Flotation Process
 

I´m simulating a flotation column in a domain with emulsion of liquid water and oil.
The simulation is semi-batch where air flows continuously through a batch layer of water-oil.
Oil concentration inside the column is 500ppm.
The distributor inlet at the bottom of the vessel has only air flowing through it (air volume fraction =1).

This are my setup options:

Steady state Simulation with 10000 iteration.

Basic Settings:

Liquid Water - Continuous Fluid ; oil - Dispersed Fluid with 20 micron diameter ;air- Dispersed Fluid with 1 mm diameter


FLUID MODELS:

Multiphase: Inhomogeneous

Heat Transfer : Isothermal


FLUID SPECIFIC MODELS:

Fluid Buoyancy Model: Density Difference

Turbulence: Liquid Water –shear stress transport; kerosene - Dispersed Phase Zero Equation ;air- Dispersed Phase Zero Equation

FLUID PAIR MODLES:

Interphase Transfer: Particle Model

Momentum Transfer: Drag Force - Grace

Turbulence Transfer: Sato Enhanced Eddy Viscosity
Mass transfer- None.
If I solve with this parameters all goes right.
My questions are:
1)If i interested in modeling the interactions between the oil droplets or gas bubbles (coalescence or breakup), what would i need to model them?
2)My supervisor said that there is mass transfer occur between the oil droplets and gas bubbles and this phenomena is the cause of separation of oil from water by the air and rising to the surface, is this correct?
3) Does the speech in the below paragraph relate to my research?
7.15.2. User Specified Mass Transfer
For advanced applications, it is possible to directly specify the interphase mass transfer sources. You may specify the interfacial mass flux or the interfacial mass flow (volumetric mass source) between any pair of phases. The latter is more appropriate for volumetric mass transfer processes such as breakup and collision. For details, see User Defined Interphase Mass Transfer in the CFX-Solver Theory Guide.

Release 16.2 - © SAS IP, Inc. All rights reserved.

Thank you for help.

1) The interactions between the two disperse phases is via the continuous phase. There is no direct interaction between the disperse phases available in CFX to my knowledge.
2) I have no idea if that is correct as I am not an expert in what you are modelling. But I would have thought that mass transfer from a liquid oil phase into a gaseous air phase would be negligibly small.
3) Yes, if you want to do a custom mass transfer source that seems to be the way to do it.

Thank you for your help
So,you said that "The interactions between the two disperse phases is via the continuous phase. There is no direct interaction between the disperse phases available in CFX to my knowledge"

1)If i want to represent or modeling the interaction between the two disperse phases via the continuous phase, how to do that?
2)As interaction between the disperse phases is not available in CFX, is this mean there is no need to modeling the interactions between the oil droplets and gas bubbles (coalescence and breakup), if it is necessary to modeling them, how to do that?
Can i modeling them by using UDFs, or this choice (UDFs) is not available in cfx?
Thank you for your help.

I am no expert on this, you should check my comments.

But you don't need to model anything special if you want the two disperse phases to interact by the continuous phase, that is what will happen by default.

But if you do want to two disperse phases to interact directly then you will probably have to write some user CEL of user fortran to model this. UDFs are Fluent's way of doing this. But CFX does allow quite a lot of flexibility in things like this. Read the documentation for details.

Previously, you mentioned you need to know if you need to add masstransfer. Based on you other questions I read, you don't want to model the bubble formation with nucelation, bubble growth, etc. (Why don't you tell that immediately. Now let us draw that conclusion ourself. That is inconvenient.)

In stead, you want to model the interaction between air and kerosene both dispersed in the water.
In CFX you can (easily) model the interaction between water as a single continuous phase and a dispersed phase like air bubbles or kerosene droplets. But the ineraction between 2 dispersed phases is not available by default. Maybe in Fluent, I don't know.
Alternatively you have to define 2 continuous phases, (water and kerosene) and add air as dispersed phase.
You have to realize that by default CFX only models physical interaction, as if the gas and oil are rigid spheres. It doesn't know you are trying to model a DAF process. CFX is a general purpose package.
So, if you want to model the interaction bewteen air and kerosene you have to add extra models that describes the attachment of air bubbles to a kerosene droplet, coalescense, growth, breakup, etc. There are some built-in models that cover the physical interactions (musig, beakup, coalescence for e.g. a bubble column with air dispersed in water) but I don't know if they are suited for the framework of 2 continuous and 1 dispered phase. If not, you have to tune the models yourself.
Alternatively you should go and use population balances. Search open literature for this, there are examples that cover this in DAF. And I think they were performed in Fluent.

And finally, coming back to mass transfer....... In CFD, we speak of mass transfer when mass is transferred from one phase to the other (it is called interphase mass transfer). In DAF processes I know, that doesn't play an important role. Dispersed air and kerosene remain dispersed air and kerosene. Kerosene doesn't disolve in water, water does not dissolve in kerosene, air the same etc........ So, no significant interphase mass transfer

I think you are mixing up physical transport of mass by coalescence from small aglomerate of bubbles and droplets to larger ones,. But that remains within a phase......

1)When i add kerosene as continuous phase this error appears to me
(In Analysis 'Flow Analysis 1' - Domain 'Default Domain': A Dispersed Phase Zero Equation model requires that only one continuous phase be present.)
2)You said:
(So, if you want to model the interaction between air and kerosene you have to add extra models that describes the attachment of air bubbles to a kerosene droplet, coalescense, growth, breakup, etc. There are some built-in models that cover the physical interactions (musig, beakup, coalescence for e.g. a bubble column with air dispersed in water).
How to model the interaction between bubbles and droplets?
How to add the extra models?

1) I think what you can do is model two continuous phases (water and kerosene) and add air as lagrangian particle track
2) 'The interaction' is too vague. You first need to find out what you want to model, what you want to include. Take a step back. If you know what you want to model then you should find out how to add the interaction. Believe me, there is not mouse-click model available. If I try to imagine what your want to do, I think you should go for population balances in Fluent. Try that forum.

I want to simulate the mechanism of the flotation process, i mean i want to represent the the adhesion of air bubble with oil drop by simulation.

If you want to model that on microscale, then you meed to do a DNS calculation and include surface tension.I would not use CFX for that.

I should also say that this conversation is falling into the "I want to model everything" trap. This is a sure way of not producing any useful results and wasting everybody's time.

The only way you are going to make progress in things like this is to do manageable steps, and validate carefully as you go along.

What is the basis of the separation of oil from water in this process (Flotation process or dissolved air flotation)
i mean how the program cfx perform the separation process?
in other words (depending on which principle)?

Have a look in the CFX documentation for the available models. They are all described in the theory and modelling documentation.

To add on that:
CFX doesn't know that you are modelling a DAF process. So if you don't do anything special, the dispersed material will just leave the domain through the outlet with the continuous phase, i.e. water.
Therefore, if you model the dispersed phase as lagrangian particle tracking, you can remove the dispersed phase by setting the restitution coefficient equal to 0 on walls or the free surfaces.
If you model the dispersed phase as eulerian dspersed phase, you can set a degassing boundary on the free surface.