Lagrangian Particle Tracking in Eulerian-Eulerian Multiphase Flow
 

Hi All,

I have a question regarding Lagrangian Particle Tracking in Eulerian-Eulerian Multiphase Flow that I hope you guys can help me with. It is loosely related to an old thread, namely this one :

http://www.cfd-online.com/Forums/cfx...-tracking.html

A quick rundown of my simulation : I am looking at a 3 phase oil-water-gas separator and i am trying to inject some sand at the inlet to see how it disperses in the gas phase only.

I know there are 2 methods in cfx - Lagrangian Particle Tracking and Euler-Euler. I want to do both cases to see the difference, but I just want to clarify 1 specific question regarding Particle Tracking in Euler-Euler multiphase simulations.

In the thread ghorrocks mentioned that we can stick a Lagrangian Particle Tracking model on top of one continuous phase of the Euler-Euler multiphase simulation. But when I do this (DomainName > BasicSetting > Sand > Morphology > Particle Transport Solid), a warning sprouts up :

"In Analysis 'Flow Analysis 1' - Domain 'ExtensionInlet': Particle tracking simulations with more than one continuous Eulerian phase are not supported."

Also under fluid pair the sand material is paired with every continuous phase, namely water, oil and gas.

According to the excerpt from the manual that ghorrocks pointed out in the aforementioned thread, this should be possible (Given that my understanding that Particle Transport Solid option means Lagrangian Particle Tracking is correct).

So my question is how do you define the particle tracking in one of the continuous phase only? I hope you guys can help me shed some light on this.

For sand particles using purely Euler-Euler approach, I think I have the settings right based on what I read from the tutorials. Basically what I do is just define the sand as a "Dispersed Solid" instead. Then I can basically define the interphase transfer models in 'Fluid Pair' to reflect what I want to see in the gas phase only. Is this correct?

Sorry for the long message. I hope my question is clear and you guys can shed some light on this for me.

Thank you very much!

Cheers
DarrenC

Particles: As the error message says you must have only one continuous phase. You have four phases (water, air, oil, sand), so only one of water, oil and air can be continuous. When you do this then the particle tracking model for sand should be happy.

Eularian: This sounds OK, but.....

Why does the sand only hang around in the air? Why does it avoid the oil and water? I would have thought sand would hang about in the water and oil phases and avoid the air. But I have no idea what you are modelling and I am just guessing.

Hi Glenn,

Now that you have put it this way, the answer to my question is terribly obvious :)

So i guess if I were to continue to have a VOF style simulation with 3 continuous phases, then defining Dispersed Solids in the Eulerian-Eulerian simulation will be the only way I can have sand particles in my multiphase simulation?

Forwarding the discussion a little for Eulerian-Eulerian, if I were to define an interphase model between the sand and gas only and leave the other sand/oil and sand/water interphase models as none, what would actually happen to the sand at the interface of the gas and liquids (oil/water)? Will the sand just deposit at the interface?

And if I were to do it in this fashion, will there be any savings in computing time?

Just to elaborate on your question, I am just interested if the sand from the inlet settles on the surface of the liquid or if it gets carried over to the gas outlet pipe, which is something that is undesirable. I am dealing with very high gas velocities so there is a possibility that the gas might carry over the sand to the gas outlet. I'm not terribly concerned about the sand content in the liquids. In fact I would very much like to have more sand in the liquid as the liquids are routed to a desander!

Thanks again for your inputs

Cheers
DarrenC

Yes you appear to be correct - so if you want to run a 3 way VOF style simulation which requires the three phases to be continuous then you cannot use particle tracking. So your only option is to use a eularian model.

Only defining interphase transfer models for a few combinations does not sound wise. Your flow sounds like a complex interaction between all the phases so I would include all combinations and not make assuptions about what is important - unless you REALLY know that an interaction is insignificant.

Thanks Glenn. Ill have a go at this.

Cheers
DarrenC

The three phase model can possibly be handeled by setting one of the 'continous' fluid phases to dispersed fluid even if it is in a connected region - set the 'particle' size to be small (1 micron). Then solid particle tracking will be happy.

Matt Campbell