[Stabum]

Hello,
I would like to create a velocity profile at the inlet side of a duct, obviously due to the presence of the boundary layer.
Since CFX works in terms of Cartesian coordinate system, how can I do it?

I've read all the previous posts I could find regarding this topic, but, unfortunately, nothing has really helped me

Thanks a lot in advance

[Thomas MADELEINE]

you can have cylindrical component in CFX or create your own component system with CEL expressions.

I don't understand why you want to model the boundary layer in the inlet face... Do you know the size of the boundary layer here ? it can vary quite a lot from what there is before...

you can always extend your domain just a little (10 cells in duct direction) and add a free slip wall to avoid the boundary layer problem, but then you will have a very small boundary layer.
Or you can use the massFlow option as inlet and let CFX do the rest, but i am not sure it will be more accurate...

[Stabum]

Originally Posted by Thomas MADELEINE

you can have cylindrical component in CFX or create your own component system with CEL expressions.

I don't understand why you want to model the boundary layer in the inlet face... Do you know the size of the boundary layer here ? it can vary quite a lot from what there is before...

you can always extend your domain just a little (10 cells in duct direction) and add a free slip wall to avoid the boundary layer problem, but then you will have a very small boundary layer.
Or you can use the massFlow option as inlet and let CFX do the rest, but i am not sure it will be more accurate...

Thanks so much for your reply
The inlet duct is extremely long. I think that I can avoid modeling the ENTIRE duct if I model a velocity profile according to the equation v/vinf = (y/d)^(1/n) being v = local velocity; vinf = velocity outside the boundary layer; y = distance from wall; d = local boundary thinckness; n =7-9

It would be extremely helpful if you could add some more details on "create your own component system with CEL expression".

Thanks again for your help

[Stabum]

Originally Posted by Stabum

Thanks so much for your reply
The inlet duct is extremely long. I think that I can avoid modeling the ENTIRE duct if I model a velocity profile according to the equation v/vinf = (y/d)^(1/n) being v = local velocity; vinf = velocity outside the boundary layer; y = distance from wall; d = local boundary thinckness; n =7-9

It would be extremely helpful if you could add some more details on "create your own system with CEL expression".

Thanks again for your help

...and of course I think I can predict the local boundary layer thickness according to the equation:

d = 0.270 * l * Rel ^(-1/6) (if n = 9), being l the duct lenght

[Thomas MADELEINE]

If you know your geometry well enough to have the equation of your inlet circle then you can create your own components:

supposing C the center of your circle Xc the X coordinate etc...

create a new expression R = sqrt( (X-Xc)² + (Y-Yc)² for a circle in the XY plane
then create the function V(R)
and use it as velocity inlet...

I don't know by heart the name of CEL function so don't copy paste...

[Stabum]

.... THANKS A LOT! It was the help I neeeded!