International Journal of Heat and Mass Transfer 65 (2013) 1–9

Contents lists available at SciVerse ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

Impingement heat transfer of a plate fin heat sink with fillet profile

Kok-Cheong Wong ⇑, Sanjiv Indran 1
Department of Mechanical, Manufacturing and Materials Engineering, The University of Nottingham Malaysia Campus, Jalan Broga, Semenyih, Selangor Darul Ehsan 43500, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: The fluid flow and thermal characteristics of an air-impinged plate fin heat sink was investigated numer-
Received 4 March 2013 ically to understand the effect of fillet profiles at the bottom of the plate fin. The numerical model is val-
Received in revised form 20 May 2013 idated against existing literature. The results of pressure drop and mean Nusselt number were obtained
Accepted 23 May 2013
for cases with and without fillet profiles for variable channel widths and inlet widths. It is found that fillet
Available online 20 June 2013
profile enhances the overall thermal performance of a plate fin heat sink, and the U-shaped channel pro-
file is capable of achieving thermal enhancement as high as approximately 13%. The fillet profiles altered
Keywords:
the flow pattern such that the flow is smoothened near the bottom of the plate fin, reducing the negative
Plate fin heat sink
Fillet
impact induced by the rebounding flow. Furthermore, the addition of fillet profiles at the bottom of the
Heat transfer plate fin reduces the constriction resistance and contributes to the overall thermal enhancement.
Air impingement Ó 2013 Elsevier Ltd. All rights reserved.
U-channel

1. Introduction
Power density in electronic components is steadily increasing
due to the advances in modern electronic devices. Thermal man-
agement for such devices is becoming more challenging as a result
of miniaturization. Heat sinks coupled with peripherals such as jets
and fans are continuously developed to meet the requirement of
higher heat dissipation. Plate fin heat sink is one of the more com-
mon types of heat sink, consisting of plate fins at regular intervals.
This research explores the possibility to further enhance the heat
transfer efficiency of the plate fin heat sink with the consideration
of a fillet profile at the bottom of the plate fin and channel.
Extensive studies have been carried out by researchers on plate
fin heat sinks. Kordyban [1] compared the performance of a pin fin
and a plate fin heat sink of similar dimensions. The temperature
difference between the heat sink base and the ambient air for
the pin fin was 50 °C, and for the plate fin it was 44 °C. Although
the pin fin had a surface area of 194 cm2, larger than the plate
fin surface area of 58 cm2, the plate fin showed a merit in heat
transfer, recording a lower temperature difference. Forghan et al.
[2] published test data of pin fin, plate fin and flared fin heat sinks.
They found that for low air velocity between 1 m/s and 4 m/s, the
thermal performance for plate fin is at least 20% better as com-
pared to other heat sinks. Some researchers [3–6] developed ana-
lytical solutions for the heat transfer of plate fin heat sink in
parallel flow. Teertstra et al. [3] developed an analytical model to
⇑ Corresponding author. Tel.: +60 3 82948117; fax: +60 3 89248017.
E-mail addresses: wongkc6@gmail.com (K.-C. Wong), kedx2sir@nottingham.
edu.my (S. Indran).
1
Tel.: +60 3 82948358; fax: +60 3 82948017.
predict thermal performance for air cooled plate fin heat sinks in
parallel flow. The average heat transfer rate predicted was ex-
pressed in terms of Nusselt number, which combines the limiting
cases of fully-developed and developing flow. The model was
found to be in good agreement with the experimental results. Cul-
ham and Muzychka [4] developed a procedure that allows design
parameters in a plate fin heat sink to be optimized. The model used
the apparent friction factor model [5], which was asymptotic be-
tween a developing and fully developed flow. The procedure is
based on the minimization of entropy generation resulting from
viscous fluid effects and heat transfer. Copeland [6] suggested a
laminar flow pressure drop model for parallel flow in rectangular
channeled (plate fin) heat sinks. The friction factor data for devel-
oping laminar flow were taken from Shah and London [7] and fit-
ted to an equation of the Churchill–Usagi form. Optimum
dimensions of fin thickness and pitch were calculated for a variety
of realistic operating conditions.
The performance of impingement heat transfer of plate fin heat
sinks was reported by researchers [8–14]. Kondo and Matsushima
[8] completed an experimental study and reported a zonal model
of a thermal resistance prediction for impingement cooling of plate
fin heat sinks. The impingement flow over the plate fins was di-
vided into six regions. A set of correlations were proposed between
the thermal resistance of the heat sink and the geometry of the
plate fins. The accuracy of the predicted thermal resistance was
found to be within ±25% of the experimental data. They showed
that optimized plate-fin heat sinks provide 40% lower thermal
resistances compared to optimized pin-fin heat sinks.
Biber [9] carried out a numerical study to determine the thermal
performance of a single isothermal channel of plate fin heat sink

0017-9310/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2013.05.059
2 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9

Nomenclature

Dh hydraulic diameter, m Pout outlet pressure, Pa

hb thickness of the heat sink base, m q00 heat flux, W/m2
hc channel height, m Re Reynolds number
he extended height at inlet, m r fillet radius, m
h heat transfer coefficient, Wm2 K1 Tin inlet temperature, K
h average heat transfer coefficient, Wm2 K1 Tw bottom wall temperature, K
H Height of heat sink, m u, v, w velocity components along x, y, z-axes, respectively,
k thermal conductivity, Wm1 K1 m s1
L half of the heat source length or half of base length of win inlet velocity, m s1
heat sink (Fig. 1), m W width of heat sink, m
Lin half of inlet width, m Wc channel width, m
Le extended length at outlet, m Ws section width of heat sink, m
Nu local Nusselt number x, y, z Cartesian coordinates, m
Nu mean Nusselt number along the heat source
Pin inlet pressure, Pa

Impinging Air Flow

Ws
Wc t
2Lin

hc
H

Fillet
radius r
hb

2L
W W
(a) (b)
Pressure
Adiabatic
Outlet

Tin, win Symmetrical

Lin
Inlet

he

Le

H Adiabatic (bottom)
z L
y
x

q”, Tw (bottom)
(c)
Fig. 1. Schematic of (a) physical model, (b) single section of channel and (c) domain.

Table 1
Comparison of various turbulence models for Le = 60 mm, Re = 693 and
with variable widths of impingement flow. Biber [9] studied many Wc = 2 mm, Lin = 20 mm.
different combinations of channel parameters and obtained corre-
Turbulence model Pressure % Discrepancy as compare to
lations for channel mean Nusselt number. These correlations cover
drop experiment [13]
a wide range of practical plate-fin heat sinks with air supplied by
Spalart–Allmaras 10.525 3.186
axial fans or other uniform or nonuniform methods. Saini and Webb
Standard k-epsilon 10.788 5.764
[10] presented a modified Biber model [9] and validated their mod- Standard k-omega 10.846 6.333
el by experiments. The predicted pressure drop is 13–31% lower
 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9 3

Table 2 than the experimental data, and higher flow rates result in a higher
Grid independence study for Wc = 2 mm, Lin = 20 mm, Re = 346 and Le = 60 mm. under-prediction of pressure drop. Duan and Muzychka [11,12] also
Grid size Pressure drop Mean Nusselt number developed a simple semi-empirical model for predicting the pres-
6376 5.832 13.28 sure drop and the heat transfer coefficient of plate-fin heat sinks.
13,932 5.437 13.26 Their model provides an accurate means to predict the mean heat
41,520 5.154 13.26 transfer coefficient for systems utilizing impingement flow. Kim
82,650 5.126 13.24 et al. [13] compared the thermal performances of plate fin and
332,160 5.102 13.22
pin fin heat sinks subjected to an impinging flow both experimen-
tally and analytically. The mathematical model based on the
volume averaging method is used to compare the optimized plate
fin heat sink and optimized pin fin heat sink under fixed pumping
Table 3 power conditions. They found that for low pumping power,
Extension length independence study for Wc = 2 mm, Lin = 20 mm and Re = 346. optimized pin fin heat sinks demonstrate relatively lower thermal
resistance. In the case of high pumping power, plate fin heat sinks
Extension length, Le (mm) Pressure drop Mean Nusselt number
demonstrate relatively lower thermal resistance. Li et al. [14]
20 5.247 13.32
conducted experimental and numerical parametric studies on the
40 5.206 13.28
60 5.154 13.26
thermal characteristics of plate fin heat sinks (6  2 array with a
80 5.148 13.26 cut off passage) subject to jet impingement. They studied the effect
100 5.14 13.26 of variable impinging distance, jet Reynolds number, fin width as
well as fin height. The lowest thermal resistance was found at an

(a)

(b)

Fig. 2. Validation of (a) pressure drop and (b) thermal resistance for different flow rate with Lin = 20 mm.
4 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9
optimum impinging distance of 20 times the nozzle diameter. The
fin widths at a specific range were found to provide better heat
transfer. Higher values of fin height gave lower values of thermal
resistance.
To the best knowledge of the authors, the effect of adding fillet
profiles to a plate fin heat sink has not been investigated. The
objective of the present study is to investigate the flow character-
istics and thermal performance of a plate fin heat sink with fillet
profiles subject to air impingement. Simulations are carried out
using the software Fluent™.
The physical model for the present study is shown in Fig. 1. The
plate fin heat sink is shown in Fig. 1a, with a base area of
40  40 mm. The impinging air enters the heat sink with an inlet
width of 2Lin. In this study, variable inlet width is considered where
Lin 6 L. For Lin < L, the top of the channel or heat sink is confined.
For Lin = L, it is not confined. The plate fins of the heat sink have
uniform thickness and the channel width Wc between the plate fins
are also uniform. Due to the uniform size of each channel and fin
Δ
Nu
Fig. 3. Results of (a) pressure drop and (b) mean Nusselt number for different fillet radii with Wc = 2 mm and Lin = 20 mm.
thickness, the heat sink is modeled with a single section of the
channel as highlighted by the dotted lines in Fig. 1b, where Ws is
the section width. Based on that, the 3D domain used for the sim-
ulation is a half symmetrical model as shown in Fig. 1c, where an
extended inlet height he and an extended outlet length Le are in-
cluded. The present model considers the impingement flow from
the top of the heat sink. The inlet flow is not expected to enter
all the channels smoothly as some of the flow will hit the top sur-
face of the fin. Hence, the inlet is modeled with an extended height
above the plate fin. Note that the channel length is the same as the
base length of the heat sink, but the computation domain considers
half of the channel length as it is half symmetry.
The material of the solid (heat sink) is set to be aluminium alloy
6061 with a thermal conductivity k of 170 W/m K. The channel
length is 40 mm, giving L a value of 20 mm. The fin height hc is
set to be 25 mm to replicate the model of Kim et al. [13]. In addi-
tion, the heat sink base thickness hb and fin thickness t are set to be
5 mm and 0.5 mm, respectively. The heat sink is subjected to a
(a)
(b)
 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9 5

uniform heat flux at the bottom of the heat sink. The effects of the The values of some parameters are set with values as follows:
physical parameters of inlet width and channel width for different
fillet radius sizes on heat transfer are investigated. This model con- r ¼ 2=3; C b1 ¼ 0:1355; C b2 ¼ 0:622; k ¼ 0:41;
siders a fillet profile at the bottom of the channel with radius r as 2
C w1 ¼ C b1 =k þ ð1 þ C b2 Þ=r; C w2 ¼ 0:3; C w 3 ¼ 3; ð12Þ
shown in Fig. 1b. The addition of the fillet is expected to change the
C v 1 ¼ 7:1; C t1 ¼ 1; C t2 ¼ 2; C t3 ¼ 1:2; C t4 ¼ 0:5
flow pattern when the flow reaches the bottom of the channel. It is
also expected to affect the thermal performance of the heat sink.
For the solid domain, the energy equation is as follows:

2. Numerical procedures  
@ 2 @T
¼0 ð13Þ
@xi @xi
The problem is modeled as a 3D fluid–solid conjugate. The flow
is assumed to be steady and incompressible. The fluid properties
The boundary conditions for the present study are illustrated in
are assumed to be constant. The continuity, momentum and en-
Fig. 1c. The geometry and boundary conditions are based on the
ergy equations for the fluid region are as follows:
conditions outlined by Kim et al. [13], Biber [9] and Xie et al. [16].
@ui The boundary conditions can be expressed as follows:
¼0 ð1Þ
@xi At x = 0, the surface is set as symmetric boundary.
" # At x = L + Le (flow exit), the outlet is set as pressure outlet.
@ui @ @ 2 ui At x ¼ Lin ðH 6 z 6 H þ he Þ, the surface is set as adiabatic and
quj ¼ pdij þ m 2  qui uj
0 0
ð2Þ non-slip wall.
@xj @xj @xj
At z ¼ 0ð0 6 x 6 LÞ, uniform heat flux q’’ is applied on the bot-
"  # tom heated non-slip wall.
@T @ m mt @T At z ¼ 0ðL < x 6 L þ Le ), the bottom surface is set as adiabatic
uj ¼ þ ð3Þ
@xj @xj Pr rT @xj and non-slip wall.
At z ¼ HðLin 6 x 6 L þ Le ), the surface is set as adiabatic and
where rT = 1 and the Reynolds stress is given by:
non-slip wall.
 
@ui @uj 2 At z = H + he, the velocities are u = v = 0, w = win, and the inlet
u0i u0j ¼ mt þ  dij K ð4Þ
@xj @xi 3 temperature T = Tin.

mt is the turbulence eddy viscosity, K is the turbulence kinetic en-

ergy and dij is the kronecker delta.
Although the range of Reynolds number investigated is in the
Wc
laminar regime, a turbulence model is used for the present work
as the formation of eddies are expected, as shown in Fig. 5. Numer-
ical investigation has been conducted to find a suitable turbulence
model. Table 1 shows the comparison between different turbu-
lence models. The Spallart Allmaras model [15] is found to produce
the closest value of pressure drop when compared to the experi-
mental value of pressure drop of 10.222 Pa [13]. The Spallart Allm-
aras model is employed to solve the turbulence eddy viscosity
based on the equations as follows:
@m
~i 1n o
uj ¼ C b1 ½1  ft2 e
Sm~þ r  ½ðm þ m~Þrm~ þ C b2 ½rm~2
@xj r
  2
C b1 m~
 C w1 fw  2 ft2 þ ft1 DU 2 ð5Þ
j d
where
v3 m~
mt ¼ m~fv 1 ; f v 1 ¼ ; v¼ ð6Þ
v3 þ C 3v 1 m
" #16
1 þ C 6w3   m~
fw ¼ g ; g ¼ r þ C w2 r 6  r ; r ð7Þ
g 6 þ C 6w3 e
S j2 d
2
r r
 i
w2 h 2 2
ft1 ¼ C t1 g t1 exp C t2 t 2 d þ g 2t dt ð8Þ
DU
 
ft2 ¼ C t3 exp C t4 v2 ð9Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
S¼ 2Xij Xij ð10Þ

The rotation tensor is given by:

  (a) (b) (c)
1 @ui @uj
Xij ¼  ð11Þ Fig. 4. Illustration of channel profiles (a) without fillet (r = 0 mm) (b) with fillet of
2 @xj @xi
small radius (r < Wc/2) and (c) with fillet size forming a U-channel (r = Wc/2).
6 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9

At y = 0 and y = Ws, both the surfaces are set as symmetrical evaluated as mean Nusselt number for different values of Reynolds
boundaries. number. The Reynolds number is defined as follows:
The interior walls and surfaces are fluid/solid interface bound-
win Dh
aries; they are set as coupled walls. In the present problem, the Re ¼ ð14Þ
heat flux is fixed as 18,750 W/m2. A uniform inlet velocity profile
t
is specified at the inlet for defining simultaneously developing sin- where Dh is the hydraulic diameter of the channel and t is the kine-
gle-phase flow conditions. The inlet temperature of the air is 300 K. matic viscosity of air (1.789  105 kg/m-s).The hydraulic diameter,
The second order upwind scheme was applied for the convec- Dh is computed from the following expression:
tion–diffusion formulation of the momentum and energy equation
4wc Hc
of the fluid. The SIMPLE algorithm for the pressure–velocity cou- Dh ¼ ð15Þ
2ðwc þ Hc Þ
pling was employed. For the solution to converge, a convergence
criterion of residuals below 1  105 is adopted. To ensure better The mean Nusselt number Nu is computed based on
accuracy, a grid refinement study was conducted and the analysis
is shown in Table 2. The grid size of 82,650 has shown negligible hDh
Nu ¼ ð16Þ
difference in pressure drop and mean Nusselt number as compared k
to a higher grid size of 332,160. In addition, the outlet extension where the mean heat transfer coefficient, h (W/m2-K) is given by:
length Le was studied to ensure sufficient length to produce a
developed flow. The results in Table 3 show that the extension q00
h¼ ð17Þ
length Le of 60 mm gives negligible difference in pressure drop ðT w  T in Þ
and mean Nusselt number as compared to higher values of Le.
where T w is the mean temperature of the bottom heated wall and
Therefore, the grid size of 82,650 and extension length Le of
Tin is the inlet temperature.The pressure drop DP is evaluated based
60 mm are adopted to generate the results presented hereafter.
on the following expression:
DP ¼ Pin  Pout ð18Þ
3. Results and discussion
where Pin is the inlet pressure and Pout is the average outlet
The present numerical model is validated against the results of pressure.
thermal resistance and pressure drop reported by Kim et al. [13] The effect of fillet size at the bottom of the channel has been
and Kim and Kim [17]. The results presented in Fig. 2 show that investigated for different values of Reynolds number, with Wc = 2 -
the results of the present model are in good agreement with the mm and Lin = 20 mm. The results of DP and Nu are presented in
experimental results [13] and exact solution [17] with the maxi- Fig. 3a and b respectively for r = 0 mm, 0.2 mm, 0.4 mm, 0.6 mm,
mum discrepancy of 9.5% for thermal resistance and 8.3% for pres- 0.8 mm, 1.0 mm. Note that the fillet radius of r = 1.0 mm with
sure drop. channel width Wc = 2 mm forms a U-channel as illustrated in
The heat transfer characteristics for a plate fin heat sink with fil- Fig. 4c. The fillet radius of r = 0.2, 0.4, 0.6, 0.8 forms the channel
let profile at the bottom of the channel are investigated for variable shapes illustrated in Fig. 4b, whereas the case for r = 0 (rectangular
channel width and inlet width. The heat transfer performance is channel) is illustrated in Fig. 4a. The results show that the increase

(a) (b)

backstep backstep

(c) (d)
Fig. 5. Streamlines for the flow of Re = 693 in the channel with Wc = 2 mm, Lin = 20 mm for (a) r = 1 mm (with fillet forming a U-channel) and (b) r = 0 (channel without fillet).
The close-up view of (a) and (b) are shown in (c) and (d), respectively.
 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9 7

(a)

Nu

(b)
Fig. 6. Results of (a) pressure drop and (b) mean Nusselt number with Lin = 20 mm for variable Reynolds numbers and channel widths, comparing the cases of a U-channel
and a channel without fillet profile.

in fillet radius size is found to increase the values of DP and Nu
such that the U-shaped channel gives the highest value of Nu. As
the size of fillet increases, the channel cross-sectional area be-
comes smaller. This leads to a higher exit velocity and lower exit
pressure and hence a pressure drop is observed as fillet radius in-
creases. Fig. 3 also shows that increasing the Reynolds number re-
sults in higher values of DP and Nu. This is in agreement with the
studies conducted on plate fin heat sinks [18,19], where an in-
crease in Reynolds number reduces the thermal resistance. The ef-
fect of the fillet is observed to be less significant at low Reynolds
numbers between 173 and 346, but more significant for higher
Reynolds numbers. The introduction of a fillet profile at the bottom
of the plate fin is expected to change the flow pattern of the flow
near the fillet. The impinging air that flows vertically downwards
will hit the bottom surface of the channel and produce a rebound-
ing flow. The higher the Reynolds number, the stronger the
rebounding flow. The flow near the middle of the channel is likely
to produce a vertical rebound, whereas the flow near the side of the
channel will hit the fillet radius to make a smooth turn rather than
a rebound. The resultant flow is a mixture between the rebound
and smooth turn flow. To better understand the effect of having
a fillet radius, the flow patterns for a channel without fillet and a
channel with fillet (r = 1 mm) are presented in Fig. 5 for the case
of Lin = 20 mm, Wc = 2 mm. It is obvious that for the channel with-
out fillet profile (Fig. 5b), the flow that hits the bottom surface of
the channel experiences a stronger vertical rebound compared to
the channel with fillet profile (Fig. 5a). This is judged by the height
of the rebound flow as evident in Fig. 5a and b. The strong vertical
rebound is in the opposing direction of the impinging flow, and ex-
pected to result in momentum loss or lower resultant velocity. The
addition of fillet at the bottom of the plate fin reduces the cross-
sectional area of the channel and hence results in higher velocity.
This is another reason that the flow becomes stronger in the chan-
nel when the size of fillet radius increases.
Close-up views for Fig. 5a and b are shown in Fig. 5c and d,
respectively. It can also be observed that at the backstep, a recircu-
lation is significant for the channel without fillet (Fig. 5d) but not
found for the channel with fillet (Fig. 5c). The channel flow with
the fillet is produced by a mixture between rebound flow and turn
flow near the fillet. Hence, the resultant flow is stronger in the
8 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9

(a)

Nu

(b)
Fig. 7. Results of (a) pressure drop and (b) mean Nusselt number with Wc = 1 mm for different inlet widths, comparing the case of a U-channel and a channel without fillet
profile.

horizontal direction (Fig. 5c) relative to the case without fillet
(Fig. 5d). This confirms that the presence of the fillet enhances
the strength of the flow in the horizontal direction above the solid
surface and enhances the heat transfer. This effect is even better for
greater sizes of fillet radii, as it’s clearly indicated by the results in
Fig. 3b, where a larger fillet radius produces better thermal perfor-
mance. When compared to a channel without fillet, the presence of
fillet profile reduces the constriction resistance as a result of addi-
tional substrate material at the bottom of fin. This gives a positive
effect on the heat transfer. Nevertheless, the change from a rectan-
gular channel to a U-channel after the addition of the fillet reduces
the contact area between the fluid and solid. Although this could
give a negative effect on the heat transfer, an overall thermal
enhancement of heat transfer is achieved due to the lower con-
striction resistance and the smoother turn flow when a fillet is
present.
The effects of the fillet on DP and Nu are also investigated for
different channel widths Wc, comparing the case without fillet pro-
file and with fillet profile forming a U-channel. The results of DP
and Nu have been obtained and are presented in Fig. 6a and b
respectively. The results show a similar trend, where the mean
Nusselt number and pressure drop are higher for the channels with
fillet profile (U-channel) as compared to channels without fillet
profile, regardless of the value of the channel width. In Fig. 6, It
can be observed that the narrower the channel the higher the mean
Nusselt number as the thermal contact between the flow and the
solid heat transfer surfaces is more effective.
This study also verified whether the inlet width has an impact
on DP and Nu. The results in Fig. 7a and b show a similar trend,
where pressure drop and mean Nusselt number are higher for a
channel with fillet profile (U-channel) as compared to a rectangu-
lar channel without fillet profile, regardless of the value of inlet
width. Fig. 7 shows that a higher inlet width results in a higher
pressure drop and a higher mean Nusselt number for flows with
the same Reynolds number.
The maximum thermal enhancement for the channel with a
fillet profile as compared to the channel without a fillet profile
was calculated from the results. Among the results obtained, the
 K.-C. Wong, S. Indran / International Journal of Heat and Mass Transfer 65 (2013) 1–9
maximum thermal enhancement was found to be approximately
13% for cases of Re = 483 for Wc = 1 mm, Re = 443 for Wc = 1.5 mm,
and Re = 520 for Wc = 2 mm, with Lin = 20mm.
The reported results show the advantage of having a fillet pro-
file in the plate fin heat sink. The plate fin heat sink with fillet pro-
file can produce better heat dissipation for heat emitting devices.
The fillet can be easily incorporated into the design of plate fin heat
sinks. Plate fin heat sinks with fillet profile can be applied in the
same way as conventional plate fin heat sinks and they can be
manufactured by the stamping process.
4. Conclusion
The effect of fillet profiles on the fluid flow and heat transfer of a
plate fin heat sink subjected to air impingement and a constant
heat flux has been investigated numerically. The presented numer-
ical model is validated with a maximum discrepancy of 9.5% when
compared to the results in the literature. This study has demon-
strated numerically the benefit of introducing fillet profiles at the
bottom of the plate fins of a multichannel heat sink. The results
confirm that the addition of a fillet profile enhances the heat trans-
fer performance as compared to the conventional design (i.e. rect-
angular channel without fillet), capable of achieving thermal
enhancement of 13% for the conditions investigated. The larger
the size of fillet radius, the better the heat transfer performance.
The U-channel gives the best heat transfer performance regardless
of the size of inlet width and channel width. The investigation on
the fluid flow revealed that the turn flow formed near the fillet ra-
dius leads to a smoother flow with better strength in the horizontal
direction, therefore enhancing the heat transfer performance. In
addition, the presence of a fillet profile reduces the constriction
resistance, which contributes to the overall thermal enhancement
of the plate fin heat sink.
References
[1] T. Kordyban, Hot Air Rises and Heat Sinks – Everything You Know about
Cooling Electronics is Wrong, ASME Press, NY, 1998.
9
[2] F. Forghan, D. Goldthwaite, M. Ulinski, M. Metghalchi, Experimental and
theoretical investigation of thermal performance of heat sinks, ISME May
(2001).
[3] P. Teertstra, M. Yovanovich, J.R. Culham, T. Lemczyk, Analytical Forced
Convection Modeling of Plate Fin Heat Sinks, in: Proceedings of 15th IEEE
Semi-Therm Symposium, iugguyfgu, 1999, 34–41.
[4] J.R. Culham, Y.S. Muzychka, Optimization of plate fin heat sinks using entropy
generation minimization, IEEE Trans. Compon. Packag. Technol. 24 (2001)
159–165.
[5] Y.S. Muzychka, and M.M. Yovanovich, Modelling Friction factors in non-
circular ducts for developing laminar flow, in: Proceedings 2nd AIAA
Theoretical Fluid Mechanics Meetings, Albuquerque, NM, Paper No. AIAA,
1998, pp. 98–2492.
[6] D. Copeland, Optimization of parallel plate heat sinks for forced convection, in:
Proceedings 16th IEEE Semi-therm Symposium, San Jose, CA, 2000, pp. 266–
272.
[7] R.K. Shah, A.L. London, Laminar Flow Forced Convection in Ducts, Academic,
New York, 1978.
[8] Y. Kondo, H. Matsuhima, Study of impingement cooling of heat sinks for LSI
packages with longiudinal fins, Heat Transfer Jpn. Res. 25 (1996) 537–553.
[9] C.R. Biber, Pressure drop and heat transfer in an isothermal channel with
impinging flow, IEEE Trans. Compon. Packag. Manuf. Technol. A 20 (1997)
458–462.
[10] M. Saini, R.L. Webb, Validation of models for air cooled plane fin heat sinks
used in computer cooling, in: Proceedings of Eighth Intersociety Conference on
Thermal and Thermomechanical Phenomena in Electronic Systems, San Diego,
CA, 2002, pp. 243–250.
[11] Z. Duan, Y.S. Muzychka, Pressure drop of impingement air cooled plate fin heat
sinks, Trans. ASME 129 (2007) 190–194.
[12] Z. Duan, Y.S. Muzychka, Experimental investigation of heat transfer in
impingement air cooled plate fin heat sinks, Trans. ASME 128 (2006) 412–418.
[13] D.-K. Kim, S.J. Kim, J.-K. Bae, Comparison of thermal performances of plate-fin
and pin-fin heat sinks subject to an impinging flow, Int. J. Heat Mass Transfer
52 (2009) 3510–3517.
[14] H.-Y. Li, K.-Y. Chen, M.-H. Chiang, Thermal-fluid characteristics of plate-fin
heat sinks cooled by impingement jet, Energy Convers. Manage. 50 (2009)
2738–2746.
[15] P.R. Spalart, S.R. Allmaras, A one-equation turbulence model for aerodynamics
flows, AIAA 90-0439.
[16] X.L. Xie, Z.J. Liu, Y.L. He, W.Q. Tao, Numerical study of laminar heat transfer and
pressure drop characteristics in a minichannel heat sink, Appl. Therm. Eng. 29
(2009) 64–74.
[17] D.-K. Kim, S.J. Kim, Averaging approach for microchannel heat sinks subject to
the uniform wall temperature condition, Int. J. Heat Mass Transfer 49 (2006)
695–706.
[18] H.-Y. Lee, S.-M. Chao, Measurement of performance of plate-fin heat sinks with
cross flow cooling, Int. J. Heat Mass Transfer 52 (2009) 2949–2955.
[19] H.-Y. Li, K.-Y. Chen, Thermal performance of plate-fin heat sinks under
confined impinging jet conditions, Int. J. Heat Mass Transfer 50 (2007) 1963–
1970.