Learning Forum

[Rémi Gérard-Marzan]

Hello everyone,

I hope this question is not a duplicate, I haven't been able to find a similar question so far.

In Ansys Mechanical, I am performing a steady-state thermal analysis involving thermal radiation. I have added radiation probes to witness the output flux. One of the surfaces to which I have ascribed radiation thus reads :

• Outgoing Net Radiation : 4.8438 W
• Emitted Radiation : 44.001 W
• Reflected Radiation : 4.3481 W
• Incident Radiation : 43.505 W

What I find baffling is the following.

For Emitted Radiation, that specific surface is ascribed a fixed surface temperature of 314.69K, an emissivity of 0.9 and a surface of 0.10474m². A very simple Excel-made computation shows me that I would expect an emitted radiation amounting to :

5.67e-8 0.9  0.10474 * POWER(314.69;4) = 52.41 W. 

This differs significantly from the 44,001W indicated by the probe, and I have no clue why.

Have a nice day all,

[dlooman]

I agree, a small 315K surface would not emit thousands of Watts!  The incicdent radition seems very high also.  Could there be a units issue?  Could you describe the model?  Are you using the "Perfect" setting in Mechanical?  

[Rémi Gérard-Marzan]

Hello Dave,

Thank you for your answer. I am afraid there has been a typo in my post. I have mixed and matched French and English number separators. The opening post should read:

• Outgoing Net Radiation 4.8438 W
• Emitted Radiation 44.001W
• Reflected Radiation 4.3481 W
• Incident Radiation 43.505 W

Seen that way, you can understand that my surprise comes from the fact the “Emitted Radiation” is 8 W lower than the expected emitted radiation that I got from a hand-made computation. It might have to do with the fact that I am not clear on what “Emitted Radiation” means in the context of a radiation probe. I assumed it was the black-body emission of the considered surface, and the Ansys Mechanical User Guide does not give any more clues. However, it might actually mean something else?

Best regards,

Rémi

[dlooman]

You are correct that emitted is the black body radiation.  What version of Ansys are you using and are you using "Perfect?"

[Rémi Gérard-Marzan]

This is Ansys 19.2 and indeed, the enclosure type I am using in this context is "Perfect". Should I expect a different outcome by selecting "Open"? As far as I can see, there would be no physics-based reason for that to happen, but I am willing to try.

[dlooman]

"Perfect" fudges the view factors so it's an additional unknown.  Also, at 19.2, it was possible for a face to "see" through the back of other faces.  This was corrected in 2020R2.  This non-physical behavior made it possible for a view factor to be greater than 1.0.  To minimize the effect of this error you should include all faces of bodies in the radiation definition (so that a face can't see inside them.)  

[Rémi Gérard-Marzan]

Alright, I will try your suggestions. However, if I understand correctly, these two parameters should have no bearing on the black-body radiation of the considered surface. The "Emitted Radiation" values does not rely on view factors. 

I will tell you what happens once I have run the calculation according to your advice.

[Rémi Gérard-Marzan]

 

 

Hello again Dave,

Here is some feedback from your suggestions.

I have modelled radiation in a very simple cylinder : one of the bottom (inner) faces models the Earth, at 15°C (so 288,15°K), while all the other inner faces of that cylinder model outer space, at 4°K. I have run a steady-state thermal analysis by applying radiation to all these surfaces, all with an emissivity of 1, and alternatively with an Open configuration or a Perfect configuration. The image below will show you an exploded view of that setup, with the radiating surfaces for “espace” (“outer space” in French) being highlighted. The disk at the foreground is the Earth.

Now, the surface of Earth in that setup is 76.905 m². Given an emissivity of 1 and a temperature of 288.15K, I would expect the “Emitted Radiation” value to be equal to 30,061.6W. However, both in the “Open” and in the “Perfect” enclosures, I can read in the thermal probe that the “Emitted Radiation” value is actually equal to 29,317.0W. So, similarly to the case I mentioned in the opening post, there is a significant difference here. 

Even more surprisingly, when reviewing the thermal probe for Space, since this is a perfect black-body, I would expect reflected radiationto be equal to 0. Maybe, due to numerical errors, I could understand an actually non-zero value, provided it is very small. However, I can read that the computed “Reflected Radiation” to be equal to 45.426W. This is a surprisingly high amount.

So that’s it. Do you have any clues as to why all this happens?

Best regards and thank you for your help so far.

Rémi

 

 

[Rémi Gérard-Marzan]

Hello Dave,

I have found the problem.

-The fact the "Emitted Radiation" did not line up with the expected blackbody radiation from theory was due to a mesh that was too coarse. Refining the mesh led to an "Emitted Radiation" much more in line with what theory tells.

-The strange nonzero "Reflected Radiation" was greatly diminished once I refined the radiosity solver tolerance. Going from 0,1 to 0,001 W/m² allowed for the "Reflected Radiation" value to decrease by two orders of magnitude.

All is good as far as I am concerned.

[dlooman]

Good to hear, but a little surprising that mesh quality affected the emitted radiation.  As you said, the view factor should not affect the emitted radiation.  Maybe it does though, numerically. 

[Rémi Gérard-Marzan]

What I suspect happened is that mesh quality does not affect the way Ansys computes the blackbody radiative flux density in W/m². However, once Ansys retrieves the actual area of the considered surface to compute its "Emitted Radiation" output, if the mesh is too coarse then Ansys retrieves an incorrect area value, and thus that means the radiative flux in W ends up being muddled by that error.