i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

Available online at www.sciencedirect.com

w w w . i i fi i r . o r g

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

Air-side heat transfer and pressure drop in spiral

wire-on-tube condensers

Jader R. Barbosa Jr.a,*, Rodrigo A. Sigwalt b

a
Polo e Research Laboratories for Emerging Technologies in Cooling and Thermophysics, Department of Mechanical Engineering,
Federal University of Santa Catarina, 88040-900 Florianópolis, SC, Brazil
b
Whirlpool S.A., Joinville, SC, Brazil

article info abstract

Article history: The air-side thermal-hydraulic performance of spiral wire-on-tube condensers is investi-
Received 4 May 2011 gated experimentally in this paper. Sixteen prototypes have been manufactured and tested
Received in revised form in an open-loop wind tunnel calorimeter. The influence of the following geometric
25 February 2012 parameters has been evaluated: the number of tube passes, the radial and longitudinal
Accepted 27 February 2012 tube spacings and the wire spacing. Measurements of the air-side thermal conductance
Available online 3 March 2012 and pressure drop were carried out for air flow rates ranging from 70 to 220 m3 h
1. The
data were correlated using empirical relationships for the Colburn j-factor and the Darcy
Keywords: friction factor. The agreement with the experimental data presented RMS deviations of
Condenser 0.9% for the air-side heat transfer and 1.3% for the frictional pressure drop dimensionless
Wire parameters. A quantitative analysis based on the core volume goodness factors for heat
Tube transfer and pressure drop is presented to provide the most viable configuration from the
Spiral point of view of application in a refrigerator.
Forced convection ª 2012 Elsevier Ltd and IIR. All rights reserved.

Transfert de chaleur côté air et chute de pression dans les

condenseurs à serpentin
Mots clés : Condenseur ; Fil ; Tube ; Serpentin ; Convection forcée

1. Introduction components, which increases the net storage volume of the

The development of more compact heat exchangers for
domestic refrigeration applications has been the focus of
several studies thanks to its impact in reducing material costs
and energy consumption. It also allows for a more effective
usage of the space occupied by the cooling system
refrigerator without changing the external dimensions of the
product. Due to its low manufacturing cost, wire-on-tube heat
exchangers are still the most advantageous condenser type
for household refrigeration applications where natural
convection is the dominant heat transfer mode (Hermes and
Melo, 2009). However, for higher cooling capacity systems,

* Corresponding author. Tel./fax: þ55 48 3234 5166.

E-mail address: jrb@polo.ufsc.br (J.R. Barbosa Jr.).
0140-7007/$ e see front matter ª 2012 Elsevier Ltd and IIR. All rights reserved.
doi:10.1016/j.ijrefrig.2012.02.010
940 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

Nomenclature T temperature [ C]

Greek
Roman
a generalized linear or angular coefficient of the
aj angular coefficient of the j correlation [e]
correlations [e]
af angular coefficient of the f correlation [e]
b surface area per unit volume [m2 m
3]
A surface area [m2]
bk empirical constants in Eq. (24) [e]
A,B,C,D normalized parameters in Eq. (24) [e]
ε heat exchanger effectiveness of the cross-flow
bj linear coefficient of the j correlation [e]
arrangement [e]
bf linear coefficient of the f correlation [e]
εcc,2 heat exchanger effectiveness of the 2-pass cross-
C thermal capacity rate [W K
1]
counter flow arrangement [e]
cp specific heat capacity at constant pressure
εcc,4 heat exchanger effectiveness of the 4-pass cross-
[J kg
1 K
1]
counter flow arrangement [e]
C* ratio between the minimum and maximum
ho overall surface efficiency [e]
thermal capacity rate [e]
m dynamic viscosity [kg m
1 s
1]
Dt tube diameter [m]
r density [kg m
3]
Dw wire diameter [m]
sf ratio between the minimum free-flow and face
er tube radial spacing [m]
areas [e]
er relative error [e]
4 normalized coefficient in Eq. (24) [e]
f Darcy friction factor [e]
J1 thermal conductance per unit volume [W m
3 K
1]
G mass flux [kg m
2 s
1]
J2 pumping power per unit volume [W m
3]
h average heat transfer coefficient [W m
2 K
1]
j Colburn j-factor [e] Subscripts
k thermal conductivity [W m
1 K
1] 2 2-pass heat exchanger
m _ mass flow rate [kg s
1] 4 4-pass heat exchanger
Np number of tube passes [e] a air
Nw number of wires [e] cf internal filling
NTU number of transfer units [e] f face
P pressure [Pa] fin fin
Pr Prandtl number [e] i internal
Q_ heat transfer rate [W] in inlet
R2 coefficient of determination [e] max maximum
Rs thermal resistance of the solid surface [K W
1] min minimum
Re Reynolds number [e] o external
U experimental uncertainty of a given variable [e] out outlet
UA overall thermal conductance [W K
1] w water
V velocity [m s
1]
Abbreviations
V_ volume flow rate [m3 s
1]
AAD absolute average deviation
t Student’s t coefficient [e]
RMS root mean square

the benefits of more compact forced convection wire-on-tube
condensers seem to overcome the costs and energy
consumption involved in pumping the external ambient air
through the condenser.
One of the first studies of forced convection in wire-on-
tube condensers to be reported in the open literature was
that of Hoke et al. (1997). They evaluated experimentally the
thermal performance of eight single layer coils made from
carbon steel in a wind tunnel, and quantified the effects of the
tube and wire diameters, tube and wire spacings, angle of
attack, flow velocity and flow orientation (normal to the wires
and normal to the tubes). The internal heat transfer fluid was
hot water. Correlations were proposed for the wire and tube
heat transfer coefficients. The proposed heat transfer calcu-
lation method was capable of predicting the air-side experi-
mental data to within 16.7% error. It was concluded that the
angle of attack is an important parameter in the overall
thermal performance, with a nearly two-fold increase in the
heat transfer coefficient when the angle of attack increased
from 0 to 20 .
A relatively small number of wire-on-tube heat exchanger
configurations other than single layer coils have been
proposed in the open literature for forced convection appli-
cations (Dasher, 1996; Petroski and Clausing, 1999; Lee et al.,
2000; Ohgaki, 2002; Jenkins, 2003). Petroski and Clausing
(1999) evaluated the air-side thermal performance of eight
saw-tooth wire-on-tube condensers (also made from carbon
steel) under forced convection. Their facility and experi-
mental procedure were similar to those employed by Hoke
et al. (1997). They investigated the influence of the following
parameters: (i) the flow orientation with respect to the wires
(in-line and cross-flow), (ii) wire spacing, (iii) amplitude of the
saw-tooth and, (iv) the distance between the test section wall
and the heat exchanger. Tests were carried out in an open-
loop wind tunnel for air velocities in the range of
0.2e2 m s
1. The air-side thermal conductance was correlated
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1 941

Fig. 1 e Spiral condenser geometry (condenser #14). (a) Unwrapped view. (b) Axial view. (c) Photograph. (d) Wires welded on
the tube.

as a combination of radiation and convection contributions,
with the latter consisting of separate terms for the heat
transfer from the wires and from the tubes. A correlation was
also proposed for the air-side pressure drop.
Lee et al. (2000) studied wire-on-tube condensers assem-
bled in the form of a multilayer stack. In their experiments,
the tubes were heated up with electrical resistances placed
inside the tubes. The rate of Joule heating was controlled by
a power regulator and measured with a power meter. Ther-
mocouples were placed on the outer surface of the coil for
temperature measurement. Lee et al. (2000) developed multi-
plying factors for the Zhukauskas (Incropera et al., 2007)
correlation in order to account for the presence of wires. The
correlation takes into account the directional effects of the air
flow, which may be either all-cross (i.e., when the air is in
cross-flow with wires and tubes) or wire-cross (when the air is
in cross-flow with the wires, but flows parallel to the tubes).
Their multiplying factor-based method predicted their own
heat transfer rate data with an absolute average deviation of
3.7%.
The purpose of this paper is to investigate the air-side heat
transfer and pressure drop characteristics of spiral wire-on-
tube condensers. The spiral condenser consists of a single
layer coil that is wound around a fixed center line parallel to
the wires (Ohgaki, 2002). In comparison to the other existing
wire-on-tube condenser geometries, the spiral condenser has
a comparatively larger heat transfer area per unit volume
(Sigwalt, 2010). The air-side heat transfer and pressure drop
942 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

characteristics of sixteen heat exchanger samples were
determined experimentally in an open-loop wind tunnel
calorimeter for air flow rates ranging from 70 to 220 m3 h
1
(face velocities were in the range of 0.57e4.00 m s
1), since
these are the flow rates typically encountered in real appli-
cations. The influence of the number of tube passes, the radial
and longitudinal tube spacings and the wire spacing was
evaluated. The data were correlated with empirical expres-
sions for the Colburn j-factor and the Darcy friction factor. The
agreement with the experimental data presented RMS devia-
tions of 0.9% and 1.3% and coefficients of determination (R2) of
0.96 and 0.94 for the heat transfer and frictional pressure drop
dimensionless parameters, respectively.
2. Experimental work
2.1. Heat exchangers
Sixteen condensers were designed and fabricated exclusively
for this study. The wires and tubes were made from carbon
steel and the wires were spot welded onto diametrically
opposing sides of the tubes. The condensers were painted
black on the outside. Carbon steel was selected as the heat
exchanger material because real wire-on-tube condensers are
made from carbon steel. In real applications, there is no risk of
corrosion since the surface exposed to the air side is covered
with paint, and the inside is in contact with pure refrigerant
that is oxygen-free. In the present experiments, as will be seen
later, the internal fluid is water. However, since an experi-
mental run for each prototype is completed over the course of
one to two days, corrosion effects are expected to be minimal.
The geometry of a typical spiral condenser is illustrated in
Fig. 1, and the basic dimensions of all condensers are presented
in Table 1. The wire diameter and the tube O.D. were 1.4 mm
and 4.76 mm, respectively. The tube wall thickness was
0.7 mm. All condensers had a fixed tube length of 4 m. Thus,
depending on the geometry parameters of Table 1, the face area
and the heat exchanger volume assumed different values for
each condenser. The air flow was parallel to the wires.
2.2. Experimental facility
The air-side thermal-hydraulic performance of the
condensers was assessed in a wind tunnel calorimeter (see
Fig. 2), which was designed and constructed according to the
ASHRAE Standards (1987, 1988). Thermal insulation was ach-
ieved via a 100-mm layer of glass wool along the 3000-mm
long tunnel. The air flow was supplied by a variable speed
centrifugal fan and was measured by a set of calibrated
nozzles, according to ANSI/ASHRAE 51 (1999). A differential
pressure transducer (Ashcroft ATE, 0e3 in H2O) was used to
measure the nozzle pressure drop. The air flow rate was fine-
tuned with electrically-controlled dampers located at the exit
of the wind tunnel. The temperature of the ambient air
entering the calorimeter was controlled to 0.5  C before
entering the fan (Sigwalt, 2010). The air temperature upstream
and downstream of the condenser was determined by aver-
aging the readings of six on-site calibrated T-type thermo-
couples at each measuring station (C and G, see Fig. 2b). The
uncertainty of the temperature measurements was calculated
as 0.2  C.
In the test section, the condenser to be tested was placed
inside a made-to-fit cylindrical polystyrene tube insert whose
internal diameter was equal to the external diameter of the
condenser, thus preventing air flow leaks. The insert was
mounted inside an external fixed section (common to all
condensers), also made from polystyrene. The air-side static
pressure drop taps were located in this external section, as
seen in Fig. 3. They consisted of perforated plastic hoses
mounted on two grooves cut into the wall of the polystyrene
section. The grooves were perpendicular to the main flow
direction and their depth was such that the pressure taps were
leveled with the wall. One end of each hose was connected to
the differential pressure transducer (Ashcroft ATE, 0e5 in
H2O) while the other end was sealed. In all tests, a small

Table 1 e Geometry parameters of the spiral wire-on-tube condensers.

Condenser Number of Radial tube Wire Wire Number of Longitudinal Heat transfer Surface area
tube passes (
) spacing (mm) spacing length wires (
) tube area (cm2) per unit
(mm) (mm) spacing (mm) volume (m2 m
3)

#1 2 11 5 50 718 25 2159 147.2

#2 2 11 5 75 828 50 3316 127.3
#3 2 11 10 50 360 25 1372 93.81
#4 2 11 10 75 416 50 1954 73.93
#5 2 15 5 50 718 25 2159 114.8
#6 2 15 5 75 828 50 3316 99.33
#7 2 15 10 50 360 25 1372 73.19
#8 2 15 10 75 416 50 1954 57.69
#9 4 11 5 100 346 25 2104 128.1
#10 4 11 5 175 342 50 3223 110.1
#11 4 11 10 100 174 25 1349 82.44
#12 4 11 10 175 172 50 1915 64.59
#13 4 15 5 100 346 25 2104 104.3
#14 4 15 5 175 342 50 3223 89.7
#15 4 15 10 100 174 25 1349 67.11
#16 4 15 10 175 172 50 1915 52.62
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1 943

Fig. 2 e The wind tunnel calorimeter. (a) Schematic view. (b) Dimensions and main components. A: fan; B: diffuser; C:
temperature measurement; D: test section; E: condenser; F: flow homogenizers; G: temperature measurement; H: calibrated
nozzles; I: dampers.

cylindrical filling (diameter of 50.8 mm) was placed inside the

71-mm diameter gap (see Fig. 1b) to eliminate any flow by-
pass.
The thermally insulated hot water loop consisted of
a speed-controlled rotary pump, a Coriolis-effect mass flow
meter, a thermostatic bath and two T-type immersion ther-
mocouples that were placed at water inlet and outlet. As
pointed by Hoke et al. (1997), the use of water as a working
fluid instead of a condensing refrigerant simplifies the
instrumentation and installation of the heat exchanger
prototypes. Moreover, it gives rise to a relatively small thermal
resistance on the inside of the tube that can be accurately
estimated and it enables an accurate experimental determi-
nation of the overall heat transfer rate.

2.3. Experimental procedure

The temperature of the air entering the wind tunnel was kept
at 21  C and the inlet water temperature was maintained at
Fig. 3 e A condenser in the test section. 40  C or higher. The fan and the damper were operated in
944 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

conjunction to set the desired air flow rate. The water flow
rate was set so that the water temperature range was larger
than 2  C. Approximately 50e120 min e depending on the air
flow rate e are needed to reach steady-state. The steady-state
criterion was established so that each parameter (tempera-
ture, pressure drop or flow rate) was averaged over a 10-min
interval and the standard deviation associated with its
signal was calculated. If the difference between the reading at
the beginning of the sampling interval (t ¼ 0) and at the end
(t ¼ 10 min), for all variables, was less than 3 times the value of
a pre-defined maximum standard deviation associated with
each variable, the test was considered steady. Fig. 4a shows
a temperature plot and the demarcation (dotted line) between
the transient and steady-state condition for a typical test.
A test is considered valid if, in addition to being steady, the
heat transfer rates calculated based on energy balances on the
water side and on the air side are within 5% of each other
(Waltrich et al., 2011), as can be verified in Fig. 4b. The
maximum standard deviations for the experimental param-
eters are as follows: (i) temperatures, 0.04  C; (ii) pressure drop,
0.5 Pa; (iii) air volume flow rate: 0.34 m3 h
1; and (iv) water
mass flow rate: 0.2 kg min
1. In total, 64 data points were
collected for heat transfer and pressure drop.

3. Data regression

3.1. Heat transfer

The overall thermal conductance can be determined from the

number of transfer units (NTU ) using the ε-NTU relationships
for cross-counter flow heat exchangers with 2 and 4 passes,
respectively. These relationships have been listed by ESDU
(1991), Wang et al. (1999) and Navarro and Cabezas-Gómez
(2005). In the present work, the experimental data were such
that the NTU ranged from 0.19 to 0.47. For such low values,
irrespective of the values assumed by the thermal capacity
rate ratio (see definition below), there is very little difference
between the ε-NTU curves for cross-counter flow heat
exchangers with 2 and 4 tube passes and the curve for single-
pass mixedemixed cross-flow, which is much simpler and is
given by (Shah and Sekulic, 2003),

1 1 C 1
¼ þ
(1)
ε 1
exp ð
NTUÞ 1
exp ð
NTU C Þ NTU

where

UA
NTU ¼ (2)
Cmin


Cmin ¼ min ra V_ a cPa ; m
_ w cpw (3)

Cmin
C ¼ (4)
Cmax

Q_
ε¼ (5)
_
Q max

where Q_ is the heat transfer rate calculated from a heat

balance in each stream and Q_ max is given by,


Q_ max ¼ Cmin Tw;in
Ta;in (6)

As can be seen from Fig. 5, which shows contour plots of the

relative errors defined as,

ε
εcc;2
er2 ¼ (7)
εcc;2
Fig. 4 e Experimental procedure verification (a)
Temperature signals during the initial transient and in the and
steady-state condition. Air flow rate: 220 m3 hL1. (b)
Comparison of the heat transfer rates measured on the ε
εcc;4
er4 ¼ (8)
water and air sides for all data points. εcc;4
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1 945

as a function of C* and NTU, the errors incurred in assuming
that NTU can be calculated from Eq. (1) are of the order
of
0.1% to
0.3% for the present experimental NTU range,
and are certainly negligible when compared with other
experimental and data correlation errors involved in the
study. Therefore, for the sake of simplicity, the data reduction
in the present paper was carried out using Eq. (1) for all sixteen
prototypes.
The air-side heat transfer coefficient is calculated sub-
tracting the internal convection and the solid wall heat
transfer resistances from the overall thermal resistance as
follows (fouling resistances on both sides were assumed zero),
  
1
1 1
ho ho ¼

Rs Ao
UA hi Ai
where Rs is the solid wall conduction resistance, Ai is the in-
tube surface area and Ao is the external (wire and tube)
surface area. The in-tube heat transfer coefficient was calcu-
lated with the Gnielinski correlation (Incropera et al., 2007).
The overall surface efficiency was calculated by,
number of tube passes, Dt is the tube diameter, Nw is the
number of wires and Dw is the wire diameter. The heat
exchanger pressure drop, DPa, was calculated by subtracting
the pressure drop measured without the condenser (i.e., that
corresponding to the internal polystyrene insert only) from
that measured with the condenser and filling in place.
The Colburn j-factor and friction factor data were corre-
lated in terms of the air-side Reynolds number defined as,
Rea ¼ Ga er =ma (16)
The tube radial spacing, er, was chosen as the characteristic
length scale in the air-side Reynolds number because it
defines, together with the tube diameter, the minimum free-
flow area on the air-side (i.e., it is a real length scale in the
(9)
problem). In the correlation presented in Section 4.2, the tube
radial spacing proved to be more efficient in correlating the
data than other parameters, such as the tube diameter. This is

Afin 
ho ¼ 1
1
hfin (10)
Ao
where hfin is the fin efficiency defined as (Incropera et al., 2007),

tanh ðmLÞ
hfin ¼ (11)
mL
and mL is the dimensionless fin length given by,
sffiffiffiffiffiffiffiffiffiffiffiffiffi
ho eL
mL ¼ (12)
kfin Dw

The Colburn j-factor is given by,

ho Pr2=3
j ¼ StPr2=3 ¼ (13)
Go cPa
where Ga is the air mass flux given by Ga ¼ raVmax. Vmax is the
maximum air velocity, which takes place at the minimum air-
side free-flow area (see Section 3.2).

3.2. Pressure drop

The Darcy friction factor that incorporates the inlet contrac-

tion and outlet expansion losses was calculated as follows
(Kays and London, 1998),
" #
Amin ra 2DPa ra;in   r
a;in
f¼
1 þ s 2

1 (14)
Ao ra;in G2a f
ra;out

where ra,in and ra,out are the air density corresponding to the
inlet and outlet sections, respectively. ra is the arithmetic
average of the inlet and outlet air densities. sf is the ratio of the
minimum free-flow area, Amin, and the face area, Af. DPa is the
heat exchanger pressure drop. The minimum free-flow area is
given by,

pD2cf L pD2
Amin ¼ Af

Dt
Nw w (15)
4 Np 4
Fig. 5 e (a) Effectiveness relative error for the 2-pass cross-
where Dcf is the diameter of the internal cylindrical filling, Dt is counter flow heat exchanger. (b) Effectiveness relative error
the tube diameter, L is the overall tube length, Np is the for the 4-pass cross-counter flow heat exchanger.
946 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

because the tube diameter remained constant throughout the
experiments, while the radial spacing was changed according
to Table 1.
3.3. Uncertainty analysis
An uncertainty propagation analysis was conducted (Sigwalt,
2010). Measurement uncertainties, U, were calculated based
on the values of the mean and of the standard deviation
associated with each measured variable and on accuracy
information provided by manufacturers of measurement
equipment. Thus, the uncertainty is given by,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
U ¼ t u20 þ s2
For any number of tube passes, the wire surface area seems to
be the most important factor in the air-side thermal conduc-
tance. As can be seen, the poorest thermal performances in
both cases (2 and 4 tube passes) were associated with the
condensers that have the largest wire spacing and the small-
est longitudinal spacing (Samples #7 and #15).
The behavior of the air-side overall surface efficiency, ho, as
a function of the air flow rate is shown in Fig. 8. The largest
values of surface efficiency are associated with the
condensers with the shortest wires (50-mm long). The surface
efficiency decreases with the air flow rate as a result of the
increase in heat transfer coefficient (see below). By comparing
Fig. 8a and b, it can be seen that the number of passes also has
(17)

where t is the Student’s t-value (Holman, 2001), assumed a

equal to 2 for 95% confidence level, u0 is the uncertainty of the
measurement equipment (obtained from manufacturer data)
and s is the standard deviation. The uncertainties associated
with the measurements of temperature and pressure were
0.2  C and 0.1% of the full-scale of the pressure transducer.
The uncertainty of the voltage measurement in the data
acquisition board of the water mass flow meter was 0.3 mV.
The expanded uncertainties associated with the variables that
were not measured directly were calculated via the following
relationship (Holman, 2001),
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u N  2
uX vy
y ¼ f ðx1 ; x2 ; .; xN Þ/UðyÞ ¼ t Uðxi Þ (18)
i¼1
vxi

The experimental uncertainties for the air volume flow rate,

pressure drop, heat transfer rate, friction factor and Colburn j-
factor were estimated at 2 m3/h (maximum), 3%, 4%, 8%
and 6%, respectively.

4. Results b
4.1. Experimental results

As seen in Fig. 6, the heat exchanger pressure drop increases

with the air flow rate. For a fixed tube length, the condensers
with the largest values of pressure drop are those with four
passes and the smallest tube radial spacing (i.e., Samples #10
and #12), since these give rise to the smallest face area and,
consequently, to the highest values of Vmax. The longitudinal
spacing is also seen to influence the pressure drop. However,
because the longitudinal spacing does not affect the face
area, its influence is not as marked as that of the number of
tube passes and radial spacing. The condensers with
a longitudinal spacing of 50 mm tend to have a slightly larger
pressure drop than their 25 mm counterparts because they
are longer and have a larger surface area (and hence more
skin friction).
The air-side thermal conductance, ho ho Ao , as a function of
the air flow rate is shown in Fig. 7. The largest values of
thermal conductance are associated with condensers #14 and
#10, which have a large external surface area (due to the small Fig. 6 e Pressure drop as a function of the air flow rate. (a)
wire spacing and large longitudinal spacing) and a significant Condensers with two tube passes. (b) Condensers with
air velocity (due to the small face area and four tube passes). four tube passes.
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1
a a
b
Fig. 7 e Air-side thermal conductance as a function of the
air flow rate. (a) Condensers with two tube passes. (b)
Condensers with four tube passes.
some influence on the surface efficiency because the face
velocity increases with the number of tube passes.
The air-side heat transfer coefficient, ho , as a function of
the air flow rate is illustrated in Fig. 9. For the 2-pass heat
exchangers (Fig. 9a), the influence of each parameter on the
heat transfer coefficient is quite clear; it is high when the
longitudinal and radial tube spacings are small (25 and 11 mm,
respectively) and the wire spacing is large (10 mm). While the
small radial spacing certainly gives rise to large air velocities
in the heat exchanger, the small longitudinal tube spacing
possibly enhances heat transfer due to positioning of the
947
b
Fig. 8 e Air-side overall surface efficiency as a function of
the air flow rate. (a) Condensers with two tube passes. (b)
Condensers with four tube passes.
downstream tube pass in a more turbulent region in the wake
of the upstream tube pass. Small wire spacing may help
organize the flow in the longitudinal direction and possibly
suppress turbulence that otherwise enhances the heat
transfer coefficients for a wire spacing of 10 mm. It must be
mentioned, however, that these are just possible explanations
that need experimental validation are beyond scope of the
present work. For the 4-pass heat exchangers (Fig. 9b), the
influence of the radial tube spacing is not as clear as in the 2-
pass case, but it is evident that the longitudinal tube spacing
still plays a major role in determining the magnitude of the
heat transfer coefficient.
948 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

A combined evaluation of the heat transfer and pumping

power performances was carried out by means of the core
volume goodness factors (Kays and London, 1998; Shah and
Sekulic, 2003). Fig. 10 shows the thermal conductance per
unit volume as a function of the pumping power per unit
volume for the condensers with the smallest wire spacing
(large wire surface area). These parameters are defined by,

J1 ¼ ho ho b (19)

and

V_ a DPa
J2 ¼ b (20)
Ao

Fig. 10 e Behavior of the air-side thermal conductance per

unit volume as a function of the pumping power per unit
volume. Only the samples with a wire spacing of 5 mm are
shown.

For a given pumping power per unit volume, the best

surface from the point of view of the heat exchanger volume is
clearly that of Sample #2. As a matter of fact, when the friction
power expenditure is taken into consideration, the
condensers with two tube passes show a consistently better
performance than their four-pass counterparts. In other
words, the increase in overall thermal conductance provided
by the higher core velocity seen in Fig. 7 for the 4-pass
condensers is surpassed by the corresponding pressure drop

Fig. 9 e Air-side heat transfer coefficient as a function of

the air flow rate. (a) Condensers with two tube passes. (b) Fig. 11 e Experimental Colburn j-factor as a function of the
Condensers with four tube passes. air-side Reynolds number.
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1 949

penalty observed in Fig. 6. It is worth noting that the core

volume goodness factors of the 10-mm wire spacing heat
exchangers are not plotted in Fig. 10 because their values of
heat transfer conductance per unit volume are about one half
of those calculated for the condensers with 5 mm wire
spacing.

4.2. Empirical correlation

Figs. 11 and 12 present the experimental Colburn j-factor and

the friction factor as a function of the air-side Reynolds
number. The expected behavior of j and f decreasing with the
Reynolds number is observed. However, the scatter of the data
(especially for the pressure drop data) is an indication of the
influence of the geometric parameters that have to be taken
into account into empirical correlations for the dimensionless
parameters for air-side heat transfer and pressure drop
(Figs. 13 and 14).
Purely empirical correlations for the Colburn j-factor and
friction factor were derived in the following form,

j ¼ aj lnðRea Þ þ bj (21) Fig. 13 e Comparison between the experimental and

calculated Colburn j-factors.
f ¼ af lnðRea Þ þ bf (22)

where aj, bj, af and bf are empirical coefficients which have

been obtained via least-squares minimization using the
methodology proposed by Montgomery (2001). A general
expression for obtaining the coefficients is given by,
 
4þ1
a ¼ amin þ ðamax
amin Þ
2 number of tube passes, radial spacing, wire spacing and
where a can be aj, bj, af or bf, and amin and amax are tabulated
constants for each parameter (see Table 2). 4 is a normalized
(i.e., between
1 and þ1) value of a given by the following
general expression,
4 ¼ b0 þ b1 A þ b2 B þ b3 C þ b4 D (24)
where bk are empirical constants specific for each coefficient
(see Table 3). A, B, C and D are normalized values of the
(23)
longitudinal spacing, respectively. The normalization is such
that A, B, C and D can only be equal to
1 and þ1, depending of
the actual value of each parameter. For example, if the
number of tube passes is equal to 2, then A ¼
1. Otherwise, if

Fig. 12 e Experimental friction factor as a function of the Fig. 14 e Comparison between the experimental and
air-side Reynolds number. calculated friction factors.
950 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1

4, and their values are sufficiently accurate for engineering

Table 2 e Empirical constants for Eq. (20).
purposes. It should be mentioned that the correlations are
aj bj af bf valid only within the range of experimental conditions for
amin
1.0322  10
2
2.9533  10
2.1808  10
2
8.1111  10
2
2 which the data were obtained.
amax
2.6530  10
3 9.8156  10
2
6.0037  10
3 3.0152  10
1

5. Conclusions

The present paper carried out an experimental evaluation of

Table 3 e Empirical constants for determining 4 in Eq.
(21).
aj bj
b0 9.2359  10
2
4.9509  10
2 9.5615  10
2
8.3907  10
2
b1 2.2435  10
1
2.0238  10
1
4.5690  10
1 3.0414  10
1
b2
2.3447  10
1 2.6351  10
1
2.6119  10
1 3.3971  10
1
b3
2.7637  10
1 2.6612  10
1
1.0911  10
1 1.8273  10
1
b4 1.7465  10
2
5.3560  10
2 6.4297  10
2
1.3107  10
1
it is equal to 4, then A ¼ þ1. A similar procedure holds for the
other variables, where a linear interpolation can be per-
formed to determine intermediate values of normalized
variables.
Figs. 13 and 14 show the overall performance of the
empirical correlations in predicting the experimental values
of the Colburn j-factor and the friction factor. In order to
evaluate the ability of the present correlations in predicting
the data, the following quantities can be defined,
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u n

100 u
t
X zexp;i
zcal;i 2
RMS ¼
n i¼1
zexp;i
2
the air-side pressure drop and heat transfer characteristics of
spiral wire-on-tube condenser. Sixteen prototypes were
af bf designed and manufactured especially for this study, and
were subsequently tested in an open-loop wind tunnel calo-
rimeter. The influence of geometric parameters like the
number of tube passes, the radial and longitudinal tube
spacings and the wire spacing on the heat transfer and pres-
sure drop was investigated systematically for air flow rates
ranging from 70 to 220 m3 h
1. Empirical correlations were
developed for the Colburn j-factor and for the Darcy friction
factor. The RMS deviations of the correlations were 0.9% for
the Colburn j-factor (associated with the air-side heat transfer
coefficient) and 1.3% for the friction factor (frictional pressure
drop). The coefficients of determination were 0.96 and 0.94 for
j and f, respectively. It was demonstrated that all geometric
parameters investigated are important for the heat transfer
performance of the heat exchanger, with the most significant
contribution being that of the wire spacing. As far as the
pressure drop is concerned, the biggest individual influences
were caused by the radial tube spacing and number of tube
passes. This is due to their direct impact on the air velocity,
(25)
since a fixed tube length has been considered.

n  
100 X 
zexp;i
zcal;i 
AAD ¼ (26) Acknowledgements
n i¼1  zexp;i 

Financial support from CNPq through grant No. 573581/2008-8

(National Institute of Science and Technology in Cooling and
100 X
n
zexp;i
zcal;i
Thermophysics) is duly acknowledged. The authors thank
Bias ¼ (27)
n i¼1 zexp;i
Prof. Cláudio Melo for his support and technical advice.

 P
P 
P  2 references
n i zexp;i zcal;i
i zexp;i i zcal;i
R2 ¼ h P
P  2
ih P
P 2 i (28)
n i z2exp;i
i zexp;i n i z2cal;i
i zcal;i

ASHRAE Standard 37, 1988. Methods of Testing for Rating

where R2 is the coefficient of determination (Holman, 2001), n Electrically Driven Unitary Air-conditioning and Heat Pump
is the total number of experimental points, zexp,i is the Equipment. Atlanta, GA, USA.
experimental friction factor or Colburn j-factor and zcal,i is the ASHRAE Standard 41.2, 1987. Standard Methods for Laboratory
corresponding calculated value using the proposed correla- Airflow Measurement. Atlanta, GA, USA.
ASHRAE Standard 51, 1999. Laboratory Methods of Testing Fans
tions. A summary of the statistical parameters for the heat
for Rating. Atlanta, GA, USA.
transfer and pressure drop correlations is presented in Table Dasher, J.F., 1996. Apparatus and method of forming a refrigerator
condenser. US Patent No. 5,502,983.
ESDU, 1991. EffectivenesseNTU Relations for the Design and
Performance Evaluation of Two-stream Heat Exchangers,
Table 4 e Summary of the statistical quantities Engineering Science Data Unit 86018 with Amendment. ESDU
associated with the proposed correlations. International PLC, London, pp. 92e107.
Hermes, C.J.L., Melo, C., 2009. A heat transfer correlation for
AAD, % RMS, % Bias, % R2 natural draft wire-on-tube condensers. Int. J. Refrigeration 33,
Friction factor 6.83 1.28
0.04 0.94 546e555.
Colburn j-factor 7.21 0.85 0.03 0.96 Holman, J.P., 2001. Experimental Methods for Engineers, seventh
ed. McGraw-Hill.
 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 3 5 ( 2 0 1 2 ) 9 3 9 e9 5 1
Hoke, J.L., Clausing, A.M., Swofford, T.D., 1997. An experimental
investigation of convective heat transfer from wire-on-tube
heat exchangers. J. Heat Transfer 119, 348e356.
Incropera, F.P., de Witt, D.P., Bergman, T.L., Lavine, A.S.,
2007. Fundamentals of Heat and Mass Transfer, sixth ed..
Wiley.
Jenkins Jr., J.H., 2003. Three-layer condenser. US Patent No.
6,640,885 B2.
Kays, W.M., London, A.L., 1998. Compact Heat Exchangers, third
ed.. Krieger.
Lee, T., Yun, J., Lee, J., Park, J., Lee, K., 2000. Determination of
airside heat transfer coefficient on wire-on-tube type heat
exchanger. Int. J. Heat Mass Transfer 44, 1767e1776.
Montgomery, D.C., 2001. Design and Analysis of Experiments,
fifth ed.. Wiley.
Navarro, H.A., Cabezas-Gómez, L., 2005. A new approach for
thermal performance calculation of cross-flow heat
exchangers. Int. J. Heat Mass Transfer 48, 3880e3888.
951
Ohgaki, M., 2002. Condenser for a refrigerator. US Patent No.
D467,946 S.
Petroski, S.J., Clausing, A.M., 1999. An Investigation of the
Performance of Confined, Saw-tooth Shaped Wire-on-tube
Condensers. Technical Report ACRC-TR153. University of
Illinois.
Shah, R., Sekulic, D.P., 2003. Fundamentals of Heat Exchanger
Design. Wiley.
Sigwalt, R.A., 2010. Experimental analysis of wire-on-tube
condensers under forced convection for domestic
refrigeration applications, M.Sc. dissertation, Federal
University of Santa Catarina (in Portuguese).
Waltrich, P.J., Barbosa Jr., J.R., Hermes, C.J.L., Melo, C., 2011.
Air-side heat transfer and pressure drop characteristics of
accelerated flow evaporators. Int. J. Refrigeration 34, 484e497.
Wang, C.-C., Lee, C.-J., Chang, C.-T., Lin, S.-P., 1999. Heat transfer
and friction correlation for compact louvered fin-and-tube
heat exchangers. Int. J. Heat Mass Transfer 42, 1945e1956.