AnsweredAssumed Answered

Heat Transfer between two solids separated by an air filled cavity

Question asked by Dan Radulescu on Apr 27, 2016
Latest reply on May 4, 2016 by Amit Katz

Like • Show 0 Likes0
Comment • 10

Hello all,

I am trying to do heat transfer simulation that involves a hot plate(surface/body) and a cold plate, separated by an air filled cavity.

The setup looks like in the below picture:

I am running a simulation considering the following features:

type : internal

* heat conduction in solids

* time dependent 120 seconds with 1 second increments

* gravity

The fluid of choice here is air.

Bodies are by default Stainless Steel 321.

Default wall condition is heat transfer coefficient of about 35 W/m^2/K and temperature of external fluid 293.2 K and roughness 0 micrometer.

First, I set up the fluid sub-domain, the air inside the cavity.

I then set up body 1 as a volume heat generation and set it to 20 000 W.

I assume that the walls have their default condition but I have also set them up with heat transfer coefficient individually towards the same result.

The results I am getting are the following:

The heat transfers to the walls in the top region as well as the air inside which heats up. HOWEVER.. the air does not seem to transfer any heat anywhere. It seems that the heat is only from the heat generator towards the bodies/fluids it comes DIRECTLY in contact with.

What I need is to see how the heat transfers from heat generator -> air -> cold plate.

Any suggestions on how to do that? Is there something I am assuming or doing wrong?

Thank you for your time and attention. Looking forward to hearing your advice and learn from it.

Best regards,

Dan

No one else has this question

Outcomes

Amit Katz Apr 27, 2016 8:52 AM

First of all, I would not use a default heat transfer flux on all the walls, unless you know for a fact (derived experimentally) that is the case in this scenario. You chose a very specific number as well, where does that number come from?

As for the air, use the geometry check tool and click "show fluid" to make sure you have air where you think you have air. You mention a subdomain, but this is an internal flow problem with only 1 domain, so there might be some confusion here.

What boundary condition do you want to apply to the side walls of this cavity? Are they adiabatic or constant temperature? How do they interact with the energy transfer between the hot/cold plates?

Actions

Dan Radulescu @ Amit Katz on Apr 27, 2016 9:57 AM

Amit,

Thank you for your reply.
Actually, I was a bit unsure what boundary condition to use on the walls.

I was looking over on this site Overall Heat Transfer Coefficient and chose an approximate value from there, just for a test run to see if something actually happens. I am willing to do proper calculations for it next time.

I understand your point about the subdomain confusion. However, the cavity is water tight an when I click on subdomain and select AIR it shows me the filled space as a fluid region.

The walls should not be adiabatic and they are connected in the top side with the hot plate and in the bottom with the cold one.

Let's just assume that air is trapped in that cavity.

Heat comes from the pink part, transfers through the captive air and warms the blue part.

I am interested to see this transfer of heat, from the pink part through the fluid region and on to the blue part.

Looking forward to your suggestions.

Thank you.

Actions

Amit Katz Apr 27, 2016 10:16 AM

Can you upload the model?

As far as the walls, you still need to think about the boundary condition on the outer edges. The question is, do you want heat to be able to escape through those walls, or will it all be transferred to the cold plate? If it's the latter, make sure you create an adiabatic condition on the outer surfaces of the walls. This will trap all the heat inside the calculation domain and still enable heat conduction through the walls as well as convection via the air.

Actions

Dan Radulescu @ Amit Katz on May 3, 2016 7:59 AM

Amit,

I did do that, however, I still cannot get the heat to transfer from the air to the cold plate.

Here is the model. If it is not too much for you I would really like to see how you set up the simulation.

Thank you!

Google drive link - Heat Transfer test.zip - Google Drive

Actions

Amit Katz May 3, 2016 10:31 AM

Dan,

Your files do not include a Flow SImulation project. Try to "pack and go" again and make sure you have the setup in your files.

Actions

Dan Radulescu @ Amit Katz on May 4, 2016 8:08 AM

Here it is Amit. Sorry for the delay.

Heat Transfer test - project.zip - Google Drive

Actions

Amit Katz May 4, 2016 9:46 AM

OK, I see you used a constant temperature of 1000K for your heat source and ideal wall conditions for your cavity walls. Why the ideal walls though? I'm not sure what you were going for there.

You had the "exclude cavities without flow conditions" option turned on, that's why there wasn't any fluid in the cavity.

The real issue here is that you don't have a heat sink. Without anywhere for the heat to go, the entire solid and fluid domain will simply heat up to 1000K and stay there.

Here is what I did instead:

Added volumetric heat generation sources (equal and opposite on the hot/cold plates)

Suppressed the wall conditions. It seems like there's an assumed adiabatic condition on the boundary of the domain, so that's fine.

Here's a snapshot of some quick 'n dirty results, proving that the boundary conditions work:

You'll note I used 20W instead of 20,000W like you have in the original posting. I'm assuming that was a typo, since that much heat will completely incinerate this steel assembly.

By the way, this simulation would be a good candidate for symmetric domain conditions. You could simulate just 1/4 of the domain with symmetry in the X and Z axes and have a much finer mesh.

Attachments

Heat Transfer test.zip275.4 KB

Actions

Dan Radulescu May 4, 2016 10:06 AM

Thank you for the tips Amit! is there any way you could upload the project with your modifications please?

Actions

Dan Radulescu May 4, 2016 10:06 AM

I am not seeing any link for download

Actions

Amit Katz @ Dan Radulescu on May 4, 2016 1:33 PM

I can see it, it's at the bottom of my post. I'll attach it to this one as well. You may need to upgrade to the latest version to open this file, I'm using SW2016 SP3.0

Attachments

Heat Transfer test.zip275.4 KB

Actions

10 Replies

Amit Katz Apr 27, 2016 8:52 AM
Mark Correct
Correct Answer
First of all, I would not use a default heat transfer flux on all the walls, unless you know for a fact (derived experimentally) that is the case in this scenario. You chose a very specific number as well, where does that number come from?

As for the air, use the geometry check tool and click "show fluid" to make sure you have air where you think you have air. You mention a subdomain, but this is an internal flow problem with only 1 domain, so there might be some confusion here.

What boundary condition do you want to apply to the side walls of this cavity? Are they adiabatic or constant temperature? How do they interact with the energy transfer between the hot/cold plates?
Like • Show 0 Likes0
Actions
- Dan Radulescu @ Amit Katz on Apr 27, 2016 9:57 AM
  Mark Correct
  Correct Answer
  Amit,
  
  Thank you for your reply.
  Actually, I was a bit unsure what boundary condition to use on the walls.
  I was looking over on this site Overall Heat Transfer Coefficient and chose an approximate value from there, just for a test run to see if something actually happens. I am willing to do proper calculations for it next time.
  
  I understand your point about the subdomain confusion. However, the cavity is water tight an when I click on subdomain and select AIR it shows me the filled space as a fluid region.
  
  The walls should not be adiabatic and they are connected in the top side with the hot plate and in the bottom with the cold one.
  
  Let's just assume that air is trapped in that cavity.
  Heat comes from the pink part, transfers through the captive air and warms the blue part.
  I am interested to see this transfer of heat, from the pink part through the fluid region and on to the blue part.
  
  Looking forward to your suggestions.
  
  Thank you.
  Like • Show 0 Likes0
  Actions
Amit Katz Apr 27, 2016 10:16 AM
Mark Correct
Correct Answer
Can you upload the model?
As far as the walls, you still need to think about the boundary condition on the outer edges. The question is, do you want heat to be able to escape through those walls, or will it all be transferred to the cold plate? If it's the latter, make sure you create an adiabatic condition on the outer surfaces of the walls. This will trap all the heat inside the calculation domain and still enable heat conduction through the walls as well as convection via the air.
Like • Show 0 Likes0
Actions
- Dan Radulescu @ Amit Katz on May 3, 2016 7:59 AM
  Mark Correct
  Correct Answer
  Amit,
  I did do that, however, I still cannot get the heat to transfer from the air to the cold plate.
  Here is the model. If it is not too much for you I would really like to see how you set up the simulation.
  Thank you!
  
  Google drive link - Heat Transfer test.zip - Google Drive
  Like • Show 0 Likes0
  Actions
Amit Katz May 3, 2016 10:31 AM
Mark Correct
Correct Answer
Dan,
Your files do not include a Flow SImulation project. Try to "pack and go" again and make sure you have the setup in your files.
Like • Show 0 Likes0
Actions
- Dan Radulescu @ Amit Katz on May 4, 2016 8:08 AM
  Mark Correct
  Correct Answer
  Here it is Amit. Sorry for the delay.
  Heat Transfer test - project.zip - Google Drive
  Like • Show 0 Likes0
  Actions
Amit Katz May 4, 2016 9:46 AM
Mark Correct
Correct Answer
OK, I see you used a constant temperature of 1000K for your heat source and ideal wall conditions for your cavity walls. Why the ideal walls though? I'm not sure what you were going for there.
You had the "exclude cavities without flow conditions" option turned on, that's why there wasn't any fluid in the cavity.
The real issue here is that you don't have a heat sink. Without anywhere for the heat to go, the entire solid and fluid domain will simply heat up to 1000K and stay there.

Here is what I did instead:
Added volumetric heat generation sources (equal and opposite on the hot/cold plates)
Suppressed the wall conditions. It seems like there's an assumed adiabatic condition on the boundary of the domain, so that's fine.

Here's a snapshot of some quick 'n dirty results, proving that the boundary conditions work:

You'll note I used 20W instead of 20,000W like you have in the original posting. I'm assuming that was a typo, since that much heat will completely incinerate this steel assembly.
By the way, this simulation would be a good candidate for symmetric domain conditions. You could simulate just 1/4 of the domain with symmetry in the X and Z axes and have a much finer mesh.
Attachments
- Heat Transfer test.zip275.4 KB
Like • Show 1 Like1
Actions
Dan Radulescu May 4, 2016 10:06 AM
Mark Correct
Correct Answer
Thank you for the tips Amit! is there any way you could upload the project with your modifications please?
Like • Show 0 Likes0
Actions
Dan Radulescu May 4, 2016 10:06 AM
Mark Correct
Correct Answer
I am not seeing any link for download
Like • Show 0 Likes0
Actions
- Amit Katz @ Dan Radulescu on May 4, 2016 1:33 PM
  Mark Correct
  Correct Answer
  I can see it, it's at the bottom of my post. I'll attach it to this one as well. You may need to upgrade to the latest version to open this file, I'm using SW2016 SP3.0
  Attachments
  - Heat Transfer test.zip275.4 KB
  Like • Show 0 Likes0
  Actions

Heat Transfer between two solids separated by an air filled cavity

Outcomes

Attachments

Attachments

Related Content

Recommended Content