# Problem with K omega boundary conditions

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

February 19, 2013, 10:59
Problem with K omega boundary conditions
#1
Member

Malik
Join Date: Dec 2012
Location: Austin, USA
Posts: 53
Rep Power: 13
Hi FOAMers !
I am studying some airfoils in 2D and before applying any turbulence model to them I test the model on a cylinder.

I tested the komega model which gave me good results with my drag and my strouhal.

However, when I want to see the k field in the domain, I see a high value at the inlet (the one I fixed) and then a fast decrease.
Do you know why I observe it ? Do you know how to fix this ?

My Reynolds is 1000
Here are my boundary conditions :

inlet
U fixed value 0.015
k fixedValue 0.00000375
epsilon fixedValue 0.0000000112
nut calculated value 0
omega fixedValue 0.037

outlet
p fixedValue 0
nut calculated value 0

up
U symmetryPlane
p symmetryPlane
k symmetryPlane
epsilon symmetryPlane
nut symmetryPlane
omega symmetryPlane

down
U symmetryPlane
p symmetryPlane
k symmetryPlane
epsilon symmetryPlane
nut symmetryPlane
omega symmetryPlane

cylinder
U fixedvalue 0
k kqWallfunction 0.00000375
epsilon epsilonWallFunction 0.0000000112
nut nutkWallfunction 0
omega omegaWallFucntion 0.037

frontAndBack
U empty
p empty
k empty
epsilon empty
nut empty

I enclose a screenshot of the kfield with komega so you can see the problem.

Thanks a lot for your help
Attached Images
 komega.jpg (67.7 KB, 213 views)

February 21, 2013, 09:52
#2
Member

Malik
Join Date: Dec 2012
Location: Austin, USA
Posts: 53
Rep Power: 13
Hi all,
since my previous post, I have tried to look at the same case but with komega SST turbulenceModel. I applied the same boundary conditions and found that we had the same jump at the inlet.

However this time we have some example of OpenFoam Cases with kOmega SST turbulence model. For example the motorbike.
Looking at the inlet, we find the same jump (see attached picture). In fact, what bothers me is that the solution depends on the size of the domain.

I don't think we should have this, should we ?

There is something that makes me a little more comfortable with this solution which is that the jump is very small.

One could say me that we fixed the inlet value quite arbitrarily and that there are no reason why the internalfield would have the same value. But should not we have zero Gradient Boudnary Condition instead ?

Attached Images
 motorbikekomegasSST.jpg (17.1 KB, 113 views)

 February 28, 2013, 05:18 #3 Member   Roland Join Date: Mar 2009 Location: Netherlands Posts: 92 Rep Power: 17 Hi, Does reducing turbulent dissipation (epsilon) help? regards, Sylvester

 February 28, 2013, 08:43 #4 Senior Member   Olivier Join Date: Jun 2009 Location: France, grenoble Posts: 272 Rep Power: 17 hello, Your k/epsilon ratio isn't ok. But why don't you try : - turbulentIntensityKineticEnergyInlet for k, - turbulentMixingLengthDissipationRateInlet for epsilon, - turbuelntMixingLengthFrequencyInlet for omega ? regards, olivier

March 1, 2013, 10:56
#5
Member

Malik
Join Date: Dec 2012
Location: Austin, USA
Posts: 53
Rep Power: 13
Actually I choosed my Espilon Value at the inlet with this formula :

epsilon = k^(3/2) * c mu / l
with l, the turbulence length scale. I choosed l = 0,05*D where D is my cylinder's diameter.

As sylvester suggested it, I lowered the epsilon value at the inlet and all disappeared. I divided my previous espilon inlet value by 100 (and my omega by 100)to get the first enclosed result. However the nut Field has significantly increased, which seems quite weird to me (2nd enclosed result)
Sylvester may be right, I should have chosen a turbulent length scale value lower than 5%.

For the boundary conditions, I did not use the turbulent BC simply because I did not know they existed. I just tried to implement it and it required to set the value for k, epsilon and omega, in addition of the turbulent intensity and mixinglength. Furthermore It does not change anything in the solution compared to the case where I only implemented the fixedValue BC.

From what you said and what i just tested I really think that I don't know how to define the turbulent length scale. Do you know some experimental formulas I could use ? (I already have looked for it on google but I found no result for external aerodynamic flows).

Attached Images
 k_lowepsilon.jpg (34.6 KB, 109 views) Nut_fixedValue.jpg (38.4 KB, 89 views) Nut_lowepsilon.jpg (46.3 KB, 98 views)

 April 18, 2015, 17:27 #6 New Member   bassam djedi Join Date: Apr 2015 Posts: 2 Rep Power: 0 I am doing simulation using open-foam on 2D aerofoil. I used komega model and put all boundary conditions and started the simulation. However, at Time= 26 i receive an error, I hope i can get help with this issue. Thank you #0 Foam::error:rintStack(Foam::Ostream&) at ??:? #1 Foam::sigFpe::sigHandler(int) at ??:? #2 in "/lib/x86_64-linux-gnu/libc.so.6" #3 Foam::GaussSeidelSmoother::smooth(Foam::word const&, Foam::Field&, Foam::lduMatrix const&, Foam::Field const&, Foam::FieldField const&, Foam::UPtrList const&, unsigned char, int) at ??:? #4 Foam::GaussSeidelSmoother::smooth(Foam::Field&, Foam::Field const&, unsigned char, int) const at ??:? #5 Foam::smoothSolver::solve(Foam::Field&, Foam::Field const&, unsigned char) const at ??:? #6 Foam::fvMatrix::solveSegregated(Foam::dict ionary const&) at ??:? #7 Foam::fvMatrix::solve(Foam::dictionary const&) at ??:? #8 Foam::SolverPerformance Foam::solve(Foam::tmp > const&) at ??:? #9 Foam::incompressible::RASModels::kOmega::correct() at ??:? #10 at ??:? #11 __libc_start_main in "/lib/x86_64-linux-gnu/libc.so.6" #12 at ??:? Floating point exception (core dumped)

 April 19, 2015, 23:25 Answer #7 Senior Member     CFD Join Date: Nov 2010 Location: United States Posts: 243 Rep Power: 16 HeyBassam, Just zip your files up, and I will fix your problem. Regards,