Laminar-turbulent transition model in OpenFOAM
 

Hello.

Does OpenFOAM have a laminar-turbulent transition model? Can't seem to find one, but I might have missed it somewhere. Is anyone working on this?

/Mads

transition model? Sorry, I have not heard the term so far.
If you mean is there turbulence model that can predict transition (free-shear and wall-shear), I would say, yes.

Uh, it's a quite common term within fluid mechanics (here is some info ) and has been implemented in commercial codes like CFX and STAR for quite some time now, but only recently they really improved on it, in my opinion.

But maybe not in OF yet. I'll have to try myself, then :-)

/Mads

Ya, I see, I don't think there's any in OF right now, perhaps you need to write it yourself. BTW, I feel the formula of transition model is too complicated. I think it's better to be "simple, but not simpler":) Luckily, I am using LES and DES, so I won't suffer it for now.

Hi Mads

I have been considering implementing the Menter extension for transition flows in the k-omega framework, i.e. an addition of two advection-diffusion equations.
The articles/book chapters covering this implementation has one drawback, namely that the implementation is not fully documented. This is due to the fact that some of the emperical formulations are stated as properitary expressions, hence it is impossible to validate the implementation except if you buy a CFX-license.
Their empirical expressions are as I recall it based on the work by Abu-Ghannam & Shaw, however modified. As it was not essential at first hand to include transition effects in my modeling, I have left the progress here.

Best regards,

Niels

Concerning the transition model based on k-omega SST which is available in Fluent and CFX (addition of equations for gamma and Retheta), some research has been done to find empirical correlations to replace those proprietary of Ansys. For example, the following reference presents correlations that seem to produce decent results.

K. Suluksna, P. Dechaumphai & E. Juntasaro. Correlations for modeling transitional boundary layers under influences of freestream turbulence and pressure gradient. International Journal of Heat and Fluid Flow. 30, 66-75. 2009.

Thank you very much for pointing that out!

Best regards,

Niels

Concerning Menter, my feeling is,

1. He is a very active and very productive man;
2. His papers are always hard to get or download completely.

Thanks for the info.

Niels, how can I encourage you to proceed with the implementation? :-)

Isn't all information needed about this gamma-Re_theta model not readily available by now? It seems that the proprietary functions are found also (link)? Or maybe not, I am just starting out on gathering information on laminar-turbulent transition.

To my understanding, we need (for the gamma-Re_theta model at least) two extra transport equations (gamma and Re_theta). Some correlation function then switches on intermittency production when Re_theta > than some critical value. Seems reasonably easy on paper, but might be a bit harder to implement, I do not know...

But I know we need it for stalling airfoil flows :-)

/Mads

Hi All

I would also thank for all the information. I cannot wait to read it, however the limiting factor is always the lack of time:( I have need to have and nice to have, and the transition model is unfortunately in the latter catagory.

@Mads: If you know someone at DTU, who is soon going to do his/her Master Thesis in the area of coastal engineering or similar or simply fluid mechanics, then we might find a way to get it implemented:o) However, as it is a advection-diffusion equation, the frame-work should be available in OF. How the boundary conditions needs to be set is, on the other hand, probably the more difficult part. Though RISØ states that the method is stable, which is a good start.

Bests

Niels

Niels,

no I am not aware of any graduate students within this topic, as of now. But I could ask, although you seem to be closer to the action at dtu, no? :-) Were you at DANSIS yesterday? I'd like to say hello at some time.

I have no idea as to how to implement...well, anything...in OF but maybe you could give me some pointers to this, Niels? I'll try to look into it myself but you are right...time is an issue here. Maybe the OF-developers wants to add to this...or maybe we can convince them into this exciting and fantastic project?

/Mads

Yes, indeed I am closer to DTU;) I'll keep my eyes and ears open so we can get this project started.
No, I was not on the DANSIS meeting the other day, however looking at the program, I realize it was highly relevant and interesting. If you come close to Copenhagen/Lyngby then drop me an email and we can figure something out.

Best regards,

Niels

A colleague of mine has successfully implemented this transition model into our in-house flow solver. He found the missing correlations by examining many standard test cases in subsonic, supersonic and hypersonic flows (this is his PhD thesis by the way). If you guys are still looking for the missing correlations, you can find them in the following paper:

Krause M, Ballmann J (2008) Application of a correlation-based intermittency transition model for hypersonic flows. DGLR Kongress 2008 Darmstadt 23-25 Sept 2008.

Alternatively you can send me an email and I will send you the paper.

We found that the transition model is very grid-sensitive (the same goes for the CFX version) and it can be tricky to get a good prediction.

If you know the exact transition location, you could also try to implement a fixed transition model where you can force the flow to be laminar upstream of this point and turbulent downstream of this point. This normally works quite well (of course if you don't know where transition is then a transition model is better).

Hope this help,
Tue

Hi all,

I received couple of emails requesting the paper. I am attending a conference at the moment and don't have the paper in my laptop, I will send you the paper when I get back to my office.

Cheers,
Tue

Dear Sirs,
my name is Claudio Comis, and I am a PhD candidate in Energetic at the Department of Mechanical Engineering of University di Padova. I deal with aerodynamic shape optimization, and recently I’ve heard about OPENFOAM as CFD solver.
Bearing in mind that I am not "well experienced", anyhow, among the different turbulent models implemented in OPENFOAM I haven’t notice one able to capture laminar-to-turbulent transition, such as the one implemented in others commercial codes (that is, Langtry and Menter’s model).
With the present e-mail, I kindly ask if it is possible to know if someone in the “OPENFOAM community” has already implemented it, or in case You don’t know, if You can suggest me someone can help me.
Yours sincerely,


Claudio Comis
--
Ing. Claudio Comis Da Ronco
University of Padova
Department of Mechanical Engineering
Via Venezia, 1 - 35131 Padova, Italy
Phone 3494552408
claudio.comis@unipd.it
claudio.comis@alice.it

I am currently working on my Master thesis on laminar turbulent transition. I am doing performance tests between different CFD codes with different transition models.

As far as I know, Menters correlations should be fully published. Menters correlation is implemented in our universities inhouse code, so it should be possible to implement it in OpenFoam.

Is anyone currently working on this? Or is there a correlation already implemented?

paper
 

How can I have the paper above since it is published in a german national congress?

Is this question still open or someone already developed a transition model for openfoam?

Dear Tue, (or any other that may have worked on that)

Do you know how to fix the transition point (laminar-turbulent boundary layer) with OpenFOAM?
Or do you have any idea of any transition model applied in OpenFOAM? (I am using k-omega SST turbulence model, but if it is applied for any other turbulence model it may also be interesting for me)

Thank you for your help.

Regards,

José

I am not familiar with OpenFOAM so I can only tell you how to fix the transition point numerically. Assuming that you know the region where the flow is laminar, you should define this region in term of x, y, z before executing the simulation. You will then need to modify your flow solver in such a way that the turbulent viscosity is set to zero in every cell within the predefined region. Downstream of this region, the turbulence model will force the flow to become turbulent and the transition point is approximately at the interface. Of course this only works if you already know the transition point.