Вы находитесь на странице: 1из 7

Energy Conversion and Management 50 (2009) 1264–1270

Contents lists available at ScienceDirect

Energy Conversion and Management


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

Thermal enhancement from heat sinks by using perforated fins


M.R. Shaeri, M. Yaghoubi *
Mechanical Engineering Department, School of Engineering, Shiraz University, Shiraz 71348-51154, Iran

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

Article history: Three-dimensional incompressible laminar fluid flow and heat transfer of a heated array of rectangular
Received 23 June 2008 perforated and solid fins attached on a flat surface are studied numerically. Perforations with rectangular
Accepted 18 January 2009 cross section are along the length of bluff plates and their number varies from 1 to 8. A SIMPLE code using
Available online 23 February 2009
control volume approach is developed for solving Navier–Stokes equations with conjugate energy equa-
tions for both fluid and solid regions. The second order upwind technique is used for discretization of
Keywords: momentum and energy equations. Computations were performed for a range of Reynolds numbers
Fin weight
100  ReD  350, based on fin thickness. Thermal performances and effectiveness as well as friction coef-
Perforated fin
Fin performance
ficient of perforated and solid fins are determined and compared. Higher performances for perforated fins
Nusselt number are observed and effectiveness increased by increasing number of perforations. Higher reduction of
Laminar flow weight of fins due to perforation is another major of utilization of the new type of fins structure. Finally
a new correlation is proposed to predict perforated fin effectiveness for the range of computation carried
out for the considered fins.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction arrays. Rouvreau et al. [9] studied laminar water flow around a
rectangular block mounted at a short distance from the leading
Advanced technologies need high performance heat transfer edge of a flat plate both experimentally and numerically. Their cal-
equipments. Methods for improving heat transfer classified in culations with obstacle in the vicinity of the inlet boundary were in
two categories: active and passive methods [1]. Active methods re- poor agreement with experiments. They reported that the up-
quire external power to enhance heat transfer and passive meth- stream boundary must be placed at a sufficient distance from the
ods do not require external power. Extended surfaces or fins are obstacle to eliminate interactions between pressure perturbation
example of passive methods that are commonly used in variety associated with this object and the boundary. Saha [10] studied
of industrial applications to enhance the rate of heat transfer be- three-dimensional laminar unsteady fluid flow and heat transfer
tween primary surface (heat sink) and ambient fluid. Due to [2,3] from a stationary cube placed in a uniform flow numerically for
straight fin’s concave parabolic profile provides the maximum heat Reynolds number range of 50–400 based on cube dimension. They
dissipation for a given profile area. For most applications, the rect- observed that for Re = 270 the flow became unsteady. In the steady
angular fins are used to reduce cost of manufacturing. Based on the regime, the drag coefficient decreased with increase of Reynolds
widespread application of rectangular fins that are commonly used number. On the other hand the drag force increased in the unstea-
for heat exchangers, the understanding of convection mechanisms dy regime. Dijk and Lange [11] studied three-dimensional laminar
and prediction of heat transfer performance on rectangular fins are flow over a block placed on the floor of a channel numerically. They
usually analyzed by the flow and heat transfer simultaneously [4]. identified several distinct flow features and reported that the
Related studies about bluff plates and thermal efficiency for vari- shape and size of these flow features are mainly dominated by Rey-
ous surfaces and geometries are extensive. Asako and Faghri [5,6] nolds number. Also they proposed a logarithmic correlation to
modeled three-dimensional periodic fully developed laminar flow determine the position of separation and attachment point with
and heat transfer over rectangular blocks. Holahan et al. [7] devel- Reynolds number.
oped analytical methodology for calculating thermal performance In addition, heat transfer rate can be improved by employing
and pressure drop in fully shrouded, laminar, parallel plate heat perforations, porosity or slots [12]. On the other hand, due to
sinks. Baskaya et al. [8] solved three-dimensional elliptic governing restrictions in setup space and economical reasons, heat transfer
equations for horizontal rectangular fin arrays by finite volume equipments have been required to be much more compact in size
based CFD code. They investigated the effects of fin spacing, fin and lighter in weight [13]. So while a simple solid baffle plate at-
height, length and temperature on performance of rectangular fin tached to the duct wall enhances heat transfer, a perforated plate
attached to the same duct wall poses less resistance to the flow
* Corresponding author. Tel./fax: +98 711 6272060. and, thus, might have a better performance [14]. Sara et al. [15]
E-mail address: yaghoub@shirazu.ac.ir (M. Yaghoubi). determined thermal performances of solid and perforated rectan-

0196-8904/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2009.01.021
M.R. Shaeri, M. Yaghoubi / Energy Conversion and Management 50 (2009) 1264–1270 1265

Nomenclature

Cf average friction coefficient WP perforation width (m)


D fin thickness (m) X,Y,Z rectangular coordinates
Gr Grashof number
h average convection heat transfer coefficient Greek symbols
(W/m2 K) a fluid thermal diffusivity (m2/s)
H fin height (m) ePF perforated fin effectiveness
HP perforation height (m) / porosity V v oid =V
L fin length (m) k fluid thermal conductivity (W/m K)
N number of perforations m fluid kinematic viscosity (m2/s)
Nu average Nusselt number ðhDÞ=k q fluid density (kg/m3)
P pressure (Pa) sw average wall shear stress over the faces of fin (Pa)
q heat transfer rate from fin surface (W)
Re Reynolds number Subscripts
ReD Reynolds number ðu1 DÞ=m b fin base
S fin spacing (m) i,j tensor notation
T temperature (K) in inlet condition
u,v,w velocity components in X, Y, Z directions (m/s) PF perforated fin
V fin volume ðL  H  DÞ SF solid fin
V v oid void volume ðL  HP  W P  NÞ 1 free stream

gular blocks that are attached on a flat surface in a rectangular duct The temperature field on the outer surface of block and perforations
in comparison with heat transfer from the same plate without were obtained by solving the Fourier’s heat conduction equation
blocks. They reported that the solid blocks generated a net energy simultaneously with forced convection in the fluid.
loss despite significantly enhanced heat transfer due to the in-
creased of heat transfer surface area. Sahin et al. [12] experimen- 3. Problem description and boundary conditions
tally investigated overall heat transfer, friction factor and the
effect of various design parameters on heat transfer and friction Typical array of fins modeled in this study are shown in Fig. 1.
factor for the heat exchanger equipped with circular cross section The air flow is considered to be steady and laminar with constant
perforated pin fins in a rectangular channel and also they obtained properties. Moreover, air velocities are such that forced convection
correlations for the enhancement efficiency. Dorignac et al. [16] is the dominant heat transfer mechanism between fins and ambi-
experimentally determined convective heat transfer on a multi ent air because corresponding Richardson number (Gr/Re2) based
perforated plate. They also proposed an empirical relation for heat on the fin height is less than 0.1. Air with inlet velocity u1 blows
exchange at the windward surface of a perforated flat plate for a across the fins array. The fin material is aluminum with thermal
large range of perforations spacing. conductivity 202 (W/m K) that are widely used for heat removal
In the present study, three-dimensional laminar fluid flow and from various surfaces. Also, fins have a length, height and thickness
convective heat transfer around an array of solid and perforated equal to 24, 12 and 4 mm, respectively and the fin spacing is
rectangular fins are analyzed numerically. Fins have a new design 20 mm. In addition the length, height and width of perforations
type of perforations with rectangular cross section and assumed to are 24, 1.2 and 2.22 mm. The fin thickness is used as the character-
have 1, 2, 3, 4, 6 and 8 straight holes along the fin’s length which istic length in the Reynolds number. Flows around three-dimen-
corresponds to porosity of 0.0555–0.444. According to the previous sional obstacles are complex and usually characterized by large
investigations there is not any detailed experimental or theoretical unsteady rotational structures, by regular vortex shedding and by
work on flow and heat transfer for the perforated fins such as illus- a strongly three-dimensional topology. The flow Reynolds number,
trated in Fig. 1. The present study seems to be the first research boundary layer thickness and nature, dimensions and position of
that considers these types of perforated fins in an array of rectan- the obstacles and the blockage ratio proved to have a major influ-
gular fins in laminar regime. For investigation, three-dimensional, ence on the nature of heat transfer and fluid flow [9]. In the present
steady, laminar incompressible Navier–Stokes equations and the study, the range of Reynolds number is taken 100–350. With these
energy equation for solid and perforated fins are solved numeri- Reynolds numbers, no instability in flow has observed and flow is
cally and the comparison between their thermal performances is completely steady three-dimensional.
elaborated. Due to flow symmetry assumed in Fig. 1, computations can ap-
ply only to one fin instead of array of fins. Fig. 2a shows the com-
2. Governing equations putational domain. Unsteady state conditions may happen at low
or high Reynolds numbers and this will conduct to unsymmetrical
For three-dimensional steady state incompressible laminar fluid flow field. For such observation of flow regime, first complete fin
flow the governing equations are: was considered and computations with Reynolds number equal
to 100 performed for full domain. For these conditions completely
oui
Continuity equation : ¼0 ð1Þ symmetric of results were obtained which indicates a steady state
oX i
flow regime. After that, the computation is applied for other Rey-
oui 1 oP o2 ui nolds numbers with assumption of steady state flow. For plane
Momentum equation : uj ¼ þm 2 ð2Þ
oX j q oX i oX i abcd, as inlet boundary, uniform flow condition is considered for
all variables using uin ¼ u1 ; v in ¼ win ¼ 0 and T in ¼ T 1 . Also the
oT o2 T same conditions are applied for plane bckj that is free stream plane.
Energy equation : uj ¼a ð3Þ
oX j oX i oX i For plane ijkl that is outlet and far from the plate, zero gradients of
1266 M.R. Shaeri, M. Yaghoubi / Energy Conversion and Management 50 (2009) 1264–1270

Fig. 1. Arrays of solid and perforated fins considered in the present investigation.

For planes abji and dckl, symmetry conditions are applied that
means zero gradients of variables in the Z direction, oðÞoZ
¼ 0. Also
in these planes, w = 0. All remaining planes are walls with no slip
boundary condition. Free stream temperature is assumed 25 °C
and the fin base, the plane efgh, has a constant temperature equal
to 70 °C. Also the planes adeh and gfli are assumed adiabatic.
According to Leung and Probert [17] radiation heat transfer rate
is less than 5% and 8% of total heat transfer rate from the polished
aluminum finned surfaces for temperature differences around
40 °C and 77.5 °C, respectively. Therefore in the present study the
effect of thermal radiative heat transfer is neglected because the
maximum temperature difference is equal to 45 °C.

4. Computational scheme
Fig. 2a. Computational domain for solid and perforated fins.
4.1. Domain geometry and grid size

variables in the X direction, oðÞ


oX
¼ 0 are imposed. This condition is The domain used in the present study consists of an entrance
for fully developed condition but it can be applied for this simula- region, an exit plane and the upper free stream surface as illus-
tion because the outlet boundary is sufficiently far from the plate. trated in Fig. 2a. These planes should be sufficiently far from the
M.R. Shaeri, M. Yaghoubi / Energy Conversion and Management 50 (2009) 1264–1270 1267

fin surfaces that the results become independent of the domain


size. For such analysis some tests were performed to obtain an
appropriate distance from the fin surfaces and by using previous
studies [9,18] finally the computational domain extends 15D up-
stream, 30D downstream, 7H in Y direction and 5D in Z direction.
For proper capture of recirculation zone and reattachment point,
the appropriate grid points in the three dimensions that are dense
near the fin and perforated walls are chosen. Also the solution
must be independent of grids. For this reason several grid points
were studied by changing the number of grids in three dimensions.
For each grid configuration the changes in flow and heat transfer
parameters were compared until independent grids result were Fig. 2b. Grid configuration for a fin with two perforations.
achieved. The related calculation as a sample for solid fin is re-
ported in Table 1.
According to Table 1, the number of grids of 180  80  37 for 0.45 Experimental result [9]
solid fin in X, Y and Z directions respectively are acceptable. For per- Numerical result [9]
forated fins the same procedure is repeated. The corresponding grid Present numerical result

configuration for a typical perforated fin is illustrated in Fig. 2b. 0.4

4.2. Computational procedure


0.35
The governing Equations (1)–(3) are discretized using finite vol-

Z/D
ume code that control volume cells for velocity components are
staggered with respect to the main control volume cells using 0.3

the SIMPLE algorithm developed by Patankar [19]. The second or-


der upwind technique is used to calculate momentum and energy
0.25
equations. The set of discretized equations are solved iteratively
line-by-line. Computation was started by first solving continuity
and momentum equations to determine the flow field and then
0.2
the energy equations to find the thermal field in the computational
-0.2 0 0.2 0.4 0.6 0.8 1
region which include solid block.
u/u ∞
4.3. Validation
Fig. 3. Comparison of velocity profile across a channel at Y/D = 0 and X/D = 0.

For the present model of perforated fins we were not able to


find any experimental or numerical study in three-dimensional with experimental and numerical results of [9] for the velocity pro-
condition. To validate the present numerical modeling, the result file across the channel.
of the present numerical modeling is compared with the computa- According to Fig. 3, the good agreement is illustrated between
tional and experimental results of [9]. The experimental setup in the present computation and those of [9]. Therefore the present
[9] was a 160 mm  160 mm closed-loop water channel. The computational scheme is reliable and we can continue to analyze
length, thickness and height of the rectangular obstacle were 60, perforated fins of Fig. 1.
60 and 18 mm, respectively, and it was mounted on a flat plate
placed at a distance of 130 mm from the upper wall of the channel 5. Results and discussion
and positioned at 40 mm from its leading edge. Flow Reynolds
number based on obstacle thickness and free stream velocity, With a valid code, numerical computations are carried out for
u1 ¼ 1:8 cm=s, was equal to 1000. Authors [9] reported that in solid fin and perforated fins. Perforation is made by considering
numerical modeling the distance of inlet boundary to obstacle rectangular duct along the fin length and number of channels
must be sufficient large to eliminate interactions between the changed from 1 to 8 as illustrated in Fig. 1. Flow Reynolds number
pressure perturbation associated with the obstacle and the bound- is changed from 100 to 350.
ary. They solved this problem numerically with adding a develop- To observe the treatment of flow over the fin’s side, the fluid
ment zone to upstream distance. The computational domain is stream lines around the solid and perforated fins at different height
selected a box of 300 mm  160 mm  130 mm in the X, Y and Z are shown in Fig. 4. It is visible that flow is symmetrical for both
directions respectively [9]. The leading edge of the plate was at solid and perforated fins and the recirculation region decreases
120 mm from the inlet of the channel. Other flow and geometry from top of the fin to the fin base due to viscous effect of fluid close
conditions were the same as experiment. In the present study to the base. In addition the length of wake behind the fin decreases
the computational domain is chosen the same as computational with increase in height. For perforated fins because of flow across
domain in [9]. Fig. 3 shows the comparison of the present results perforations, the shape and size of formed wake is different from
solid fin.
Table 1 Average friction coefficient and average Nusselt number over
Grid independent studies for solid fin, ReD = 250.
the faces of fins are computed by Eqs. (4) and (5) and results are
Grid Nu Cf illustrated in Figs. 5 and 6.
125  56  30 3.19 0.021 sw
140  64  34 3.22 0.022 Cf ¼ 1 ð4Þ
150  72  37 3.24 0.024 2
q u21
180  80  37 3.25 0.026 hD
210  90  50 3.25 0.027 Nu ¼ ð5Þ
k
1268 M.R. Shaeri, M. Yaghoubi / Energy Conversion and Management 50 (2009) 1264–1270

Fig. 4. Fluid path lines at various sections of different fins, ReD = 250.

In all cases, the average friction coefficient decreases with increase such a manner the solid fin has the maximum value of the average
in Reynolds number and also for any specific Reynolds number, the overall Nusselt number at each Reynolds number. Reduction of heat
value of C f is largest for the solid fin and decreases with increase in transfer coefficient from perforated fins is due smaller flow velocity,
the amount of perforations. In addition one can observes that aver- very short entrance region and laminar fully developed flow in
age Nusselt number increases with increase in Reynolds number channels of perforated fins.
and also decreases with increase in the amount of perforations. In The important design consideration of fins is rate of heat trans-
fer. For describing performance of an individual perforated fin in
comparison with solid fin at the same operation conditions, a rela-
tion as perforated fin effectiveness (PFE) is used as:
0.05 qPF  qSF
ePF ¼  100 ð6Þ
qSF
0.045
Fig. 7 shows the PFE according to Eq. (6) for all cases considered. For
low values of Reynolds number, the value of PFE changes linearly
0.04 Solid fin but for high Reynolds numbers, its value increases exponentially.
1 perforation However PFE has the positive value that indicates by use of perfo-
0.035 2 perforations rated fins, the rate of heat transfer increases in comparison with so-
Cf 3 perforations lid fin.
0.03 4 perforations The performance of perforated fins with higher amount of per-
6 perforations forations is considerable such that perforated fin with eight perfo-
0.025 8 perforations rations has the largest PFE. According to the numerical results, a
correlation that is fitted to all numerical values is obtained to de-
0.02 scribe PFE with respect to ReD and porosity (/), as follow:
100 6 ReD 6 350
0.015 ePF ¼ 8:133 þ 0:095ðReD Þ0:868 ð1  /Þ1:62 for
0:0556 6 / 6 0:444
50 100 150 200 250 300 350
ð7Þ
Re D
The value of correlation accuracy coefficient of the above relation is
Fig. 5. Variation of average friction coefficient over the faces of various types of fins. 0.947.
M.R. Shaeri, M. Yaghoubi / Energy Conversion and Management 50 (2009) 1264–1270 1269

4 6. Conclusion

3.5 Three-dimensional investigation of laminar fluid flow and con-


vective heat transfer over an array of solid and perforated fins is
3 studied numerically. Perforations are made by rectangular cross
Solid fin section and arranged along the fins attached over a heat sink. Com-
2.5 1 perforation putations for a range of Reynolds numbers from 100 to 350 are per-
2 perforations formed and the following results are obtained:
Nu 2 3 perforations
4 perforations
1. Due to perforations, the shape and position of formed wakes in
1.5 6 perforations
the back of perforated fins in any height of fin are different with
8 perforations
1
wakes for solid.
2. Average friction coefficient, pressure drop and average Nusselt
0.5 number decrease with increasing perforation and solid fin has
the largest value of C f ; sw and Nu.
0 3. For describing the operation of perforated fins, a parameter as
50 100 150 200 250 300 350 perforated fin effectiveness (PFE) is defined that shows the
Re D enhancement of heat transfer rate from individual perforated
fin in comparison with corresponding solid fin. The values of
Fig. 6. Variation of average Nusselt number over the faces of various types of fins. PFE are always positive that show an increase in the heat trans-
fer rate due to use of perforated fins. For low Reynolds numbers
the operation of perforated fins are alike the solid fin. For higher
Reynolds numbers and in perforated fins with more number of
40
perforations, the rate of enhancement of heat transfer is consid-
erable. Finally a correlation for PFE is derived:
35

30
ePF ¼ 8:133 þ 0:095ðReD Þ0:868 ð1  /Þ1:62

25
1 perforation 4. One of the most important benefits of utilization of perforated
2 perforations fins is reduction of fin’s weight. Low weight certifies saving
3 perforations material of fins and related equipments such as heat sinks.
ε PF 20
4 perforations From the results, practically one goal of fin optimization for
6 perforations low Reynolds numbers and both goals of fin optimization for
15
8 perforations higher Reynolds numbers are achieved.
10
References
5
[1] Bergles AE. The implications and challenges of enhanced heat transfer for the
0 chemical process industries. Chem Eng Res Des 2001;79:437–44.
[2] Kern QD, Kraus DA. Extended surface heat transfer. New York: McGraw-Hill;
50 100 150 200 250 300 350
1972.
Re D [3] Malekzadeh P, Rahideh H, Karami G. Optimization of convective-radiative fins
by using differential quadrature element method. Energ Convers Manage
2006;47:1505–14.
Fig. 7. Perforated fin effectiveness. [4] Suksangpanomrung A, Chungpaibulpatana S, Promvonge P. Numerical
investigation of heat transfer in pulsating flows through a bluff plate. Int
Commun Heat Mass Transfer 2007;34:829–37.
[5] Asako Y, Faghri M. Three-dimensional heat transfer and fluid flow analysis of
rectangular blocks encountered in electronic equipment. Num Heat Transfer
Table 2 1988;13:481–98.
Percentage of weight reduction of solid fin due to creation of perforations. [6] Asako Y, Faghri M. Three-dimensional heat transfer analysis of arrays of heated
square blocks. Int J Heat Mass Transfer 1989;32:395–405.
Number of perforations 1 2 3 4 6 8 [7] Holahan MF, Kang SS, Bar-Cohen AA. Flow stream based analytical
model for design of parallel plate heat sinks. In: Proceedings of the 31st
Percentage of weight reduction 5.56 11.11 16.67 22.22 33.33 44.44 ASME national heat transfer conference, USA 1996, vol. 329, No. 7, p. 63–
71.
[8] Baskaya S, Sivrioglu M, Ozek M. Parametric study of natural convection heat
Inserting perforations in solid fin, the weight of fin decreases transfer from horizontal rectangular fin arrays. Int J Therm Sci
and this is the second goal of fin optimization. Table 2 shows the 2000;39:797–805.
[9] Rouvreau S, David L, Calluaud D, Joulain P. Laminar junction flow at low
percent of reduction of fin’s weight by considering perforations. Reynolds number: influence of the upstream region on the comparison
The optimization of fins can be made by either minimizing the between experiments and calculations. C. R. Mecanique 2005;333:265–72.
volume (weight) for a required heat dissipation or maximizing the [10] Saha AK. Three-dimensional numerical study of flow and heat transfer from a
cube placed in a uniform flow. Int J Heat Fluid Flow 2006;26:80–94.
heat dissipation for any given fin volume (weight) [3]. It is clear [11] van Dijk A, de Lange HC. Compressible laminar flow around a wall – mounted
that reduction of weight by increase in the number of perforations cubic obstacle. Comput Fluids 2007;36:949–60.
is considerable. As a result of high reduction in weight, cost reduc- [12] Sahin B, Demir A. Thermal performance analysis and optimum design
parameters of heat exchanger having perforated pin fins. Energ Convers
tion of manufacturing of fins and related equipment such as heat Manage 2008;49:1684–95.
sinks will be highly remarkable. The benefit of this weight reduc- [13] Fujii M, Seshimo Y, Yamanaka G. Heat transfer and pressure drop of perforated
tion is accompanied of heat transfer enhancement by considering surface heat exchanger with passage enlargement and contraction. Int J Heat
Mass Transfer 1988;31(1):135–42.
the present slots along the fins.
[14] Molki M, Hashemi-Esfahanian A. Turbulent convective mass transfer
Accordingly by use of perforated fins, especially with more per- downstream of a perforated baffle blockage. Int J Heat Fluid Flow
forations, both goals of fin’s optimization will be improved. 1992;13(2):116–23.
1270 M.R. Shaeri, M. Yaghoubi / Energy Conversion and Management 50 (2009) 1264–1270

[15] Sara ON, Pekdemir T, Yapici S, Ersahan H. Thermal performance analysis for [17] Leung CW, Probert SD. Heat exchanger performance: effect of orientation. Appl
solid and perforated blocks attached on a flat surface in duct flow. Energy Energ 1989;33:35–52.
Convers Manage 2000;41:1019–28. [18] Velayati E, Yaghoubi M. Numerical study of convective heat transfer from an
[16] Dorignac E, Vullierme JJ, Broussely M, Foulon C, Mokkadem M. Experimental array of parallel bluff plates. Int J Heat Fluid Flow 2005;26:80–91.
heat transfer on the windward surface of a perforated flat plate. Int J Therm Sci [19] Patankar SV. Numerical heat transfer and fluid flow. Washington
2005;44:885–93. DC: Hemisphere; 1980.

Вам также может понравиться