Simulation Mechanical Forums (Read-Only)
Welcome to Autodesk’s Simulation Mechanical Forums. Share your knowledge, ask questions, and explore popular Simulation Mechanical topics.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

2D Heat Transfer Issues

6 REPLIES 6
Reply
Message 1 of 7
jrager
474 Views, 6 Replies

2D Heat Transfer Issues

First of all:   how do I fix a non-conformal mesh (the meshes of adjacent surfaces in a 2D model of mine do not join together and this is throwing off the temperature contours when I run the simulation)?  I tried changing the element type to triangular and mixed but this did not help.  I also tried meshing the adjacent parts all at once, then one at a time but neither approach worked.

 

Related to a different 2D problem:  suppose I have an idealized cubic heater in the middle of a metal box surrounded by air.  I can simplify this in 2D by simulating a square heater in the middle of a "thick" rectangle (surrounded by a layer of air).  My physical heater puts out 150W of heating power, so I add surface heat loads to my square heater so that the total output over its 4 faces sums to 150W.  I run the simulation with conduction through air, convection, and a radiation enclosure (as well as convection and radiation loads from the box's exterior to the ambient) but the values of temperature and heat flux that I obtain are way to high.  By changing the convection and radiation multipliers in the Analysis Parameters dialogue box, I can get the model to give me the right values of flux and temperature but this seems a little fishy to me.  Are these multipliers merely "fudge factors" that have to be empirical fit to the problem, or do they (as I suspect) represent something else?  Do they weight the respective importance of convection and radiation in the problem?  Is it okay to change them from their default values?

6 REPLIES 6
Message 2 of 7
Joey.X
in reply to: jrager

For conformal mesh, before doing meshing, make sure enforcing "surface knitting" at CAD import which makes one2one mapped surfaces between adjacent parts, and the part contact should be bonded(default). 

For your cube heater problem, note that the 2D model (2D planar) assumes infinite depth, and then you can image the identical 3D models. So a cubic heater in large cube does not have an identical 2D (planar) model and the results can't be compared.

Jianhui Xie, Ph.D
Principal Engineer
MFG-Digital Simulation
Message 3 of 7
jrager
in reply to: jrager

Ok, but I am not importing a CAD model.  I constructed my 2D geometry within Multiphysics using its sketch features and then just specified its element definition as Axisymmetric.  How do I enforce surface knitting?  When you say bonded do you mean fine bonded to course mesh or course bonded to fine?  The default disables bonded contact.

 

For the square heater:  if the 2D planar regime assumes infinite depth then what does specifying the thickness under the "Element Definition" option in the navigation tree do?

Message 4 of 7
Joey.X
in reply to: jrager

For meshing issue, better to see your model, please post if possible.

The infinite depth assumption is always valid for 2D planar models, the thickness option in 2D element definition is postprocessor only, such as (a) calculate some physical area based quantities in depth direction (b) visualization. 

Jianhui Xie, Ph.D
Principal Engineer
MFG-Digital Simulation
Message 5 of 7
John_Holtz
in reply to: jrager

Hi jrager,

 

It sounds like you need to change the wireframe to get the proper connectivity between parts for the 2D mesh. Please see "2D Mesh Generation" , especially Figure 2 and description for the requirements.

 

For the second problem, a 2D model of a cubic heater might be too much of an approximation. For example, the center of the heater is affected by the heat being drawn off of all 6 sides of the heater. In the 2D approximation, you are only drawing heat from 2 sides (presumably), so the center temperature would calculate to be hotter than reality.

 

What is the purpose of the analysis?

Calculate the temperature of the heater? This could be done with a model setup as follows:

  1. model the heater as a cube.
  2. Add a "Setup > Thermal Loads > Internal Heat Generation" and enter a value equal to 150 W / volume of the heater.
  3. Apply the "Setup > Thermal Loads > Convection" to the 6 sides of the heater. This obviously requires some estimation using emperical formulas or test data.
  4. Apply the "Setup > Thermal Loads > Radiation" to the 6 sides of the heater. Since this is radiation to the ambient, it ignores the re-radiating effects of the enclosure.

If the purpose of the analysis is to calculate the temperature of the enclosure, then I think it will be difficult to use a regular heat transfer analysis to accurately simulate the buoyancy effects that occur in the air inside the enclosure. (You might be able to use orthotropic conduction to simulate the transport of heat that occurs through natural convection, but it would be an educated guess at best.)  The best solution would be to perform a "Multiphysics > Transient Coupled Fluid Flow and Thermal" analysis. The model would include the air, the heater (optional but convenient way to apply the 150 W), and the enclosure (optional).  One thing you should check first is whether body-to-body radiation can be used in a coupled analysis. I do not remember.

 

 

 



John Holtz, P.E.

Global Product Support
Autodesk, Inc.


If not provided already, be sure to indicate the version of Inventor Nastran you are using!

"The knowledge you seek is at knowledge.autodesk.com" - Confucius 😉
Message 6 of 7
jrager
in reply to: John_Holtz

Ok, thanks!  But would such a model be able to give me the correct values of heat flux inside the enclosure?  This is my primary concern.

Message 7 of 7
John_Holtz
in reply to: jrager

Hi,

 

Before I forget, I just noticed a mistake in my previous reply. I said that in the real life situation, the heater is being cooled from 6 sides, and in the 2D model it is cooled on 2 sides. Two sides was a mistake; I meant to type that the 2D model can only provide cooling on 4 sides of the heater, so the heat flow in the X direction is ignored. That assumption may or may not be accurate enough.

 

To answer your new question, the coupled analysis should be able to give you accurate heat fluxes. The difficulty (as you may have read in other posts) is getting the analysis converge. Fluid flow is highly nonlinear which requires iterations to solve which leads to potential convergence issues. If the enclosure is sealed, the motion of the air is due solely to temperature differences/density differences/buoyancy effects. These are the most difficult simulations to solve. On the other hand, if there is a fan circulating/blowing the air through the encloser, these are easier to solve. If the airflow is independent of the temperature, because of fans moving the air, the solution becomes one level easier to solve. The page "Multiphyics" in the wiki Help touches on some of the advantages of each of these analysis types.

 



John Holtz, P.E.

Global Product Support
Autodesk, Inc.


If not provided already, be sure to indicate the version of Inventor Nastran you are using!

"The knowledge you seek is at knowledge.autodesk.com" - Confucius 😉

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums  

Autodesk Design & Make Report