# "LRR Turbulent Model" Problems

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

 March 11, 2011, 13:01 "LRR Turbulent Model" Problems #1 Senior Member     Join Date: Jan 2010 Posts: 347 Blog Entries: 2 Rep Power: 9 Dear experts, By changing KEpsilon to LRR in constants/RASProprties OpenFOAM works fine apparently but the problem is in this model: k and epsilon which is set for inlet has a sharp decrease toward internal domain for example in my case which is a backward facing step it reach from k=7.5 to 0.0001, also same for epsilon. I think it is a reason which it is not possible to reach true answer. Settings are same as pitzDaily tutorial. i used pisoFoam. What is the problem? Any suggestion will be appreciated. Regards.

 March 12, 2011, 04:43 #2 Senior Member   Claus Meister Join Date: Aug 2009 Location: Wiesbaden, Germany Posts: 241 Rep Power: 10 Hard to day: Did u set the inlet conditions for R correctly, meaning everywhere zero apart from the main diagonal and on the k=... I haven't the conditions formula on me, but you can loop up in Versteeg.

 March 12, 2011, 04:49 #3 Senior Member     Join Date: Jan 2010 Posts: 347 Blog Entries: 2 Rep Power: 9 inlet R is set zero for all 9 elements.

 March 12, 2011, 05:46 #4 Senior Member   Claus Meister Join Date: Aug 2009 Location: Wiesbaden, Germany Posts: 241 Rep Power: 10 You must set the BC for R correctly: Compute k=3/2*(U_ref*Ti)^2 For U_ref take the velocity at the inlet, look in "U", set Ti = 0.05 (for other cases you muss adjusted it). If you have concrete value for k than set for R at the inlet: type fixedValue; value uniform (k/2 0 0 k/2 0 k); as you can see the k-values are divided by 2; one is untouched, this one points in flow direction (here z-axis). Since you integrate a epsilon equation with these, you must set the BC suitable. Somewhere in the user guide or programmer's guide is an entry, you gotta go for it.

March 12, 2011, 07:55
#5
Senior Member

Join Date: Jan 2010
Posts: 347
Blog Entries: 2
Rep Power: 9
Quote:
 Originally Posted by idrama You must set the BC for R correctly: Compute k=3/2*(U_ref*Ti)^2 For U_ref take the velocity at the inlet, look in "U", set Ti = 0.05 (for other cases you muss adjusted it). If you have concrete value for k than set for R at the inlet: type fixedValue; value uniform (k/2 0 0 k/2 0 k); as you can see the k-values are divided by 2; one is untouched, this one points in flow direction (here z-axis). Since you integrate a epsilon equation with these, you must set the BC suitable. Somewhere in the user guide or programmer's guide is an entry, you gotta go for it.
Dear Claus,

i have calculated k and epsilon. k=7.5 and epsilon=500.
But by seeing incompressible/simpleFoam/pitzdaily tutorial decide to use (0 0 0 0 0 0) for inlet R.
By changing inlet R to what you told answers are seemed to be better.
Two questions:
1- How can R be calculated to what value you told? i mean if x direction be streamwise or if it be 3D and ...
2- Is T=(a b c d e f) means:
a d e
d b f
e f c
or it has another format for symmetric tensors. i searched in user guide and programmers guide but nothing found about it.

 March 12, 2011, 11:36 #6 Senior Member   Claus Meister Join Date: Aug 2009 Location: Wiesbaden, Germany Posts: 241 Rep Power: 10 (1) Basically, when you consider a canonical reference frame where the inlet is pointing into the x-direction than you must set: R=(k 0 0 k/2 0 k/2), i.e. in flow direction k and into the others k/2. The reason why k appears here is due to the relationship trace(R)=2*k (look Versteeg). R is a symmetric tensor; to reduce memory consumption the symmetry property is exploited, i.e. 6 entries instead of 9, so (2) is right! R is called the Reynolds Stress Tensor and appearce by reynolds averaging procedure. My suggestion: Try to get Versteeg. Many times posted here and by simply entering it in google; you find in google books, definitely. Cheers and good luck

 March 12, 2011, 12:31 #7 Senior Member     Join Date: Jan 2010 Posts: 347 Blog Entries: 2 Rep Power: 9 I have Versteeg. thanks for your suggestion. if trace of reynolds stress tensor should be 2k and if your told R could be true, format of 6 elements tensor instead of 9 elements couldn't be what i told. it should be: T=(a b c d e f) means: a b c b d e c e f Isn't it? songwukong likes this.

 March 12, 2011, 13:58 #8 Senior Member   Claus Meister Join Date: Aug 2009 Location: Wiesbaden, Germany Posts: 241 Rep Power: 10 You got it, foamer!

 Thread Tools Display Modes Linear Mode

 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 OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post Edy OpenFOAM 7 August 10, 2011 12:00 truman Main CFD Forum 6 March 7, 2011 14:44 Olga FLUENT 2 October 11, 2002 15:05 David Yang FLUENT 3 June 3, 2002 06:13 Christian Holm Main CFD Forum 4 June 23, 2001 22:04

All times are GMT -4. The time now is 18:45.