CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   Boundary Layer (

Fabio March 16, 2009 07:30

Boundary Layer
Hi all, i'm studying a 3d duct flow i got both Tgrid and Gambit, is there anyway to do a y^{+} refinement near the wall boundaries?
The gemetry has some sharp edges and create a 3d boundary layer ad usual is very difficult i got in Tgrid some intrfaces issues when back to Fluent.
If necessary i shoul post some images of my domain.
Thanks for any answer.

Fabio March 19, 2009 04:09

I will try to put the question in a more easy way, using Gambit or Tgrid is possible to create a good mesh for a viscous fluid simulation without using the Boubdary Layer tool?
I got problems with this tool due to internal corners less than 45°

Adam S March 20, 2009 06:46

I for one donīt understand what you mean. A pic may help.

Fabio March 20, 2009 07:33

Hi and thanks, this is the image is a flap inside a tube,it's a 3d case i'm trying to run a viscous fluid simulation, but i got problems with mesh generation aronud the flap due tot the small angles of the geometry.
If i use a 3d boundary layer i got poor skewness and i useless even to try run the simulation, so i'm thinking to get a mesh good for a viscous case without usin BL but with an eye to y^{+} parameter.
I hope now is more clear if not i try to do best.

Adam S March 20, 2009 07:51

your tube is circular?
how is your meshing procedure?
you have notion of icem hexa blocking?

Fabio March 20, 2009 11:11

Yes the tube circular.
First i have meshed the edges to get a finer mesh where required, the i have meshed the surfaces with Tri/Pave because other tipes are not applicable or give poor skewness, then in BL panel i use:
Algorithm = Uniform
First row = 0.1
Growth factor = 1.1
Rows = 4
Depth = 0.46
Internal contimuity = Yes
Wedge conrner shape = No
Transition pattern = 1:1
i got those results
I now Gambit and Tgrid but never used ICEM

Adam S March 23, 2009 05:57

meshing around this flap isnīt easy. and my knowledge about gambit isnīt best, sorry. what does "wedge corner shape" do?
try to adjust the parameters for the specific corner to get better result.

Fabio March 23, 2009 06:07

Well actually i'm working around the flap with size function, skewness has improved but i need to tune the parameters in order to not get too much cells in mumber.
Anyway if you got some tricks about ICEM tell me, maybe some friends with that program could help me.
Thanks anyway ;)

Fabio April 1, 2009 05:37

Even if i got a better skewness seems that much work is needed, so my final question is: with a finer mesh near the wall boundries without using the BL tool is possible to get convrgence in a viscous fluid case?
Thank to anyone who will answer.

hassan79 September 5, 2010 00:02

HI , Good Day
for you case
First row = 0.1
Growth factor = 1.1
Rows = 4
Depth = 0.46

I aske about any relation ship between first row and groth factor and now ofv row and depth

my case it is expansion pipe i fond formala from gird spacing calculator to find first row ? iam not sure about this formal it is correct or no ?

could u help me
Re 8112
d= 0.0254

D = 0.09525
fluid = air
temprature inlet = 20
constant heat flux = 720
i want to calculate surface nusel no

thank u

Fabio September 6, 2010 09:34

Hello, that formula is reffred to a fal plate in free stream, i don't remeber if is needed som sort fo correction factor to apply the results in a duct case.
Anyway what i don't know with which number have you filled the y^+ field, it makes difference for the turbulence model that you could use appropriately.
Which program are you using to create your mesh? In the Gambit manual there is description for every factor involved in the boundary layer creation.

hassan79 September 6, 2010 09:52

i use gambit ? i want simulate heat transfer in expansion pipe with turbulent fluid flow i need to calclate surface nusselt no my problem in y+ i make bondary layer but i get y+= 30 how can i maked y+=1

how can to optain first row and growth factor no of row to get y +=1

Adam S September 6, 2010 11:36

when you now have y+ = 30 then simply refine your mesh.

in general you can not adjust it before, cause itīs depending on model size and flow conditions.

in your (simple) duct geometrie you can pre-calculate it analytically.

Fabio September 6, 2010 12:11

If i rember well is possible to have a preview of the boundary layer that will be created and the relative y+.
Anyway i have found this paper maybe it helps

All times are GMT -4. The time now is 16:07.