Optimization Study of Heat Dispersion in Additively Manufactured Heat SinksJohn Patten, Holden Hyer, Ed Dein, Dr. Yongho Sohn

University of Central Florida, Orlando, FL

Abstract

References

Experimentation Discussion

Acknowledgements

Manufacturing

ASME1

Heat sinks are commonly used today; virtually every largecomputational device incorporates at least one heat sink in itsdesign. This is because computer processors experiencelosses in performance as they accumulate heat. The additionof a heat sink can disperse this heat and subsequentlyincrease performance.

The purpose of this project is to determine if the design ofheat sinks can be improved upon. Specifically, the way theprotrusions, or “fins,” are arranged on the surface of the heatsinks will be tested in order to determine which design mosteffectively disperses heat. Two designs will be tested: atraditional rectangular fin design and one with more optimizedtopology. Due to the complex geometry of the second design,Selective Laser Melting (SLM) additive manufacturing is usedto create both prototypes, and the heat dispersion isevaluated through experimentation. The results of thisexperiment will conclude which design is most effective.

Results

Simulation

Conclusion

The design of the optimized heatsinkwas based on a 2015 study by theAmerican Society of MechanicalEngineers1 (ASME). The purpose of theunconventional topology was to allowfor increased air flow from the top of theheat sink. The construction of such apart would be cumbersome orimpossible through normal machiningmethods due to the complex geometry;therefore SLM additive manufacturingwas employed. Solidworks was used tomodel both parts. It is important to notethat the surface areas and volumes ofthe respective parts were made to beessentially the same, in order to testonly the geometry of the designs.

The results would indicate that the optimized heat sink designis more effective at dispersing heat than traditional designs.The optimized heat sink remained significantly hotterthroughout the trials when compared to the normal heat sink,suggesting that it is able to conduct heat more effectively.Additionally, the temperatures of the base and the fins of theoptimized design were significantly closer together than thoseof the normal design. This2 would suggest that the optimizedheat sink is more efficiently able to disperse heat from its baseto its fins, which allows for more effective dispersal of heatfrom, for example, a processor. Future research could bedone to better quantify how much more effective this type oftopology can be, and whether or not it would be cost effectiveto produce on a large scale.

Six trials were conducted for each heat sink, three with a fanand three without. The averages of these trials are plottedbelow. These results show that the optimized heat sinkreached a higher temperature in both the base and the fins; atthe peak temperatures, the optimized heat sink wasapproximately 10°C hotter than the normal heat sink.Additionally, the difference in temperature between the baseand the fins of each heat sink is plotted with respect to time. Itis shown that the temperature of the base and the fins of theoptimized heat sink were much closer together than thenormal heat sink’s temperatures.

[1] Dede, Ercan M., Shailesh N. Joshi, and Feng Zhou."Topology Optimization, Additive Layer Manufacturing, andExperimental Testing of an Air-Cooled Heat Sink." Journal ofMechanical Design 137 (November 2015): 1-9.[2] Wong, M., I. Owen, C.J. Sutcliffe, and A. Puri."International Journal of Heat and Mass Transfer." ConvectiveHeat Transfer and Pressure Losses across Novel Heat SinksFabricated by Selective Laser Melting 52, no. 1-2 (January 15,2009): 281-88.

Special thanks to Duke Energy, for providing financial supportfor the duration of this project. Additional thanks to HoldenHyer, for providing assistance in the additive manufacturingprocess, Michael Pickering for providing technical support andmaterials, and Ed Dein and Dr. Yongho Sohn, for providingmuch-needed guidance throughout the entire process.

Heat transfer simulation was conductedusing Solidworks as preliminary testing.Under conditions similar to the practicalexperiment, which consisted of a largeheating power (92 W) applied to thebase and minor convection heating (20W) from the environment, Solidworksgenerated the figures below. With aquick glance, it can be seen that theoptimized design has a more uniformand hotter profile than the traditionaldesign. This would predict better heattransfer when heat is applied directly tothe bottom of the base.

In order to determine the efficiency ofthe designs, a Peltier cooler was appliedto the base of the heat sink with thermalpaste in order to increase heat transfer.The voltage through the cooler was keptconstant at 5 V. Heating was applied for10 minutes from room temperature(24°C), then removed. Thermocoupleswere attached to the center of the baseas well as the top of the most centralfins (shown in red in the figures below)in order to analyze the difference intemperature between the two.

10

20

30

40

50

60

70

80

00:0

0

01:0

0

02:0

0

03:0

0

04:0

0

05:0

0

06:0

0

07:0

0

08:0

0

09:0

0

10:0

0

11:0

0

12:0

0

13:0

0

Average Temperatures With Fan

Normal Base T (°C) Normal Fin T (°C) Optimized Base T (°C) Optimized Fin T (°C)

10

20

30

40

50

60

70

80

00:0

001

:00

02:0

003

:00

04:0

005

:00

06:0

007

:00

08:0

009

:00

10:0

011

:00

12:0

013

:00

14:0

015

:00

Average Temperatures Without Fan

Normal Base T (°C) Normal Fin T (°C) Optimized Base T (°C) Optimized Fin T (°C)

0

1

2

3

4

5

6

7

00:0

0

01:0

0

02:0

0

03:0

0

04:0

0

05:0

0

06:0

0

07:0

0

08:0

0

09:0

0

10:0

0

11:0

0

12:0

0

13:0

0

Temperature Difference Without Fan

Normal Heat Sink (°C) Optimized Heat Sink (°C)

-2

0

2

4

6

8

10

12

00:0

0

01:0

0

02:0

0

03:0

0

04:0

0

05:0

0

06:0

0

07:0

0

08:0

0

09:0

0

10:0

0

11:0

0

12:0

0

13:0

0

Temperature Difference With Fan

Normal Heat Sink (°C) Optimized Heat Sink (°C)