Internal heat genearation

[jkmalagi]

All

I have a PCB with a component rejecting ~20W of thermal power.  The component  itself is ~1 inch Wide  X 1 inch long X0.1 inch thick which I am trying to cool.

How does one apply "internal heat generation" for heat transfer studies? Since the unit for internal heat generation is Heat rate/volume, I am confused as to how do I apply this to my processor. Dividing 20W by the volume of the component gives a strange number that I am not sure is right.

Can someone clarify please?

thanks

Jay

ADESK Algor 2010

[AstroJohnPE]

Hi Jay,

Your formula is correct: heat generated in part X / volume of part X.

[jkmalagi]

Hello John,

Reading thru the help in detail, I noticed that we are required to enter heat generation per unit volume & the software will do the final calculation.

Q= Heat generated/unit volume * multiplier factor * volume of the part 

The volume is calculated by the s/w, while user needs to input items 1 & 2. What I ended up doing was to divide the total heat generated by the no. of elements created in the component & entering that info (eg: 20W divided by 200 elements will 0.1, which I entered into heat generation tab). I guess this approach is more intutive...

Thanks for the feedback.

Jay

[AstroJohnPE]

I am not entirely clear what values you entered for the heat generation value (J/s/m^3, for example) and the heat generation multiplier (normally = 1). But it is clear to me that unless the 200-element part has a volume of 200 units^3, then you are entering the wrong value. (Each element would need to be 1x1x1 cube.)

Keep in mind that the volume of each part in the model can be calculated with the command "Analysis > Analysis > Weight and Center-of-Gravity".

[jkmalagi · Accepted solution]

Hello John,
I am confused.

The part volume is 8.2E-7 cubic metre.
The part heat generation is 20W (or 20Joules/Sec)
So the internal heat generation is : 20W/ 8.2E-7 = 108932461 W/m^3!!

This is way too high... what I am doing wrong?

Thanks for your time.
Jay

[AstroJohnPE · Accepted solution]

What you are doing wrong is thinking that it is wrong just because you have a "large number".

But 108932461 W/m^3 times 8.2E-7 cubic metre gives 20 W -- the amount of heat that you want to be applied (except that 20 W/8.2E-7 m^3 = 24E6 W/m^3, not the 10.89E7 W/m^3 that you typed! 24E6 W/m^3 times 8.2E-7 m^3 = 20 W.)

[jkmalagi]

Hello John,

Now it makes sense. I mis-typed the values from another heat geneating component.

And I agree- 'too big' is just relative & the values actually worked just fine in the analysis

Thanks for your feedback & time

Jay