[CFDGERMANY]

Hi, guys.

I am simulating a centrifuge which is full of air. How can I capture one of the components e.g. O2 in the simulation?


I have already done the CFX-Pre with multicomponent O2 and N2. Do I have to write a Expression to post the properties of O2?



Thanks.

[ghorrocks]

I don't think CFX can model the separation of a multicomponent mixture by gravity/acceleration. So unless something else is happening, CFX will model the air as the same composition everywhere.

But to answer your question, are you looking to get the mass fraction of O2? Refer to the CFX reference manual, but it is simply a CEL variable like O2.mass fraction (assuming you called the O2 mass fraction "O2").

[CFDGERMANY]

Thanks for your Reply.

Yes, I am trying to get the mass fraction of O2 at the wall of the centrifuge, cause O2 is heavier than N2.

I have already known the differences of definition between Multicomponent and Multiphase.
Could you tell me the differences of calculating process between Multicomponent and Multiphase in cfx?

[CFDGERMANY]

Thanks for your Reply.

Yes, I am trying to get the mass fraction of O2 at the wall of the centrifuge, cause O2 is heavier than N2.

I have already known the differences of definition between Multicomponent and Multiphase.
Could you tell me the differences of calculating process between Multicomponent and Multiphase in cfx?

Originally Posted by ghorrocks

I don't think CFX can model the separation of a multicomponent mixture by gravity/acceleration. So unless something else is happening, CFX will model the air as the same composition everywhere.

But to answer your question, are you looking to get the mass fraction of O2? Refer to the CFX reference manual, but it is simply a CEL variable like O2.mass fraction (assuming you called the O2 mass fraction "O2").

[Opaque]

Have you looked at the theory documentation for the modeling of multicomponent flow?

You can see the transport equation for the component is only due to main flow advection, and mass diffusion; therefore, there is no information about additional body forces that will drive a given mass fraction away/towards anything.

For components/species separation due to centrifugal forces, you will need those body forces in the definition of the mass diffusion velocity. In addition, the discretization details are tricky then to differentiate the proper gradients.

I am not clear what you meant by

"Yes, I am trying to get the mass fraction of O2 at the wall of the centrifuge, cause O2 is heavier than N2."

Are you asking how to access the mass fraction value for O2 at the wall? (See Glenn's response),

Are you trying to solve for the mass fraction distribution/field for O2 & N2? and you expect to find O2 at the outer wall because it is heavier than N2. Then, my description above addresses that.

A quick search will lead you to
https://www.fsb.unizg.hr/hydro/web_p...a_(2nd_Ed).pdf

Reading Chapter 24 describes the different mechanisms for mass transport in multicomponent systems.

[lordluan]

If I have a multicomponent fluid and I want to use the pressure of one component in a expression it would be in CEL: ethanol.p?

[EDIT] I used the expression ethanol.p in a source term expression, and CFX-pre replied with this: The variable 'p' referenced by parameter 'Source' in object '/FLOW:Flow Analysis 1/DOMAINefault Domain/SUBDOMAIN:Subdomain 1/SOURCES/EQUATION SOURCE:continuity' is not allowed to have a component name prefix.

I tried to use ptot, but I've got the same error message.

[EDIT2] I've found the answer in CFX error: Variable is not allowed to have a component name prefix

[Antanas]

Originally Posted by lordluan

If I have a multicomponent fluid and I want to use the pressure of one component in a expression it would be in CEL: ethanol.p?

[EDIT] I used the expression ethanol.p in a source term expression, and CFX-pre replied with this: The variable 'p' referenced by parameter 'Source' in object '/FLOW:Flow Analysis 1/DOMAINefault Domain/SUBDOMAIN:Subdomain 1/SOURCES/EQUATION SOURCE:continuity' is not allowed to have a component name prefix.

I tried to use ptot, but I've got the same error message.

[EDIT2] I've found the answer in CFX error: Variable is not allowed to have a component name prefix

There are no partial pressure fields in CFX. The pressure field is common for all phases and components. You should evaluate partial pressures by yourself based on this common pressure and molar fractions of components.

[Opaque]

Originally Posted by Antanas

There are no partial pressure fields in CFX. The pressure field is common for all phases and components. You should evaluate partial pressures by yourself based on this common pressure and molar fractions of components.

Following up on Antanas' clarification:

MyEthanolPressure = Pressure * ethanol.Molar Fraction

or

MyEthanolPressure = p * ethanol.molf

[lordluan]

Originally Posted by Opaque

Following up on Antanas' clarification:

MyEthanolPressure = Pressure * ethanol.Molar Fraction

or

MyEthanolPressure = p * ethanol.molf

Many thanks, Opaque and Antanas!

[Antanas]

Originally Posted by lordluan

Many thanks, Opaque and Antanas!

Don't forget that Pressure or p is relative pressure, not absolute.

[boreal]

Hi ghorrocks

To dig up this thread again. Does your statement really also apply to a separation of species by gravity? In my case i have a variable gas mixture with strongly differing molar masses (1:10). I understand that i need an additional term in the Maxwell-Stefan equation if i want to consider centrifugal forces. But how's about if the gas separates just through gravity. Wouldn't this be taken into account by the body force in the momentum balance because gravity applies to all species by the same way? Unfortunately, this was not entirely clear to me from the solver theory guide.

Best regards,
boreal

Originally Posted by ghorrocks

I don't think CFX can model the separation of a multicomponent mixture by gravity/acceleration. So unless something else is happening, CFX will model the air as the same composition everywhere.

But to answer your question, are you looking to get the mass fraction of O2? Refer to the CFX reference manual, but it is simply a CEL variable like O2.mass fraction (assuming you called the O2 mass fraction "O2").

[Opaque]

Have you read a chapter describing generalized diffusion transport? generalized driving force?

https://www.researchgate.net/file.Po...01497887539930

Ansys CFX solves the momentum transport equation for the mixture (all species have the same bulk velocity), not the species momentum transport equation (where each species has its own velocity).

The only way for such a model to work is that the species transport equation includes all the external forces of relevance; therefore, forcing a species to migrate from one location to another. Those forces can be
1 - Thermal gradient
2 - Pressure gradient
3 - Rotational effects
4 - Electro-magnetic effects (like charged species/ions)
5 - gravity
6 - Concentration gradient

Ansys CFX only includes (6), anything else must be modeled/customized.

[boreal]

I read a bit and tried to understand everything...but i'm not right there ;-).

But you answer was very helpful. At least i know now the limits of my model.



Custumizing such a thing would be new territory for me. I would add this with CEL, right? I'm not sure yet if I do this approach or if i use COMSOL for this simulation.

[ghorrocks]

CEL is just the expression language - to calculate some maths.

To get CFX to model separation of gas species you are going to have to develop models of Brownian motion, differential forces due to gravity on the different species and probably implement it as a source term. There may also be additional effects like the ones Opaque lists. This will all be in the CFX multicomponent model, which is a Eularian framework.

You could also look at it this way - if it was easy to implement, then it would already be in CFX, wouldn't it?

[boreal]

Originally Posted by ghorrocks

You could also look at it this way - if it was easy to implement, then it would already be in CFX, wouldn't it?

Good argument! Well thanks for your answers. I have to think about how i/we will proceed.



By the way: I'm sure you already know it, but you guys are very helpful. For me, CFX is one of several tools I use. I'm far away from the insights you have into the software and math behind it. So thanks a lot.

[ghorrocks]

I am happy to help! Good luck with your analysis.