CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   multi fluid mixture model issue (

rystokes August 5, 2009 07:43

multi fluid mixture model issue
HI all,

I was trying to establish a model to simulate how two different kinds of liquid mixing with each other within an enclosed domain which is under rotating.

I use two liquids "water" and "ink" here to be mixed(I just duplicate the physical property of "water" to create "ink"). And in the simulation, gravity can be ignored and it's basically a laminar flow.

There are couples of options in the "fluid models" tab of the domain setting up in Pre-CFX, "Free surface model" has two options in its dropping down list: "none" and " standard".I assume the two fluids shouldn't have a distinct interface, since there should be a kind of mixing effect into each other, so "free surface model" shouldn't be utilized, and "none" should be toggled, is it correct?

I noticed " homogeneous model" refers to that the multi fluids will share the flow fields in the simulation, and it's often used together with free surface model. So my question is that once I didn't choose free surface model, does it mean I needn't toggle the "homogeneous model" option please?

Besides, in the "fluid pair models" tab of the domain setting, there is an option called" interphase transfer", and its dropping down list contains two: "none" and "miture model". My another question is should I use this "miture model" in my case? I didn't find any help from CFX tutorial regarding this.

Thank you very much for your kind help in advance!

ghorrocks August 5, 2009 21:53

If the two fluids just mix by diffusion and advect around with each other then it might be better to model the ink as a passive scalar. Much simpler and will run much faster. Have a look under additional variables in the documentation. This approach is not valid if the fluids have different properties.

Glenn Horrocks

rystokes August 9, 2009 09:03

Hi Glenn,

Thanks for your kind reply. Since the mixing is supposed to be generated at the molecular level between "ink" and "water", it's a kind of microscalar mixing, I noticed that multiphase model is at macroscalar level, so perhaps it's not suitable.

Do you mean I should use multicomponent model to simulate it? I don't know how to establish a passive scalar for ink, do you have any reference or tutorial that could help me to fix it? Or is there any documentation introducing how to establish a multicomponent mixing case? Thanks a lot.

By the way, when you said that "This approach is not valid if the fluids have different properties.", which approach do you refer to , the passive scalar in multicomponent model?

Have a great day.


ghorrocks August 9, 2009 20:13

The passive scalar approach just has a scalar variable which gets advected/diffused into the flow. This approach is very good if the stuff is truly passive, that is it has not effect on the bulk fluid properties. It is also good when the quantity of ink is small enough that its effect on the properties of the mixed fluid is insignificant.

If the ink and water have significantly different properties and there is enough ink around to change the properties then you have to be more careful. You might be able to use a additional variable with the concentration of the variable linked into material properties, not sure as I have not done this. The alternative is to go to multi component mixtures and then the bulk material properties comes from the local composition.

Your first line is correct - if the mixing is on the molecular level then a multiphase approach is not appropriate.

Have a look in the tutorials for additional variables. I think a few examples use them.


All times are GMT -4. The time now is 13:28.