Is homogenous viscosity a good indicator for general homogeneity in mixing?

[Pádraig]

Hi all,

 

I'm mixing two fluids, and want to find when they become homogenous. I'm using a mixing programme called Ansys, and I'm checking the mixing via the molecular viscosity. It gives this kind of result (this is after 2550 s of mixing):

 

 

In this mix, I have the initial viscosities of the fluids set at 100 cSt (99 % of mix) and 500 cSt (1 % of mix) respectively. I'm mixing 60-180 s per step, until I'm at homogenous viscosity (so I can see at what time this occurs). Is this OK for finding the general homogeneity of the fluid? Here is the initial mix after 300 s, for interest:

 

[studiot]

 

42 minutes ago, Pádraig said:

I'm mixing two fluids, and want to find when they become homogenous. I'm using a mixing programme called Ansys, and I'm checking the mixing via the molecular viscosity. It gives this kind of result (this is after 2550 s of mixing):

As this is outside my experience, I am simply going to add  @exchemist  , @OldChemE and @sethoflagos  who may welll be better placed to help you.

[sethoflagos]

3 hours ago, Pádraig said:

Is this OK for finding the general homogeneity of the fluid?

In the absence of any explanation of how Ansys attempts to solve the advection equation in particular, then how can we be expected to judge how reliable it is?  

[Pádraig]

15 hours ago, sethoflagos said:

In the absence of any explanation of how Ansys attempts to solve the advection equation in particular, then how can we be expected to judge how reliable it is?  

Sorry, I probably should have added - Assuming Ansys Fluent is accurate (and I am running validation tests on it), is homogenous viscosity an OK meter for finding general homogeneity in mixed fluids? 

[sethoflagos]

5 hours ago, Pádraig said:

Sorry, I probably should have added - Assuming Ansys Fluent is accurate (and I am running validation tests on it), is homogenous viscosity an OK meter for finding general homogeneity in mixed fluids? 

Many viscosity models for mixtures employ terms that resemble in form those that characterise entropy of mixing. So if your model indicates a homogenous viscosity, it implies that the entropy is also homogenous which is as good an indicator of complete mixing as you could wish for.

You will be on fairly safe ground when dealing with say mixtures of liquid alkanes where the components are fully miscible in all proportions and where much of the research on this subject has been focussed. Don't rely on it for say a suspension of cellulose fibres.

[Pádraig]

3 hours ago, sethoflagos said:

Many viscosity models for mixtures employ terms that resemble in form those that characterise entropy of mixing. So if your model indicates a homogenous viscosity, it implies that the entropy is also homogenous which is as good an indicator of complete mixing as you could wish for.

You will be on fairly safe ground when dealing with say mixtures of liquid alkanes where the components are fully miscible in all proportions and where much of the research on this subject has been focussed. Don't rely on it for say a suspension of cellulose fibres.

 

Thanks very much - I'll look into those terms you've given me, that's a big help.

 

Edit: @sethoflagos, this is in the second link: "mixed without chemical reaction". I'm mixing with crosslinking, will that heavily impact whether the viscous homogeneity is general homogeneity?

Edited March 13 by Pádraig

[sethoflagos]

4 hours ago, Pádraig said:

Edit: @sethoflagos, this is in the second link: "mixed without chemical reaction". I'm mixing with crosslinking, will that heavily impact whether the viscous homogeneity is general homogeneity?

It will heavily impact whether your software's predicted homogeneity is realistic, utter nonsense or any point between those two extremes. Really. 

If you're wondering what could possibly go wrong, just think 'lumpy gravy'.

Bottom line is that software simulators are no substitute for laboratory work and pilot plant tests however much employers may wish that they were.   

Edited March 14 by sethoflagos

[Pádraig]

19 hours ago, sethoflagos said:

Bottom line is that software simulators are no substitute for laboratory work and pilot plant tests however much employers may wish that they were.   

 

I completely agree with this, I'd actually be tempted to shout it even louder again. I'm just saying, even in a lab situation, when crosslinking is included - does that affect the overall homogeneity, when the molecular viscosity is homogenous throughout after mixing, and why/why not?

 

(Sorry if this seems like I'm asking the same thing again - it's just that the link for entropy of mixing gives "when several initially separate systems of different composition, each in a thermodynamic state of internal equilibrium, are mixed without chemical reaction"

[sethoflagos]

13 minutes ago, Pádraig said:

 

I completely agree with this, I'd actually be tempted to shout it even louder again. I'm just saying, even in a lab situation, when crosslinking is included - does that affect the overall homogeneity, when the molecular viscosity is homogenous throughout after mixing, and why/why not?

 

(Sorry if this seems like I'm asking the same thing again - it's just that the link for entropy of mixing gives "when several initially separate systems of different composition, each in a thermodynamic state of internal equilibrium, are mixed without chemical reaction"

Depends what you mean by 'overall homogeneity'. If this includes the assumption that your cross-linked product has a uniform molecular weight with exactly the same degree of cross-linking across all particles, then at least in the general case, I see no guarantee of that.

Furthermore, I'd tend to view this as a multiphase mixture rather than an ideal fluid. It's more of a dispersion of solids in a liquid medium and what would the viscosity of that solid phase actually mean?

But I'm straying a fair bit outside my field here. Perhaps @exchemist or @John Cuthber could add a more expert chemist's perspective.    

[npts2020]

Unless the two fluids you are mixing have the same viscosity to begin with, I don't see how it is possible to have homogenous viscosity without also having homogenous mixing. Whether your program will predict the correct outcome or not is a different matter...

[John Cuthber]

8 hours ago, npts2020 said:

Unless the two fluids you are mixing have the same viscosity to begin with, I don't see how it is possible to have homogenous viscosity without also having homogenous mixing. Whether your program will predict the correct outcome or not is a different matter...

That seems sensible to me.
Even if the materials have the same viscosity to start with, a mixture may have a different viscosity (anyone who has seen an engine where water has got into the oil will know that; the product is much more viscous than the components even though no chemical reaction takes place).

So, if the viscosity is homogeneous, the mixture must also be.

But there's a catch 22 here.
If your model doesn't work, then using a modeled property- like viscosity- won't be a good test.
If you could model the density (which is usually "better behaved"than viscosity) it would be better.

[Pádraig]

On 3/15/2023 at 2:07 AM, npts2020 said:

Unless the two fluids you are mixing have the same viscosity to begin with, I don't see how it is possible to have homogenous viscosity without also having homogenous mixing. Whether your program will predict the correct outcome or not is a different matter...

 

That's exactly the logic I've been using, I've just been worried that as I haven't got a high level/"expert" level of chemistry that I am/was overlooking something