CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   radiation in CFX (

george September 26, 2005 18:21

radiation in CFX
I'm using CFX for a simulation including radiation. For every solid surface, there is a solid-fluid interface. I was told that radiation should be applied to the fluid side of solid-fluid interface. This sounds wired to me. Does not radiation occurs on solid surface? Anybody has experience with radiation in CFX? Thank you.


opaque September 26, 2005 21:48

Re: radiation in CFX
I imagine you meant weird instead of wired.

Imagine typical textbook example of radiation between infinite opaque parallel plates at temperatures T1 and T2.. The solution is given as a function of emissive power and surface emissivities. There is no mention of the solid material the plates are made of. Since opaque surfaces absorb radiation within a layer of zero thickness (specially if coated/rusted/treated), the base solid material properties are not important.

As a second example, take now two semitransparent solid separated by an infinitely thin opaque material (glue,tin foil, etc). Radiation will not travel through the interface since it is opaque, and the emissivity on either side is a function of the thin material characteristics. Again, the solid base materials on either side are not important when setting the boundary conditions.

Summary, for opaque materials radiation is absorbed at the interface and it can be different on either side. Then, being able to set different values on either side is an advantage if needed.

I hope I did not confused you with this, Good luck, Opaque..

Glenn Horrocks September 26, 2005 23:17

Re: radiation in CFX

Don't forget also that if the solid is not opaque (eg glass) then you can still model radiation in the solid if it is important.

Glenn Horrocks

george September 27, 2005 10:14

Re: radiation in CFX10
Thanks a lot.

So for a opaque surface, do I need to set the same emissivity on both solid & fluid side, or just the fluid side?

Furthermore, under what situation do you want set different values on solid & fluid side?

BTW, I'm using CFX10.


opaque September 27, 2005 10:45

Re: radiation in CFX10
You only set the emissivity of an opaque surface on the side the radiation model is active. If it is not active on a given side, there is nothing to set.

On the previous example, imagine the thin material is tin foil, and it is specially painted on one side. Now one side is metallic, and the other is coated. If radiation is active on both sides, then the emissivity is different on either side.. That is the beauty of surface coating/treatment to improve/change thermal radiation properties..

Good luck Opaque..

george September 27, 2005 11:12

Re: radiation in CFX10
My application is a little different from your example. Let me try to explain it:

I have a solid part with certain volume, which is a solid domain in CFX. It's surrounded by air, which is my fluid domain.

To model radiation, do I turn on radiation model for both solid domain and fluid domain, and set same emissivity for both solid side & fluid side of soild-fluid interface?

Or I just turn on radiation model for fluid domain? this is what I was told. This is where I got confused: trun on radiation for solid domain sounds more logic to me:)

Thanks! george

opaque September 27, 2005 13:50

Re: radiation in CFX10
Dear George,

The examples are exactly what you are doing assuming your solid is opaque, is it? You did not say what solid you have, is it something like glass (semi-transparent), or an opaque metal.

Radiation will travel with small attenuation within Air, and it will be absorbed, reflected and re-emitted at an opaque surface based on the previous discussion.

Also,if you have a non-opaque domain interface, that is semi-transparent, no need of specifying an emissivity.

I am still not sure yet why you insist that radiation should be turned on in the solid material (again assumming is opaque). The radiation is being absorbed at the interface in a so short distance within the solid that the phenomena become a surface phenomena (from the modeling point of view) than a volumetric one (as you seem to infer).

However, in CFX you for sure can make the radiation active in the solid, turn the boundary condition type to Conservative Interface Flux on both sides, set the Absorption Coefficient for the solid (which will be huge since it is opaque) and refine your mesh beyond believe at the interface on the solid side to capture the absorption. The mesh scale will be of the order of (1./Absorption coefficient) * ln(attenuation fraction). The attenuation fraction will be much smaller than 1%.. That is why you model an opaque material as a surface phenomena, and not a volumetric one.

Hope this help, Opaque..

george September 27, 2005 14:17

Re: radiation in CFX10
It really helps!

Yes, you are right. The solid part is opaque metal. So I only need to acticve radiation for fluid domain and set emissivity on fluid side at the fluid-solid interface.

What I'm interested is to see how much radiation will affect the temperature on the solid part. I assume the radiation energy is generated from the solid surface, then transmitted to fluid domain. That's why I thought emissivity 'd applied to solid surface.

Thanks for the help! george

Yingchun October 12, 2005 06:44

Re: radiation in CFX10
Thanks to Opaque, your comments are very helpful to me as well.

I am using CFX5.7.1 to model double skin façades incorporate with blinds. Since the blinds are suspended within the façade cavity, I need to turn on the radiation effects for both sides of the blinds. However, I am not able to provide a fixed temperature or a fixed heat flux on the blinds surface. Also the thickness of the blinds is very small (less than 1.0mm) I am therefore not going to model the heat conduction through them. It seems the only option left for me is the 'adiabatic' surface property, but if I use it I assume I am going to miss the heat conduction through it. One question for this is that â€" will the code take the adjacent fluid temperature as the surface temperature of the blinds (as the emissive power)? If the fluid (air) temperature at either side of the blinds doesn't have a big difference, can the 'adiabatic' boundary be sufficient to use for this case?

Thank you for your help

Regards Yingchun

topsedar April 26, 2013 18:43

how to model vacuum apace ?
I am modeling an absorber tube of parabolic trough collector using CFX.
the simplified geometry is as follows: (like the attached photo)
a steel tube that is enveloped by a concentric glass tube . between glass tube and absorber tube, a vacuum space exists. and inside the steel tube the heat transfer fluid, flows. my goal is to calculate the distribution of temperature in the glass envelope
my question is:

1- how could i model radiation in the vacuum space?

2- in which domains should i active the radiation model?
fluid domain? absorber domain? vacuum? glass??

absorber tube

ghorrocks April 27, 2013 07:06

Wow, a post of mine from 7.5 years ago. I suspect I had a few less grey hairs back then.

It depends what you want to do for the vacuum space. If you want to model the thin gas which is present there then model it as a normal fluid domain. If you want to assume it as a pure vacuum and have no fluid flow or heat transfer (other than radiation) then I would model it as a solid, but with zero thermal conductivity and no opacity. Then the radiation will pass through unaffected, but no fluid flow of heat transfer will occur.

You would have to post a drawing for me to suggest how this is to be modelled.

topsedar April 28, 2013 16:11

wow !! the start of this topic in 2005 !!
I tried to set a solid domain with thermal conductivity of zero for vacuum domain but as soon as Starting the solver, the solver prints floating error (divided by zero).

my domain is something like the picture in my previous post. an absorber tube which Surround by a glass envelope and to maintain a vacuum space between this two tube (absorber tube and, glass envelope), at end sides, two accordions are utilized.

i am trying to model an exact vacuum space between this two tube.(not filed with a thin gas)
this is my first experience with radiation modeling and i am really confused !!

i want to know in which domain should i active radiation, to obtain the temperature distributions in glass envelope due to emission of radiation from absorber tube surface?? and also the effect of emission of radiation from glass envelope surface on absorber tube temperature?

ghorrocks April 28, 2013 18:51

CFX might not like zero conductivity. Then just set a small number - and then do a sensitivity analysis to check that your number is small enough to be effectively zero.

Before jumping into radiation modelling you need to have a think about whether you actually need to model it. If you know the heat load on the surface it heats then you can just apply it as a heat flux and not use radiation modelling at all. You would only model the radiation if:
* The radiation establishes some form of equilibrium (and even then many forms of this equilibrium can be done with simple heat loads)
* The radiation interacts with the solids and fluids in a more complex way than just heating up a surface.
* The path the radiation takes is important

topsedar May 4, 2013 15:42

Dear ghorrocks Thanks For Your attention.
I was Reading different post related to modeling vacuum @

altogether several options was mentioned for modeling vacuum space :

1- filling the space between two solids with air and turning of the solver by means of export parameters.

2- filling the space between two solids with a solid with near Zero Thermal conductivity

a- Which of above methods would be helpful and accurate?

b- activating Radiation For a Steel Tube is meaningful?

c- assume That We have an Ideal Vacuum (there is not any fluid).
What Should I do in order to model The Vacuum?

d- assume That We have a Rare gas in Vacuum Space.
In Which domains Should I Active The Radiation? ( suppose that We are not allowed to ignore The effect of Radiation )

Here A attached a picture of my Case:

ghorrocks May 5, 2013 07:45

You always start from the physics you want to model. Does this mean you want no thermally conducted heat, but radiation heat condition is taking place? Anything else?

topsedar May 5, 2013 09:56


Originally Posted by ghorrocks (Post 425313)
Does this mean you want no thermally conducted heat, but radiation heat condition is taking place?

Yes . in Vacumm the only heat transfer mechanism is Radiation. :)

ghorrocks May 5, 2013 18:47

Do you need to model the outer envelope and vacuum space at all? Can the radiation heat load be just applied as a heat source on the outer face of the absorber tube?

In fact, what are you trying to achieve with this model at all? What are you trying to learn?

topsedar May 7, 2013 17:32


Originally Posted by ghorrocks (Post 425422)
In fact, what are you trying to achieve with this model at all? What are you trying to learn?

I am trying to find temperature distribution in the absorber tube and glass tube considering the effect of Radiation.

ghorrocks May 7, 2013 19:25

I assume then you know the radiation intensity, wavelengths and/or directions at the outside of the glass envelope. But then do you know the radiation impinging on the absorber tube? If yes then this simplifies the model a lot. If not then you will have to run a radiation model in the vacuum space.

topsedar May 8, 2013 09:48


Originally Posted by ghorrocks (Post 425968)
I assume then you know the radiation intensity, wavelengths and/or directions at the outside of the glass envelope. But then do you know the radiation impinging on the absorber tube?

Yes I calculated the Flux Distribution of Sun Radiation on the glass and absorber by SolTrace Software. this means that i know the radiation Flux on outer surface of glass and absorber. you are right. i guess in this situation activing Radiation model only in vacuum space and setting some sources on the absorber and glass would complete radiation modeling in my case.

All times are GMT -4. The time now is 13:20.