CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   modeling turbulence (https://www.cfd-online.com/Forums/fluent/49580-modeling-turbulence.html)

 Gabriella October 21, 2008 09:06

modeling turbulence

I'm modeling the flow of a liquid inside a reactor. I choose for the inlet the boundary condition: Velocity inlet. I determined the Reynolds number and I found a value equal to 2195. I selected the RNG k-epsilon model. Is it correct? And, what is the better model I have to select for describing this type of flow condition?

thank you very much

 John S. October 21, 2008 14:45

Re: modeling turbulence

Unfortunately this isn't a simple yes/no question and there's no absolute "correct" way to model turbulence. Each model has its own advantages and disadvantages and you should pick a model based on the conditions you're running and what you're trying to resolve. For example, k-w based models tend to perform better in the near-wall resolution but lose accuracy as you move into the freestream. k-e models tend to be accurate in the freestream, while employing wall functions to resolve the viscous layer. Reynold's stress modals are formally more accurate but are computationally more expensive and tend to be unstable.

If you're just concerned with the behavior of the fluid in the reactor and not the near-wall solution, I suspect any of the k-e models will be fine.

John

 devesh October 22, 2008 10:05

Re: modeling turbulence

Hi Gabriella,

This is Laminar or Turbulent Flow ?,,,,,because as u said Reynolds number is 2195.....If I am wrong, Please correct me.

 Gabriella October 22, 2008 11:30

Re: modeling turbulence

Hi Devesh, I know that the value of Reynolds number which defines the passage from a laminar to turbulent flow is 2100. So, I think that Reynolds equal to 2195 means that the flow in the transition regime (not laminar and not fully turbulent). Moreover, I determined this value in relation to the inlet surface but inside the reactor (the object of my study) there are regions where both velocity and reynolds number increase and there are a turbulent flows. For this reasons I don't know the correct model! I hope I describe you my problem

 sam October 23, 2008 01:55

Re: modeling turbulence

As discussed, the flow is intransition regime.

If u cannot afford high cost computation try the k omega sst with transition.

otherwise, also try the LES wale or kinetic energy model! gd luck

LES wale, kinetic energy and SMAGORINSKY-LILLY mod

Dear ALL,

Anyone know what is the different between these: wale, kinetic energy and SMAGORINSKY-LILLY model??

Help help here... thankz

 Paolo Lampitella October 24, 2008 18:34

Re: LES wale, kinetic energy and SMAGORINSKY-LILLY

In the eddy viscosity approach to the closure of the LES equations (very similar to the rans one) these are just three models of the eddy-viscosity.

Will not be of much help if i put here the equations, but just to give you a taste:

the smagorinsky model was the first one. It is based on the equilibrium assumption of the inertial range of the turbulent spectrum.

Kinetic energy model uses an equation for the subfilter kinetic energy and so it can handle more complex situations of nonequilibrium turbulence.

A major drawback of these models is the necessity of defining some costants whose value is not so universal and that, being costant, do not provide the necessary near wall behaviour for the eddy viscosity.

Instead of using some damping functions, the wale model of nicoud and ducros adress the issue defining the eddy viscosity in a more complex (and less intuitive) way which gives it the proper behaviour.

Actually the more common way to handle it is to use a dynamic model. It can be shown to give the proper eddy viscosity behaviour near walls + a lot of other interesting things. Both the kinetic energy and smagorinsky models have a dynamic counterpart.

Hope this helps.

 All times are GMT -4. The time now is 11:03.