International Journal of Heat and Mass Transfer 55 (2012) 54495458

Contents lists available at SciVerse ScienceDirect

International Journal of Heat and Mass Transfer

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

Numerical study of heat-transfer enhancement by punched winglet-type

vortex generator arrays in n-and-tube heat exchangers
Y.L. He a,, H. Han a, W.Q. Tao a, Y.W. Zhang a,b
a
Key Laboratory of Thermo-Fluid Science and Engineering of MOE, School of Energy and Power Engineering, Xian Jiaotong University, Xian, Shaanxi 710049, China
b
Department of Mechanical and Aerospace Engineering, University of Missouri, Columbia, MO 65211, USA

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

Article history: The potential of punched winglet type vortex generator (VG) arrays used to enhance air-side heat-trans-
Received 25 November 2011 fer performance of nned tube heat exchanger is numerically investigated. The arrays are composed of
Received in revised form 26 April 2012 two delta-winglet pairs with two layout modes of continuous and discontinuous winglets. The heat
Accepted 29 April 2012
transfer performance of two array arrangements are compared to a conventional large winglet congu-
Available online 18 June 2012
ration for the Reynolds number ranging from 600 to 2600 based on the tube collar diameter, with the
corresponding frontal air velocity ranging from 0.54 to 2.3 m/s. The effects of different geometry param-
Keywords:
eters that include attack angle of delta winglets (b = 10 deg, b = 20 deg, b = 30 deg) and the layout loca-
Fin-and-tube heat exchanger
Vortex generator array
tions are examined. The numerical results show that for the punched VG cases, the effectiveness of the
Heat transfer enhancement main vortex to the heat transfer enhancement is not fully dominant while the corner vortex also shows
signicant effect on the heat transfer performance. Both heat transfer coefcient and pressure drop
increase with the increase of attack angle b for the side arrangements; the arrays with discontinuous
winglets show the best heat transfer enhancement, and a signicant augmentation of up to
33.870.6% in heat transfer coefcient is achieved accompanied by a pressure drop penalty of 43.4
97.2% for the 30 deg case compared to the plain n. For the front arrangements of VGs higher heat trans-
fer enhancement and pressure drop penalty can be obtained compared to that of the side arrangement
cases; the case with front continuous winglet arrays has the maximum value of j/f, a corresponding heat
transfer improvement of 36.781.2% and a pressure drop penalty of 60.7135.6%.
2012 Elsevier Ltd. All rights reserved.

1. Introduction streamwise vortices are generated in the ow eld due to ow sep-

aration on the leading edge of vortex generators (VGs), causing
Fin-and-tube heat exchangers are widely used in various engi- bulk ow mixing, boundary-layer modication, and ow destabili-
neering elds, such as heating, ventilation, air conditioning, and zation; heat transfer is enhanced due to these vortices. The longi-
refrigeration (HVACR) systems. High heat exchanger performance tudinal vortex generators applied in various heat exchangers
is very important in meeting efciency standards with low cost have received considerable attentions for the advantage of heat
and environmental impact. For liquid-to-air and phase-change transfer enhancement, accompanied by a modest pressure drop
heat exchangers that are typical for HVAC&R systems, the air side penalty.
convection resistance is usually dominant due to the thermophys- Fiebig et al. [4] experimentally investigated three tube rows
ical property of air. Thus many efforts have been made to enhance heat exchanger element with delta winglets. They reported a
air-side heat transfer performance and variants of n patterns like 5565% heat transfer augmentation with corresponding 2045%
wave, louver and slit n have been adopted [13]. However, with increase in the apparent friction factor for an inline arrangement.
signicant heat transfer enhancement, the associated penalty of Biswas et al. [5] performed numerical investigations of ow struc-
pressure drop is also tremendous for those conventional heat ture and heat transfer characteristics in a channel with a built-in
transfer enhancement methods. tube and a winglet type vortex generator. Fiebig et al. [6] measured
In recent years, a very promising strategy of enhancing air-side a 10% local heat transfer enhancement for round tubes but a 100%
heat transfer performance is using ow manipulator, known as enhancement for at tubes under certain geometry. Deb et al. [7]
vortex generators. When the uid ows through vortex generators, numerically investigated heat transfer characteristics and ow
structure in laminar and turbulent ows through a rectangular
Corresponding author. Tel.: +86 29 8266 5930; fax: +86 29 8266 5445. channel containing built-in vortex generators. Jacobi and Shah
E-mail address: [email protected] (Y.L. He). [8] reviewed vortex-induce heat transfer enhancement and

0017-9310/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2012.04.059
5450 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458

Nomenclature

Ac minimum ow cross-sectional area (m2) U velocity vector (m s1)

Af n surface area (m2) x, y, z Cartesian coordinates (m)
Afr frontal area (m2) Z heat transfer power per unit temperature and per unit
A0 total surface area (m2) volume (Wm3 K1)
Dc tube outer diameter (m)
Dh hydraulic diameter, 4AcL/A0 (m) Greek symbols
E fan power per unit core volume (W m3) b attack angle of the delta winglet (deg)
Fp n pitch (m) df n thickness (m)
f friction factor l dynamic viscosity (kg m1 s1)
h heat transfer coefcient (W m2 K1), or the height of q density (kg m3)
the delta winglet (m) k thermal conductivity (W m1 K1)
j Colburn factor r Ac/Afr, contraction ratio of cross-sectional area
L n length along the main ow direction (m) gf n efciency
l chord length of the large winglet (m) g0 surface efciency
Nu Nusselt number
p pressure (Pa) Subscripts
Pl longitudinal tube pitch (m) a air
Pt transverse tube pitch (m) f n
Pr Prandtl number in inlet
Re Reynolds number based on tube collar diameter m mean
Q heat transfer rate (W) out outlet
T temperature (K) w tube wall
u, v, w velocity components in x-, y-, z-directions (m s1)
um mean velocity at the minimum ow cross-sectional area
(m s1)

discussed the theoretical basis for the method. Biswas et al. [9] The foregoing literature reviews show that most works in vor-
found that the ow behind a winglet type vortex generator is con- tex generators have employed ideal geometries, and generally
sists of a main vortex, a corner vortex and induced vortices. Lee the number of winglet-type VGs in each row tube dose not exceeds
et al. [10] carried out a numerical study of heat transfer character- one pair. One general thought is that more VGs pairs can be added
istics and turbulent structure in a three-dimensional boundary in each row tube to explore the true potential of the vortex-
layer with longitudinal vortices. enhancement strategy, because the associated pressure drop pen-
Chen et al. [11] numerically investigated heat transfer and ow alty of using winglet type VGs is relatively lower than that of the
in a oval-tube heat transfer element in both in-line and staggered louvered n or slit n. Chen et al. [21] compared the thermo-
arrangements with punched winglets type vortex generators. hydraulic performance of nned oval tube heat exchanger element
Wang et al. [12] utilized a dye-injection technique to visualize with one to three rows of winglets. If more winglet-type VGs were
the ow structure for enlarged plain n-and-tube heat exchanger employed, another question that needs to be answered is that how
with annular and delta winglet vortex generators. Torri et al. [13] we can organize the winglets to achieve the best performance.
proposed a common-ow-up arrangement strategy which can aug- More recently, He et al. [22] proposed a VG array deployed in a
ment heat transfer while reducing pressure penalty in a n-and- V-like conguration by imitating the formations of animals group
tube heat exchanger at relatively low Reynolds number. Leu et al. movement and their experiments showed that this conguration
[14] numerically and experimentally studied the heat transfer only caused very low pressure drop penalty in the case of nearly
and ow in the plate-n and tube heat exchangers with inclined heat transfer performance. However VGs were mounted on the
block shape vortex generators mounted behind the tubes. Wu n surface in their experiment, not punched. In the practical heat
and Tao [15] conducted a numerical simulation to investigate the exchanger design, punched winglets may be easier to be adopted.
heat transfer performance of a rectangular channel with a pair of Hirokazu [23] indicated that continuous punched hole can lead
rectangular winglets VGs punched out from the wall of the chan- to reduction of n efciency, thus the impact of the punching
nel. They analyzed the results from the view of eld synergy prin- should be assessed. Moreover, if the back-row small winglet adja-
ciple [1618]. Joardar and Jacobi [19] experimentally assessed the cent to the preceding one in a VG array, the vortices generated by
potential of winglet-type vortex generator arrays for multi-row different winglets may affect each other and the interactions of the
inline-tube heat exchangers. They found that the air-side heat vortices should be considered.
transfer coefcient increased from 16.5% to 44% for the single- In this work, numerical study of n-and-tube heat exchanger
row winglet arrangement with an increase in pressure drop of less with two kinds of VG arrays and a conventional VG conguration
than 12%; for the three-row vortex generator array the heat trans- in common-ow-up arrangement are performed. For two different
fer coefcient increases from 29.9% to 68.8% with a pressure drop VG arrays, the effect of VG spacing is investigated. Detailed study
penalty from 26% to 87.5%. Joardar and Jacobi [20] also numerically was focused on the complex ow and heat transfer interactions
investigated the ow and heat transfer enhancement using an ar- resulting from employing delta winglet VGs. Meanwhile, effects
ray of VGs in a n-and-tube exchanger with common-ow-up of punching, attack angle and placement locations of delta winglet
arrangement. They observed that the impingement of winglet redi- on the uid ow and heat transfer characteristics were examined.
rected ow on the downstream tube is an important heat transfer Two different performance evaluation criteria (PEC) of area-good-
augmentation mechanism for the inline-tube geometry. ness factor and volume-goodness factor are employed to assess
 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458 5451

(a) Schematic of core region of a plain fin-and-tube heat exchanger

135 135 135

6 mm

6 mm

6 mm
2 mm

air flow 10.2 5.115 mm 5.11
3m 5mm
m

Large winglet Array-A Array-B

(b) Large winglet and two kinds of delta winglet arrays geometry and placements
Fig. 1. Schematic diagram of n-and-tube heat exchanger with three kinds of VG placements.

the heat transfer and pressure drop performance of n-and-tube

0.13 mm

heat exchanger.

3.2 mm
2. Model descriptions y

2.1. Physical model x periodic

Fig. 1(a) shows the schematic diagram of core region of a plain (a) Side view
n-and-tube heat exchanger with two-rows tubes along the ow
direction. The tube outside diameter Dc is 10.23 mm, the trans- 22 mm
verse tube pitch Pt is 25.4 mm, the longitudinal tube pitch Pl is
22 mm, the n pitch Fp is 3.2 mm, and the n thickness df is
0.13 mm. The delta winglet pairs are punched out from the n sur-
face and placed symmetrically on both sides of each round tube in
an inline arrangement. The base chord length l and height h of large outflow
winglet are 10.23 mm and 2.56 mm, respectively. The attack angle inlet
b of VG, in this investigation has three different values (b = 10 deg,
25.4 mm

b = 20 deg and b = 30 deg) with respect to the coming ow direc- x

tion. The VG arrays are composed of two small delta winglets; each symmtry
z
delta winglet has the same height, same placement position, and
half chord length of the large winglet. For discontinuous winglet
array 2 mm spacing between winglets can be observed. For conve-
nience, array-A and array-B are used to represent the continuous (b) Top view
winglet VG array and discontinuous winglet VG array, respectively.
Fig. 2 presents the computational domain for n-and-tube heat Fig. 2. Computational domain and coordinate system.
exchanger with delta winglets. The side and top views of the com-
putational domain are shown in Fig. 2(a) and (b), respectively. Due
to symmetry, the region outlined by the dashed lines in Fig. 2 is se- 2.2. Governing equations and boundary conditions
lected as the computational domain, and the two neighboring ns
centric surface are taken as the upper and lower boundaries of the The uid is considered to be incompressible with constant ther-
computational domain. The actual computational domain is ex- mophysical properties. Due to the low air inlet velocity and the
tended ten times of the n spacing from the entrance to ensure a small n space, the ow in the n channel of compact heat exchan-
uniform velocity distribution. At the exit, 30 times of n spacing ger is assumed to be laminar and steady state. The thermal contact
is also extended downstream to avoid recirculation and ensure resistance between the tube and n collar is ignored, and the tube
the outow boundary condition can be applied [24]. surface is set as constant temperature. Effects of n thickness on
5452 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458

the heat conduction in the ns are taken into account. The temper- Table 1
ature distribution in the ns will be determined by solving the con- Results of different grid numbers.

jugated heat transfer problem between uid and solid in the Grid no. 600,000 840,000 1170,000
computational domain. The governing equations and boundary Nu 24.00 23.81 23.74
conditions for numerical simulation in this paper can be found in f 0.0452 0.0456 0.0456
Ref. [24] and will not be restated here for brevity.

2.3. Numerical methods and grid independence validation performance. The area goodness factor is dened as j/f, while
the volume goodness factor is described as follow:
The computational uid code FLUENT is used to solve the  
NavierStokes and energy equations with the corresponding 4r _ a 4r
m
Z g0 h ; E Dp 8
boundary conditions. The multi-block hybrid method is adopted Dh A0 q D h
to generate the computational grid, and structured grids are pref-
erential where it is feasible. To improve the accuracy of the simu- where Z represents the heat transfer per unit temperature differ-
lation results, the grids around the tubes and vortex generators are ence and per unit core volume and E represents the fan power per
rened with unstructured grids as shown in Fig. 3. Grid indepen- unit core volume.
dence is examined to ensure the accuracy and reliability of the
numerical methods. Three different grid systems, which include 3. Validation of model and numerical method
about 600,000, 840,000 and 1170,000 hybrid cells, are adopted
for the case of array-B at attack angle b = 30 deg and Re = 2600. In order to validate the reliability of the numerical method,
The difference in averaged Nusselt number for the three grid sys- numerical simulation is conducted on a n-and-tube heat transfer
tems is less than 3%, as shown in Table 1. The nal adopted grid exchanger with the same tube diameter as presented in Ref. [26].
number is 840,000. Similar validations are also performed for the The inlet air velocity ranges from 0.54 to 2.33 m/s and the corre-
large winglet and array-A cases. sponding Reynolds number ranges from 600 to 2600. The Colburn
j-factor and friction factor are shown in Fig. 4. The maximum dis-
2.4. Parameter denitions crepancy between the predicted Colburn j-factor and the experi-
mental results is less than 9.5% and the maximum discrepancy
Some parameters are dened as follows: between the predicted friction factor and experimental values is
less than 10%. Almost all the simulation results located within
qum Dc
Re 1 the permitted error range of experiments. The excellent agreement
l
between the numerical results and experimental data indicates the
Q present numerical results are reliable to predict the heat exchanger
h 2
A0 DT g0 performance.
A
g0 1  f 1  gf 3
A0
4. Results and discussion
hDc
Nu 4
k 4.1. Flow pattern and heat transfer
Dp pin  pout 5
h The delta winglets are punched out from the heat transfer ns
j Pr2=3 6
qum cp surface. When the air ows through the winglets, the longitudinal
Dp Ac vortices were generated due to the pressure difference between
f 1  7 the front surface and the back surface of the winglets. To investi-
2
qu2m A0
gate the effect of punched vortex generators on the uid ow
where um is the uid velocity at the minimum cross-section of the structure of the uid in n channel, taking the array-B case with
tube row, Dc is the hydraulic diameter based on the n collar out- the attack angle b = 30 deg as an example, the path line is numer-
side diameter, A0 is the total heat transfer area, Ac is the minimum ically generated for Re = 2600 in Fig. 5. When one uid parallel to
free ow area, and gf is obtained using Schmidts method [25]; and the n channel (xz plane) encounters the punched delta winglet,
DT is the log-mean temperature difference. part of the uid ows over the winglet (from y-direction) and
The PECs of area goodness factor and volume goodness forms the main vortex, another part of uid ows through the
factor are used to evaluate the heat transfer and pressure drop punching hole, mixes with the uid bypassing the winglet (from

Fig. 3. Typical computational mesh.

 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458 5453

0.030 Induced vortex Corner vortex Main vortex

Region behind tube
Experimental
0.025
Numerical

0.020

0.015
j

(a) Large winglet

0.010
Corner vortex Main vortex Induced vortex
Region behind tube
0.005

0.000
500 1000 1500 2000 2500 3000
Re
(a) Couburn j factor

(b) Array-A
0.08
Experimental
Numerical Induced vortex Corner vortex Main vortex
Region behind tube
0.06
f

0.04

0.02
(c) Array-B

0.00 Fig. 6. Local velocity vectors of secondary ow on the cross section (x = 20 mm).
500 1000 1500 2000 2500 3000
Re
(b) Friction factor f the same rotation axis with the mainstream (x-direction). In gen-
eral, the strength of corner vortex is weaker than the main vortex,
Fig. 4. Comparison of numerical and experimental results. but the opposite conclusion can be observed from the partially en-
larged view in Fig. 5, which looks different but reasonable. The
main reason can be attributed to that the present winglets are
z-direction) from the trailing edge and forms the corner vortex. punched, not mounted or attached in the n surface. A consider-
The main vortex, located directly downstream of the delta winglet, able portion of uid does not turn over the winglet, but streams
is formed by ow separation at the leading edge of the winglet, through the punching hole, which reduces the pressure difference
while the corner vortex is usually formed by the deformation of caused by winglet and also reduces the quantity of uid to form
the near wall vortex lines at the pressure side of the winglet. How- the main vortex.
ever, the corner vortex here is different from the conventional Then a detailed study of the ow structure is carried out to get a
corner vortex described by Biswas et al. [9], and shows more char- better understanding of the role of VGs. Fig. 6 presents the local
acteristics of main vortex and stronger disturbance to uid due to velocity vectors of secondary ow on the cross section

Main vortex

Winglet

Tube

Corner vortex
Fin

Fig. 5. Path lines around the delta winglet array.

5454 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458

plain fin
70 large winglet
array A
array B
60

-1
h Wm K
-2
(a) Large winglet 50

40

30

500 1000 1500 2000 2500

Re
(a) Heat transfer coefficient versus Re
(b) Array-A
25
plain fin
large winglet
20 array A
array B

15

(c) Array-B P (Pa)

10
Fig. 7. Local temperature distribution on the cross section (x = 20 mm).

5
(x = 20 mm) at Re = 2600. The vortices behind the large winglet are
typical with a main vortex, a corner vortex and an induced vortex.
Compared to the large winglet case, the swirling motion associated 0
500 1000 1500 2000 2500
with the VG arrays is obviously more complicated. Two pairs of Re
main vortices and corner vortices can be observed. The main vorti-
(b) Pressure drop versus Re
ces are located in the upper side of the n channel and the corner
vortices are located in the lower side. The generated vortices Fig. 9. Heat transfer coefcient and pressure drop versus Re.
squeeze each other, and the geometry dimension and the inu-
enced range of the vortices are smaller than that of large winglet
case. More vortices with smaller dimension make the local uid distribution on the cross section (x = 20 mm). All VGs arrange-
mixing stronger and more uniform. In addition, for array-A ments obviously change the temperature distribution in the n
arrangement, an induced vortex appears in the center of two pairs channel, and signicant distortion occurred in the thermal bound-
of counter-rotating vortices; the existence of induced vortex may ary layer due to the strong swirling ow effect of mainstream vor-
not enhance the local heat transfer of the n channel but will tex. In the up-wash region of the mainstream vortex the thermal
weaken mainstream vortex and causes extra pressure drop. boundary becomes thicker while in the down-wash region the
To further investigate the inuence of uid ow on the heat thermal boundary shows the opposite tendency. Moreover, the
transfer performance, Fig. 7 shows the local temperature mainstream vortex also enhanced the mixing of the uid, espe-
cially in the wake region of the tube. For the large winglet case
shown in Fig. 7(a), the disturbance of corner vortex to the uid
50 downstream is obviously stronger than that caused by the main
45 large winglet vortex. For the two kinds of VG arrays shown in Fig. 7(b) and (c),
array A the main vortex produced by the back-row small winglet shows
40 array B little impact on isotherm distribution, and the corner vortex in-
35 duced by the front-row small winglet of array-B has a greater effect
on the isotherm distribution than that of array-A. Compared to the
30 large winglet case, the array arrangement can effectively change
Nu

25 local temperature distribution and enhance the local heat transfer,

but the effect to the wake region is relatively weak.
20 In order to quantitatively analyze the effects of different VGs
15 arrangements on the heat transfer behavior along the ow direc-
tion, the span averaged local Nusselt number distributions for
10 Tube 1 Tube 2 three VG cases are shown in Fig. 8. The averaged local heat transfer
5 shows periodicity in the stream-wise direction. Two kinds of array
0 5 10 15 20 25 30 35 40 45
cases present better heat transfer performances at the axial loca-
x (mm)
tion of tubes because of more uniform cross sectional temperature
Fig. 8. Distribution of spanwise averaged local heat transfer along the main ow distribution, while the large winglets presents better heat transfer
direction at Re = 2600. performance in the wake region due to strong mixing effects
 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458 5455

0 38. 80
plain fin
plain fin large winglet - 10
0.36
large winglet 70 array A - 10
array B - 10
array A
0.34 large winglet - 20
array B array A - 20
60

-1
h Wm K
array B - 20
0.32 large winglet - 30

-2
j/f

array A - 30
0.30 50 array B - 30

0.28
40

0.26
30
0.24
500 1000 1500 2000 2500 500 1000 1500 2000 2500
Re Re
(a) Area goodness factor (a) Heat transfer coefficient

4
4.5x10 20 plain fin
plain fin
large winglet - 10
large winglet
4 array A - 10
4.0x10 array A array B - 10
array B 15 large winglet - 20
4
-1

3.5x10 array A - 20
Z Wm K

P (Pa)
array B - 20
-3

large winglet - 30
4
3.0x10 10 array A - 30
array B - 30
4
2.5x10
5
4
2.0x10

2 3 500 1000 1500 2000 2500

10 10
-3 Re
E Wm (b) Pressure drop
(b) Volume goodness factor
Fig. 11. Effects on heat transfer coefcient and pressure drop at different attack
Fig. 10. Area goodness and volume goodness factors for different VG arrangements angles.
at attack angle b = 30 deg.

Fig. 10 presents the area goodness and volume goodness factors

for three different VGs arrangements compared with the plain n
between the main ow and tube wake region. For the array case. The j/f ratio of the cases of array-A and array-B are lower than
arrangements, two peaks of Nusselt number on each tube can be the plain-n case as shown in Fig. 10(a). The case of large winglet
clearly observed. Compared to array-A case, the peak value of Nus- shows the highest value of j/f in the three different VGs arrange-
selt number for array-B arrangement is higher and the position is ments. Another criterion is the volume goodness factor where
closer to the upstream. In the whole ow domain array-A shows the heat transfer power per unit temperature and per unit volume
the worst heat transfer capability due to the mutual interference (Z) is plotted against the fan power per unit core volume (E). From
and offset of vortices generated by the two adjacent small winglets. the viewpoint of the heat exchanger volume required, geometries
Fig. 9(a) presents the variation in the air-side heat transfer coef- with larger value of Z will require less core volume for a given
cient h versus Reynolds number. With a higher value of Reynolds air-side heat transfer capacity. It is evident that for a given E value,
number, a global augmentation in heat transfer can be observed. all the three arrangements with VGs have higher Z value than the
Among all three VGs layouts, the array-B case presents the best plain n case, and the large winglets case and array-B case have
heat transfer performance due to the best ow organization and relatively higher Z value than the array-A case.
the front delta winglets of the array-B are placed closer to up-
stream position relative to the other layouts; the large winglets 4.2. Effect of attack angle
shows relatively lower pressure penalty which can be explained
as a better wake region management and reducing the form drag Effects of attack angle of delta winglets for three different VG
by introducing high momentum uid to the wake and delaying arrangements are investigated and the attack angle has three dif-
the uid ow separation on the tube surface. Comparing to the ferent values of 10 deg, 20 deg, and 30 deg. All other geometrical
plain n at the range of Reynolds number of the present study, parameters of n-and-tube heat exchanger remain the same as
the heat transfer coefcient and pressure drop increase by the baseline case. Fig. 11 shows the effects of the attack angle on
28.768.3% and 35.274.6%, respectively, for the case of large heat transfer coefcient and pressure drop penalty against the Rey-
winglets; the heat transfer coefcient and pressure drop increase nolds number. The heat transfer coefcient and pressure drop in-
by 29.959.1% and 42.689.6%, for the case of array-A; the heat crease with the attack angle increase. For different attack angles,
transfer coefcient and pressure drop increase by 33.870.6% and the array-B arrangement always has the highest value of heat
43.497.2%, respectively, for the case of array-B. transfer coefcient and pressure drop. The array-A arrangement
5456 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458

m m
15
1 mm
m 5. 1
2m

30
10
.23

1mm
mm
air flow 5.115 mm

Large winglet Array-A Array-B

Fig. 12. Schematic diagram of front VG congurations.

0.36
80 plain fin plain fin
0.34 large winglets
large winglet
arrya A
70 array A
0.32 array B
array B
-1

60 0.30
h Wm K
-2

j/f
0.28
50
0.26

40 0.24

0.22
30
0.20
500 1000 1500 2000 2500 500 1000 1500 2000 2500
Re Re
(a) Heat transfer coefficient (a) Area goodness factor

30 50000
plain fin
45000 plain fin
large winglet
25 large winglet
array A
array A
array B 40000
array B
20
-1
Z Wm K

35000
P (Pa)

-3

15
30000

10
25000

5 20000

0 10
2 3
10
500 1000 1500 2000 2500
-3
Re E Wm
(b) Pressure drop (b) Volume goodness factor

Fig. 13. Heat transfer coefcient and pressure drop for front VG congurations. Fig. 14. Area goodness and volume goodness factors for front VG congurations.

shows better heat transfer performance than the large winglet case
n-and-tube heat exchanger. For the ns with three different
when the attack is less than 20 deg, while the large winglet case
VGs mentioned above, a similar geometry is adopted, reference
shows better for 30 deg attack angle. It should also be noted that
to the design of He et al. [22]. The distance between the leading
the large winglet case always has the lowest pressure drop penalty
edge of the n and the tip of the winglets is 1 mm; the transverse
for different VG arrangements and the two kinds of VGs array
spacing between the tips of two adjacent winglets is also 1 mm;
arrangements have relatively higher pressure loss.
the attack angle of VGs is selected as 30 deg.
Fig. 13 shows the variations in heat transfer coefcient h and the
4.3. Inuence of placement positions pressure drop Dp against Reynolds number. Compared to the plain-
n, the heat transfer coefcient and pressure drop increase by 38.9
In order to explore a more effective placement of VGs for the 86.3% and 70.3149.2%, for the case of front large winglet case;
aforementioned three different VG n patterns, a more upstream 36.781.2% and 60.7135.6% for the case of front array-A; and
placement is employed since the longitudinal vortices generated 30.684.1% and 62.4146.7%, for the case of front array-B, respec-
by VGs can affect a larger downstream heat transfer domain. tively. By comparing Figs. 9 and 13, it is clear that using the front
Fig. 12 shows the schematic of the front arrangements of VGs for VG arrangements can get greater heat transfer argumentation and
 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458 5457

also higher pressure drop penalty. The results of front VG arrange- (5) The front arrangements of VGs can achieve larger heat transfer
ments are obviously different from that of side VG arrangements, enhancement and pressure penalty. Among all front VG cases,
and the large winglet has the best heat transfer performance and the array-A arrangement shows best ratio value of j/f and large
also highest pressure drop. The array-A case shows higher value winglet case shows best performance in volume compactness.
of heat transfer coefcient at a lower range of Reynolds number Furthermore, it is reasonable to believe that front arrange-
than array-B, while the array-B performs better when the Reynolds ments can be more effective in low Reynolds number regions.
number is above 1500 as shown in Fig. 13(a). It can be explained
that the array-A case places closer to upstream which is benet to
enhance the air-side heat transfer, but the continuous winglets Acknowledgements
arrangement will degrade the heat transfer performance. In the
lower Reynolds number range, the effect of former is stronger while The present work is supported by the Joint Funds of NSFC-
with the increase of Reynolds number the effect of latter becomes Guangdong of China under Grant (No. U0934005) and National Ba-
larger and larger. Thus it can be thought that a placement of VGs sic Research Program of China (973 Program) (2011CB710702).
closer to the upstream can get higher heat transfer augment in
low Reynolds number regions, which can also be understood from References
the view of boundary layer. The thickness of boundary at one loca-
[1] D.T. Beecher, T.J. Fagan, Effects of n pattern on the air-side heat transfer
tion is inversely proportional to Reynolds number and the bound- coefcient in plate nned-tube heat exchangers, ASHRAE Trans. 93 (1987)
ary layer is growing rapidly along the ow direction under low 19611984.
Reynolds numbers. It is also worth noticing that the front array-A [2] C.C. Wang, C.J. Lee, C.T. Chang, S.P. Lin, Heat transfer and friction correlation for
compact louvered n-and-tube heat exchangers, Int. J. Heat Mass Transfer 42
case always has the lowest pressure drop among all the front VG (1999) 19451956.
cases in the Reynolds number range considered. [3] J.Y. Yun, K.S. Lee, Investigation of heat transfer characteristics on various kinds
The area goodness and volume goodness factors of the plain n of n-and-tube heat exchangers with interrupted surfaces, Int. J. Heat Mass
Transfer 42 (1999) 23752385.
with three different kinds of VG arrangements are presented in [4] M. Fiebig, A. Valencia, N.K. Mitra, Wing-type vortex generators for n-and-tube
Fig. 14. With the increase of Reynolds number, the ratios of j/f grad- heat-exchangers, Exp. Thermal Fluid Sci. 7 (1993) 287295.
ually reduce. Among all VG arrangements, array-A has the highest [5] G. Biswas, N.K. Mitra, M. Fiebig, Heat-transfer enhancement in n-tube heat-
exchangers by winglet type vortex generators, Int. J. Heat Mass Transfer 37
ratio of j/f; the large winglet case and array-B have similar ratio of (1994) 283291.
j/f. Fig. 14(b) presents the volume goodness factor for different n [6] M. Fiebig, A. Valencia, N.K. Mitra, Local heat transfer and ow losses in n-and-
patterns, and it can be seen that the large winglets case has a tube heat exchangers with vortex generators: a comparison of round and at
tubes, Exp. Thermal Fluid Sci. 8 (1994) 3545.
slightly higher value than other arrangements. Additionally, the ar- [7] P. Deb, G. Biswas, N.K. Mitra, Heat-transfer and ow structure in laminar and
ray-A shows better performance in a lower range of Reynolds num- turbulent ows in a rectangular channel with longitudinal vortices, Int. J. Heat
ber, in contrast the array-B has a better performance in a relatively Mass Transfer 38 (1995) 24272444.
[8] A.M. Jacobi, R.K. Shah, Heat-transfer surface enhancement through the use of
higher Reynolds number range.
longitudinal vortices a review of recent progress, Exp. Thermal Fluid Sci. 11
(1995) 295309.
5. Conclusions [9] G. Biswas, K. Torri, D. Fujii, K. Nishino, Numerical and experimental-
determination of ow structure and heat-transfer effects of longitudinal
vortices in a channel ow, Int. J. Heat Mass Transfer 39 (1996) 34413451.
In present paper, three dimensional numerical simulations are [10] S.H. Lee, H.S. Ryou, Y.K. Choi, Heat transfer in a three-dimensional turbulent
performed to investigate the heat transfer characteristics and ow boundary layer with longitudinal vortices, Int. J. Heat Mass Transfer 42 (1999)
15211534.
structure in a n-and-tube heat exchanger with delta winglet vor- [11] Y. Chen, M. Fiebig, N.K. Mitra, Heat transfer enhancement of nned oval tubes
tex generators. Two kinds of VGs array arrangements and a con- with staggered punched longitudinal vortex generators, Int. J. Heat Mass
ventional arrangement are comprehensively and comparatively Transfer 43 (2000) 417435.
[12] C.C. Wang, L. Jerry, Y.T. Lin, C.S. Wei, Flow visualization of annular and delta
investigated. The major conclusions are drawn as follows:
winlet vortex generators in n-and-tube heat exchanger application, Int. J.
Heat Mass Transfer 45 (2002) 38033815.
(1) For the punched VGs, the corner vortex shows the charac- [13] K. Torii, K.M. Kwak, K. Nishino, Heat transfer enhancement accompanying
pressure-loss reduction with winglet-type vortex generators for n-tube heat
teristics of main vortex, and plays a more important role on
exchangers, Int. J. Heat Mass Transfer 45 (2002) 37953801.
the uid ow behavior and heat transfer characteristic in the [14] J.S. Leu, Y.H. Wu, J.Y. Jang, Heat transfer and uid ow analysis in plate-n and
channel of n-and-tube heat exchanger than that of the tube heat exchangers with a pair of block shape vortex generators, Int. J. Heat
main vortex. Mass Transfer 47 (2004) 43274338.
[15] J.M. Wu, W.Q. Tao, Numerical study on laminar convection heat transfer in a
(2) The VG arrays generate more vortices and the vortices inu- rectangular channel with longitudinal vortex generator, Part A: verication of
ence each other which weaken the swirling movement of eld synergy principle, Int. J. Heat Mass Transfer 51 (2008) 11791191.
the uid, especially for the continuous small winglet array. [16] Y.L. He, W.Q. Tao, F.Q. Song, W. Zhang, Three-dimensional numerical study of
heat transfer characteristics of plain plate n-and-tube heat exchangers from
In addition, the vortices generated by the large winglets view point of eld synergy principle, Int. J. Heat Fluid Flow 26 (2005) 459473.
can affect the whole region of n channel height (y-direc- [17] Z.Y. Guo, W.Q. Tao, R.K. Shah, The eld synergy (coordination) principle and its
tion) while the VG arrays can only inuence the uid ow applications in enhancing single phase convective heat transfer, Int. J. Heat
Mass Transfer 48 (2005) 17971807.
structure of local n channel. [18] W.Q. Tao, Z.Y. Guo, B.X. Wang, Field synergy principle for enhancing
(3) For side arrangements of VGs, the discontinuous small delta convective heat transfer its extension and numerical verications, Int. J.
winglets case shows best heat transfer enhancement, and a Heat Mass Transfer 45 (2002) 38493856.
[19] A. Joardar, A.M. Jacobi, Heat transfer enhancement by winglet-type vortex
signicant augmentation of up to 33.870.6% in heat trans-
generator arrays in compact plain-n-and-tube heat exchangers, Int. J. Refrig.
fer coefcient is achieved accompanied by a pressure drop 31 (2008) 8797.
penalty of 43.497.2%. Meanwhile the large winglets case [20] A. Joardar, A.M. Jacobi, A numerical study of ow and heat transfer
enhancement using an array of delta-winglet vortex generators in a n-and-
represents the lowest pressure loss and highest value of j/f.
tube heat exchanger, J. Heat Transfer 129 (2007) 11561167.
(4) For VGs arrangements with three different values of attack [21] Y. Chen, M. Fiebig, N.K. Mitra, Heat transfer enhancement of a nned oval tube
angles, the discontinuous small winglet array always shows with punched longitudinal vortex generators in-line, Int. J. Heat Mass Transfer
best heat transfer improvement and also largest pressure 41 (1998) 41514166.
[22] J. He, L. Liu, A.M. Jacobi, Air-side heat-transfer enhancement by a new winglet-
loss. The large winglet case show relatively worse heat type vortex generator array in a plain-n round-tube heat exchanger, Trans.
transfer enhancement when the attack angle below 20 deg. ASME 132 (2010).
5458 Y.L. He et al. / International Journal of Heat and Mass Transfer 55 (2012) 54495458

[23] F. Hirokazu, Research and development on heat exchangers for air [25] T.E. Schmidt, Heat transfer calculations for extended surfaces, Refrig. Eng. 4
conditioners with the alternative winglet, in: Seventh International (1949) 351357.
Conference on Enhanced, Compact and Ultra-Compact Heat Exchangers, [26] C.C. Wang, Y.J. Chang, Y.C. Hsieh, Y.T. Lin, Sensible heat and friction
Heredia, Costa Rica, 2009, pp. 201207. characteristics of plate n-and-tube heat exchangers having plane ns, Int. J.
[24] P. Chu, Y.L. He, W.Q. Tao, Three-dimensional numerical study of ow and heat Refrig. 19 (1996) 223230.
transfer enhancement using vortex generators in n-and-tube heat
exchangers, J. Heat Transfer 131 (2009) 09190310919039.