Standard K-epsilon turbulence model with y~+1

anon_q · June 10, 2019, 14:12

Hello,
I have a 2D mesh with

of a computational domain around a cylinder with [inlet, outlet, top, bottom, cylinder].
I want to start two simulations: using the $k-\omega$ SST and the standard $k-\varepsilon$ turbulence models without changing the mesh.

As far as I know, for the $k-\omega$ SST, the mesh is ok because y+ = 1. But for the $k-\varepsilon$ model, it requires wall functions.

Can I use a mesh that has

with the standard $k-\varepsilon$ model? if Yes which wall functions should I use at the wall (cylinder) for

(turbulent kinetic energy), and $\varepsilon$ (turbulent dissipation)?

CFD_10 · June 15, 2019, 19:19

The standard k-epsilon turbulence model cannot be used with Low-Re formulation. However you can use a Low-Re implementation, LaunderSharmaKE (it uses damping functions)

sivakumar · June 26, 2019, 02:13

Hi Linda,
I dont know still you are looking for help regarding your Y+ and BCs.

It seems that you are clear about the wall function approach and resolving the boundary layers.

In your case, you are suppose to use Low Re k-Epsilon models.
for low Re k-Epsilon I have posted BCs long back (2013). I hope still you can use it (Please check it),

The boundary conditions are as follows (I used it for OF-2.1.1 Version),

Low Re model:

turbulence model : LaunderSharmaKE

nut ----> fixedValue (0)
k -----> fixedValue (1e-12)
epsilon -----> fixedValue (1e-8) it can be zeroGradient as well
nuTilda -----> fixedValue (0)

for high Re model:

nut ----> nutkWallFunction value (0)
k -------> kqRWallFunction value uniform (*******)
epsilon ----> epsilonWallFunction value uniform (*******)
nuTilda -----> zeroGradient

I hope it will be useful.

If there is any issue, please post it here,

Thanks,
Sivakumar

June 10, 2019, 14:12	Standard K-epsilon turbulence model with y~+1	#1
anon_q Senior Member Join Date: Mar 2018 Posts: 115 Rep Power: 8	Hello, I have a 2D mesh with of a computational domain around a cylinder with [inlet, outlet, top, bottom, cylinder]. I want to start two simulations: using the $k-\omega$ SST and the standard $k-\varepsilon$ turbulence models without changing the mesh. As far as I know, for the $k-\omega$ SST, the mesh is ok because y+ = 1. But for the $k-\varepsilon$ model, it requires wall functions. Can I use a mesh that has with the standard $k-\varepsilon$ model? if Yes which wall functions should I use at the wall (cylinder) for (turbulent kinetic energy), and $\varepsilon$ (turbulent dissipation)?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Antti Hellsten model: New Advanced k–ω Turbulence Model	purnp2	OpenFOAM Programming & Development	0	May 9, 2019 20:16
Turbulence model choice	dave13	CFX	3	December 22, 2015 08:12
Overflow Error in Multiphase Modelling with Two Continuous Fluids	ashtonJ	CFX	6	August 11, 2014 15:32
Wrong calculation of nut in the kOmegaSST turbulence model	FelixL	OpenFOAM Bugs	27	March 27, 2012 10:02
how modify standard turbulence model	Angmar	CFX	4	September 29, 2010 19:48

June 15, 2019, 19:19		#2
CFD_10 Member CFD USER Join Date: May 2019 Posts: 40 Rep Power: 7	The standard k-epsilon turbulence model cannot be used with Low-Re formulation. However you can use a Low-Re implementation, LaunderSharmaKE (it uses damping functions)

June 26, 2019, 02:13		#3
sivakumar Senior Member sivakumar selvaraju Join Date: Mar 2009 Location: India Posts: 205 Rep Power: 18	Hi Linda, I dont know still you are looking for help regarding your Y+ and BCs. It seems that you are clear about the wall function approach and resolving the boundary layers. In your case, you are suppose to use Low Re k-Epsilon models. for low Re k-Epsilon I have posted BCs long back (2013). I hope still you can use it (Please check it), The boundary conditions are as follows (I used it for OF-2.1.1 Version), Low Re model: turbulence model : LaunderSharmaKE nut ----> fixedValue (0) k -----> fixedValue (1e-12) epsilon -----> fixedValue (1e-8) it can be zeroGradient as well nuTilda -----> fixedValue (0) for high Re model: nut ----> nutkWallFunction value (0) k -------> kqRWallFunction value uniform (*****) epsilon ----> epsilonWallFunction value uniform (*****) nuTilda -----> zeroGradient I hope it will be useful. If there is any issue, please post it here, Thanks, Sivakumar