CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Help on granular model (http://www.cfd-online.com/Forums/fluent/32273-help-granular-model.html)

Lohen October 7, 2003 03:16

Help on granular model
 
Hi, does someone evers use the granular model ?If so. I wanted to know if the viscoity in Properties for the solid corresponds to the granular solid viscosity !!

Thanx

ap October 7, 2003 14:16

Re: Help on granular model
 
If you use the eulerian granular model based on the kinetic theory, the granular solid is considered as a fluid phase.

Its properties are specified in the Define->Phases panel, by selecting the granular solid phase and clicking on Set. Activating the Granular option, a group of parameters appears: Diameter, Granular viscosity, Granular Bulk viscosity, Frictional viscosity, Packing limit.

Hi :)

ap

Lohen October 8, 2003 11:33

Re: Help on granular model
 
ok !! in the menu Display>Contours>Properties>molecular viscosity ..., for the solid in this menu, i wanted to if it is the computed solid viscosity from the different laws previously set in the granular model .

thanx again

ap October 8, 2003 12:17

Re: Help on granular model
 
Sorry, I misunderstood your question.

The solid molecular viscosity you find in the Properties drop down menu should be the sum of all components of granular viscosity (collisional, kinetic, frictional).

In the solid phase definition panel, you specify

- Granular viscosity, which actually is the kinetic component of viscosity.

- Frictional viscosity

The collisional viscosity is always calculated according to Gidaspow et al. (1992) and Syamlal et al. (1993).

Hi :)

ap

Lohen October 10, 2003 03:26

Re: Help on granular model
 
Yeah, ok thanks, that was what i was thinking !! It seems that you have ever used the granular model, have you already tested in a gas-liquid-solid flow and particualarly the value of the coefficient of restitution ess in such a flow !!

Thanks again !

ap October 10, 2003 12:19

Re: Help on granular model
 
Yes, I'm working on the granular model for gas-solid flows because my goal is to develop a better model for risers.

I've never tested FLUENT granular model in a gas-liquid-solid system, anyway, as a first attempt, if e_ss is the particle-particle restitution coefficient, it should be the same you'd use in a gas-solid calculation, being related to solid material properties.

Hi :)

ap

Lohen October 11, 2003 13:22

Re: Help on granular model
 
ok, it seems to me that the e_ss coefficient was used to tune the value of the dissipation of granular temperature (gamma_ls). I don't know the consequence of varying for a gas-liquid-solid flow where i think there are not so much collision between the particle, and the granular model did not take into account the properties of the continuous phase . What do you think ??

ThanX !

ap October 12, 2003 18:44

Re: Help on granular model
 
Yes, you're right. The restitution coefficient takes into account the dissipation of the granular energy due to anelastic collisions between particles. It's also involved in the calculation of granular viscosity.

I think e_ss should be considered more as a physical property of the solid material rather than a model parameter. However, many models are too sensitive to the value of e_ss because the dissipation terms are bigger than the generation ones.

An important thing was noticed (Agrawal et al., Zheng et al.): the high sensitivity to e_ss is due to a not proper use of the kinetic theory, mainly related to the use of too coarse grids.

In order to obtain good results with the kinetic theory alone, cells of several particle diameter in size should be used, which is not possible with nowadays calculation resources.

So, to obtain good results on coarse grids, it's necessary to describe the particle phase turbulence using "sub-grids" models. In these models the micro-scale turbulence is described using the kinetic theory, while the meso-scale turbulence is described using a turbulence model based on mixing lenght or on the k-eps model (Hrenya-Sinclair; Zhang et al).

So I think e_ss is used as a tuning parameter only because models are not capable to properly describe the real behaviour of granular flows.



All times are GMT -4. The time now is 00:55.