CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Mass Transfer Between AIR and WATER (

Math13570 October 23, 2013 16:26

Mass Transfer Between AIR and WATER
1 Attachment(s)

Here's my case : I have one square duct of 1 meter with two inlets. First injection of water (5m/s) and Air (0.5m/s) inlet on a hydrofoil NACA0015.

I'm able to see the volume fraction of Air in my domain.

Second step : Simulate the mass transfer between AIR and WATER, I tried differents things but i don't have results yet...

1. Do I need to set a multicomponent flow? Because the fluid are mixed at the molecular level, so i don't think i'll able to see the dissolve gas...

2. In the CFX-Pre, Do i just need to set the mass transfer correctly in the fluid pair model between AIR and WATER? Specified mass transfer - Interfacial Mass flux...

3. An additionnal variable name Dissolved gas is it a interesting way to my problem? But i'm not very familiar with this setup yet.

I Attach my simulation for support

Thanks for your reading


Attachment 26324

ghorrocks October 23, 2013 17:11

Multicomponent flows are mixed at the atomic level. Multiphase flows are mixed at the microscopic level - so bubbles, droplets but can also be an interface between a fluid and a gas.

This simulation sounds like it could be a free surface multiphase model, but you also talk about dissolved air and that would be multicomponent.

The key thing is to make sure you have made the correct choice of approach first. Are you sure you are using the correct multicomponent/multiphase model?

Math13570 October 24, 2013 10:06

1 Attachment(s)
Thanks for your reply!

For now, my model is very simple : The AIR is a dispersed fluid of a mean diameter of 1 mm and the WATER is a continuous fluid. In my fluid pair model ; i use the particule model, grace drag model, FAD and Sato for turbulence. I used the tutorials for references. I don't use the free surface model. So my model is inhomogeneous multiphase, where the AIR and WATER have their own velocity field.

I use the particle model to eventualy import the MUSIG model (Break-up and coalescence) , because the mass transfer is function of the bubble diameter. But i'm not at this state for the moment!

So, you talked about multicomponent for the dissolved gas. I tried to define my AIR as a fixed composition mixture of Nitrogen (0.79) and Oxygen (0.21), both are pure substance. The WATER is define as a pure substance too. I didn't try to simulate mass transfer yet. But i'm not able to see the oxygen and the nitrogen in the Post...probably because they are mixed at the molecular level? In the best world, I will simulate the dissolved nitrogen and oxygen in the water.

I'm not sure if a describe my case very well, do you think free surface model would be more appropriate?

Any help would be welcome :)


I have the same simulation but the AIR injection is at the trailing edge of the blade! See picture

ghorrocks October 24, 2013 16:49

That makes a bit more sense. So you are ultimately looking for a change in the N2/O2 composition of the air bubbles because of the different diffusivities in water?

From what you describe the approach you are taking sounds appropriate. Free surface models are for where a distinct free surface exists and you wish to resolve that interface. This is not what you want from your description.

Math13570 October 25, 2013 11:53

Exactly Glen, looking the change in N2/O2 composition of the air bubble is one of my goals. At the end, I would like to see the concentration of my gas phase : O2(g), N2(g), and the concentration of the dissolved gas in the liquid phase : O2(l) and N2(l) in the water.

A little brainwash made me change my parameters :

1. I define two multicomponent flow ; GAS is a variable composition mixture of nitrogen (N2g) and oxygen (O2g). LIQUID is a variable composition mixture of water, liquid nitrogen (N2l) and liquid oxygen (O2l).

2. The LIQUID is a continuous phase and the GAS is a dispersed phase.

3. The Fluid Pair (GAS|LIQUID) have the same settings mentions in the preview post.

4. Setting two variable composition mixture allow to have some Component Pair Details in the Fluid Pair Models like : N2g|O2l, N2g|Water, N2g|N2l, O2g|N2l, O2g|O2l, O2g|water. I'm thinking to use the Henry's Law between N2g|N2l and O2g|O2l in the Interfacial Equilibrium Model for the mass transfer, is it appropriate...??

5. I read about the option Component Models in the Fluid Specific Models: Transport equation, Algebraic equation etc. But i'm not sure how to set each the option of my component...Does it matter if a set my N2g to transport equation and the O2g to constraint or the inverse?! I have to do the same thing with my LIQUID phase.

All advice and tip is appreciate!!



ghorrocks October 26, 2013 06:56

I have not done a simulation like this so cannot say exactly how to set it up. What you describe sounds like it is heading in the correct direction.

I would start simple, and make sure a simple water/air multiphase model works OK. Once that is working then make the liquid a multicomponent fluid and try to get air dissolving in it. Only once all that works would I then split air into N2 and O2.

Taking he simualtion setup one step at a time like this means you are much more likely to get things working properly.

Math13570 October 28, 2013 08:20

Many thanks for your help and time!

I've see that others person were interested on the subject, so if anyone have the same simulation and you want to discuss about it, just send me a private message!


Math13570 October 31, 2013 10:37

For users interested in the subject, i contact CFX support and that's the advice they gave me for that kind of simulation!

The problem you are describing involves sources and sinks which can be numerically unstable. For an example of phase change in the presence of other gases, ie water evaporating into Steam and air, have a look at the steamjet tutorial in the CFX Help documentation. The partial pressure of the other gas (in this case air) must be accounted for, and it may be necessary to linearize the user source terms that are required to model the phase change. )User source terms are required because single component phase change rules no longer hold. )

santu November 11, 2013 12:22

i have similar type problem multiphase and multicomponent how can i proceed can anyone help me

ghorrocks November 11, 2013 17:16

Mathieu's post #8 is about as thorough description as you are going to get.

Ineedyourhelp May 11, 2016 21:21

Hey, guy, did you finished your project? I have similar type problem multiphase and multicomponent, Can I ask your some questions?

hilde May 18, 2016 08:16


Originally Posted by Ineedyourhelp (Post 599716)
Hey, guy, did you finished your project? I have similar type problem multiphase and multicomponent, Can I ask your some questions?

I am not a part of the project mentioned above, but am currently working with something very similar. What are your questions? Maybe I can answer then :)

All times are GMT -4. The time now is 22:23.