cfx nanofluid

bitak · December 20, 2012, 16:15

Hello.
How I can simulate nano fluid ( fluid + nano particle) in CFX?
I should use particle transport model or dispersed solid model?
if I use lagrangian model, I should solve my case unsteady?
thanks...

cdegroot · December 21, 2012, 10:35

How many particles do you want to model? I am guessing if they are nano particles, it is a lot? Lagrangian gets more expensive with large numbers of particles. In the Modelling Guide there are tables with the advantages and disadvantages of each. You should weigh these factors when deciding which model to use.

In terms of running steady or unsteady, that depends on if your problem is steady or unsteady, although sometimes time-stepping to a steady-state solution can improve convergence for tough multiphase problems.

bitak · December 21, 2012, 12:08

number of particles are 100.

cdegroot · December 21, 2012, 12:09

I would use Lagrangian then.

ghorrocks · December 22, 2012, 07:18

Whoa there, not so fast.

We should put a FAQ up for nanoparticles, it has been asked quite a bit recently....

What physics is acting on the nanoparticles? The whole thing about nanoparticles is they do not have viscous drag acting on them, instead they have Brownian motion and possibly other forces which are usually ignored for multiphase simulations. So Lagrangian particle tracking is in general NOT suitable for nanoparticles as Lagrangian particle assume viscous drag. So unless you want to define your own particle force routine to put the correct forces on the particles, the approach I generally recommend is to use a passive scalar with diffusivity. The diffusivity can be used as a simple analog for Brownian motion.

Bitak: What is the important physics acting on the particles? That will help decide what approach is best.

cdegroot · December 22, 2012, 08:38

Very interesting Glenn. I was not aware of the special physics behind nano particles.

ghorrocks · December 23, 2012, 04:37

It all comes from CFX being a Navier Stokes solver and that assumes a continuum. Somewhere in the range from micro to nano scale flows (1e-6 to 1e-9 metres) the continuum assumption becomes inaccurate and atomic level stuff starts becoming important. Ideally you should use a molecular kinetics solver for nano scale stuff, but depending on exactly what you are doing you might be able to get away with a NS solver like CFX.

That is why I asked for futher details - whether CFX can do it depends on what is to be modelled.

bitak · December 23, 2012, 13:55

Thanks all.
I want to simulate ferrofluid ( nanoparticles have 10^-9 m diameter) in the tube (5 mm dimeter) and apply a magnetic fieild on this flow.
my governing equations are navier stokes.
I have no slip flow in my domain.
I don't know I should simulate traansient or steady my problem?
and which model in multiphase flow is useful for my problem and also is accurate?
particle transport? or solid disperssion?
Thanks...

cdegroot · December 23, 2012, 15:02

I think Glenn has already answered most of these questions for you. In terms of steady or unsteady that depends what you are trying to model. Are you seeking a steady state solution or are you interested in observing something that changes with time?

ghorrocks · December 23, 2012, 16:26

CFX can model the water bit easily. But dealing with the nanoparticles is a bit trickier. If you use a lagrangian particle tracking approach you are going to have to do something so Brownian motion can be modelled - although if you only have 100 particles the "random walk" approach sounds more appropriate. Tuning this so that it matches the true diffusion due to Brownian motion will be tricky. Adding a body force to the particles for the magentic field is relatively easy.

Alternately if you take the additional variable approach the diffusion is easy, but adding the magnetic body force to it will require some thought.

Can I ask why do you care about 100 particles of diameter 1e-9m in a tube of water? There are going to be far more particles bigger than that just from contamination and shed off the tube wall and other sources. What makes the iron particles so important?

December 20, 2012, 16:15	cfx nanofluid	#1
bitak Member bitak Join Date: Jul 2009 Posts: 56 Rep Power: 16	Hello. How I can simulate nano fluid ( fluid + nano particle) in CFX? I should use particle transport model or dispersed solid model? if I use lagrangian model, I should solve my case unsteady? thanks...

December 22, 2012, 07:18		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Whoa there, not so fast. We should put a FAQ up for nanoparticles, it has been asked quite a bit recently.... What physics is acting on the nanoparticles? The whole thing about nanoparticles is they do not have viscous drag acting on them, instead they have Brownian motion and possibly other forces which are usually ignored for multiphase simulations. So Lagrangian particle tracking is in general NOT suitable for nanoparticles as Lagrangian particle assume viscous drag. So unless you want to define your own particle force routine to put the correct forces on the particles, the approach I generally recommend is to use a passive scalar with diffusivity. The diffusivity can be used as a simple analog for Brownian motion. Bitak: What is the important physics acting on the particles? That will help decide what approach is best. Far and amin_gls like this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFX Treatment of Laminar and Turbulent Flows	Jade M	CFX	18	September 15, 2022 07:08
High Resolution (CFX) vs 2nd Order Upwind (Fluent)	gravis	ANSYS	3	March 24, 2011 02:43
CFX pressure in Simulations problem	nasdak	CFX	1	April 14, 2010 13:22
PhD using CFX	Rui	CFX	9	May 28, 2007 05:59
FSI using CFX and ANSYS	Bi Chang	CFX	2	May 10, 2005 04:47

December 21, 2012, 10:35		#2
cdegroot Senior Member Chris DeGroot Join Date: Nov 2011 Location: Canada Posts: 414 Rep Power: 17	How many particles do you want to model? I am guessing if they are nano particles, it is a lot? Lagrangian gets more expensive with large numbers of particles. In the Modelling Guide there are tables with the advantages and disadvantages of each. You should weigh these factors when deciding which model to use. In terms of running steady or unsteady, that depends on if your problem is steady or unsteady, although sometimes time-stepping to a steady-state solution can improve convergence for tough multiphase problems.

December 21, 2012, 12:08		#3
bitak Member bitak Join Date: Jul 2009 Posts: 56 Rep Power: 16	number of particles are 100.

December 21, 2012, 12:09		#4
cdegroot Senior Member Chris DeGroot Join Date: Nov 2011 Location: Canada Posts: 414 Rep Power: 17	I would use Lagrangian then.

December 22, 2012, 08:38		#6
cdegroot Senior Member Chris DeGroot Join Date: Nov 2011 Location: Canada Posts: 414 Rep Power: 17	Very interesting Glenn. I was not aware of the special physics behind nano particles.

December 23, 2012, 04:37		#7
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	It all comes from CFX being a Navier Stokes solver and that assumes a continuum. Somewhere in the range from micro to nano scale flows (1e-6 to 1e-9 metres) the continuum assumption becomes inaccurate and atomic level stuff starts becoming important. Ideally you should use a molecular kinetics solver for nano scale stuff, but depending on exactly what you are doing you might be able to get away with a NS solver like CFX. That is why I asked for futher details - whether CFX can do it depends on what is to be modelled.

December 23, 2012, 13:55		#8
bitak Member bitak Join Date: Jul 2009 Posts: 56 Rep Power: 16	Thanks all. I want to simulate ferrofluid ( nanoparticles have 10^-9 m diameter) in the tube (5 mm dimeter) and apply a magnetic fieild on this flow. my governing equations are navier stokes. I have no slip flow in my domain. I don't know I should simulate traansient or steady my problem? and which model in multiphase flow is useful for my problem and also is accurate? particle transport? or solid disperssion? Thanks...

December 23, 2012, 15:02		#9
cdegroot Senior Member Chris DeGroot Join Date: Nov 2011 Location: Canada Posts: 414 Rep Power: 17	I think Glenn has already answered most of these questions for you. In terms of steady or unsteady that depends what you are trying to model. Are you seeking a steady state solution or are you interested in observing something that changes with time?

December 23, 2012, 16:26		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	CFX can model the water bit easily. But dealing with the nanoparticles is a bit trickier. If you use a lagrangian particle tracking approach you are going to have to do something so Brownian motion can be modelled - although if you only have 100 particles the "random walk" approach sounds more appropriate. Tuning this so that it matches the true diffusion due to Brownian motion will be tricky. Adding a body force to the particles for the magentic field is relatively easy. Alternately if you take the additional variable approach the diffusion is easy, but adding the magnetic body force to it will require some thought. Can I ask why do you care about 100 particles of diameter 1e-9m in a tube of water? There are going to be far more particles bigger than that just from contamination and shed off the tube wall and other sources. What makes the iron particles so important?