Learning Forum

[Eric Steimle]

I am trying to do my first simulation in icepak of what I think is a simple board. It's a single LED on top of a small circuit board. I imported my model, set my materials, and then set my surface materials. I wasn't sure what to set my wall's too though so for the first simulation I just chose one at a in the image below.  Then I assigned a small 0.001W thermal source to the LED (has red corners in the image below. This gave me reasonable temperature rises to about 30C.  So I thought, okay to get more accuracy I should asign walls to the other materials that will radiate into the air.  I assigned a wall to the little dome of the LED and re-ran the simulation and this time it told me my temperature rise was 274C.  Obviously I'm not doing something right, I thought you had to define "walls" wherever your model is going to interact with the air and radiate heat? Do I have to do something special beause this is natural convection?  I did turn radiation on in the simulation setup.  Sorry if these are basic questions, this is my first attempt at using Icepak for anything.  Thanks!

 

[Rabindra Paul]

Hi Eric, 

Thanks for asking question through AIS forum. As you mentioned, you are new to Icepak, I wanted to know whether you went through introductory training. Going through introductory notes and workshop should be of help.

In any case, the wall boundary condition needs to be at the boundary of the domain. I believe you have used the suggested domain size for the natural convection. Most of the cases, the domain sides should be free openings for natural convection setup. Hoewever, if there are physical wall for the problem then you can use wall bc and define the appropriate bc. You should not use wall boundary type within the domain. The heat trasnfer from the objects to air will be handled by the software within the domain.  You can access training material from Ansys Learning Hub. Please let us know if this helps or if you have any comments.

Regards,

 

Rabindra

[Eric Steimle]

Hah! Okay I understand I think. I guess I was cooking it in a sealed box before. Now I assigned all faces of the "air" box around the LED to be free openings and the two simulations I ran are pretty close to a dT = Q/(h*A) that I did in excel.  Thanks!

[Rabindra Paul]

 

Thanks for the update and good to know you are getting expected results. 

 

Rabindra