam6789 (Post 809992)

Hi CFD users, for my CFD cource at the university I have to simulate flow into a 2D diffuser. Section of the diffuser is reptangular. Fluid can be treated as incompressibile. As a data, I know only Reynolds number based on hydraulic diameter, at the outlet, and the exit of the diffuser is at the atmosferic pressure. I calculate the mean velocity at the outlet using the Reynolds number given me as a data, and then the mass flow rate.
As boundary conditions, since I know nothing about the inlet, I calculate the velocity at the inlet based on the mass flow rate,because the fluid can be treated as incompressibile.
Then I imposed velocity inlet as a boundary condition at the inlet, and pressure outlet(gauge pressure equal to zero) at as boundary condition at the outlet.
I simulate the flow for two different Reynolds numbers: 50 and 8200.
However, a strange thing appears: in both cases I have flow separation, but the reattach point in case Re=8200(turbulent flow) is much more far away from the diffuser outlet than in case Re=50 (laminar flow).
Is something wrong with my boundary condition?
Someone can help me? Is very important

Originally Posted by FMDenaro (Post 810033)

If you say "strange thing" I immagine you already know the solution of your flow problem. What references are you using?



Owing to the 2D model, you are using for sure a RANS formulation, your solution is therefore strongly dependendent on the model parameters.

Originally Posted by Sam6789 (Post 810041)

I don't have any references about this case, I'm assuming that the reattach point in case Re=50 should be much more far away than in case Re=8200, because turbulent flow gives more resistance to separation than laminar flow.
I'm using a k-omega SST model

Originally Posted by FMDenaro (Post 810042)

Have you ever studied the test case of the backward facing step in laminar and turbulent regimes?