|August 21, 2012, 13:22||
How to measure eddy viscosity in a wind tunnel?
Partly because it is related to turbulence modeling... But I still do not quite understand this.
Since eddy viscosity is not a physical property of fluid, and I want to know how "physical" this quantity implies?
When people say in external flow simulations, one of the most common practices is to set the inflow eddy viscosity ratio, beta=nu_T/nu. Many say beta should be around 0.1~0.2....
1. Does beta<1 mean the flow is laminar, or what? Then what 0.1~0.2 actually means?
2. Does it what actually means much dependent on the turb model we are using, or does this eddy viscosity ratio more close to have a physical meaning and hence independent of turbulence model?
3. Suppose I have a grid generated turbulence in a wind tunnel, which has a higher turbulence inflow, then without a priori turbulence model to use, or even perhaps out of the framework of eddy viscosity hypothesis, is it possible or how to calculate the turbulent viscosity then?
Many thanks in advance!
|August 22, 2012, 03:01||
Join Date: Nov 2009
Posts: 198Rep Power: 9
turbulent viscosity is related to turbulence modeling and based on some assumptions but it still has a physical meaning.
In Laminar flows diffusion in flow field is characterized by molecular viscosity.
But, turbulent flows in its nature consists random velocity fluctuations which intensifies diffusion in the flow. So, one method to account for this intensified diffusion is adding additional viscosity to the molecular one. this added viscosity is simply known as turbulent viscosity.
its definition ,however, depends on the model that you have selected
in k-epsilon turbulence models it is function of k and epsilon.
in spalart-Allmaras it is directly calculated from a transport equation.
higher order turbulence models does not use this concept explicitly and calculate this added diffusion in other ways.
Now, if you have turbulence intensity and turbulent integral length scale at the inlet, you can calculate k and epsilon and consequently the turbulent viscosity. But keep in mind that what you calculate is not a pure physical quantity like velocity or temperature. It is deeply dependent to the model you have used. in other words such a turbulent viscosity calculation can only assess your turbulence model and its underlying assumptions.
I hope it helps!
|August 22, 2012, 11:08||
And I am very interested and grateful, if anyone can accurately or roughly give me some numbers of (1) k, (2) epsilon, and (3) eddy viscoity, that are typical in
* Wind tunnel
* Real ABL
That would be great. Thanks
|August 26, 2012, 04:15||
Join Date: Jan 2011
Blog Entries: 5Rep Power: 9
clean wind tunnel 10-12
average wind tunnel 9
dirty wind tunnel 4-8
The turbulence length depends on typical size of the object in the flow. All the quantities of interest can be then derived using intensity and length. The relationships between these quantities can be found e.g; in the Fluent manal: "7.2.2 Determining the turbulence parameters"
|August 27, 2012, 14:55||
Hmm, what did you mean by "e^n transition theory"? Could you share a link of it?
QUOTE: "The turbulence length depends on typical size of the object in the flow" ......
1. I think turbulence intensity and turbulence length scale are not linear correlated;
2. I think turbulence intensity is easy to measure, but length scale is more complex to measure;
3. My point is inflow turbulence length would change from case to case, strictly speaking, it has nothing to do with the object. Wouldn't you agree?
|Thread||Thread Starter||Forum||Replies||Last Post|
|Wind Tunnel Turbulent Viscosity Ratio||robinfisichella||Main CFD Forum||0||March 14, 2010 09:40|
|setting value of turbulent intensity and turbulent viscosity ratio in wind tunnel||nuimlabib||Main CFD Forum||0||August 4, 2009 00:05|
|how to consider perforated wind tunnel wall in CFD simulation||littlelz||CFX||4||June 11, 2009 05:51|
|Virtual Wind Tunnel in FLUENT||ND||FLUENT||0||April 7, 2006 07:43|
|Wind Tunnel Website now online||Mike Worthey||Main CFD Forum||0||June 6, 2000 02:27|