CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   problem with SST k-omega model (http://www.cfd-online.com/Forums/fluent/96523-problem-sst-k-omega-model.html)

avi031 January 25, 2012 05:10

problem with SST k-omega model
 
Hi,
I have used SST k-omega model, Realizable and RNG k-epsilon models for a problem involving heat transfer in a duct (internal flow) with a lot separation, reattachment and re-circulations at near-wall regions. Significant differences in wall temperatures are seen between the turbulence models.
Unfortunately very high wall temperatures are obtained with the usage of SST k-omega model compared to the k-epsilon models, which has resulted in bad-comparison with the literature. This finding is contrary to other numerical simulation papers. The re-circulation size in case of SST k-omega model is also very different from the k-epsilon models.
y+ less than 1 was maintained for all the cases.

The reason for high temperatures with the usage of SST k-omega model evades me. Please let me know of your suggestions, as to what can be wrong.

Regards
Avinash

swiftaircraft January 26, 2012 15:38

What other options within the sst model have you turned on?

avi031 January 27, 2012 00:43

SST model with Low-Re corrections was used for the simulation

swiftaircraft January 27, 2012 04:48

I would turn off this option as by default the k-omega model can resolve through the boundary layer without any corrections being required. My experience with this option in the past has not been that good.

shaoda169 January 27, 2012 06:53

Quote:

Originally Posted by swiftaircraft (Post 341455)
I would turn off this option as by default the k-omega model can resolve through the boundary layer without any corrections being required. My experience with this option in the past has not been that good.

The low-reynolds correction option is borrowed from the original k-w model rather than being in the sst k-w equations. It is said that this option should be used with cautions.

heidar January 29, 2012 18:24

Very nice
 
Hello

i Have used the all turbulent model for a ribbed roughened square duct. my reynolds number was from 10000 to 100000 in low reynolds number SST model was very bad and for higher than 25000 RNG, RM, SST, and etc exhebited good results
could you tell me more about your dimensions and reynolds number.
As you know the SST model uses K-w in near the wall and uses K-e in far feild if you give more details we could merge in it;)

Best Regards
Heidar Jafarizadeh


All times are GMT -4. The time now is 21:17.