Hi Krzysztof,

Your model is wrong, if I understood correctly. If you have miscibility, you do not have phases, you have A single fluid with variable composition. As water+alcohol,. Hence you can have two or more streams with different proportions, e.g. inlet 1: misture of 5%water+95%alcohol and inlet 2: mixture of 50%water+50%alcohol.  And the streams will interact and the species will be transferred. So, you have species! You do not have phases, because you need an interface, as water + oil. And each phase can be a variable composition fluid, but you do need to set the MASS TRANSFER model among the phases though interface. It is far more complex.

Now you have fluids that nor is able to mixture, you can see, small parts of fluid, but each part is a single fluid and does not have mixture in the chemical sense. 

I strongly advice you to review de modeling of each aporoach to understand what each models do and what is each physics

Bests!

Hello Ricardo,

thank you for reply. I know that mixing is really kind of simultaneous mass and momentum transport (possibly heat too). I wanted to omit  any kind of diffusion mechanisms, as main goal is to obtain a data on predicted velocity and pressure fields. And the reason I included multiphasicity is that Venturi tube is inserted into a tank, initially filled with one liquid and air in the upper section, so I do have multiple phases. It's the idea provided by supervisor of my thesis, to asses a performance of Venturi tube as a mixer with respect to time and influence of various parameters on whole process, including e.g. depth of Venturi tube.

I'll describe the general idea. We have this forementioned tank, filled with let's say benzene, enough to completely cover Venturi under its surface. Then the stream of acetone is pumped through the inlet pipe. It collides with benzene already present inside, thus marking the first contact. When flowing through the section with lower diameter, static pressure decreases, thus additional portion of benzene should be sucked through the orifices, again coliding and intermixing with feed stream. Together they leave Venturi tube through it's outlet into the volume of the tank. And so on, for a certain amount of time (or quantity of acetone pumped inside). 

I tried this also for only one fluid and my main problem is that I don't observe Venturi effect (suction into mixing chamber). Pressure is lower, velocity is higher, so it should occur, but it doesn't. I modeled Venturi tube after real life injectors, so I don't know why this happens. Or rather doesn't happen.

Still, thank you for showing an interest in my question. I'll review the posibilities you presented.

Regards,

Christopher

Use this forum Good luck.

https://www.cfd-online.com/

Hum, ok, more clear now.. so you have a other phase, but it is the air. Of course you can think  each  material such as one phase and so modeling the mass transfer. But if the materials are miscible the two-phase model will fail. hence we can have a Phase 1:  mixture benzene + acetone - a  multicomponent fluid. And Phase 2 Air. We can  neglected the mass transfer from and into the air for simplicity. A good simplification. 

To omit any diffusion we have a bad situation:

1) the effective diffusion coefficient is  the molecular (from correlacion or material data) +  turbulent dispersion. Often, the turbulent dispersion is higher than molecular diffusion (10, 100 1000x sometimes). hence, you will not have any kind of diffusion only if you have no turbulence model and give set a zero molecular diffusion coefficient. 

2)There are numerical problems and have no diffusion, that is,  with a Pe->oo , the Peclèt number. Hence,  be prepared to numerical instabilities/oscillation, such as negative values for mass fractions of components.

Regarding to the suction, it must be a boundary condition mistake. Is benzene is flowing out of tank, something must enter ( air, I guess) , You need some condition that be able to let the air inflow to domain, a 'opening' condition, and the pressure drop must be higher than weight of liquid column, with Bernoulli relation we can estimate a ideal condition, the values of velocity in the throat must be next to that magnitude. 

2D to 3D can give wrong  results due to the face area of element. A ventury is a quadric region, with cylindrical coordinates, and the normal face to the flow is a sector as a pizza peace. If you draw a 2D ventury thenormal face is a quadrilateral element. The pressure drop, flow rates and velocities will not match. It is a common mistake. 

I hope help you

That's a really wonderful insight you gave me here. I am very grateful.

As for Venturi effect not occuring, you were also right. The throat was still too wide in comparison with initial diameter of the pipe, and the pressure drop was either so low that flowing acetones pressure was greater than hydrostatic pressure of benzene so it was flowing out of the orifices, or the sucking effect was very much negilibile because pressure gradient between mixing chamber and volume of the tank was set in right direction, but still a minuscule.

I will try to do as you said, to introduce air as a 1st phase, and benzene-acetone mixture with varying compsition as a 2nd phase. While retaining as much of intended simplicity as it is possible, let's sum it up. Sorry if my questions sound stupid, but I am really a beginner in CFD. Very much a beginner, a green guy, a newbie.

Multiphase - Mixture or Eulerian? With possible diffusive transport coefficients between liquid phase and gaseous phase set to 0, while retaining mechanical interaction coefficients like interfacial tension, slip or drag.

Energy - Disabled. Still assumed to be isothermal.

Species - ENABLED? As species transport, but again with any coefficients of diffusion as equal to 0? No reactions, no combustion. This option gives me an additional sub-tab for Mixtures in Materials, where I can define benzene-acetone as a mixture.

Phases - Either Air (gas) as primary and Benzene-acetone (liquid) as secondary or otherwise, but with subsequent consideration in the set-up.

Materials - Air, benzene and acetone as fluids, solids as default, introduced acetone-benzene mixture in third tab.

And because of diffusion mechanisms ommited, I must make sure that boundary conditions for initial inflow and further, predicted flow characterstic in a system will have turbulent nature.

Did I get all of it correctly or there is still something to consider or change in my train of thought?

Well, first of all I would try to reproduce the venturi effect. Run a two fluid flow with air and the liquid with the misture properties.  It is good to setup boundary conditions, and test some options available on the solver and see wich give the beste results. 

The simples model is VOF or the Mixture model. You will have a ONE velocity field shared by the phases - in the mixture model you will have the slip velocity to inferred the secondary velocity phase. 

ON two-fluid model the physics is more complex  and shave more details. Is computational expensive and use if the VOF or mixture models do fail. 

Energy OK! Is not a isseu . This is no coupling with Temperature field.

If you need two materials, you need turn-on the species, if not, there is no EDP calculating the materials transport.

To consider or not the superficial tension will impact in the pressure drop if you have some kind of capilar tube. And it is a source term - more numerical difficulties. 

Hence, VO or Mixture with transport of species is simplest model to you. But try without the species first. 

AVOID set up all physical features. Analyse each of them isolate on of another. And try combine step-by-step new features to the simplest first approach.  It is a good numerical practice.