Law of the wall-----Y++++++++++++

steveivy · February 12, 2015, 12:52

Hello, everyone

I'm new to CFD and I want to know and how to use the law of the wall and determine the height of first inflation layer. Here is my little knowledge about it and I have no idea whether it is correct or not: To find out the height of the first inflation layer, we have to determine the Y+ value we want to use. If we expect the flow to be separate, then we need a large Y+ so it is in the log-law region(but I dont know what does the magnitude depend on, ie. how to pick a Y+ value). If the flow is smooth, then the Y+ should be around 1 since the BL is thin. Then we can use equations to back out y value based on the Y+ and the y value will be the height of the first layer.

So please correct my statement above if you find any problem. Also, in Flunt, how to check whether I have the appropriate Y+ value or not based on the solution I got. For example, some one mention to use eddy dissipation rate and see if all the prism layer cover all the large gradient region.

Also, if you know somewhere online has a good explanation about Y+, please tell me.

Thank you in advance. Appreciate it.

spl · February 13, 2015, 03:57

Hi,

The law of the wall is made up of the viscous sub layer (Y+<5), the blending region (Y+=5-30), log law region and the outer region. The Y+ is a non dimensional height from the wall. The height of the initial cell is dependent on the turbulence model you use and the detail you want within the boundary layer. For instant SST requires a Y+~1 as this model solves the flow within the viscous sub layer and blending region. However if you use k-e the Y+ should be larger as a wall function bridges the gap to the wall through the inner region.

This link has some further info - http://www.arc.vt.edu/ansys_help/flu..._overview.html

Regards

steveivy · February 13, 2015, 11:01

Hi,
Thanks for your reply. Can I say that the mesh should be fine near the wall if I use low-Re model(SST) and the mesh can be coarse(because of high Y+ value) when using wall function model?
Thanks

spl · February 13, 2015, 13:13

Yes, the initial cell height (and corresponding y+) can be larger when using wall functions. Don't confuse a low Reynolds number turbulence models with low Reynolds number flows though.

steveivy · February 14, 2015, 11:38

Hi

So I want to model the flow in a passage with 90 deg turn(right angle). In the theory guide, it says the wall function approach is not recommended and a near-wall approach should be used if there is expected to be strong separation. So my question is: in fluent, those viscous models like k-e,k-w, spalart-allamars,etc should be categorized as wall function approach,right? If so,how can I choose near wall approach(I know near-wall needs a fine mesh in bl)? Thanks

spl · February 16, 2015, 07:18

Yes, both k-e and k-w use wall functions. Have a look at this page - http://www.comsol.com/blogs/which-tu...d-application/

Examples of turbulence models that don't use wall functions are k-w SST or low Reynolds number k-e. K-w SST should give good results but you could investigate the effect of different turbulence models for your particular case.

Regards

February 12, 2015, 12:52	Law of the wall-----Y++++++++++++	#1
steveivy New Member Liang Chen Join Date: Feb 2015 Posts: 7 Rep Power: 11	Hello, everyone I'm new to CFD and I want to know and how to use the law of the wall and determine the height of first inflation layer. Here is my little knowledge about it and I have no idea whether it is correct or not: To find out the height of the first inflation layer, we have to determine the Y+ value we want to use. If we expect the flow to be separate, then we need a large Y+ so it is in the log-law region(but I dont know what does the magnitude depend on, ie. how to pick a Y+ value). If the flow is smooth, then the Y+ should be around 1 since the BL is thin. Then we can use equations to back out y value based on the Y+ and the y value will be the height of the first layer. So please correct my statement above if you find any problem. Also, in Flunt, how to check whether I have the appropriate Y+ value or not based on the solution I got. For example, some one mention to use eddy dissipation rate and see if all the prism layer cover all the large gradient region. Also, if you know somewhere online has a good explanation about Y+, please tell me. Thank you in advance. Appreciate it.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Wall treatment with OpenFOAM	roby	OpenFOAM Running, Solving & CFD	48	May 28, 2021 11:38
[Commercial meshers] Fluent3DMeshToFoam	simvun	OpenFOAM Meshing & Mesh Conversion	50	January 19, 2020 15:33
[Commercial meshers] tmerge utility creates unwanted interface/walls comes in the final mesh	Shoonya	OpenFOAM Meshing & Mesh Conversion	11	January 20, 2012 06:23
law of the wall	alansari	FLUENT	2	February 2, 2010 09:32
Wall Function Approach - log law	Paul	CFX	2	June 30, 2005 06:06

February 13, 2015, 03:57		#2
spl Member spl Join Date: Nov 2014 Posts: 33 Rep Power: 11	Hi, The law of the wall is made up of the viscous sub layer (Y+<5), the blending region (Y+=5-30), log law region and the outer region. The Y+ is a non dimensional height from the wall. The height of the initial cell is dependent on the turbulence model you use and the detail you want within the boundary layer. For instant SST requires a Y+~1 as this model solves the flow within the viscous sub layer and blending region. However if you use k-e the Y+ should be larger as a wall function bridges the gap to the wall through the inner region. This link has some further info - http://www.arc.vt.edu/ansys_help/flu..._overview.html Regards

February 13, 2015, 11:01		#3
steveivy New Member Liang Chen Join Date: Feb 2015 Posts: 7 Rep Power: 11	Hi, Thanks for your reply. Can I say that the mesh should be fine near the wall if I use low-Re model(SST) and the mesh can be coarse(because of high Y+ value) when using wall function model? Thanks

February 13, 2015, 13:13		#4
spl Member spl Join Date: Nov 2014 Posts: 33 Rep Power: 11	Yes, the initial cell height (and corresponding y+) can be larger when using wall functions. Don't confuse a low Reynolds number turbulence models with low Reynolds number flows though.

February 14, 2015, 11:38		#5
steveivy New Member Liang Chen Join Date: Feb 2015 Posts: 7 Rep Power: 11	Hi So I want to model the flow in a passage with 90 deg turn(right angle). In the theory guide, it says the wall function approach is not recommended and a near-wall approach should be used if there is expected to be strong separation. So my question is: in fluent, those viscous models like k-e,k-w, spalart-allamars,etc should be categorized as wall function approach,right? If so,how can I choose near wall approach(I know near-wall needs a fine mesh in bl)? Thanks

February 16, 2015, 07:18		#6
spl Member spl Join Date: Nov 2014 Posts: 33 Rep Power: 11	Yes, both k-e and k-w use wall functions. Have a look at this page - http://www.comsol.com/blogs/which-tu...d-application/ Examples of turbulence models that don't use wall functions are k-w SST or low Reynolds number k-e. K-w SST should give good results but you could investigate the effect of different turbulence models for your particular case. Regards