Abe-Kondoh-Nagano low-Re k-epsilon model
 

I recently implemented the low-re k-epsilon model by Abe, Kondoh and Nagano.
(Abe K., Kondoh T., Nagano Y., 1994. A new turbulence model for predicting fluid flow and heat transfer in separating and reattaching flows-I. Flow field calculations. Int. J. Heat and Mass Transfer 37(1), pp 138-151).

It's probably the low-re ke model that is best suited for heat transfer problems.
This version should be the most stable among those by Nagano et al, as it uses an easy wall condition for epsilon.

At the wall:
k=1e-10
epsilon=2*nu*k1/y1^2, where k1 and y1 are the cell-center values at the wall.
This wall function for epsilon is already implemented in openfoam: epsilonLowReWallFunction.
It was intended for the v2f model, but it's the same, at least for low-re cases.
(there may be a problem with the code of epsilonLowReWallFunction (see here), but it doesn't seem to affect this model too much).

If anyone is interested, you can find the model attached to this post.
it is called AbeKondohNaganoKE. just compile it and then add this to controldict:
I haven't used it that much, I tested it on simple channel flows and flat plates.
So let me know if I missed anything.

Greetings Maff!

I've created a basic wiki page for to help getting your turbulence model a bit better known to the community that uses OpenFOAM technology: http://openfoamwiki.net/index.php/Co...-epsilon_model

May you or anyone else feel free to update that wiki page!

Best regards,
Bruno

Hi,

This thread is kinda old, but I am just wondering whether this AKN K-E model is an anisotropic (like V2F or EB) model or not. I used it in Start -CCM+ for the problem of past a cylinder and the results are so good.

Thanks,
A E

I would say that it is anisotropic only in the sense that near the wall the turbulent viscosity is actually a directional (wall-normal) turbulent viscosity. So if the phenomenon is mainly driven by a wall-normal behaviour (i.e. wall-normal heat transfer), it should be good enough.
The same happens in the v2f model without the need for wall functions.
The EB model is fully anisotropic, of course.

Thanks Maff for your reply. I am quite new to turbulence modeling and just wondering if a k - epsilon model uses wall functions, do these functions help to introduce anisotropy to the model?

Thanks,
A E

Most k-epsilon models are eddy viscosity models based on the Boussinesq hypothesis, which assumes that turbulent diffusion is isotropic.
Wall functions can help reproduce the effect of the anisotropy near the wall (e.g. in the way I wrote in my previous post), but your model will remain isotropic.

Oh, okay! Thanks for your explanations. So at least for flow past a cylinder, any anisotropy introduced to k - epsilon is helpful.

Hi Maff.

I just downloaded turbulence model. I would like to do some tests. I have compiled it without problems but my question is how I have to define the model into RASProperties file?. I'm trying to call turbulence model writing RASModel AbeKondohNaganoKE; but this fails.


Thanks

Have you added the libs line to controldict?

Hi all,
I've decided to bring this turbulence model into the modern age.
Attached is code compatible with v1806, I believe it will also work with v1712 and possibly earlier.
I compile my code by integrating fully in the src directory, you can sort out the options and make files if you prefer libraries.
Add the attached files to a directory called AKN in:

src/TurbulenceModels/turbulenceModels/RAS/


add the lines:

#include "AKN.H"
makeRASModel(AKN);



to:
src/TurbulenceModels/incompressible/turbulentTransportModels/incompressibleTurbulenceModel.C
src/TurbulenceModels/compressible/turbulentFluidThermoModels/turbulentFluidThermoModels.C


go to:
src/TurbulenceModels
open a terminal and run
I don't think it's actually necessary to rebuild the whole turbulence library but it only takes a minute...


Since the original code was written we have an updated epsilonWallFunction.
You can now use

Thank you for sharing this library.

This turbulence model is suitable for both compressible and incompressible case, is it right? I want to use low-Re turbulence model to simulate the heat transfer deterioration. Thank you.

It's implemented for compressible flow. I can't comment on it's accuracy in complicated flows as I've only used it in compressible code with no compressibility (for the heat code).