Learning Forum

[Patrickodon]

Hi all,

 

I'm mixing two fluids in Fluent, and I'm wondering

 

1) Does Ansys take crosslinking (https://www.sciencedirect.com/topics/engineering/crosslinker#:~:text=Cross%2Dlinker%20increases%20the%20viscosity,by%20linking%20multiple%20molecules%20together.) into account when you're mixing them?

2) If not, can crosslinking be programmed into Ansys Fluent when mixing?

[Rob]

Only if you tell the solve what's going to happen. So, in this case, you may want to have a careful look at the available viscosity models, and then how you'd work out what the mixture composition effects are going to be. If there's a history effect you may need to look into Polyflow. 

[Patrickodon]

OK - so I need to look into the settings of Models -> Viscous and find which ones give crosslinking results?

[Rob]

The term you're looking for is non-Newtonian. First hint, turn off turbulence. 

[Patrickodon]

The only turbulence setting I coud find in Fluent was Models -> Multiphase (Eulerian) -> Phase Interaction -> Forces -> Turbulence Interaction and Turbulent Dispersion (already set to none). I set both of these to none, but the test yielded the same results. I'm assuming there's either a Turbulence setting I can't find, or I'm going about this incorrectly?

[Rob]

Have a look in the viscous panel - it's where the turbulence models are set. Why have you turned on multiphase modelling?

[Patrickodon]

Sorry for delayed response - when I was initially setting up the model, one of the tutorials I used had multiphase modelling set for its mixing.

 

I'm comparing my mixing to a Widman calculator (https://www.widman.biz/English/Calculators/Mixtures.html) and getting fairly accurate results (error under 2 % when the difference between the two viscosities is a factor of 10 or less)

[Rob]

OK. 

There are two methods for mixing. Species where the materials are in the same phase and can diffuse (as well as convect) and multiphase where the materials may mix in a cell but do not mix at a molecular level. The main difference is in species the cell density is taken from the mixture so mixed fluids won't then stratify. With multiphase each phase (material) retains it's own material properties so separation can occur.  We'd typically use species in most situations, but there are instances where phases may be required. 

[Patrickodon]

That's very interesting, thanks - I'll change to species mixing to try and get my chemical reaction occuring.

 

Will the chemical reaction be inherent in the mixture then, or is that something that needs its own separate setup?

[Rob]

No, reactions will need switching on. Fluent is a solver where you turn models on as you need them, so we start with single species but can then turn on multiple species, and from there reactions. It's often sufficient to look at the mixing, but here I suspect you need to go that step further. 

[Patrickodon]

OK thanks, what I've done is I've taken my working multiphase model and changed it to species:

 

 

I've set the hydrogel phase (i.e. 99 % of my mix) to this species, and I still patch the calcium section (1 %) at 0.5 in initialization. So I've gotten this after 60 seconds of mixing (I'll have more running overnight, this is just what I started for now):

 

 

Do I need to modify any of the settings I've shown in the species window (or elsewhere in Physics/User-Defined), or am I good with running a long section of these settings?

[Rob]

Patch is fine, and you should see the material diffusing. I'd look at using planes to see what's happening in the middle of the domain. Check the mesh too, I shouldn't be able to see faceting of the top edge like that. 

[Patrickodon]

I've used the default mesh, and I'll update it once I have crosslinking occuring.

 

Here is my mixing after 300 s:

 

 

The viscosity will even out at about 102 cSt when homogenous, just like the non-crosslinking mixing. How do I check whether crosslinking should be occuring? I'm already in species mixing etc. as described above.

 

From what I know, with my 99 % 100 cST and 1 % 500 cSt fluids, the homogenous viscosity should be higher again, due to the chemical crosslinking that should be occuring as the fluids mix.

[Rob]

Diffusing species will show the cell mean based on the volume (mass or whatever) weighted average and is set via the mixture properties. Cross linking is a chemical effect, so I may mix two low-ish viscosity materials that then react to form the high viscosity near liquid that may then cure to become a solid. 

Which are you after? A non-reacting mixture with properties based on it's component species or a reacting mixture where A+B=C and C's properties are not related to A or B in anything other than mass. 

[Patrickodon]

a reacting mixture where A+B=C and C's properties are not related to A or B in anything other than mass. 

 

This is what I'm after; while the mixture may not solidify, it could stay liquid with the two fluids mixed, but certainly there will be a chemical reaction that changes the fundamental chemistry of the solution.

[Rob]

So you'll need to add in some chemistry. At present you're looking at pure mixing. 

[Patrickodon]

Perfect, sorry for being ignorant here but how do I add crosslinking chemistry in Fluent? I assumed the species model would allow it, but I obviously need to do more?

[Rob]

Cross linking is (I think) where you form a clathrate type gel with some other fluids: gas hydrates being a common one. At the basic level it's chemisty, but the mechanism is much more complex. So, you may be looking at chemistry, or you may need UDFs and sink/source terms.  I can guide, but only in general terms or solver terms: the PhD you may be embarking on to model clathrates is for you to do! ;) 

[Patrickodon]

I can guide, but only in general terms or solver terms

 

What terms should I be starting with? My problem is that I'm doing a Masters, not a Ph.D., and have only a few months left in my work; ideally I will get some start at the crosslinking (with a Future Work section detailing what subsequently needs to be done).

[Rob]

What information have you got on the reaction rates and effects? That's the starting point. In Fluent it's then either reactions or sink/source terms using a UDF to remove the two materials and add the crosslinked material. 

Another starting point is to simply look at the fluid mixing: if the reaction is slow can you infer an effect based on how quickly the two materials mix? 

[Patrickodon]

Well what I'm doing is taking two fluids and "just" mixing them (for now). I was hoping I could set up Ansys such that it would react appropriately, given my inputs:

 

-> 99 % water

-> 1 % calcium (fluid; literally water, non-granular, with calcium's density & viscosity [looked up online from a paper])

-> 360 rpm stirrer bar (bottom of vessel)

-> cylindrical vessel

 

With my current mixing, I have no chemical reactions; I'm wondering how I go about programming these? I'm sure once I know what area (i.e. species, viscous, discrete phase etc.) to look into, I can go down a rabbit hole on YouTube etc. for programming the chemical crosslinking reaction.

[Rob]

Does 1% calcium in water really have that much of an effect? Other than making tea a little chewy... 

Species assumes everything mixes. Multiphase will disperse (looks like mixing) but may then settle out. I don't know enough about the cross-linking mechanism to really advise, but do know it's poorly understood and more complicated than most realise.