CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   -nano-scale treatment in CFD (http://www.cfd-online.com/Forums/main/111814-nano-scale-treatment-cfd.html)

jgg January 16, 2013 05:06

-nano-scale treatment in CFD
 
Dear all,
I am new in the forum and I would appreciate any help about the acceleration process of iron-oxide
nanoparticles caused by a flow of helium gas in a 1 mm diameter tube . We are interested in getting to know
the velocity of the nanoparticles at a fixed length (L) accelerated by the
continous phase.
Sample input data:
Tube diameter = 1 mm
Inlet velocity of helium gas = 20 m/s
Size of iron-oxide nanoparticles (between 10 and 100 nm)
inlet mass flow rate of discrete phase = 1mg/min (10.E10 particles aprox.)
Can CFD treat the "nano-scale"?
Multiphase models implemented in Fluent are going to be used?
Which one would be the most suitable??
Any idea about how to do it if possible?
** This is nothing to do with the proper use of an Euler or Lagrange approach as far as multiphase modelling
is concerned. My question or doubt (or darkness in which I am involved) is concerning about if the general
multiphase models implemented in FLUENT -can- take into account the interaction forces between the discret &
continous phase (or among the different phases) when we are in a -nano- scale instead of -mili/micro- scale.
I have read something and it seems to be that in case of -nano- particles, a special
phenomenon or motion or force called -Brownian- appeared, but I know -nothing- about
this effect neither a special treatment to be done when a -nanoparticles- aree treated.
I would need any help to get to know if this application could be treated with a general-purpose
soft like FLUENT or the problem is not as easy like it seems??
Thanks for all in advance and I will be waiting for sb to give some idea.

cfdnewbie January 16, 2013 08:49

The navier-stokes equations are derived assuming continuum assumption (Knudsennumber -> 0). If your flow has a high Knudsen number, the validity of this assumption becomes weak, and any code solving the navier stokes equations will get into trouble. There are some extensions to rarefied or extended gas dynamics, but I doubt that they have been included in a commercial solver, in fact, most of this topic is still very current research.

flotus1 January 16, 2013 10:16

Fluent comes with a boundary condition suitable for extending the applicability of the NS-equations to about Kn=0.1.
See "slip boundary formulation for low pressure gas systems" in the fluent documentation.

michujo January 16, 2013 13:09

Hi, I don't see such problem. The nano particles are transported by a continuous phase (helium), which is not rarefied, so that NS holds perfectly.

The transport of nano particles can be modelled by seeding the flow with a discrete number of particles and tracking their trajectories individually (Lagrangian approach) or using an Eulerian formulation, which also holds if the nano particles are diluted in the main gaseous phase.

This will allow you to obtain the velocity of the particles.

As far as I know, FLUENT allows you to include the Brownian motion effect.

I would suggest that you calculate the Stokes number of the particles beforehand. If it is sufficiently small the particles will accelerate right away and attain the velocity of the gaseous stream. Otherwise, if St~1 or higher, the inertia of the particles is not negligible.

Cheers,
Michujo.

ghorrocks January 16, 2013 17:37

Hi JGG:

You asked me to comment on this thread but the comments here already are a good summary of the issues to consider. Other than to say that if you are looking at the length by which the particle accelerate to flow velocity - this is easily worked out through the stokes number and you do not need CFD for that.

flotus1 January 17, 2013 06:00

Quote:

Originally Posted by michujo (Post 402272)
The nano particles are transported by a continuous phase (helium), which is not rarefied, so that NS holds perfectly.

I disagree with that. If the Knudsen number evaluated from the mean free path of the fluid and the characteristic length of the particles is high, the NS-equations are no longer valid.

michujo January 17, 2013 08:36

Quote:

Originally Posted by flotus1 (Post 402405)
I disagree with that. If the Knudsen number evaluated from the mean free path of the fluid and the characteristic length of the particles is high, the NS-equations are no longer valid.

Hi, thanks for your comment.
Out of curiosity I calculated the Knudsen numbers of the gas relative to the tube diameter (1 mm) and to the particles diameter (1e-7 to 1e-8 m). The mean free path of the gas is calculated for p=1 bar and T=273 K. I get:

Kn_tube ~ 2e-4

Kn_particles ~ 2e1 to 2e0

In light of these results I think we can say that NS hold for the gas, since the Kn relative to the tube characteristic dimension is small. However, for the particles it is not.

Do the concept of Stokes number or drag force still hold for this case, where the particles are travelling through the interstices of the helium gas particles? How would you compute the drag force of the solid particles then?

Thanks,
Michujo.

flotus1 January 17, 2013 09:14

Quote:

Originally Posted by michujo (Post 402432)
In light of these results I think we can say that NS hold for the gas, since the Kn relative to the tube characteristic dimension is small. However, for the particles it is not.
Michujo.

We can definitely agree on that.

Quote:

Do the concept of Stokes number or drag force still hold for this case, where the particles are travelling through the interstices of the helium gas particles
The drag force experienced by a particle travelling through a rarefied gas is lower than the force one would expect for creeping flow.
But it should still be possible to calculate the drag through some kind of algebraic expression.
I havent done any research on this particular topic, but i am sure that there exists literature about it.

ghorrocks January 28, 2013 17:45

You definitely should check the Knudsen number is in the range where the NS equations apply. If that is OK, then you should check the stokes number of your particles and work out the terminal velocity/range of the particles and see if that is much smaller than your region of interest. For nanoparticles this is usually the case.


All times are GMT -4. The time now is 09:23.