THE EFFECT OF BLOWING DIRECTION ON HEAT SINK PERFORMANCE BY THERMAL IMAGING

Heat sinks (HSs) are designed for the mechanical, electrical and electronic components that generate heat in considerable amount. For this purpose, an aluminum conical pin fin heat sink is designed. Aluminum conical pin-fins geometry has been experimentally investigated for the blowing direction (pushing or pulling) which is the energy efficient option for the heat sink. The heat sink was tested at the same fan power for pushing and pulling conditions for 25, 50, 75 and 100 W resistance heater power. Designed aluminum conical pin fin heat sink can be easily used in heat sweeping processes. It has found that pushing configuration of the fan is more efficient for this design.

HSs under natural convection conditions and stated that perforated pin fin HS and sideward facing orientation was better than that of solid pins and upward facing orientation. Sahin and Demir [7] investigated a flat surface equipped with circular perforated pin fin HS in terms of heat transfer enhancement and pressure drop. The results indicated that circular pin fins increase the heat transfer between 1.4 and 2.6 depending on clearance ratio and fin spacing ratio. Deshmukh and Warkhedkar [8] investigated the comparative thermal performances of elliptical pin-finned and circular pin fin HSs. They found that the elliptical inline is 24% more effective (eff.) than circular inline, elliptical staggered is 50% more eff. than circular staggered and the elliptical staggered is 63% more eff. than elliptical inline. Yüncü and Anbar evaluated the plate-fin HSs and they obtained the optimal fin spacing decreases as fin height increases [9]. The failure rate of electronic components decreases with the decrease in temperature of the component. Therefore, a conic pin-finned HS was designed to reduce the temperature of a component connected and the effect of fan orientation on the performance of a conical pin-finned HS was investigated experimentally. Zhou and Rau [10] studied a heat sink to come up with a framework to correlate hand calculations and numerical simulations with experimentally obtained results. They investigated many parameters such as power levels, orientation (flat-vertical-horizontal), gravity, and altitude under natural convection conditions. Belhadj et al. [11] studied numerically the forced convective flow in micro channels with periodic expansion-constriction cross-section under laminar conditions. They focused the effect of cross-sections on the thermal behavior of water flow used for cooling systems. They stated that the Nu number is increasing (36% of maximum) with the rise of Reynolds number for all micro channels modifications and convective heat transfer rate was improved significantly via the periodic expansion-cross section. They concluded that micro channels with triangular cavities are more efficient than cylindrical grooves. Zunaid et al. [12] studied numerically the conjugate heat transfer in a heatsink with semi cylindrical geometry under the single phase flow condition. Reynolds number from 200 to 1000 with constant heat flux of 106 W/m 2 selected for the heat sink in the simulation. The heat transfer in microchannel heat sink with semi cylindrical geometry was better than that of heat sink with rectangular one.

MATERIALS AND METHODS
Aluminum conical pin fin geometry was selected for heat transfer enhancement also relatively lowpressure drop. In this study, a new design conical pin fin HS designed and manufactured via CNC machining. The HS made of aluminum. The dimension of HS base plate is 80x80x5 mm. The base plate has 39 conical pin fins. Each conical fin has 6 mm base, 1 mm tip diameter, and 10 mm fin height.
The experimental setup consisted of the air duct, fan, heat sink, plate type resistance heater, insulation, fire brick and test instruments. The air duct is made of transparent Plexiglas and measured temperature in the air duct by PT1000 type sensor (Comet SN234 with the precision of ±0.15°C). The system can be operated in pushing and pulling modes. The air flow direction is changed by reversing the fan. The test setup is from bottom to top; fire brick, insulation, resistance heater, heat sink, fan and air duct. The fire brick is opened to a bed for resistance heater and thermally insulated in accordance with the heat sink scale. The lower horizontal and side surfaces of the heater are insulated by a ceramic wool blanket in 5 mm thickness. The base of the HS was heated by an electrical resistance plate heater with 190 W (DC 24 V). The heater is 80x80x3 mm and mounted on the HS with M4 screw, and powered by a DC power supply (Sayntech 23003 power supply). A usual 12 VDC fan is used in 80x80x25 mm for providing air flow and worked by a power supply (TT Technic RXN-303D power supply). In order to measure the air outlet temperature in the push mode and the air inlet temperature in the pull mode, T type thermocouple sensor is mounted on each side between the heat sink and the fan also a thermocouple temperature sensor (Teflon coated, Elimko T type with precision of ±0.5°C) embedded within the center of the base plate for to measure temperature of the heat source. Hot wire type (Delta ohm HD403TS1 with precision of ±0.2 m/s + 3% f.s.) and hand type anemometer (Kestrel 3000 multifunction anemometer with a resolution of 0.1 m/s) were used to measure the air velocity in the air duct. A thermal camera (Flir SC325 with the precision of ±2%) was used to measure the surface temperature of the heat sink and analyzed by Researcher 2.10 software. Monitoring and recording of test data were done with Comet MS6D universal data logger. The test recording time interval was set to five seconds. Two electrical multimeters (Brymen BM807 multimeter) were used to measure the electrical values of the heater and the fan.
The HS geometry sketch is shown in Fig. 1 and photographic representation of the heat sink and, test bed in Fig. 2. The thermo-physical properties aluminum and air are given in Table 1. The experimental set up was performed. The test apparatus was placed in the flow channel.

THEORETICAL MODEL
The heat transfer modeling of a heat sink considered in this study is schematically showed in Fig. 3. Assuming that the air flow and ambient temperature is steady state also the bottom and the side surface of the heater is perfectly insulated.
The multiplication of total surface area of a cone and fin number equals summed surface area of n fins. Volume of cone is; where, (L) length of the fin and (D) diameter of the fin at the base. Total amount of heat supplied by heater is calculated as where, (Cosθ) is power factor 1 (no inductive and capacitive load), (V) voltage supplied and (I) current. The rate The thermal contact resistance can be defined as where, (Ac) the contact area and (hc) the contact conductance. The thermal resistance for conduction can be expressed as: The thermal resistance for convection can be expressed as:

RESULTS AND DISCUSSION
The stages of the experiments consisted of the first five minutes of non-heating operation, the heating of 25, 50, 75 and 100 W each for ten minutes and finally the observing of the cooling of the HS for 15 minutes after shutting the heater.
The variation of the heating resistance plate surface temperature with time according to the pushing and pulling of the fan conditions is given in Fig. 4. The resistance heater surface temperatures (Tres) under the pushing and pulling fan states are very similar to each other and an average temperature difference of 3.5°C is measured. The surface temperature of the pin fins of the HS was measured by the thermal camera and the averages were taken. Table 2 shows the surface temperature average values for the push and pulling mode. Change of surface temperature versus heater power is given in Fig. 5. As the heater power increases, the fin temperatures increase during pulling and pushing. The transferred heat energy was higher because the air flow rate in the pushing mode was about twice as high. As a result, the fin surface temperature was lower than expected. The average surface temperatures of the HS in the pushing mode were lower by 2.0-3.9-5.1-6.9°C, respectively for 25-50-75-100 W resistance heat power.  The blowing direction of the air in the test bed is changed by reversing the fan. Air velocity is measured with a hotwire type anemometer placed in the duct. The air velocity was also measured with a vane type anemometer every five minutes to compare. The air velocity in the pulling and pushing situations is given in Fig.  6. On average, the air velocity was approximately 1 m/s in the pulling mode and 2 m/s in the pushing mode for same power source. The air velocity was twice as high in the pushing mode comparing to the pulling mode. This can be explained by the fin density of the HS. The air velocity in the push mode is very slightly fluctuating. The air flow is perpendicular to the fins in the pushing mode as it flows parallel to the fins in the pulling mode. The fluctuation is thought to be due to turbulence, which is caused by the air flowing perpendicular to the fins. The surface temperature of the HS was about 5 °C lower in overall average in pushing mode and it can be said that this is caused by the difference of the air flow rate. Considering the present HS design, it has been concluded that the operation of the fan in the pushing mode is more appropriate in terms of thermal efficiency.
The thermal images in the pulling mode are shown in Fig. 7 and the pushing mode in Fig. 8 for four heater power.

CONCLUDING REMARKS
In this study, HS with the conical surface was designed, fabricated and the effect on the thermal performance of the fan orientation was investigated experimentally. The surface temperatures of the HS for thermal performance are measured and taken with the thermal camera. Tests were conducted at 25, 50, 75 and 100 W heating powers. The fan was run at the same electrical values and the fan orientation was reversed for to change the blowing direction. The air velocity in the pushing mode was twice as high, which can be explained by the fin density of the HS. The air velocity in the push mode is very slightly fluctuating. The air flow is perpendicular to the fins in the pushing mode as it flows parallel to the fins in the pulling mode. The fluctuation is thought to be due to turbulence, which is caused by the air flowing perpendicular to the fins. As a result of the experiments in the same conditions, the surface temperature in the pushing mode as average; 2°C at 25 W, 3.9°C at 50 W, 5.1°C at 75 W and 6.9°C at 100 W was lower than pulling. The surface temperature of the HS was about 5°C lower in overall average in pushing mode and it can be said that this is caused by the difference of the air flow rate. As a result, it has been found that the use of fan in pushing mode for the conical surface HS gives better results in terms of thermal performance.