Mixing of two fluids (mashed potato and pigment-colour), singlephase or multiphase

Zeljko · February 22, 2018, 05:10

Hi,
I would like to simulate the mixing of two fluids (mashed potato and pigment-colour which is used to observe the intensity of mixing of the potato). Those fluids are mixed with stirrer. I am wondering if I have to use multicomponent singlephase simulation or multiphase simulation. I am also interested if one of the fluids (potato) should be set as continuous phase while other fluid is dispersed phase or both fluids should be set as continuous phases.

Thanks,
Zeljko

Gert-Jan · February 22, 2018, 17:41

Main considerations are:
- how much pigment colour is added
- are the fluid properties (rheology, density, etc) of both liquids equal or very different? (Provided mashed potato can be seen as a fluid).

Then:
1) If the amount is small, I would model the colour by scalar mixing as a additional variable (Similar to multicomponent single phase).
2) If the amount is large and the fluid properties are similar to the mashed patato, then I would also by model it by scalar mixing as a additional variable (Similar to multicomponent single phase).
3) If the amount is large and the fluid properties are very different, then you are in trouble, because you have to treat it as multiphase. Then, if one phase is dispersed, you need to define a particle/droplet size. If both are continuous, you need to define a Interphase length scale. both approaches are similar, so it does not really matter.

What question do you want to answer with the modelling? I think mashed potato makes the mixing laminar. Then using CFX, you can only get an idea of the macro mixing. Micromixing will be almost impossible.

Lance · February 23, 2018, 03:50

Quote:

Originally Posted by Zeljko

Hi,
I am also interested if one of the fluids (potato)

A bit off topic, but this is why I love cfd. Some people do simulations of airfoils, pumps, heat exchangers, and so on. You are doing potatoes.

ghorrocks · February 23, 2018, 05:14

Coloured potatoes in fact. Even better.

Gert-Jan · February 23, 2018, 05:33

I did CFD on a potato washing machine.......

Zeljko · February 23, 2018, 05:58

Hi Gert-Jan ,

thanks for your help.

There is around 1l of potato's volume (that is around 1,1 kg) and only around 5 g of the colour which is then dispersed through potato. Concerning your recommendation, I should model the mixing as a additional variable.

What I would like to get as a result is comparison of the two stirrers, ie. which stirrer better and faster mixes the given quantity of mashed potato within given time period.

PS: How would you define this additional variable? I tried with variable with [g] units, ie. I "pour" 5 g of the colour into the potato at time t=5 s for example, and then continue to mix it for another 60 s. So, I run it transient. Do you have an idea, which parameter should I monitor to get a contour of mixing homogeneity throughout the potato's volume?

Regards,
Zeljko

LorenzoSaintDubois · February 23, 2018, 08:18

You should set some monitoring points at different coordinates in your geometry and record the value of the additional variable. Given enough time, the monitored values should all converge at the average concentration (5 gramps per liter). This time-point defines your mixing time.

Since this hasn't been mentioned yet, I think there are some additional points to consider: Is your mixing vessel actualy mixing or rather kneading the potato mash? Because if the potato is kneaded, this is a much trickier simulation, especially if the piment is added on the freely moving (kneaded) surface of the potato phase. I would also expect the potato mash to be strongly non-newtonian with all that starch molecules floating around, so there's also that. If it is non-newtonian *and* highly viscous *and* being kneaded (with a free surface) rather than stirred, then I think we slowly depart from what it possible to simulate using CFX...

Zeljko · February 23, 2018, 08:58

Hi Lorenzo,

Thanks for the provided help. Yes, this is what I monitor at the moment. Actually, I have defined new variable based on the variable "colour". This new variable is defined as: volumeInt(colour)@domain / mass()@domain. So, I would like to monitor mass fraction of the colour as well, where variable "colour" is defined as additional variable (transport equation) with units [kg/m3]. Currently, I consider mashed potato to be newtonian fluid (I will make it non-newtonian in the next run). Mixing vessel is currently only stirring the potato, this is its only function at the moment, it doesn't have to knead it.
I hope this will work properly, particularly the monitoring of the mass fraction of the colour as I plan to use that parameter to make a decision which stirrer is more efficient.

Regards,
Zeljko

February 22, 2018, 05:10	Mixing of two fluids (mashed potato and pigment-colour), singlephase or multiphase	#1
Zeljko New Member Join Date: May 2011 Posts: 8 Rep Power: 15	Hi, I would like to simulate the mixing of two fluids (mashed potato and pigment-colour which is used to observe the intensity of mixing of the potato). Those fluids are mixed with stirrer. I am wondering if I have to use multicomponent singlephase simulation or multiphase simulation. I am also interested if one of the fluids (potato) should be set as continuous phase while other fluid is dispersed phase or both fluids should be set as continuous phases. Thanks, Zeljko

February 23, 2018, 05:14		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Coloured potatoes in fact. Even better. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

February 22, 2018, 17:41		#2
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,928 Rep Power: 28	Main considerations are: - how much pigment colour is added - are the fluid properties (rheology, density, etc) of both liquids equal or very different? (Provided mashed potato can be seen as a fluid). Then: 1) If the amount is small, I would model the colour by scalar mixing as a additional variable (Similar to multicomponent single phase). 2) If the amount is large and the fluid properties are similar to the mashed patato, then I would also by model it by scalar mixing as a additional variable (Similar to multicomponent single phase). 3) If the amount is large and the fluid properties are very different, then you are in trouble, because you have to treat it as multiphase. Then, if one phase is dispersed, you need to define a particle/droplet size. If both are continuous, you need to define a Interphase length scale. both approaches are similar, so it does not really matter. What question do you want to answer with the modelling? I think mashed potato makes the mixing laminar. Then using CFX, you can only get an idea of the macro mixing. Micromixing will be almost impossible.

February 23, 2018, 05:33		#5
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,928 Rep Power: 28	I did CFD on a potato washing machine.......

February 23, 2018, 05:58		#6
Zeljko New Member Join Date: May 2011 Posts: 8 Rep Power: 15	Hi Gert-Jan , thanks for your help. There is around 1l of potato's volume (that is around 1,1 kg) and only around 5 g of the colour which is then dispersed through potato. Concerning your recommendation, I should model the mixing as a additional variable. What I would like to get as a result is comparison of the two stirrers, ie. which stirrer better and faster mixes the given quantity of mashed potato within given time period. PS: How would you define this additional variable? I tried with variable with [g] units, ie. I "pour" 5 g of the colour into the potato at time t=5 s for example, and then continue to mix it for another 60 s. So, I run it transient. Do you have an idea, which parameter should I monitor to get a contour of mixing homogeneity throughout the potato's volume? Regards, Zeljko

February 23, 2018, 08:18		#7
LorenzoSaintDubois New Member Join Date: Mar 2015 Posts: 5 Rep Power: 11	You should set some monitoring points at different coordinates in your geometry and record the value of the additional variable. Given enough time, the monitored values should all converge at the average concentration (5 gramps per liter). This time-point defines your mixing time. Since this hasn't been mentioned yet, I think there are some additional points to consider: Is your mixing vessel actualy mixing or rather kneading the potato mash? Because if the potato is kneaded, this is a much trickier simulation, especially if the piment is added on the freely moving (kneaded) surface of the potato phase. I would also expect the potato mash to be strongly non-newtonian with all that starch molecules floating around, so there's also that. If it is non-newtonian and highly viscous and being kneaded (with a free surface) rather than stirred, then I think we slowly depart from what it possible to simulate using CFX...

February 23, 2018, 08:58		#8
Zeljko New Member Join Date: May 2011 Posts: 8 Rep Power: 15	Hi Lorenzo, Thanks for the provided help. Yes, this is what I monitor at the moment. Actually, I have defined new variable based on the variable "colour". This new variable is defined as: volumeInt(colour)@domain / mass()@domain. So, I would like to monitor mass fraction of the colour as well, where variable "colour" is defined as additional variable (transport equation) with units [kg/m3]. Currently, I consider mashed potato to be newtonian fluid (I will make it non-newtonian in the next run). Mixing vessel is currently only stirring the potato, this is its only function at the moment, it doesn't have to knead it. I hope this will work properly, particularly the monitoring of the mass fraction of the colour as I plan to use that parameter to make a decision which stirrer is more efficient. Regards, Zeljko