
[Sponsors] 
June 10, 2005, 03:34 
doubt about beta PDF

#1 
Guest
Posts: n/a

Hi Dear friends,
I got one probem. In non premixed LES combustion, f(mixture fraction SGS variance)) variance is assumed to be Beta distribution. I have DNS dat of the 3D reacting(non premixed) jet, I would like to compare actual PDF with the Beta PDF with my DNS data. in beta PDF calculation , I need two parameters one is f average and its variance. as i have only mixture fraction,how to get the variance? please help me Thankyou 

June 10, 2005, 04:45 
Re: doubt about beta PDF

#2 
Guest
Posts: n/a

In DNS you may try it with statistics.


June 10, 2005, 05:05 
Re: doubt about beta PDF

#3 
Guest
Posts: n/a

Hi DES, can u please explain with more details


June 10, 2005, 05:30 
Re: doubt about beta PDF

#4 
Guest
Posts: n/a

Can you please be more specific with question? What do you want to know?


June 10, 2005, 06:30 
Re: doubt about beta PDF

#5 
Guest
Posts: n/a

my doubt is , I want to calculate pressumed PDF for mixture fraction..its formula is involving of mixture fraction f and its variance f',, i got f from DNS, i would like to know how to get f' in the LES cell.. i,e in DNS we are resolving the small scales,we r going to up to kolmogorov scale,then where is the question of variance again..in prior evaluation of the models for the variance, how we will get exact(DNS )variance? i hope u understood my doubt


June 10, 2005, 09:46 
Re: doubt about beta PDF

#6 
Guest
Posts: n/a

I understand nothing. I am sorry.


June 10, 2005, 10:16 
Re: doubt about beta PDF

#7 
Guest
Posts: n/a

Go back to the basics lets call Z the mixture fraction. In DNS you have Z(x,t) in very point
You have to build a filter function that <Z> filtered value of Z, MEAN=<Z> 1) and the variance is VARIANCE=<Z^2><Z>^2 2) The definition of the filter function is all you need. If you assumed that the filter is a box filter basically averaging inside the LES cell of width delta (usual approach, not the most correct but the most usual without doubt) Then a function f is filtered as <f> = SUM(f_i)/N (3) where f_i are the DNS points inside a given LES cell. N is the number of those points. The gretaer the N the better the statistics. This is teh simplest way to do it. Use (3) to compute (1) and (2) The beta PDF will depend on your MEAN and VARIANCE, and you will have a PDF in every LES cell. 

June 10, 2005, 10:20 
Re: doubt about beta PDF

#8 
Guest
Posts: n/a

Sorry the equations did not display properly
[Z] indicates filtered value of Z 1) MEAN= [Z] 2) VARIANCE= [Z*Z][Z]*[Z] 3) [Z]= SUM(Z_i)/N In doubt check Poinsot book, or paper by Jimenez et al (2001) in Physics of fluids, 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Problems with non adiabatic PDF!  Christoph_84  FLUENT  0  November 17, 2009 10:55 
transported PDF for combustion (total energy PDF)  mathieu ourliac  Main CFD Forum  0  April 10, 2008 05:38 
PDF Model in compressible flows with shock waves  Hanxsi  FLUENT  0  February 26, 2006 21:51 
PDF  conserved scalar approach in KIVA  vega  Main CFD Forum  0  July 5, 2005 02:49 
Determining alpha and beta for porous baffle  Liaqat Khan  CDadapco  1  October 27, 2000 04:44 