CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   Self preserving boundary layer ansys cfx (https://www.cfd-online.com/Forums/cfx/151221-self-preserving-boundary-layer-ansys-cfx.html)

davidwilcox April 6, 2015 23:57

Self preserving boundary layer ansys cfx
 
Hi,
I am having difficulties simulating a self-preserving boundary layer in ANSYS CFX. i used the loglaw profile at the inlet for velocity and k and epsilon. This is a 2D simulation. the domain is basically a rectangle ( nothing inside). i cant seem to get a self preserving boundary layer. near the inlet, I have du/dx NOT equal to 0. which is weird. HOW do i tell CFX that the flow is already fully developed at the inlet?

Thomas MADELEINE April 7, 2015 05:55

there is an option for that in turbulence at inlet ("zero gradient" normally) look at the help for that option but I think this is what you are looking for.

ghorrocks April 7, 2015 21:28

A nice way to get fully developed flows is to replace the inlet and outlet with periodic pairs. Then the converged solution is the fully developed profile.

You need something to make the flow happen - either use the options in the periodic pair, or use a momentum source to drive the flow.

davidwilcox April 8, 2015 23:11

Thanks for the suggestion. Even after using a zero gradient for turbulence, the flow is still NOT self preserving. This is worrying because the based on the log law, production equals dissipation but I do not see this with an empty box domain under 0 pressure gradient. I also noticed that at the inlet, AFTER calculations, there are velocity gradients, i.e. dU/dx not equal to 0. ANY fix to this?

Thomas MADELEINE April 9, 2015 04:13

If I understand your problem correctly you want to simulate a portion of an infinite duct where the boundary layer is fully generated so we don't have any changes among the duct direction (X here ?).

And so the boundary layer should be the same along your wall...
First question here should be what is the boundary conditions you used (specially for walls) ?

I understood your set the velocity at inlet with a velocity profile (logarithm law)
I suppose you did with a function.

What about the outlet and the wall ?

Are you sure of the theoretical thickness of your boundary layer (the one you simulate on your function)

Are you sure of your wall condition ? if you want something that precise I would define the roughness of the wall. but since there is only the sand-grain equivalent option...


All times are GMT -4. The time now is 18:41.