LES tutorial

anishtain4 · September 8, 2015, 10:48

Hi all,
I'm trying to simulate a channel flow with CFX and LES model but can't get any fluctuations. I've used Smagorinsky and dynamic S but couldn't reproduce fluctuations in any of them. My runs were long enough but the velocity profiles turn out to be the average and the maximum fluctuates in every output.
Anyone has any experience here?

ghorrocks · September 8, 2015, 18:31

First of all - the basics - are you in a flow regime which generates turbulence? If the flow is laminar you won't get turbulence no matter what you do.

Assuming your flow should be turbulent then the fact you are not getting fluctuations suggests you have too much dissipation in your model. This is normal - you need to be very careful to get the dissipation low enough for LES to work.

Things to consider (in order of importance, most important first):
* Finer mesh
* advection scheme, use the bounded central differencing for LES. Or if you have to hybrid with a blend factor of 1.0, but you will need to run a finder mesh to get over the additional dissipation of this over central differencing.
* Time differencing - second order
* Something to trip the turbulence, either an edge or artificial oscillator
* double precision numerics

anishtain4 · September 23, 2015, 14:25

Thank you Ghorrocks. My Reynolds number was 8800 which for a pipe is fully turbulent. Also I was using the central difference (best possible in CFX) with fine enough grid and time steps. It seems still everything is too dissipative so I increased the Reynolds number to 40000 and now I'm getting the turbulence.

raunakjung · June 4, 2017, 03:22

Quote:

Originally Posted by anishtain4

Thank you Ghorrocks. My Reynolds number was 8800 which for a pipe is fully turbulent. Also I was using the central difference (best possible in CFX) with fine enough grid and time steps. It seems still everything is too dissipative so I increased the Reynolds number to 40000 and now I'm getting the turbulence.

Could you share your experience using LES and how you solved the problems ?

September 8, 2015, 10:48	LES tutorial	#1
anishtain4 Senior Member Mahdi Hosseinali Join Date: Apr 2009 Location: NB, Canada Posts: 273 Rep Power: 18	Hi all, I'm trying to simulate a channel flow with CFX and LES model but can't get any fluctuations. I've used Smagorinsky and dynamic S but couldn't reproduce fluctuations in any of them. My runs were long enough but the velocity profiles turn out to be the average and the maximum fluctuates in every output. Anyone has any experience here?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
LES Tutorial using fluent	anindya	FLUENT	7	October 17, 2016 09:17
LES tutorial pitzDaily pressure	hswzzz	OpenFOAM Running, Solving & CFD	0	December 22, 2014 09:34
LES basic and simple questions.	dshawul	Main CFD Forum	0	October 4, 2010 10:56
Turbulence dampening due to magnetic field in LES and RAS	eelcovv	OpenFOAM	0	June 8, 2010 11:35
LES tutorial file	tomosada	Siemens	1	November 23, 2003 15:17

September 8, 2015, 18:31		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	First of all - the basics - are you in a flow regime which generates turbulence? If the flow is laminar you won't get turbulence no matter what you do. Assuming your flow should be turbulent then the fact you are not getting fluctuations suggests you have too much dissipation in your model. This is normal - you need to be very careful to get the dissipation low enough for LES to work. Things to consider (in order of importance, most important first): * Finer mesh * advection scheme, use the bounded central differencing for LES. Or if you have to hybrid with a blend factor of 1.0, but you will need to run a finder mesh to get over the additional dissipation of this over central differencing. * Time differencing - second order * Something to trip the turbulence, either an edge or artificial oscillator * double precision numerics

September 23, 2015, 14:25		#3
anishtain4 Senior Member Mahdi Hosseinali Join Date: Apr 2009 Location: NB, Canada Posts: 273 Rep Power: 18	Thank you Ghorrocks. My Reynolds number was 8800 which for a pipe is fully turbulent. Also I was using the central difference (best possible in CFX) with fine enough grid and time steps. It seems still everything is too dissipative so I increased the Reynolds number to 40000 and now I'm getting the turbulence.