[sbvawt]

Thanks for your prompt reply. Felix, would you kindly respond to my question?

[puplu]

Originally Posted by FelixL

Hello, everybody,


so I just did some of the flat plate simulations with the corrected version of the turbulence model I posted yesterday. Cf distributions along the plates for two different cases (T3A - moderate freestream turbulence intensity (FSTI), T3B - high FSTI) and as you can see the now correct F1 blending function only affects the skin friction at the transitional and turbulent regime.

In T3A the skin friction is reduced which might explain the improvement of Cd values for Alex' airfoil cases. Still, using LANGTRY and MENTER's correlation for ReThetac the transition onset location is way too early. So either there are still some errors somewhere or the correlation needs calibration.

In T3B the transition onset seems to be predicted quite well, the skin friction is much too high, though. This is a flaw of the model itself caused by the high turbulent viscosity (approx 100 times nu) in the freestream diffusing into the boundary layer. Still, the results in the paper of LANGTRY and MENTER look much better - and I don't know why.


Greetings,
Felix.

The distance ahead of the flat plate may affect the CFD results, So adjust the value of the inlet nut/nu to get proper turbulence decay along the plate just as Exp, but not ro alter the correlation for ReThetac .

[sbvawt]

Hi Artur,

According to my understanding, the following correlation, proposed by Suluksna et. al 2009 is missing in the code last updated by you. I am proposing the following lines which can be added.

// CORRELATION: SULUKSNA et al. 2009
/* Flength = min(
scalar(0.1)*exp(scalar(-0.022)*ReThetatTilda_+scalar(12))+scalar(0.45),
scalar(300)
);
return Flength; */

Would you please verify (I am not an c++ expert - SORRY) and add these lines in your modified codes? Thanks.

[FelixL]

Hello everyone,

Sorry about being so slow again. The last few years I kind of drifted away from CFD and my current research focuses more on acoustics, but I will try to keep on working to finally finish the implementation of the turbulence model.

Tthanks for your comments, especially Artur and Mazharul. I updated the code to work in OpenFOAM 2.3.0 and also I found some implementation errors for the correlations by MALAN et al. and SULUKSNA et al. You can find the updated code in the attachements.

    2014-11-12 : - coding style updates
                 - ensured compatibility with OpenFOAM 2.3.x (thanks Artur)
                 - corrected correlations of SULUKSNA et al. (2009) and
                   MALAN et al. (2009) (thanks Mazharul)

Also, I created a git repository for the turbulence model where I will commit future updates. You can find the repository under https://github.com/langfeldt/gammaReThetatSST

In the next time I will try to validate the model on the ERCOFTAC-testcases (T3A, T3AM, T3B). This seems logical to me, before applying the model to more complicated cases.

As always, your help is deeply appreciated. If you find something "fishy" in the code, I am more than happy to fix the bugs.


Many Greetings
Felix

[sbvawt]

Hi Felix,

Thanks a lot for the latest update. We are benefiting from your codes and you deserve credit for that.

With best wishes,

Mazhar

[FelixL]

Hello everyone,


just a quick update: following puplu's hint and also by modifying the initial conditions in my flat plate cases (Sidenote: somehow, the solution of simpleFoam depends on the initial conditions in my cases. Does anyone have an idea why that is? A steady state solution shouldn't depend on the initial values) I obtain much better results for the T3A/T3AM/T3B validation cases.

You find my resulting skin friction coefficient distributions in the attachements. Especially the T3A results look promising. In T3AM, the transition begins a bit too early, in T3B it is too late. This could be a grid/boundary conditions issue, however. Therefore, the implementation of the turbulence model seems to be quite satisfying.

During the next days I will look into the non-zero pressure gradient cases.


Many Greetings
Felix

[FelixL]

Hello everyone,


unfortunately the results I presented above weren't sufficiently converged (that's why they were depending upon the initial conditions). I am sorry about that - I fell for the good looking results and didn't think further.

Anyways, since the last time I posted I did some improvements on the code, as you can see in the changelog:

    2014-09-22 : - add correlation selection in RASProperties dictionary
    2014-09-19 : - corrected typo in correlation by MALAN et al. (2009)
    2014-09-17 : - made equations consistent with SST model in OpenFOAM 2.3.x

You can download the most recent code of the turbulence model from the git repository:
https://github.com/langfeldt/gammaReThetatSST

One major addition is the ability to select correlations in the RASProperties dictionary. Please find an example RASProperties file in the git repository.

Using the most recent version of the code I did some (more) flat plate simulations with zero pressure gradient comparing the four different correlations I implemented in the code. You can find the resulting plots in the attachements.

As you can see, the correlations lead to quite different results in the T3A and T3A- cases. The original correlation by LANGTRY and MENTER (2009) seems to perform quite poorly while the correlations by MALAN et al. (2009) or SULUKSNA et al. (2009) seem to work well in case of T3A and T3A-, respectively.

Right now I can't really tell what's causing these differences, since - according to the published papers where I have these correlations from - the results should by equally good for all four correlations. I quadruple-checked my implementation of the correlations and I also checked the turbulence model implementation many times - all to no avail. Maybe there is a OpenFOAM-specific thing which I have overlooked when implementing the model or the correlations are simply sensitive to the solver. Hopefully we will find out one day

I will continue now to test the model on the non-zero pressure gradient cases. Also I plan to implement a boundary condition for ReThetatTilda, so you don't have to calculate the inlet value by hand anymore.

All comments and suggestions are highly appreciated!


Many greetings
Felix

[DLuo]

Hi Felix,
I was wondering how you're extracting the Cf value for your Cf v REx plots? Also are you just calculating REx as pVL/mu?

[FelixL]

Hello Andy,


I'm calculating the Cf values by using the wallShearStress utility. This yields tauW at the wall patches. From that, Cf is calculated using

Cf = |tauW|/(0.5*rho0*U0^2)

where U0 is the reference velocity, rho0 the reference density. Please remember, that for incompressible simulations, tauW is already divided by the desntiy.

The local reynolds number is calculated as Re(x) = U0*x/nu0, where nu0 is the kinematic viscosity and x the coordinate on the plate.


Greetings
Felix

[DLuo]

Thanks Felix,

I was normalizing with the patch area as well which was throwing off the magnitude of my calculations. And thank you for the clarification about rho0.

Regards,
Andy

[killsecond]

Dear Filex Langfeldt and Formers,


I have downloaded the latest version of gammaReThetatSST model and attempted to implement it into the OpenFOAM, however something went wrong.

zzwjcy@ubuntu:~/OpenFOAM/zzwjcy-2.3.0/src/turbulenceModels/incompressible/RAS$ wmake libso
wmakeLnInclude: linking include files to ./lnInclude
Making dependency list for source file gammaReThetatSST/gammaReThetatSST.C
make: *** No rule to make target `derivedFvPatchFields/ReThetatTildaInlet/ReThetatTildaInletFvPatchScalarField.dep', needed by `Make/linuxGccDPOpt/dependencies'. Stop.

I don't know how to solve this problem, help me. Thanks!

[JackW]

Hey Guys,

I understand this is a few months old now, but I am very grateful for the turb model!

I am just wondering if people have a link to an online database for the experimental data, or are you digitising plots?

Thanks a lot!

Jack

[JackW]

Originally Posted by killsecond

Dear Filex Langfeldt and Formers,


I have downloaded the latest version of gammaReThetatSST model and attempted to implement it into the OpenFOAM, however something went wrong.

zzwjcy@ubuntu:~/OpenFOAM/zzwjcy-2.3.0/src/turbulenceModels/incompressible/RAS$ wmake libso
wmakeLnInclude: linking include files to ./lnInclude
Making dependency list for source file gammaReThetatSST/gammaReThetatSST.C
make: *** No rule to make target `derivedFvPatchFields/ReThetatTildaInlet/ReThetatTildaInletFvPatchScalarField.dep', needed by `Make/linuxGccDPOpt/dependencies'. Stop.

I don't know how to solve this problem, help me. Thanks!

Also for your question, take a look at the first few pages. You need to add a Make folder.

I guess Felix has an overarching structure for his turbulence models and includes the Make folder in a higher directory

[Artur]

Hi All,

First of all, here's a good set of validation results for a zero pressure gradient flat plate at high Reynolds numbers (the pdf by Wieghardt et al. at the bottom is particularly useful but you have to digitise the plots for it, I recommend Plot Digitiser http://plotdigitizer.sourceforge.net/ and Shutter http://shutter-project.org/):

http://www.grc.nasa.gov/WWW/wind/val...rb/fpturb.html

For transition modelling the T3 test case from ERCOFTAG is very useful. If you don't have access to it here's plenty of plots for it plus some discussion on transition models:

http://uknowledge.uky.edu/cgi/viewco...dschool_theses

Finally, for the error in compilation you have to follow the advice given in the previous post by JackW, I've attached the contents of my Make folder fyi:

Make.tar.gz

Happy Foaming,

A

[JackW]

Thanks Artur,

I actually already had this data, just didn't read the readme very well. Ooops!

[mamly]

Hello my dear friends;
I am trying to write SST Gamma-Retheta turbulent code. It seems it is a 4 equation model including K-w-gamma-Re. but every where mentioned to SST 2-equation.
What is the relation between k- w and gamma Re?

[Artur]

Hi,

Here's a good discussion of how the model you are talking about has been implemented in OpenFOAM (by people from Chalmers, unsurprisingly):

http://www.tfd.chalmers.se/~hani/kur...transition.pdf

As for your question, k & omega are quantities used to provide the spatial and temporal scales for turbulence in the k-omega SST model, described here for instance:

http://turbmodels.larc.nasa.gov/sst.html

In transition modelling the other quantities are used to determine whether the flow has become turbulent or whether it's still laminar or transitional. Re-theta is the momentum thickness Reynolds number. Other quantities depend on the model but as far as I know, they are typically independent variables. Gamma is this case is the intermittency, governed by a scalar transport equations with some source terms on the right-hand-side, see section 3.2 from the first link. Also, have a look here for some interesting discussion on some of the concepts:

http://www2.coe.pku.edu.cn/tpic/201172015638907.pdf

If you are not familiar with turbulence modelling in general here's where I'd start reading up on the topic

http://www.cfd-online.com/W/images/3..._Beginners.pdf

Hope that helps,

A

[Jack001]

I was wondering if someone could advise me on the following, (hopefully Felix, because you mentioned you were working more with acoustics at the moment):

I wish to use the gammaReTheta model to initialise an LES simulation of an airfoil (maybe NACA0012 to start with). My flow is at very low Reynolds number (30,000 - 50,000 based on chord) and so I expect my suction side BL to perhaps undergo laminar separation and turbulent reattachment. I am interested in capturing these features.

I am completely new to OpenFoam and CFD in general, but with the help of people like you guys I have quickly been able to run some test case, and somewhat familiarise myself with the workflow of OpenFoam.

Any help would be much appreciated, and please keep forums like this one alive, they are invaluable to people like me!

[klausb]

Hello,

how are the values for /0/ReThetatTilda and gamma calculated?

Case: Airfoil, Re=1e6, Tu=0.06, U=14.6

Klaus

[Artur]

Hi,

It's generally quite tricky as the model is quite sensitive to the BC values. I don't have any ready formulae for your case but thesis by Langtry gives some pretty good discussion (e.g. see page 52):
http://elib.uni-stuttgart.de/opus/vo...Kap01_05_1.pdf

There's some more discussion here for airfoil cases:
http://citeseerx.ist.psu.edu/viewdoc...=rep1&type=pdf

I suggest starting here and looking at the related references. Probably best to start your simulations from looking at a canonical test case (the T3A mentioned in this thread and one of the published airfoil ones, e.g. SD7003:
http://www.as.dlr.de/hiocfd/case_c3.3.pdf )

Hopefully someone using this model more will also be kind enough to point out more things and narrow it down for you.

All the best,

A