first layer thickness & y+

lukas' · June 11, 2020, 01:23

Hey everybody!

My current project is the simulation of a heatet installation space that is filled with argon. For good results I want to do inflation meshing. Therefore I want to calculate the first layer thickness via y+. My problem is that I have only little information about the freestream velocity and the dynamic viscousity (because it is based on the temperature). If I estimate values that might be realistic I get very different results for the first layer thickness. Some of them are far to fine others are to coarse in comparison to the rest of my grid.

How can I deal with such a situation?

Thanks in advance!

K.C. · June 16, 2020, 03:31

What do you mean with unrealistic values?
Sometimes the boundary layer has to be very fine.

In general, I would take the minimal possible viscosity and the first estimation for your velocity and just build a first mesh. Test it, and calculate y+ from your turbulence model. Then you will see, if refinement is necessary.

lukas' · June 18, 2020, 03:27

Thanks for your reply!

For some parts I got fist layer thickness in the area of µm. I have a quit big model consisting of several bodies and I might get problems with the elements limitation of the licence.

colinda1 · June 22, 2020, 03:19

In any case it is good practice to make a mesh dependency study. Start with a rather coarse mesh and have a look at the results: e.g. plot the velocity and temperature profile in the boundary layer. It may give you a more clear image of the number of layers needed to capture your physical boundary layer. After all the y+ is not the only value to look at.
Then you make two or more finer meshes and study the impact of the refinements.

It would be a pity if a limitation in the license would prevent you from performing a proper mesh dependency study. It might be good to look at alternatives that do not restrict the number of elements. Then you may still have limitations on the available CPU resources to take into account in how fine you can get.

June 11, 2020, 01:23	first layer thickness & y+	#1
lukas' New Member Lukas Join Date: Jun 2020 Posts: 13 Rep Power: 5	Hey everybody! My current project is the simulation of a heatet installation space that is filled with argon. For good results I want to do inflation meshing. Therefore I want to calculate the first layer thickness via y+. My problem is that I have only little information about the freestream velocity and the dynamic viscousity (because it is based on the temperature). If I estimate values that might be realistic I get very different results for the first layer thickness. Some of them are far to fine others are to coarse in comparison to the rest of my grid. How can I deal with such a situation? Thanks in advance!

June 16, 2020, 03:31		#2
K.C. Member Join Date: Mar 2015 Posts: 36 Rep Power: 11	What do you mean with unrealistic values? Sometimes the boundary layer has to be very fine. In general, I would take the minimal possible viscosity and the first estimation for your velocity and just build a first mesh. Test it, and calculate y+ from your turbulence model. Then you will see, if refinement is necessary. colinda1 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] Help with Snappy: no layers growing	GianF	OpenFOAM Meshing & Mesh Conversion	2	September 23, 2020 08:26
[snappyHexMesh] layer not added	Rasmusiwersen	OpenFOAM Meshing & Mesh Conversion	1	January 2, 2020 09:43
[snappyHexMesh] snappyHexMesh Boundary Layer at Corner	panpanzhong	OpenFOAM Meshing & Mesh Conversion	5	July 3, 2018 05:53
Questions about Boundary Layer Thickness and Turbulence Models	famerfamer	STAR-CCM+	3	July 12, 2012 09:47
errors	Fahad	Main CFD Forum	0	March 23, 2004 13:20

June 18, 2020, 03:27		#3
lukas' New Member Lukas Join Date: Jun 2020 Posts: 13 Rep Power: 5	Thanks for your reply! For some parts I got fist layer thickness in the area of µm. I have a quit big model consisting of several bodies and I might get problems with the elements limitation of the licence.

June 22, 2020, 03:19		#4
colinda1 Senior Member Colinda Join Date: May 2012 Location: Brussels Posts: 151 Rep Power: 14	In any case it is good practice to make a mesh dependency study. Start with a rather coarse mesh and have a look at the results: e.g. plot the velocity and temperature profile in the boundary layer. It may give you a more clear image of the number of layers needed to capture your physical boundary layer. After all the y+ is not the only value to look at. Then you make two or more finer meshes and study the impact of the refinements. It would be a pity if a limitation in the license would prevent you from performing a proper mesh dependency study. It might be good to look at alternatives that do not restrict the number of elements. Then you may still have limitations on the available CPU resources to take into account in how fine you can get.