Energy 35 (2010) 101–110

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Heat transfer and pressure drop in corrugated channels

E.A.M. Elshafei*, M.M. Awad, E. El-Negiry, A.G. Ali
Mansoura University, Faculty of Engineering, Mechanical Power Dept., Mansoura, Dakahlia 35516, Egypt

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

Article history: The convective heat transfer and pressure drop characteristics of flow in corrugated channels have been
Received 22 April 2009 experimentally investigated. Experiments were performed on channels of uniform wall temperature and
Received in revised form of fixed corrugation ratio over a range of Reynolds number, 3220  Re  9420. The effects of channel
27 August 2009
spacing and phase shift variations on heat transfer and pressure drop are discussed. Results of corrugated
Accepted 28 August 2009
channels flow showed a significant heat transfer enhancement accompanied by increased pressure drop
Available online 12 October 2009
penalty. The average heat transfer coefficient and pressure drop enhanced by a factor of 2.6 up to 3.2 and
1.9 to 2.6 relative to those for parallel plate channel, respectively, depending upon the spacing and phase
Keywords:
Compact heat exchanger shift. The friction factor increased with increasing channel spacing and its phase shift. The effect of
Narrow channel spacing variations on heat transfer and friction factor was more pronounced than that of phase shift
Corrugation variation, especially at high Reynolds number. Comparing results of the tested channels by considering
Heat transfer characteristics the flow area goodness factor (j/f), it was better for corrugated channel with spacing ratio, 3  3.0 and of
Pressure drop phase shift, Ø  90 . Comparisons of the present data with those available in literature are presented and
discussed.
Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.

1. Introduction coefficients in a corrugated passage using the naphthalene tech-

In recent years, due to the increasing demand by industries for
heat exchangers that are more efficient, compact and less expen-
sive, heat transfer enhancement has gained great momentum. For
this purpose, two techniques have been identified: ‘passive’ and
‘active’ [1]. Passive technique uses special surface geometries, or
fluid additives. Because of more costs involved, active technique
has attracted relatively little attention in research and practice, and
passive technique through the use of various surface geometries
such as corrugated channels to be preferred [2].
The heat transfer characteristics of flow through such corru-
gated channels are quite different than parallel plate channels. In
corrugated channel, the main flow direction is parallel to the
channel waviness, but the local flow direction is always changed
due to channel waviness. The thermal boundary layer formed on its
walls is periodically interrupted by flow recirculation, separation
and reattachment and thereby increase of convective heat transfer
coefficient. However, such gains in heat transfer are invariably
accompanied by increased pressure drop penalty.
Several studies on heat transfer enhancement using such
passages have been reported [3]. Goldstein and Sparrow [4] were
the first to report local and average convection heat transfer
* Corresponding author. Tel.: þ20(50)346179, þ20(50)344426, þ20(50)344105;
fax: þ20(50)324426, þ20(50)344690.
E-mail address: eelshafei@mans.edu.eg (E.A.M. Elshafei).
nique. Their passage included two different corrugations with an
angle of 30 . Comparison of their results with parallel plate passage
data showed an enhancement in the average convection heat
transfer by a factor of 3 in turbulent flow regimes. O’Brian and
Sparrow [5] obtained an empirical correlation of the average Nus-
selt number for turbulent flow in corrugated channels with sharp
edged corrugation peaks. They reported that friction factor was
almost independent of Reynolds number. The promoted flow
separation accompanied with sharp edged corrugations however,
causes large pressure drops. Mendes and Sparrow [6] performed an
experimental study on converging–diverging tubes turbulent flow
in the entrance and fully developed regions. They investigated the
effect of different aspect ratios and taper angles on the heat and
mass transfer behaviors. The largest taper angle resulted in highest
heat and mass transfer rates and accompanied by larger pressure
drop. They related that to the promotion of strong circulation zones
and flow separation. Comparing their results with the straight
tubes revealed an overall enhancement in heat transfer rates for the
converging–diverging channel.
The effect of corrugated channel height on flow and heat
transfer characteristics for water was also investigated by Sparrow
and Comb [7]. They reported that the increase of the channel height
resulted in a valuable increase in Nusselt number but the friction
factor substantially increased. Islamoglu and Parmaksizoglu [8]
investigated the heat transfer coefficients and friction factor for air
flowing in corrugated channel of two different spacing and single

0360-5442/$ – see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2009.08.031
102 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110
Nomenclature
A amplitude/surface area, m2
cp specific heat, j/kg K
Dh hydraulic diameter, m
f friction factor
h heat transfer coefficient W/m2 K
j colburn factor, (Nu/Re Pr1/3)
k thermal conductivity, W/m K
L pitch of corrugated channel waviness, mm
Lt axial length of the channel, mm
Nu Nusselt number, (h Dh/k)
Pr Prandtl number, (cp m/k)
p pressure, Pa
Qin heat input rate to channel’s walls, W
Re Reynolds number, (u Dh/y)
S channel spacing, mm
T temperature, K
V volume flow rate, m3/s
u flowing air velocity, m/s
W channel width, mm
corrugation angle of 20 . It is reported that over a flow range of
1200 < Re < 4000, a substantial increase in both fully developed
Nusselt number and friction factor was obtained due to changing of
channel spacing from 5-mm to 10-mm, however the flow area
goodness factor slightly decreased. Ali and Ramadhyani [9]
demonstrated heat transfer and flow visualizations in riangular
corrugated channel with water as working fluid for two different
channel spacing. They concluded that the performance of corru-
gated channel with larger spacing is much better than that of
smaller spacing. Comparing their results with that of parallel plate
channel, they pointed out that the corrugated channel had much
higher Nusselt number. Islamoglu et al. [10] measured heat transfer
and pressure drop for turbulent air flow in a corrugated duct having
an angle of 30 . They presented an empirical correlation of cycle-
average fully developed Nusselt number.
Sang and Hyung [11] conducted experimental and numerical
study on flow and local heat/mass transfer characteristics of wavy
duct using a naphthalene sublimation technique. The flow visual-
ization and CFD simulation were used to predict the overall flow
structures inside the duct. The results showed that complex
secondary flows and transfer processes exist inside the wavy duct
resulting in non-uniform distributions of the heat/mass transfer
coefficients on the duct side walls. They reported that the average
heat/mass transfer coefficients are higher than those of the smooth
circular duct, and. that enhancement was accompanied by
increased pressure drop penalty. Naphon [12] investigated the heat
transfer characteristics and pressure drop in a channel of different
corrugation angle under constant heat flux. The experiments were
performed for laminar and turbulent flow. It is concluded that heat
transfer and pressure drop were tremendously enhanced compared
to those of parallel plate channel. They related that to the promoted
recirculation zones and flow separation in corrugated channel flow.
Wang and Vanka [13] conducted numerical studies of flow and
heat transfer in periodic wavy passages having phase angle of 180 .
The steady-flow pattern was observed for Re  180, after which
there was self-sustained oscillatory flow, leading to destabilization
of laminar thermal boundary layers and thus provide a natural
mechanism of heat and mass transfer enhancement. In the steady-
flow regime, the average Nusselt number for the wavy channel was
slightly larger than that for a parallel plate channel. Meanwhile, the
Greek symbols
D difference
3 channel spacing ratio, (S/2A)
g channel corrugation ratio, (2A/L)
Ø phase shift
q corrugation angle
r density, kg/m3
n kinematic viscosity, m2/s
Subscripts
a air
c channel cross-section
i at inlet conditions
m mean value
o at outlet conditions/orifice
w at wall conditions
superscripts
– average
heat transfer enhanced by a factor of 2.5 in the transitional-flow
regime. Friction factors for the wavy channel were about twice
those for the parallel plate channel in the steady-flow region, and
remained almost constant in the transitional regime. Forced
convection for flow through a periodic array of a wavy-wall channel
has been investigated numerically by Wang and Chen [14]. It is
reported that the major region of flow reversal occurs at higher
Reynolds numbers. As the wavelength ratio and the Reynolds
number increase, the local Nusselt number increases manifestly in
the converging section of the wavy-wall channel and show a small
change in the diverging section.
Zhang et al. [3] numerically investigated the effect of fin waviness
configuration and its spacing on flow and heat transfer character-
istics in uniform temperature channels with sinusoidal wall corru-
gations for laminar flow regime (Re < 1000). It is reported that the
strength of developed recirculation cells depends on Reynolds
number, spacing ratio (3) and corrugation ratio (g). As the spacing
ratio decreases, viscous forces dominate and dampen the swirl; with
large spacing, the impact of wall waviness diminishes and the core
fluid flows largely undisturbed. The extent of swirl increases with
flow rate, when multiple pairs of helical vortices are formed. This
significantly enhances the overall heat transfer coefficient and
pressure drop penalty, when compared to those in a straight channel
of the same cross-section. It is also reported that the flow area
goodness factor (j/f) increases with small spacing. The optimum j/f
enhancement was obtained in the swirl flow regime (Re > 100) with
corrugation ratio, g > 0 and spacing ratio, 3 range of 0  3  1.2.
From the preceded survey, it is obvious that the enhancement of
heat transfer using corrugated passages is one of interested
subjects to the researchers. Although several studies for steady-
and unsteady-flows have been reported, little knowledge is avail-
able on the flow and heat transfer characteristics in such passages.
This work will contribute the heat transfer and pressure drop
characteristics in channels with corrugated walls.
2. Test rig
A schematic diagram of the experimental apparatus is shown in
Fig. 1. Air, the working fluid was drawn from the laboratory by the
induced blower (6) via a bell mouth inlet (1) into the upstream
 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110 103

7 mm

11
25 mm
50 mm
100 mm
14

3 5
15
2 13 6
1 4

10 9 8
11 7

12

(1) Bell mouth, (2) Duct, (3) Pressure tabs, (4) Steam inlet, (5) Pressure tabs
(6) Blower, (7) Orifice meter, (8) Thermocouples, (9) Temperature recorder
(10) Selector switch, (11) Test section, (12) Condensate tank, (13) Control gate
(14) Honey comb, (15) Electric motor

Fig. 1. Schematic diagram of the experimental apparatus. (1) Bell mouth, (2) Duct, (3) Pressure tabs, (4) Steam inlet, (5) Pressure tabs, (6) Blower, (7) Orifice meter, (8) Ther-
mocouples, (9) Temperature recorder, (10) Selector switch, (11) Test section, (12) Condensate tank, (13) Control gate, (14) Honey comb, (15) Electric motor.

wooden duct (2) of 1300-mm length to assure uniform flow with The pressure difference across the orifice plate was measured by
a negligible turbulence at the entrance of the test section (11). The a water U-tube glass manometer (5) which was connected with two
upstream duct (2) configuration is characterized by the spacing, static pressure taps; one upstream and the other downstream of
S and width, W of the tested channel. After leaving the test section, the orifice plate, respectively. The pressure drop across the test
the air flow passed over an orifice plate (7), proceeded by a control section was also measured by another water U-tube glass
gate (13) which controls the volume flow rate of the air, and then manometer (3) which was connected to a two static pressure taps
discharged into the atmosphere. The air flow rate was measured by placed 15 mm upstream and downstream of the tested channel,
detecting the pressure drop across the orifice plate. respectively.
The side walls of the channel were made of aluminum and heated
by steam. The steam was generated at an atmospheric pressure with 2.1. The test section
two degrees of superheat by an electric boiler, and flows through an
insulated pipe to the test section where it was uniformly distributed The test section is shown in Fig. 2. It consists of a pair of plates,
into two side boxes that contain the two plates of tested channel. The forming the flow channel. These two plates are made of aluminum
two side boxes were completely insulated by a 25-mm thick sheet of alloys; each has 100-mm axial length, 125-mm width, and 10-mm
glass wool to eliminate the heat loss. The channel walls tempera- thickness. The inner surface of each plate has been machined, and
tures were assured to be uniform at 99  0.5  C by the good distri- the test section was adapted to give different configuration for the
bution of steam inside these insulated boxes. The upstream and corrugated channel.
downstream ducts of the test section were made of wood to serve as
a thermal barrier for axial heat flow. The condensate was collected in
a condensate tank (12). A L
The temperature of the two principal walls of the tested channel S
W
was measured by 1.5-mm diameter precalibrated copper- Ø
constantan thermocouple (type T) with an accuracy of 0.2  C. Five
thermocouples were fitted into each wall. These thermocouples
were installed in holes drilled from the rear face and centered of the 1st corrugated 3rd corrugated channel,
o
channel walls with the respective junctions positioned within channel, Ø = 0 Ø = 180o
2 mm apart from the inner channel walls. Nevertheless, the
measured temperatures were used to detect the average surface
temperature of the inside walls of the channel because of the high
Θ
thermal conductivity for the channel material. The entering and 21 ˚
leaving bulk air temperatures were measured by two T-type ther-
mocouples placed at a close distance from inlet and exit of the test
section, respectively. All thermocouples were connected to a multi-
2nd corrugated 4th channel,
channel temperature recorder (9) (with accuracy of 1% and of channel, Ø= 90o parallel plate
0.1  C resolution) via a multi-point selector switch (10), as shown in
Fig. 1. Fig. 2. Configuration of the tested channels.
104 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110
Three corrugated channels and one with parallel plates were
investigated. All corrugated walls are symmetrical, and variable
configuration can be created according to phase shift; Ø: In the first,
the crest of the lower wall corresponds to the crest of the upper
wall, Ø ¼ 0 ; in the second, one of the walls considered in the first,
has a phase-advanced/lag of half pitch, Ø ¼ 90 ; in the third, the
crest of the lower corrugated wall corresponds to the trough of the
upper corrugated wall of the channel, Ø ¼ 180 ; and the fourth is
the conventional parallel plate channel.
The three corrugated channels have the following: fixed corru-
gation angle, q of 21.80 , each corrugated plate has ten crests
separated by distance/pitch, L of 10-mm; fixed wave amplitude, A of
1-mm and hence fixed amplitude-to-length ratio (g ¼ 2A/L) of 0.2;
mean spacing, S, and with a spacing ratio, 3 (3 ¼ S/2A).
3. Data reduction
The present experimental investigation is aimed to study the
characteristics of heat transfer and pressure drop for fully
developed turbulent flow in narrow corrugated channels. The
independent parameters included are Reynolds number and
geometries of the channels including its spacing and phase shift.
The Reynolds number based on the channel hydraulic diameter is
given by
um Dh
Re ¼
n were of uniform temperature exposed to a turbulent air flow over
where Dh is expressed as
4SW
Dh ¼
2ðS þ WÞ
The mean velocity of air flowing through the channel was
determined by the measurements of volume flow rate using the
existed orifice plate. The principal walls of the channel were heated
by steam and supposed to be of uniform temperature. The
measured temperatures of the side walls were used to evaluate the
average wall temperature, Tw. The rate of heat input, Qin to channel
walls is supposed to be totally dissipated to the air flowing through
the channel passage, raising its temperature from inlet bulk
temperature, Tai to exit bulk air temperature, Tao. The rate of heat
dissipated to the bulk air is given by
Qin ¼ ra V$cpa ðT ao Tai Þ
With the heat input, Qin, the average heat transfer coefficient is
then evaluated from the following equation:
Q with respect to its phase shift and spacing.
hm ¼
Aw DTm
where DTm is the difference between the averaged heated principal
walls temperature and that for flowing air, expressed as
DTm ¼ Twm  Tam
The term Aw is the convective heat transfer area of both inner
walls of the tested channel; Aw ¼ 2LtW/cosq. The Nusselt numbers
for lower and upper surfaces of the channel were averaged to
evaluate the average Nusselt number, given by
hm Dh
Nu ¼
k to the promotion of unstable mode for formed vortices in the core
The measurements of pressure drop across the tested channels
over the operating range of Reynolds number were employed to
evaluate the friction factors, given by

DpðDLth
f ¼ (7)
1ru2
2 m
For the sake of performance assessment of such narrow chan-
nels, regarding their heat transfer characteristics and pressure drop,
flow area goodness factor; the ratio of the Colburn factor to friction
factor, j/f is almost considered, expressed as
1
Nu=Re Pr3
j=f ¼ (8)
f
The uncertainties of measurements in the aforementioned data
were evaluated by the procedures described by Kline and McClin-
tock [15]. The uncertainties were 7% in Reynolds number, 10% in
friction factor, 9% in Nusselt number, and 12% in flow area
goodness factor.
4. Results and discussion
Heat transfer and pressure drop measurements in corrugated
channels are presented in the present article. The effect of channel
configuration regarding both interwall spacing and phase shift
variations on turbulent flow pressure drop and convective heat
transfer coefficients are discussed. The channels were tested at
variable channel spacing, S ¼ 4, 6, 8, and 10 mm, and variable phase
shift, Ø ¼ 0 , 90 , and 180 . The side walls of the tested channels
(1)
a range of 3230 < Re < 9420. Tests were also conducted for parallel
plate channel of the same spacing to establish a baseline for
comparison.
(2)
4.1. Pressure drop and friction factor results
4.1.1. Parallel plate channel
The measured pressure drop for flow in tested channels enabled
the prediction of the friction factor described in Eq. (7). The varia-
tion of the evaluated friction factor, f with Re at different spacing of
parallel plate channel is shown in Fig. 3. As can be noticed, for all
channel spacing, f decreases with increasing Re. At the same time, at
any value of Re, f for larger channel spacing is higher. This is due to
the effect of increasing channel hydraulic diameter according to the
Darcy equation, where the friction factor is proportional to D0.2.
(3)
4.1.2. Corrugated channel
The measured pressure drops among the tested corrugated
channels are also used for friction factor predictions. This factor is
mainly affected by Reynolds number and channel configurations
(4)
Friction factors as a function of the Reynolds number for flow in
relatively narrow corrugated channel (S ¼ 4 mm) and in relatively
wide channel (S ¼ 10 mm) at different phase shift, Ø ¼ 0 , 90 , and
180 are plotted in Figs. 4 and 5, respectively. The figures show that
(5) for any value of phase shift, the friction factor gradually increases
with increasing Reynolds number, up to nearly Re ¼ 7000, after
which it tends to decrease. Over the tested range of Reynolds
number, the friction factor for both spacing (S ¼ 4 and 10 mm)
seems to have a little bit higher values as the phase shift increases.
In the mean time, as the channel becomes wider, the friction factor
gets higher values. The friction factor for all corrugated channels is
(6) much higher than that for parallel plate channel. This is attributed
of flow associated with corrugated channel’s flow [16], leading to
a significant increase of pressure drops relative to those associated
with parallel flow channels.
 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110 105

Parallel plate channel

s=10 mm
s=4 Parallel plate channel

s=6 1.2 ϕ=00

0.8
s=8 ϕ=900
s=10 ϕ=1800

0.6
0.8

f
f

0.4

0.4

0.2

0
0 2 4 6 8 10
2 4 6 8 10 Re x 103
Re x 103
Fig. 5. Friction factor versus Reynolds number for flow in corrugated channel of
Fig. 3. Friction factor versus Reynolds number for parallel plate channel of different different phase shift, S ¼ 10 mm.
spacing.

The friction factor as a function of Reynolds number for flow in
corrugated channels of 0 phase shift at different channel spacing,
S ¼ 4, 6, 8, and 10 mm is shown in Fig. 6. It can be observed that the
friction factor for flow in all corrugated channels gradually
increases up to Re z 7000, after which it tends to decrease again. It
is also noticed that the friction factor gets higher values for larger
spacing. However, the flow in the channel of 8-mm spacing offers
higher friction factor values than those for 10-mm channel spacing
at higher values of Reynolds number. This is may be attributed to
uncertainties in measurements.
The friction factor variations with Reynolds number for chan-
nels of phase shifts; Ø ¼ 180 at different channel spacing is shown
in Fig. 7. It is observed that the f–Re behavior has the same trends as
that for Ø ¼ 0 (Fig. 6) over the whole range of Reynolds number. In
all tested channels, the friction factor gets higher values with
increasing the spacing. The noticed increase in friction factor over
the range of Re < 7000 may be attributed to the undeveloped flow
in the short core length of the tested channels. It has been exper-
imentally found that after 3–5 cycles, the fluid flow and heat
transfer become periodically fully developed [18]. The disruption of

φ=0ο
s=4 mm S=4
1.2 smooth 1.2 S=6
ϕ=00
S=8
ϕ=900
S=10
ϕ=1800

0.8 0.8
f
f

0.4 0.4

0 0
2 4 6 8 10 2 4 6 8 10
Re x 103 Re x 103
Fig. 4. Friction factor versus Reynolds number for flow in corrugated channel of Fig. 6. Friction factor versus Reynolds number for flow in corrugated channel of
different phase shift, S ¼ 4 mm. different spacing, Ø ¼ 0 .
106 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110

80
φ=180
s=4 S=10 mm
1.2 s=6 Ø=0o
s=8 Ø=90o
s=10
60 Ø=180o

0.8

Nu
40
f

0.4
20

0
0
2 4 6 8 10 2 4 6 8 10
Re x 10 3 Re x 103
Fig. 7. Friction factor versus Reynolds number for flow in corrugated channel of Fig. 9. Averaged Nusselt number versus Reynolds in corrugated channel of different
different spacing, Ø ¼ 180 . phase shift, S ¼ 10 mm.

the main flow in the transitional region resulted in a significant
increase in the pressure drop leading to an increase of friction
factor. In Re > 7000 region, the fully developed flow might occur
and the friction factor tends to decrease.
4.2. Heat transfer results
Corrugated channels are normally used to enhance the heat
transfer. The configuration of such channels with respect to its
spacing and flow passage shape are important parameters that
influence their heat transfer performance. The effects of the vari-
ation of phase shift and spacing of the tested corrugated channels
on the average Nusselt number are discussed in the following
subsections.
4.2.1. Effect of phase shift
The concomitant influence of the flow structure in corrugated
channels with different phase shift on the averaged Nusselt number
ðNuÞ for 4-mm and 10-mm spacing is illustrated in Figs. 8 and 9,
respectively. The variations of Nu with Re for a corrugated channel
of 4-mm spacing at three different phase shifts (Ø ¼ 0 , 90 , and
180 ) are shown in Fig. 8. For a certain phase shift, Nu is seen to
enhance significantly with increasing Re relative to that for parallel

100
100
S=4mm Ø=0o
Parallel plate channel S=4 mm
Ø=0o 80 S=6 mm
80 Ø=90o S=8 mm
Ø=180o S=10 mm
60
60
Nu
Nu

40
40

20
20

0
2 4 6 8 10 12 2 4 6 8 10 12
Re x103 Re x 103
Fig. 8. Averaged Nusselt number versus Reynolds number in corrugated channel of Fig. 10. Averaged Nusselt number versus Reynolds number in corrugated channel of
different phase shift, S ¼ 4 mm. different spacing, Ø ¼ 0 .
 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110 107

100 4
Ø=180o
S=4 mm Re=3230
S=6 mm Ø=0o
80 S=8 mm Ø=90o
S=10 mm 3
Ø=180o

60

j/f x 103
Nu

40

20

0
2 4 6 8 10 12 0 2 4 6
Re x 103 ε

Fig. 11. Averaged Nusselt number versus Reynolds number in corrugated channel of Fig. 12. Flow area goodness factor versus corrugation ratio, Re ¼ 3230.
different spacing, Ø ¼ 180 .

channel. This is attributed to lengthen of the flow path and the
presence of recirculation cells accompanied with corrugated
channel. The strength of such promoted cells depends on Reynolds
number and channel spacing [3]. The discrepancy between Nu
values for flow in corrugated channel and those in parallel plate one
of the same spacing (4 mm) becomes bigger with increasing Rey-
nolds number. Over the tested range of Reynolds number, Nu
generally gets a little bit higher values with 0 and 180 than those of
90 phase shift. The effect of changing phase shift for this narrow
passage on Nu seems to be very small. The heat transfer in corru-
gated channels of 0 , 90 and 180 phase shift enhanced by a factor
of 2.8, 2.6, and 3.2 to those of parallel plate channel, respectively in
such transitional-flow region.
The results for corrugated channel of 10-mm spacing and 3-
different phase shifts (Ø ¼ 0 , 90 , and 180 ) is shown in Fig. 9. It is
noticed that, the Nu—Re plots have the same trends as those foe 4-
mm spacing. However, the effect of changing phase shift on Nu is
more pronounced, especially as the rate of flow increases. For
corrugated channels of 90 and 180 phase shifts are very close to
each other and lower than those for 0 phase shift. Over the tested
range of Reynolds number, Nu for corrugated channel of 0 phase
shift increased by about 20% and 12% higher than those of 90 and
180 phase shift, respectively.
From the above discussion, it can be said that the effect of phase
presented. Measurements were made with four different spacing,
(S ¼ 4, 6, 8, and 10 mm) while keeping fixed phase shift, Ø over
a range of Reynolds number.
For flow in corrugated channels of Ø ¼ 0 , the average Nusselt
number, Nu for flow in channels of 8- and 6-mm spacing are very
close to each other and a little bit higher than those of 4- and 10-
mm spacing as shown in Fig. 10. Over the tested range of Reynolds
number, Nu for corrugated channel of 8-mm spacing increased by
about 21% of that belongs to the channel of 4-mm spacing. The heat
transfer for flow in the parallel plate channel of 4 mm is much
lower than those of corrugated channels.
In the convergent–divergent passage (Ø ¼ 180 ), the enhance-
ment in heat transfer with the channel of 6-mm spacing is obvious
4
Re=5880
Ø=0o
Ø=90o
3 Ø=180o
j/f x 103

shift variation on heat transfer for corrugated channels is change-

able and related to how much is the channel spacing. Good
enhancement in heat transfer was achieved with the channel of
2
Ø ¼ 180 for relatively narrow channel (S ¼ 4), and of Ø ¼ 0 for
relatively wide channel (S ¼ 10 mm). As discussed earlier and based
on data reported by Zhang et al. [3], the effects of flow disruption
and promoted vortices accompanied with corrugated channel’s
flow on heat transfer are influenced by its spacing. With conver-
gent–divergent passage (Ø ¼ 180 ), these effects are more 1
pronounced in narrow channel (S ¼ 4 mm), and with Ø ¼ 0 in
wider channel (S ¼ 10 mm).

0 2 4 6
4.2.2. Effect of channel spacing ε
In this subsection, experimental measurements of heat transfer
for flow in corrugated channels of different interwall spacing are Fig. 13. Flow area goodness factor versus corrugation ratio, Re ¼ 5870.
108 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110

4 100
Present data, Parallel plate, S=10 mm
Present data, Ø=0o, S= 10mm
Re=9420
Kilicalsan & Sarac [17], Parallel plate, S=14 mm
Ø=0o 80 Kilicalsan & Sarac [17], Ø=0o, S= 14 mm

3
Ø=90o
Ø=180o
60
j/f x 103

Nu
2
40

1 20

0
0 2 4 6 8 10
0
1 2 3 4 5 6 Re x 103
ε
Fig. 16. Comparison of the average Nusselt number for the present investigation with
other results, Ø ¼ 0 .
Fig. 14. Flow area goodness factor versus corrugation ratio, Re ¼ 9420.

with respect to other channel spacing as can be seen from Fig. 11.
Over the tested range of Reynolds number, the average Nusselt
number for corrugated channels of 6-mm spacing enhanced by
a factor of about 1.4 times that for the 4-mm spacing.
Based on the above results and its discussion, it appeared that
the variation of corrugated channel spacing has a relatively greater
effect on heat transfer performance more than that due to phase
shift variation. Both the interwall spacing and phase shift are
important geometrical parameter affecting the fluid flow and heat
transfer characteristics in corrugated channels. These parameters
with the aid of the present experimental findings are used to detect
the area goodness factor for flow in such channels.
4.3. Flow area goodness factor
Improvements in heat transfer performance of swirling flow in
corrugated channels are also accompanied with an increase in the
pressure loss. Therefore, it is important to evaluate the net benefits
of using such channels. One method of assessing the relative
thermal–hydraulic performance enhancement is to consider the
area goodness factor, j/f [19,20].
The performances of these corrugated channels are graphed as
j/f versus channel spacing ratio, 3, for three phase shifts over three

80

Present data, Ø=180o, S= 10mm

Naphon [12], Ø=180o, S=20 mm

60
Nu

40

20

0
2 4 6 8 10
Re x 103
Fig. 15. Comparison of the pressure drop for the present results with those in the Fig. 17. Comparison of the average Nusselt number for the present investigation with
available literature, Ø ¼ 0 . other results, Ø ¼ 180 .
 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110
Reynolds number regimes (Re ¼ 3230, 5870 and 9420) and depic-
ted in Figs. 12–14. For the tested three regimes of Reynolds number,
j/f decreases with increasing the spacing ratio, 3. The highest area
goodness factor was predicted for flow in corrugated channel of
spacing ratio within the range 3  3 and of 0 and 90 phase shift in
the three tested Reynolds number regimes. In low Re regime
(Re ¼ 3250), channel of 0 phase shit offered the highest j/f,
whereas for Re ¼ 5870 and 9420, the channel of Ø ¼ 90 seems to
have the best goodness facto, j/f.
4.4. Comparisons with previous work
Pressure drop (DP) and heat transfer results for air flow
through parallel and corrugated channels are compared with
experimental results reported in literature [17,12]. The geometries
considered by Kilicalsan and Sarac [17] have a corrugated channel
of cylindrical and triangle ribs at zero phase shift and of 14-mm
spacing over Reynolds number up to 7500. It can be seen from
Fig. 15 that DP of the present investigation and theirs have the
same trend. For corrugated channel flow, the present data are
a little bit higher than those reported in [17]. However, for
parallel channel flow, Dp are smaller than those of the present
results. These discrepancies may be related to the differences in
geometries of the present tested channel (S ¼ 10 mm) and theirs
(S ¼ 14 mm).
The heat transfer data reported by Kilicalsan and Sarac [17] were
used for comparison with the present results. The Nu—Re behavior
of the present investigation had similar trends as theirs as illus-
trated in Fig. 16. Their results however, are a little bit higher than
the present ones. These discrepancies may be due to the difference
in channel’s geometries and the uncertainty of measurements. As
expected, the heat transfer in parallel plate channel was much
lower and the percentage enhancement due to corrugation in both
results was assured.
Heat transfer results over the operating Reynolds number range
for one of the present corrugated channel (Ø ¼ 180 , S ¼ 10 mm) are
also compared with that reported by Naphon [12] in Fig. 17. The
geometries considered by Naphon [12] have two opposite corru-
gated plates on which all configuration peaks lie in a staggered
arrangement (Ø ¼ 180 , corrugation angle, q ¼ 20 ) and of 20-mm
spacing. As is depicted in Fig. 17, Nu—Re characteristics of the
present data are higher than those reported by Naphon [12], but
have similar trend. The discrepancies in both results may be argued
to the differences in the compared channel’s geometries as well as
due to the uncertainty of measurements. Since there are no enough
experimental data with similar geometries for further comparison
and due to the complexity of such flow, it is difficult to discuss the
reliability of the data.
5. Conclusions
Corrugated passages are one of the popular passive techniques
that are extensively used in compacted heat exchanger
manufacturing. Lengthen of the flow path and the presence of
recirculation cells with such channel enhanced the heat transfer
with pressure drop penalty compared to that of parallel plate
channel. In this study, three parameters were systematically varied,
including Reynolds number, phase shift, and channel spacing of the
corrugated channel modules. The gained benefits of such passages
were assessed by area goodness factor. The following conclusions
may be drawn:
1. A significant heat transfer enhancement was associated with
the corrugated channel as compared to the conventional
parallel plate channel. The average Nusselt number, Nu
109
increased by a factor of 2.6 up to 3.2 relative to that of parallel
plate channel, depending upon phase shift and spacing of the
corrugated channel.
2. The enhancement of heat transfer was accompanied with an
increase in the pressure drop. The friction losses enhanced by
a factor of 1.9–2.6 relative to that for parallel plate channel,
depending on both phase shift and spacing of the corrugated
channel. The pressure drop was magnificently decreased with
increasing channel spacing.
3. The f–Re relation for the tested corrugated channels was
changeable; in low Reynolds number region, f gradually
increased with increasing Re up to Re z 7000 after which it
tends to decrease again. For all corrugated channels the friction
factor is much higher than that for parallel plate channel.
4. Phase shift variation seemed to have a little effect on friction
factor compared with the effect of spacing variation, especially
at high Reynolds number. The value of f increased with
increasing Ø, and generally gets higher values at larger channel
spacing for any phase shift.
5. The effect of phase shift variation on heat transfer for
corrugated channels is related to how much the channel
spacing is. The channel of convergent–divergent passage;
Ø ¼ 180 had the highest Nu for lower spacing. However,
Nu for channel of Ø ¼ 0 was better for relatively large
spacing.
6. The channel of 6-mm spacing almost gave the highest values of
Nu for Ø ¼ 0 and 180 over the tested range of Reynolds
number.
7. For the tested corrugated channels with geometrical
parameters: 2  3  5, and of Ø ¼ 0 , 90 , and 180 over
three flow regions; Re z 3230, Re z 5870, and Re z 9420,
the flow area goodness factor, j/f decreased with increasing
the spacing ratio, 3 and less sensitive to the phase shift
variation.
8. Comparison with the available literature showed fair agree-
ment; the data of the present investigation had similar trend as
those reported in literature and discrepancies may be due to
differences in compared channels geometries and uncertainties
in measurements.
References
[1] Webb RL. Principals of enhanced heat transfer. New York: John Wiley & Sons
Inc.; 1994.
[2] Kuppan T. Heat exchanger design handbook. New York: Marcel Dekker Inc.;
2000.
[3] Zhang J, Kundu J, Manglik RM. Effect of fin waviness and spacing on the lateral
vortex structure and laminar heat transfer in wavy-plate-fin cores. Interna-
tional Journal of Heat and Mass Transfer 2004;47(17):1719–30.
[4] Goldstein L, Sparrow EM. Heat/mass transfer characteristics for flow in
a corrugated wall channel. Transaction of the ASME, Journal of Heat Transfer
1997;99:187–95.
[5] O’Brian JE, Sparrow EM. Corrugated-duct heat transfer, pressure drop, and
flow visualization. Transaction of the ASME, Journal of Heat Transfer
1982;104:410–6.
[6] Mendes P, Sparrow EM. Periodically converging–diverging tubes and their
turbulent heat transfer, pressure drop, fluid flow, and enhancement charac-
teristics. Journal of Heat Transfer 1984;106:55–63.
[7] Sparrow EM, Comb JW. Effect of interwall spacing and fluid flow inlet
conditions on a corrugated-wall heat exchanger. International Journal of Heat
and Mass Transfer 1983;26(7):993–1005.
[8] Islamoglu Y, Parmaksizoglu C. The effect of channel height on the enhanced
heat transfer characteristics in a corrugated heat exchanger channel. Applied
Thermal Engineering 2003;23:979–87.
[9] Ali M, Ramadhyani S. Experiments on convective heat transfer in corrugated
channels. Experimental Heat Transfer 1992;5:175–93.
[10] Islamoglu Y, Halici F, Parmaksizoglu C. Experimental study of heat transfer and
pressure drop in a wavy channel with sharp corners. In: Proceedings, 4th
international thermal energy congress, Turkey; 2001.
[11] Sang I, Hyung N. Experimental study on flow and local heat/mass transfer
characteristics inside corrugated duct. International Journal of Heat and Fluid
Flow 2006;27:21–32.
110 E.A.M. Elshafei et al. / Energy 35 (2010) 101–110
[12] Naphon P. Heat transfer characteristics and pressure drop in channel with
V-corrugated upper and lower plates. Energy Conversion and Management
2007;48(5):1516–24.
[13] Wang G, Vanka SP. Convective heat transfer in periodic wavy passages.
International Journal of Heat and Mass Transfer 1995;38(17):3219–30.
[14] Wang CC, Chen CK. Forced convection in a wavy-wall channel. International
Journal of Heat and Mass Transfer 2002;45:2587–95.
[15] Kline SJ, McClintock FA. Describing uncertainties in single-sample experi-
ments. Mechanical Engineering 1953:3–8.
[16] Nishimura T, Yano K, Yoshino T, Kawamura Y. Occurrence and structure of
Taylor–Goertler vortices induced in two-dimensional wavy channels for
steady flow. Journal of Chemical Engineering of Japan 1990;23(6):697–703.
[17] Kilicalsan I, Sarac I. Enhancement of heat transfer in compact heat exchanger
by different type of rib with holographic interferometry. Experimental
Thermal and Fluid Science 1998;17:339–46.
[18] Yuan Z, Tao W, Wang Q. Numerical prediction for laminar forced convection
heat transfer in parallel plate channels with stream-wise-periodic rod
disturbances. International Journal for Numerical Methods in Fluids
1998;28:1371–88.
[19] Manglik RM, Zhang J, Muley A. Low Reynolds number forced convection
in three-dimensional wavy-plate-fin compact channels: fin density
effects. International Journal of Heat and Mass Transfer 2005;48(8):
1439–49.
[20] Shah RK. Compact heat exchangers. 1st ed. Taylor & Francis; 1990.