CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Multiphase or Multicomponent (

Pandu Sattvika November 2, 2001 15:51

Multiphase or Multicomponent
Hi cfd folks,

I have a bunch of solid particles each has different particle sizes. I can represent this solids by having 7 different diameters. This solid flows through a pipe bends(blown by constant velocity of air), the place I want to model its particle distribution based on their diameter, since there is centrifugal forces acting in there. Is it best to model this as a multicomponent or multiphase model? Can anyone tell me the difference between multicomponent and multiphase thermodynamic phases?



Jan Rusaas November 2, 2001 16:54

Re: Multiphase or Multicomponent
What do you mean with a multicomponent? An important question is, what is the volumefraction of the particles - is it a dense or dilute flow. Jan

keny November 3, 2001 04:56

Re: Multiphase or Multicomponent
if you talk about multiphase, I guess you mean gas and solid, for which the solid phase can be modeled on a lagrangian scheme because the particles are few, (i guess global particle hold up is low). Gas being the continuous phase and the the solids dispersed phase. Some people wil call this Eulerian-Langrangian scheme.

However, when you talk about multicomponents, how many compents do you have and which ones are they?

In a multicomponent modelling, the species are mixed at molecular level, share the same pressure, velocity and temperature fields. As the relative size of the species become bigger, there is deviation from the uniform flow field and hence Multiphase modelling takes over from multicomponent modelling.


Ribeiro November 5, 2001 14:54

Re: Multiphase or Multicomponent

I think Pandu should talking about particles size distribution. In the CFX-4 there are a model that treats a solid-gas flow with multiple particle size. It is MUSIG. I try to use in a FCC Riser (dilute gas-solid flow) and it seems very well.


Jan Rusås November 6, 2001 04:04

Re: Multiphase or Multicomponent
I think Pandu maybe should explain himself :)and the question was about solid particles, so the MUSIG model is not appropriate.

Just a question about the MUSIG model. Did you use the default MUSIG model to get the result that seems very well. Have you tried to adjust the constants in the MUSIG model, I am very interessed in how sensitive the model is on a real test case.


Ribeiro November 6, 2001 12:38

Re: Multiphase or Multicomponent

Indeed I saw a co-worker to use just as a test case. It was a riser with a dilute gas-solid flow. The solid phase was supposed to have 2 size groups. Without beakup or coalescence (BC). Thus, a very simple case. But we could enter some breakup-coalescence model because in the real equipment, the solid particles agglomerate and break up.

I will try soon to model a real test case for solids (the suggestions of use appearing in the manual is because the BC model)

Pandu Sattvika November 6, 2001 13:15

Re: Multiphase or Multicomponent
Dear all,

I apoplogize for not being clear enough. Let me clarify this.

The particle is just sand. The sand can be categorized into 7 different sizes by their diameter. At some length in a pipe, I want to look at cross sectional particle distribution, especially during the pipe bends.

I havent consider the MUSIG model yet, but I certainly will.


Jan Rusås November 6, 2001 14:47

Re: Multiphase or Multicomponent
I have never considered using the MUSIG model for solid particle, but I guess it is never to late to learn something new.

I imagine that the MUSIG model is only relevant when coalence or breakup occurs. For a dilute flow case it must be a waste of time and inaccurate (I guess). I do think the MUSIG model was developed for liquid and gas bubbles - gas riser. (correct me if I am wrong) If that is true, one probably has to adjust the constants in the model to a gas solid flow, which require good skills and high quality measurements.

Any experience, I have tried using the MUSIG model for steam atomization of oil, and atomization of water with pressurized air. I was not happy with the results, and belive the reason is the problems with the constants. I am going to do some experiments and will from that dig more into the problem. But if any out there already has some experience please share.

Regards Jan

Pandu Sattvika November 8, 2001 02:01

Re: Multiphase or Multicomponent
Hi Riberio,

I do not understand how to set up this MUSIG model with no coalescence. In >>CFX4 >>OPTION and in
:>MODEL BOUNDARIES, how do you set it up for my case. I still dont understand the musig volume fraction. Is it the fraction for the solid component only or the fraction as a whole components like air + solid. And also my aim is to look at the pipe cross section at a given length and see the solid distribution at that cross section . The solid particles, again can be categorised into 7 different diameters. Solid is just sand grain.

Sorry to ask you the detail, but the manual did not specify much for this model. I am still confused, but I am sure this will work.

Thanks for your suggestion,


Ribeiro November 12, 2001 18:22

Re: Multiphase or Multicomponent
Ok, man. Let's go

The Musig model without Coalescence and Breakup results in a equation with two terms: transient and convective. Thus, the position of each group will depend on only of the fluid velocity. Of course, this is no longer a "Musig model". But an equation to describe the effects of bulk flow over particles.

In >>OPTIONS, I would set the number of groups like the number of diameters. In >> Model Boundaries, I'd set the fraction of each size group that enters the equipment, i.e., I MUST know the grain size distribution.

The Musig volume fraction is based in the dispersed phase. So, if you have 50% of a 50 microns particle and 50% of a 70 microns particle, you have musig volume fraction1 = 0.5 and musig volume fration2 = 0.5. In other words, is the fraction of dispersed phase that has the same diameter.

I don't know if I was clear...

Pandu January 7, 2002 05:40

Re: Multiphase or Multicomponent
Hi all,

I finally looked up at the musig model, and I followed the solver example (the model is not define in the cfx build), and it is working. However, when I want to use the model which i construct in cfx build, error message appear in CFX visualise. I still not have clear grasp how to appropriately model structure in cfx build (CFX 4.4). For example, do we make patches after meshing? How do you join 3 different solids, etc

Thanks for the help


All times are GMT -4. The time now is 02:48.