CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   radiation in glass plate (http://www.cfd-online.com/Forums/cfx/119799-radiation-glass-plate.html)

keeper June 25, 2013 07:51

radiation in glass plate
 
1 Attachment(s)
Hello guys.

I have a problem with radiation simulation. I simulate flow in the winter garden including sun heat flux. I want to compare different types of glasses (three-plate vs. two-plate etc.).

If I use Discrete Transfer radiation model, everything is OK if I set radiation up only on fluid domain. But I want to simulate radiation through the windows so I suppose that I need to define radiation model also for solid domain so I have to use Monte Carlo for both domains. But as soon as I run the calculation with MC model, I get strange results (temperatures between -50 and 300C).

The radiation source is on grey wall with mesh displayed (600 W/m^-2).

I defined material properties (scattering and absorbtion ceoffs and refractive index) for glass and I left all the interfaces as a default. And this is what I don't know how to set up. Can anyone help me with that? Any idea could help...

Thanks in advance.

Best Regards

ghorrocks June 25, 2013 18:58

I presume then you want to model transmission, reflection and absorption in the glass. I do not think the discrete transfer model can handle this, I think you will need the full Monte Carlo model to do this.

If the Monte Carlo model is giving strange results then I suspect it is not fully converged yet, or you do not have enough rays to accurately model the heat transfer.

Can you post your CCL?

keeper June 27, 2013 01:10

3 Attachment(s)
Thank you for reply.

Yes you are right that is what I want to define and check. And because MC is the only radiation model available for solids I have to use it. I attached CCL and two pictures to see what I obtained. The contours are identical, the only difference is in variable range (post1.png - user specified range, post2.png - local range). Just to see nonrealistic values.

In CCL please note that "okna" domain represents windows.

Thank you in advance for reply.

Best Regards

ghorrocks June 27, 2013 01:16

That is exactly what happens when the radiation model is too coarse. Try reducing the radiation ray tracing coarsening factor.

keeper June 27, 2013 01:57

Excuse me, what do you mean exactly? Where I can find it please? Cause I don't know what you mean exactly right now

ghorrocks June 27, 2013 02:06

Code:

    THERMAL RADIATION CONTROL:
      Diagnostic Output Level = 1
      Iteration Interval = 3
      COARSENING CONTROL:
        Target Coarsening Rate = 32
      END
    END

Target coarsening rate - reduce it from 32, and maybe try a shorter iteration interval. Anything which look like it will make the radiation modelling more accurate.

keeper June 27, 2013 02:07

OK I'll try. Thank you very much Glenn.

keeper June 27, 2013 08:13

So I tried to decrease coarsening rate to 8 and calculated radiation each 2 steps but it is still the same. What is wierd to me is that when I check radiation coarsening information just in partitioning it shows me that I get only 1 radiation element from 2.4M elements. Is it possible to force it somehow or just to have "power" to affect this coarsening? I know that there are few other options but I don't know which one could be appropriate.

ghorrocks June 27, 2013 18:41

There are a lot of radiation options and to be honest I do not know them all well. But I can say the image you showed was typical of poor radiation resolution and/or accuracy, so it looks like adjusting the radiation defaults is required in your case. So I would adjust all relevant looking radiation controls and look at their impact.


All times are GMT -4. The time now is 16:50.