CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   turbulent model k-epsilon (http://www.cfd-online.com/Forums/fluent/34045-turbulent-model-k-epsilon.html)

 yann June 25, 2004 05:49

turbulent model k-epsilon

hi,

i am trying to look at the flow of air around a rectangle (length 7m and width 2.5m) (in turbulent regime). the inlet velocity is 16.66m/s. i solved in 2nd order implicit segregated. i have tried 3 different models of turbulent regime in order to compare them: k-epsilon standard, K-epsilon realisable and K-epsilon RNG. i did focus my study on the recirculation zone generated after the rectangle and the drag coefficient.

the results i have are quite disturbing: length of recirculation zone: k-espilon standard: 4.44m k-epsilon realisable: 4.36m k-epsilon RNG: 7.59m drag coefficient k-espilon standard: 2.3811 k-epsilon realisable: 2.2554m k-epsilon RNG: 2.5053m

my problem is that i dont understand why the length of the recirculation zone is that big for the case of the k-epsilon RNG. it is almost the double of the length for the realisable or the standard.

if someone has an explanation that would be nice!

thanks

 gopal June 27, 2004 14:20

Re: turbulent model k-epsilon

Hi, It depends on which model you are using. You have to select apporpriate model for your geometry from the literature available. Each model will treats the N-S equations in different way. So, you should be careful before selecting the Turbulence model.

 Iason Hatztheodorou June 29, 2004 13:40

Re: turbulent model k-epsilon

Hi there,

I am doing something pretty similar right now. You can look up a paper by Yao, Savill and Sandham (in Flow, Turbulence and Combustion 68, 2002); they compare DNS and RANS for the truncated trailing edge flow. RANS always gave approximatelly half the re-circulation region. This might suggest that the RNG variant is actually closer to the truth. What is more peculiar is that the RNG and the realizable model do not agree well, but instead the realizable agrees with the standard. The solution will also depend on the wall functions used. Are you using the same wall functions in all 3 runs? Are the mesh densities the same? Also since you are using essentially variants of the same model, try using the S-A model or the k-omega model and compare with that as well. Good luck!

 ramezani July 7, 2004 07:37

Re: turbulent model k-epsilon

 ramezani July 7, 2004 07:39

Re: turbulent model k-epsilon

 Surena July 25, 2004 15:51

k-epsilon method

Hi Im Surena Norouzi. Im from iran. Im a fluid mechanic engineering. I try to solve internal flow in a gas turbine blade. I must to use the K-Epsilon Method to simulate flow. I use finite difference method to solve Navier Stocks equations. But i have not discreat form of K-Epsilon equation in staggered mesh. please help me. regards

 Ravi August 7, 2004 08:18

Re: k-epsilon method

hello,

My project is also regarding internal flows of a turbine blade. I am trying to find the internal heat transfer coefficints and trying to cool the blade by means of impingement and film cooling. I am dealing with totally conjugat heat transfer.

Can you please tell me on what model should i select and what boundary conditions should i apply. You please tell me as to what journals should i refer to or send some if you have already any.

thanks Ravi

 All times are GMT -4. The time now is 10:04.