pimpleFoam LES y+ and wall functions

Pj. · January 16, 2014, 04:43

Hi everybody,
i did a quick search but i could find an answer to this pretty basic question.
I'm running a case trying to reproduce the flow inside a wind tunnel. I'm running it with pimpleFoam and LES turbulence model.

the first problem is about the correct way to compute y+. Using the yPlusLES tool it says that the maximum y+ on my object (a building inside the WT) is about 200.
Since my firt cell is 0.7mm thick I did some calculation and, using

this would mean that at the centre of the first cell u is 420 m/s while u is never bigger than 24m/s (paraview says) in my domain.

Reversing those formulas to compute y+ i found it to be about 20-30 in the worst case scenario.

Which y+ is the correct one? Mine or the yPlusLES one?

______________________________________

Once I figure which y+ is correct, which are the corrects boundary conditions for U, p, nuSgs and nuTilda on a wall patch?

I found this topic, but it only says to try to have y+<7 wich is quite hard for me to obtain since it would mean to have a layer approximately 0.15mm thick that would cause courants to be huge (moreover i'm having some troubles with snappyHex adding more than 8 layers, but this is discussed in another topic)

Thank you very much

okroud · January 16, 2014, 07:00

Have you already taken a look at the y+ values with paraFoam?
1. Are all y+ values on your Object of this magnitude?
2. Is it just one single spot and where is it located?
In the second case you should check the mesh at this point.

PS:
I get the same results for u in the first cell center for y=0.0007 in the first formula.
But shouldn't you set

y = 0.0007/2 = 0.00035

since the wall distance is not the size of your cell, but the distance to the first cell centre? I'm not sure about that... (This would just reinforce your problem anyway).

okroud

Pj. · January 16, 2014, 10:18

The problem only occurs in a small spot (see picture) where the velocity isn't too big and where the mesh looks pretty much ok.

PS: the real maximum shown by the yPlusLES utily is indeed in a very small spot. In the picture the red area has an y+ of about 100-110

okroud · January 16, 2014, 11:31

Regarding the pictures the problem seems to be global. Y+ not 100 everywhere but still much more than expected.

An alternative way to calculate the gradient normal to the wall $\frac{\partial{u}}{\partial{y}}$ is to use the tool wallGradU. Execute this tool, then open your case in paraView and calculate y+ with your cell-size and the formulas of your first post. This should be a good comparison to the yPlusLES tool.

January 16, 2014, 04:43	pimpleFoam LES y+ and wall functions	#1
Pj. Member Luca Join Date: Mar 2013 Posts: 68 Rep Power: 13	Hi everybody, i did a quick search but i could find an answer to this pretty basic question. I'm running a case trying to reproduce the flow inside a wind tunnel. I'm running it with pimpleFoam and LES turbulence model. the first problem is about the correct way to compute y+. Using the yPlusLES tool it says that the maximum y+ on my object (a building inside the WT) is about 200. Since my firt cell is 0.7mm thick I did some calculation and, using this would mean that at the centre of the first cell u is 420 m/s while u is never bigger than 24m/s (paraview says) in my domain. Reversing those formulas to compute y+ i found it to be about 20-30 in the worst case scenario. Which y+ is the correct one? Mine or the yPlusLES one? ______________________________________ Once I figure which y+ is correct, which are the corrects boundary conditions for U, p, nuSgs and nuTilda on a wall patch? I found this topic, but it only says to try to have y+<7 wich is quite hard for me to obtain since it would mean to have a layer approximately 0.15mm thick that would cause courants to be huge (moreover i'm having some troubles with snappyHex adding more than 8 layers, but this is discussed in another topic) Thank you very much

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Wall functions - questions about implementation	AlmostSurelyRob	OpenFOAM Programming & Development	44	May 6, 2020 10:11
Enhanced Wall Treatment	paduchev	FLUENT	24	January 8, 2018 11:55
Water subcooled boiling	Attesz	CFX	7	January 5, 2013 03:32
LES automatic wall function	Sungho Yoon	CFX	7	October 9, 2008 07:00
Wall functions	John Grimm	Main CFD Forum	13	February 1, 2002 18:34

January 16, 2014, 07:00		#2
okroud New Member Join Date: Oct 2011 Posts: 19 Rep Power: 14	Have you already taken a look at the y+ values with paraFoam? 1. Are all y+ values on your Object of this magnitude? 2. Is it just one single spot and where is it located? In the second case you should check the mesh at this point. PS: I get the same results for u in the first cell center for y=0.0007 in the first formula. But shouldn't you set y = 0.0007/2 = 0.00035 since the wall distance is not the size of your cell, but the distance to the first cell centre? I'm not sure about that... (This would just reinforce your problem anyway). okroud

January 16, 2014, 11:31		#4
okroud New Member Join Date: Oct 2011 Posts: 19 Rep Power: 14	Regarding the pictures the problem seems to be global. Y+ not 100 everywhere but still much more than expected. An alternative way to calculate the gradient normal to the wall $\frac{\partial{u}}{\partial{y}}$ is to use the tool wallGradU. Execute this tool, then open your case in paraView and calculate y+ with your cell-size and the formulas of your first post. This should be a good comparison to the yPlusLES tool.