Learning Forum

[hagos.gebru]

How can I define the phases of a mixture of water, ethanol and a stillage at the bottom of the distillater ? should i use evaporation condensation or species transport for the mass transfer case? i appreciate your responses.
regards,

[Rob]

I'd probably use Eulerian multiphase with species and evaporation/condensation. I'm not sure what a stillage is though, it's not a term I remember from my Uni days designing columns. 

[hagos.gebru]

Thank you for the response Rob, stillage is a residual of alcohol distillation. And I do have the following questions;

1. Does the species use only for creating a mixture or does have other uses?
2. Should I use the species mass transfer or evaporation condensation for the mass transfer mechanism?
3. Since it is batch distillation the system doesn't have an inlet so how could I specify the volume fraction of the components of the mixture? I try to patch the volume fraction of the mixtures, but Ansys results doesn't show the correct analysis.
4. Lastly, I think the computational time for multiphase is very high so how could I minimize it? Regards,

[Rob]

Ah, we called it "bottoms" product as it was then sent off to the cracker: oil distillation! 

You probably want two phases. Liquid phase, which will be a mixture of ethanol and water (I assume stillage is mostly water). Similarly a gas phase with ethanol-vapour and water-vapour. Phase change is handled in the Phase panel. 

For the initial condition you will want Standard initialisation and then patch the phase/species for which ever phase wasn't in the initial condition. Use a cell zone or register for the patch. You may need to think a bit about species & phase as the species fraction is on the phase, it's a little confusing so I'd patch and post process to understand what's going on. 

Multiphase models are often quite long runs. Think about 2d or 3d sectors depending on the geometry. Also look at parallel processing. There are cheats to get around some of the time issues, but you then need to understand (and defend) those assumptions. 

[hagos.gebru]

Thank you very much Rob.

[Rob]

Oops, 2d, 2d axi-symmetric and 3d sector. 

[hagos.gebru]

I am not clear with your idea Rob, my geometry isn't axi-symetric and it is 2d. What do you mean 3d sector and Oops?

[Rob]

The oops is because I always forget to explicitly mention 2d axi-symmetric on here. It's second nature to pick the most suitable option as I've been using the code for a while. 

A 3d sector would be just that in a cylindrical domain. One of the ways of speeding up a simulation is to reduce the number of cells used; we still need to resolve the flow so instead we reduce the amount of domain that we model. 

[hagos.gebru]

And when i see the heat flux in 2D is changed, since it only considers surface area as a line so should i change the dimension so that to have the correct heat flux? your help is appriciated.  

[Rob]

I would leave the thickness reference alone. As it's one metre (default) it makes the maths easier: it's PI or 2PI (it's early!) for 2d axi-symmetric. Additionally some options use the reference thickness and others may be hard coded to 1m: it's all in the manual but I never want to check the umpteen locations to ensure everything works as expected. 

[hagos.gebru]

Thank you for the replay Rob, but you know, let say the volume of mixture liquid in 3D was 8L but when we come to 2D the volume goes to 32L as the volume is equals to surface area in 2D. Simillarly for heat flux in 2D and 3D is different so shall i change the dimension such that the volume and heat flux are tobe equal in 2D and 3D? I hope you understand my point. 

[Rob]

2d and 3d are different, I wouldn't alter Fluent's reference to suit, I'd rethink how you use the data.