CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Turbulence model (http://www.cfd-online.com/Forums/cfx/18456-turbulence-model.html)

Jose Caridad February 6, 2001 10:55

Turbulence model
 
I am trying to simulate two-phase flow (air-water) through a centrifugal impeller but when i look at the residual plot in the canvas window the epsilon variable is in the order of 1E5 or higher. It is that correct? It is possible to estimate a "good value" for this variable? I really appreciate all the help you can give me.

Regards,

Jose Caridad

Jan Rusås February 7, 2001 05:46

Re: Turbulence model
 
Very difficult question,

The residuals are not non-dimensionalised, so if you want a lower value you can reduce your massflow ! You can also read in the manual in the section about convergence problems, some of the hints there might help you reducing the residuals, especially activating deferred correction. Be careful with how you set the inlet conditions for k-epsilon, check if there is a high jump in the values at the inlet and into your flow domain.

Regards Jan

Jose Caridad February 9, 2001 14:33

Re: Turbulence model
 
Thanks Jan,

I have tried all the hints recommended in the manual (relaxation factors, more accurate solution of the inner iterations, deferred corrections...).

All the variables show excelent convergence but the residuals for epsilon remain high (about 1E3 and 1E5). I have not tried to give explicit values of k and epsilon at the inlet. Instead, i especified values for turbulent intensity and dissipation lengh scale. Is it the mistake?

Regards,

Jose Caridad


Jan Rusås February 9, 2001 20:58

Re: Turbulence model
 
Hi Jose No probably not, It is a wild guess. I just remembered years ago someone told me that a value of k and epsilon which was decades higher than just downstream the inlet could lead to numerical troubles. Your values for epsilon is not uncommon. A quick, and mayby dirty, way to exaimine your values are by scaling them, i.e. eps_residual/eps_ave_inlet compare with the other scaled values.

Regards Jan


All times are GMT -4. The time now is 06:00.