CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Main CFD Forum

Fluent-OpenFoam pipe flow comparison

Register Blogs Members List Search Today's Posts Mark Forums Read

Like Tree1Likes
  • 1 Post By RodriguezFatz

Reply
 
LinkBack Thread Tools Display Modes
Old   January 7, 2014, 04:28
Default Fluent-OpenFoam pipe flow comparison
  #1
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
Dear all,

I am trying to get the same results in Fluent and OpenFoam for a simple pipe flow (periodic). Unfortunately, the results are quite different and I am wondering if that is just normal for two solvers or if I am doing something wrong.
My settings are: DN25 pipe, standard k-e-model, mean velocity = 7 m/s, water, y+=1. I use "enhanced wall treatment" in Fluent and lowRe-wall functions in OpenFoam.
These are some results:
U-comparison-ske.png

k-comparison-ske.png

epsilon-comparison-ske.png
My feeling says something is wrong...
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   January 7, 2014, 08:26
Default
  #2
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 2,972
Rep Power: 30
-mAx- will become famous soon enough
which Turbulence Model did you use in OpenFOAM?
If it can help you: Wall treatment with geometrical restriction
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Old   January 7, 2014, 08:27
Default
  #3
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
It's both the standard k-epsilon without any addition.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   January 7, 2014, 09:26
Default
  #4
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 2,972
Rep Power: 30
-mAx- will become famous soon enough
As far as I know there is no "enhanced wall treatment" option in OF like in Fluent for standard k-epsilon turbulence model.
But you can use k-Omega SST model with nutUSpaldingWallFunction.
Then set k and omega with uniform value 1e-10 instead of zeroGradient, especially if you have low y+ (y+<=1)
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Old   January 7, 2014, 09:30
Default
  #5
Senior Member
 
flotus1's Avatar
 
Alex
Join Date: Jun 2012
Location: Germany
Posts: 1,098
Rep Power: 19
flotus1 will become famous soon enoughflotus1 will become famous soon enough
I can confirm your results in fluent. Is the "standard" k-epsilon model exactly the same in both solvers, i.e. are the same values used for the coefficients?
Messing with the coefficients in fluent, one can approximately reproduce the results from OpenFoam
flotus1 is offline   Reply With Quote

Old   January 7, 2014, 09:34
Default
  #6
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
Hi, I think the coefficients are the same for the standard k-epsilon. Most likely the boundary conditions are different, also I can't find exactly what Fluent does with all different turbulence values at the walls.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   January 7, 2014, 09:37
Default
  #7
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
Max, I also wrote in the thread you linked: I don't understand why you would set omega to zero for y+<=1 case. It should be some very high value instead.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   January 7, 2014, 09:46
Default
  #8
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
K-Omega-SST comparison:

U-comparison-komegasst.jpg

k-comparison-komegasst.jpg

omega-comparison-komegasst.jpg
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   January 7, 2014, 10:21
Default
  #9
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 2,972
Rep Power: 30
-mAx- will become famous soon enough
Quote:
Originally Posted by RodriguezFatz View Post
Max, I also wrote in the thread you linked: I don't understand why you would set omega to zero for y+<=1 case. It should be some very high value instead.
Thanks for correcting me.
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Old   January 8, 2014, 08:05
Default
  #10
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
U-comparison-SA.png

nuTilda-comparison-SA.png
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Old   January 10, 2014, 02:38
Default Standard k-epsilon model
  #11
New Member
 
Geon-Hong Kim
Join Date: Feb 2010
Location: Ulsan, Republic of Korea
Posts: 27
Rep Power: 7
Geon-Hong is on a distinguished road
Although you applied low Re BC on the wall, the standard k-epsilon model can not resolve appropriate turbulent field near the wall since it does not include a damping function.

Setting y+ to be close to the unity near the wall for the "standard" k-epsilon model is quite nonsense and I don't expect the solution to be correct with such settings.

It seems that you are trying to tighten a screw using pliers instead of screwdrivers. You can tighten screw somewhat using the pliers but it will not be tight enough.

Meanwhile the k-omega sst model can inherently resolve the viscous sub layer, and it is natural to estimate reasonable solution with y+ ~ 1.

If you want to maintain the y+ near the wall and use a k-epsilon family model, I'd recommend you to apply low Re k-epsilon model with low Re BC's, such as Launder-Sharma k-epsilon model.
Geon-Hong is offline   Reply With Quote

Old   January 10, 2014, 02:45
Default
  #12
Senior Member
 
RodriguezFatz's Avatar
 
Philipp
Join Date: Jun 2011
Location: Germany
Posts: 1,097
Rep Power: 16
RodriguezFatz will become famous soon enough
Hey Geon-Hong. Great! Thank you for the answer. That sounds pretty logical. I thought that maybe the wall function will patch all near boundary cells to apply the damping. But I didn't have a look at the source code, which should have revealed what you wrote... Thanks again!
AshwaniAssam likes this.
__________________
The skeleton ran out of shampoo in the shower.
RodriguezFatz is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
setup problems - LES pipe flow with cyclic BC (1) and direct mapped inlet (2) florian_krause OpenFOAM 22 June 13, 2013 21:25
Different flow patterns in CFX and Fluent avi@lpsc FLUENT 4 April 8, 2012 06:12
[GAMBIT] meshing in GAMBIT, a flow through a pipe having complex inflow geometry mazhar1613 ANSYS Meshing & Geometry 1 January 12, 2012 00:18
Can OpenFOAM generate flow at the speed of light? Michel_sharp OpenFOAM 6 October 24, 2009 04:09
flow in perforated pipe distributor pertupd ANSYS 0 August 12, 2009 08:36


All times are GMT -4. The time now is 21:24.