wallShearStress, wallGradU are inconsistent

dancfd · June 18, 2013, 23:57

Hello all,

I just tried to calculate the wallShearStress entries of my case from wallGradU (dU/dy), rho (all found using OF postprocessors and sample utility), and nu (from transportProperties). Since tau = mu * (dU/dy) = rho * nu * (dU/dy), I figure that tau / (rho * dU/dy) should be constant, however this is not the case.

Can anyone explain why there is a discrepancy? I feel like I should be able to calculate wallShearStress somehow using its constituent parameters. This is important to me since the wallShearStress utility does not work for laminar flow and I want to compare the forces on an airfoil in laminar and turbulent flow for the same case.

Thanks in advance,
Dan

dancfd · July 14, 2013, 15:07

Issue resolved - rather, issues:

1) for compressible cases, you must run "wallShearSress -compressible" with that flag in place. The only way to know to do this is to read the code for the wallShearStress utility. Suggestion: the utility should be able to distinguish between compressible and incompressible cases by itself, or force the user to use a flag for compressible and incompressible, to make this utility more user friendly.

2) by looking at the turbulence model (kwSST) code and the wallShearStress code, I found that OF calculates wallShearStress as follows:
tau_w = -(mu + mut) * (wallGradU)

When I made that change, the data matched.

Therefore, to calculate laminar, tau_w = -mu * (wallGradU) since mut = 0.

May others learn from my mistakes!

dancfd · May 25, 2014, 10:13

To continue this monologue, I noticed that when running wallShearStress with and without the compressible flag on the same case, the results differed by a factor of rho yet the dimensions in both wallShearStress files were the same. This suggests that the equation for tau_w (wallShearStress) changes as follows for compressible/incompressible cases:

wallShearStress:
$\tau_w = \nu_{eff} \frac{\partial U}{\partial y}$

wallShearStress -compressible:
$\tau_w = \mu_{eff} \frac{\partial U}{\partial y}= \rho \nu_{eff} \frac{\partial U}{\partial y}$

Therefore, if I understand correctly:
$\tau_{wCompressible} = \rho\tau_{wIncompressible}$

Which would not be a problem if the dimensions in both cases did not read the same (m2/s2):

Code:

dimensions      [0 2 -2 0 0 0 0];

This looks to me like a bug. Thoughts?

Regards,
Daniel

wyldckat · December 6, 2014, 05:30

Greetings Daniel,

Sorry for the (very) late reply, but only now did I trip over your thread, while looking for the exact answer you've given in post #2

The explanation to your latest question is simple: OpenFOAM uses a simplification of omitting rho to save up some CPU cycles in pointless multiplications by rho, when handling incompressible simulations... at least in most cases. This should also increase accuracy, since with each multiplication, there is always an error involved, no matter how small.

This to say that wallShearStress in incompressible mode does what it can with the data we provide it, namely since rho was not provided, it cannot multiply by an unknown rho

I also explained this with other details a few months ago (which was a few months after your last post), on this post: http://www.cfd-online.com/Forums/ope...tml#post509408 post #6

Best regards,
Bruno

vikramaditya91 · May 4, 2016, 04:52

If you are not happy with the yplusRAS utility provided by OpenFOAM which has it's problems documented here Official OpenFOAM and here http://www.cfd-online.com/Forums/ope...earstress.html, you might want to be aware of the fact that in the equation
tau_w = -(mu + mut) * (wallGradU)
mut is almost 0 for y+<3, so it can be ignored. This means that you will not have to call the mut, but just the
tau_w= mu*wallGradU, will do the trick.

randolph · July 18, 2018, 08:33

Quote:

Originally Posted by vikramaditya91

If you are not happy with the yplusRAS utility provided by OpenFOAM which has it's problems documented here Official OpenFOAM and here http://www.cfd-online.com/Forums/ope...earstress.html, you might want to be aware of the fact that in the equation
tau_w = -(mu + mut) * (wallGradU)
but is almost 0 for y+<3, so it can be ignored. This means that you will not have to call the mut, but just the
tau_w= mu*wallGradU, will do the trick.

_______________________

I feel this would not work for the wall functions. Does the mut automatically index the corrected mut on the patch?

June 18, 2013, 23:57	wallShearStress, wallGradU are inconsistent	#1
dancfd Senior Member Daniel Join Date: Jul 2009 Location: Montreal, Canada Posts: 156 Rep Power: 16	Hello all, I just tried to calculate the wallShearStress entries of my case from wallGradU (dU/dy), rho (all found using OF postprocessors and sample utility), and nu (from transportProperties). Since tau = mu * (dU/dy) = rho * nu * (dU/dy), I figure that tau / (rho * dU/dy) should be constant, however this is not the case. Can anyone explain why there is a discrepancy? I feel like I should be able to calculate wallShearStress somehow using its constituent parameters. This is important to me since the wallShearStress utility does not work for laminar flow and I want to compare the forces on an airfoil in laminar and turbulent flow for the same case. Thanks in advance, Dan immortality likes this.

May 25, 2014, 10:13		#3
dancfd Senior Member Daniel Join Date: Jul 2009 Location: Montreal, Canada Posts: 156 Rep Power: 16	To continue this monologue, I noticed that when running wallShearStress with and without the compressible flag on the same case, the results differed by a factor of rho yet the dimensions in both wallShearStress files were the same. This suggests that the equation for tau_w (wallShearStress) changes as follows for compressible/incompressible cases: wallShearStress: $\tau_w = \nu_{eff} \frac{\partial U}{\partial y}$ wallShearStress -compressible: $\tau_w = \mu_{eff} \frac{\partial U}{\partial y}= \rho \nu_{eff} \frac{\partial U}{\partial y}$ Therefore, if I understand correctly: $\tau_{wCompressible} = \rho\tau_{wIncompressible}$ Which would not be a problem if the dimensions in both cases did not read the same (m2/s2): Code: dimensions [0 2 -2 0 0 0 0]; This looks to me like a bug. Thoughts? Regards, Daniel sina.s likes this.

December 6, 2014, 05:30		#4
wyldckat Retired Super Moderator Bruno Santos Join Date: Mar 2009 Location: Lisbon, Portugal Posts: 10,975 Blog Entries: 45 Rep Power: 128	Greetings Daniel, Sorry for the (very) late reply, but only now did I trip over your thread, while looking for the exact answer you've given in post #2 The explanation to your latest question is simple: OpenFOAM uses a simplification of omitting rho to save up some CPU cycles in pointless multiplications by rho, when handling incompressible simulations... at least in most cases. This should also increase accuracy, since with each multiplication, there is always an error involved, no matter how small. This to say that wallShearStress in incompressible mode does what it can with the data we provide it, namely since rho was not provided, it cannot multiply by an unknown rho I also explained this with other details a few months ago (which was a few months after your last post), on this post: http://www.cfd-online.com/Forums/ope...tml#post509408 post #6 Best regards, Bruno vikramaditya91 and gebab like this. __________________ OpenFOAM: FAQ \| Getting started Forum: How to get help, to post code/output and forum guide Read this before sending me PM

May 4, 2016, 04:52	Just a side note while calculating yPlus	#5
vikramaditya91 New Member eu sou cfd Join Date: Jun 2012 Location: Brazil Posts: 18 Rep Power: 13	If you are not happy with the yplusRAS utility provided by OpenFOAM which has it's problems documented here Official OpenFOAM and here http://www.cfd-online.com/Forums/ope...earstress.html, you might want to be aware of the fact that in the equation tau_w = -(mu + mut) * (wallGradU) mut is almost 0 for y+<3, so it can be ignored. This means that you will not have to call the mut, but just the tau_w= mu*wallGradU, will do the trick. lucasbraganca likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
wallShearStress Monitoring	Mojtaba.a	OpenFOAM Running, Solving & CFD	6	July 30, 2019 05:14
Problems with YPlusRAS and wallShearStress	Peter85	OpenFOAM Post-Processing	21	May 4, 2016 05:15
wallGradU components	amin66	OpenFOAM Post-Processing	9	March 25, 2016 20:09
getting (0 0 0) wallGradU value when i m probing wallGradU at some points	suryawanshi_nitin	OpenFOAM	1	December 9, 2013 11:36
WallShearStress for settlingFoam	Elise	OpenFOAM Post-Processing	1	August 28, 2012 05:50

July 14, 2013, 15:07		#2
dancfd Senior Member Daniel Join Date: Jul 2009 Location: Montreal, Canada Posts: 156 Rep Power: 16	Issue resolved - rather, issues: 1) for compressible cases, you must run "wallShearSress -compressible" with that flag in place. The only way to know to do this is to read the code for the wallShearStress utility. Suggestion: the utility should be able to distinguish between compressible and incompressible cases by itself, or force the user to use a flag for compressible and incompressible, to make this utility more user friendly. 2) by looking at the turbulence model (kwSST) code and the wallShearStress code, I found that OF calculates wallShearStress as follows: tau_w = -(mu + mut) * (wallGradU) When I made that change, the data matched. Therefore, to calculate laminar, tau_w = -mu * (wallGradU) since mut = 0. May others learn from my mistakes! nimasam, eternityboy, kiddmax and 9 others like this.