ln
i
a;in
i
r;out
i
a;out
i
r;in
_ _ : 5
Myers [6] derived the enthalpy-based overall heat transfer coecient U
o;w
to individual
resistance as
1
U
o;w
b
0
r
A
o
h
i
A
p;i
b
0
p
x
p
A
o
k
p
A
p;m
1
h
o;w
A
p;o
b
0
w;p
A
o
A
f
g
f;wet
b
0
w;m
A
o
_ _; 6
air
S
t
S
l
Staggered
Array
f
h
f
t
f
s
d
o
d
i
d
f
Crimped Spiral fins
Fig. 3. Relevant denitions of the geometrical parameters of crimped spiral ns.
where
h
o;w
1
C
p;a
b
0
w;m
h
c;o
y
w
k
w
: 7
Note that the ratio of water lm thickness and thermal conductivity of water y
w
=k
w
is very
small compared to other terms [7] and it is neglected in this study.
The tube side heat transfer coecient can be calculated from Gnielinski correlation [8] as
h
i
f
i
=2Re
Di
1000Pr
1:07 12:7
f
i
=2
_
Pr
2=3
1
k
i
d
i
_ _
; 8
where
f
i
1
1:58 ln Re
Di
3:28
2
: 9
The four quantities in Eq. (7) can be estimated following the method of Wang et al. [7] based on
the enthalpy-temperature ratios. b
0
r
and b
0
p
can be calculated as
b
0
r
i
s;p;i;m
i
r;m
T
p;i;m
T
r;m
; 10
b
0
p
i
s;p;o;m
i
s;p;i;m
T
p;o;m
T
p;i;m
: 11
The quantity b
0
w;p
is the slope of the saturated enthalpy curve evaluated at the outer mean water
lm temperature at the base surface and can be approximated at the slope of saturated enthalpy
curve evaluated at the base surface temperature of tube [7]. However, the quantity b
0
w;m
, which
denes as the slope of saturated enthalpy curve evaluated at the outer mean water lm temper-
ature at the n surface, cannot be calculated directly. Consequently, a trial and error procedure of
iteration is needed [7]. The detailed procedures are as follows:
1. Assume a value of T
w;m
and calculate the quantity b
0
w;m
.
2. Calculate h
o;w
from Eq. (6).
3. Calculate the quantity i
s;w;m
by the following relation:
i
s;w;m
i
C
p;a
h
o;w
g
f;wet
b
0
w;m
h
c;o
1
_
U
o;w
A
o
b
0
r
h
i
A
p;i
_
x
p
b
0
p
k
p
A
p;m
__
i i
r;m
: 12
4. Determine the new T
w;m
at i
s;w;m
and repeat the procedure again until the tolerance is met.
The wet n eciency is calculated as [7]
g
f;wet
2r
i
M
T
r
2
o
r
2
i
K
1
M
T
r
i
I
1
M
T
r
o
K
1
M
T
r
o
I
1
M
T
r
i
K
1
M
T
r
o
I
0
M
T
r
i
K
o
M
T
r
i
I
1
M
T
r
o
_ _
; 13
where
M
T
2h
o;w
k
f
f
t
2h
c;o
k
f
f
t
b
0
w
C
p;a
: 14
In this research work, the sensible heat transfer coecient h
c;o
and the pressure drop of air
stream across tube bank are presented in terms of the Colburn factor j and the friction factor
f factors,
j
h
c;o
G
max
C
p;a
Pr
2=3
; 15
f
A
min
A
o
q
i
q
m
2q
i
DP
G
2
c
_
1 r
2
q
i
q
o
_
1
__
: 16
4. Results and discussion
4.1. Sensible heat transfer coecient
The related heat transfer coecients and the air stream pressure drop vs. frontal velocity for all
the test samples are shown in Figs. 48. For comparison purpose, the relevant heat transfer
coecient in dry condition is also shown in the gure. It is found that the heat transfer coecient
of the wet surface is slightly lower than that of the dry surface. Actually, there are many studies
showing the comparison of the heat transfer coecients between wet and dry surface heat ex-
changer. Some studies indicated that the heat transfer coecient is augmented in wet surface
conditions, such as Myers [6], Elmahdy [10] and Eckels and Rabas [11] who reported results for
the continuous plate nned tube. These investigators argued that the presence of water condensate
0
10
20
30
40
50
60
70
80
0 0.5 1 1.5 2
h
c
o
(
W
/
m
2
K
)
0
10
20
30
40
50
60
70
80
90
P
(
P
a
)
Frontal Velocity (m/s)
fs = 3.85 mm, fh = 10 mm, St = 50.0 mm, Sl = 43.3 mm
do = 17.3 mm, wet
do = 27.2 mm, wet
do = 27.2 mm, dry
Fig. 4. Comparisons of the heat transfer coecient and the pressure drop in dry and wet conditions for samples 1
and 10.
may roughen the heat transfer surface, leading to higher heat transfer coecient. However, some
of the studies showed a drop of heat transfer coecient of wet surface, such as the wavy nned
tube heat exchanger by Mirth and Ramadhyani [12] who reported about 1750% decreasing of
heat transfer coecient of wet surface. One possible cause of the degradation is due to water lm
resistance and condensate blocking. Moreover Wang et al. [7] shows the decreasing of the Colburn
j factor of plate nned tube heat exchangers when the Reynolds number is lower than 2,000.
0
10
20
30
40
50
60
70
80
90
0 0.5 1 1.5 2
h
c
o
(
W
/
m
2
K
)
0
10
20
30
40
50
60
P
(
P
a
)
Frontal Velocity (m/s)
fs = 3.85 mm, do = 21.7 mm, St = 55.6 mm, Sl = 48.2 mm
fh = 10 mm, wet
fh = 15 mm, wet
fh = 10 mm, dry
fh = 15 mm, dry
Fig. 5. Comparisons of the heat transfer coecient and the pressure drop in dry and wet conditions for samples 8
and 9.
0
10
20
30
40
50
60
70
80
90
0 0.5 1 1.5 2
h
c
o
(
W
/
m
2
K
)
0
5
10
15
20
25
30
35
40
P
(
P
a
)
fh =10mm, do =21.7mm, S1=36mm
Frontal Velocity (m/s)
fs = 6.10 mm, wet
fs = 3.85 mm, wet
fs = 2.85 mm, wet
fs = 6.10 mm, dry
fs = 3.85mm, dry
fs = 2.85 mm, dry
St = 72 mm,
Fig. 6. Comparisons of the heat transfer coecient and the pressure drop in dry and wet conditions for samples 2, 3,
and 4.
However, at higher Reynolds number, the j factor of wet surface is slightly higher than that of the
dry surface. The present results are generally in agreement with the trend of Wang et al. [7].
The eect of tube diameter on the heat transfer performance is shown in Fig. 4, it is found that
the larger tube diameter (d
o
27:2 mm) has lower heat transfer coecient than that of the smaller
one (d
o
17:3 mm). This phenomenon is attributed to the fact that the ineective area behind the
tube increases with the tube diameter. Wang et al. [13] performed ow visualizations via dye
injection technique for n-and-tube heat exchangers having inline arrangement. Their visual re-
sults unveil a very huge ow circulation behind the tube row. Consequently this huge recirculation
0
10
20
30
40
50
60
70
80
90
0 0.5 1 1.5 2
h
c
o
(
W
/
m
2
K
)
0
5
10
15
20
25
30
35
40
P
(
P
a
)
Frontal Velocity (m/s)
fh = 10 mm, do = 21.7 mm, St = 84 mm, Sl = 24.2 mm
fs = 6.10 mm, wet
fs = 3.85 mm, wet
fs = 2.85 mm, wet
fs = 6.10 mm, dry
fs = 3.85mm, dry
fs = 2.85 mm, dry
Fig. 7. Comparisons of the heat transfer coecient and the pressure drop in dry and wet conditions for samples 5, 6,
and 7.
0
10
20
30
40
50
60
70
80
0 0.5 1 1.5 2
h
c
o
(
W
/
m
2
K
)
fs = 6.10 mm,St = 72 mm,Sl = 36 mm
fs = 3.85 mm,St = 72 mm, Sl = 36 mm
fs = 2.85 mm,St = 72 mm,Sl = 36 mm
fs = 6.10 mm,St = 84 mm,Sl = 24.2 mm
fs = 3.85 mm,St = 84 mm,Sl = 24.2 mm
fs = 2.85 mm,St = 84 mm,Sl = 24.2 mm
Frontal Velocity (m/s)
fh = 10 mm, do = 21.7 mm
Fig. 8. Comparisons of the heat transfer coecient in dry and wet conditions for samples 2, 3, 4, 5, 6, and 7.
not only contributes to the decrease of heat transfer coecient but also to the rise of pressure drop
as shown in the next discussions. In addition, the huge recirculation may also block the sub-
sequent tube row and degrade the heat transfer performance hereafter.
Fig. 5 shows the eect of n height on the heat transfer coecient. As seen in the gure, the
inuence of the n height is negligible in wet condition. For the fully dry case, Nuntaphan and
Kiatsiriroat [4] reported the 15 mm n height gives lower heat transfer coecient than that of 10
mm. This result comes from the airow by-pass eect. Actually the airow is prone to owing the
portion where the ow resistance is small. In case of f
h
15 mm, the airow resistance across the
nned tube portion is larger than that of f
h
10 mm. Therefore, part of the directed airow just
by-pass the tube row without eective contribution to the heat transfer, hence showing a lower
heat transfer coecient. However, in case of wet surface, the condensate of water vapor covering
the surface of heat exchanger becomes dominant and it increases the airow resistance around the
tube. Therefore the eect of n height is comparatively reduced.
The eect of n spacing is shown in Figs. 6 and 7. As seen in the gures, the eect of the n
spacing is small. However, small n spacing tends to have lower heat transfer coecient, and this
is particularly pronounced in Fig. 7 where S
t
84 mm and S
l
24:2 mm. This result may be
related to the water condensate eect. The presence of water condensate increases the air ow
resistance into the heat exchanger, thereby causing more airow to by-pass. This phenomenon
becomes more signicant with S
t
84 mm. In this regard, one can see a moderately decrease of
heat transfer coecient at smaller n spacing. The reports of McQuiston [14,15] also show the
decreasing of sensible heat transfer coecient when the n density is higher than 10 ns per inch.
Fig. 8 shows the eect of tube arrangement on the sensible heat transfer coecient. It is found
that higher transverse tube pitch tends to have lower heat transfer coecient and again this result
is attributed to the airow by-pass eect. In case of S
t
84 mm and S
l
24:2 mm, more airow is
prone to owing across the space between adjacent tube. Therefore, the amount of air stream
contributing to the heat transfer is decreased when compared to those of S
t
72 mm and S
l
36
mm.
The associated eect of geometric parameters on the heat transfer performance are correlated in
terms of the Colburn j factor, and is given as
j 0:0208Re
m
D
d
o
S
t
_ _
2:5950
f
t
f
s
_ _
0:7905
S
l
S
t
_ _
0:2391
d
o
d
f
_ _
0:2761
; 17
where
m 0:2871 0:5322
d
o
S
t
_ _
1:2856
f
t
f
s
_ _
0:1845
S
l
S
t
_ _
: 18
In Fig. 9, one can see the proposed j correlation can predict 95% of the experimental data
within 15% accuracy.
4.2. Pressure drop
The associated pressure drops for all the test samples are also shown in Figs. 47. In Fig. 4, the
inuence of tube diameter is examined. As seen in the gure, the pressure drops for wet condition
is only slightly higher than that of dry condition. This is because the n spacing in this gure is
comparatively large (3.85 mm). In this respect, the condensate can easily drain without adhering
to the interspacing of the ns, thereby giving only a slight increase in pressure drops of the wet
surface relative to dry conditions. However, one can nd a considerable inuence of tube size on
the total pressure drop. For the same frontal velocity of 1.5 m/s, the associated pressure drop for
d
o
27:2 mm is roughly 2.5 times higher that of d
o
17:3 mm.
The eect of n height on the total pressure drops is shown in Fig. 5. The pressure drops in-
crease with the n height because more n surface is provided. The eective surface area of
f
h
0:015 m is roughly 30% higher than that of f
h
0:01 m and the corresponding increase
pressure drop is also around 3040% which indicate a linear relationship of the n height and total
pressure drop. Conversely, one can go back to Fig. 4 where the eective surface area increase
caused by the tube size is less than 10% because the surface area is dominated by the secondary
surface (ns). However, the pressure drop is greatly increased with the tube size. The excess
pressure drop is attributed to (1) the drag of the large tube; and (2) the ineective ow separation
zone behind the tube which may increase its inuence to further downstream and results in more
pressure drops associated with it. This phenomenon is analogous to the continuous n pattern
reported by Wang et al. [9].
The eect of n spacing on the pressure drops is shown in Figs. 6 and 7. Notice that there is
very small dierence between dry and wet conditions since the tube diameter and n height is
relatively small which in term helps to drain the water condensate. Again smaller n spacing
increases the eective surface area and correspondingly higher pressure drops. The pressure drop
increases with tube diameter d
o
, n height f
h
and decreases for a smaller n spacing f
s
.
Among them, the inuence of the tube diameter is most pronounced. The eect of tube
arrangement is also found in Figs. 6 and 7. Higher transverse pitch of tube bank S
t
gives rise to
lower pressure drop. Again this is also attributed to the increase of surface area. The relevant
inuences of geometric parameters on the friction characteristics are correlated in terms of friction
factor, and is given as
0
0.01
0.02
0.03
0.04
0.00 0.01 0.02 0.03 0.04
+15%
-15%
Experimental data
P
r
e
d
i
c
t
e
d
j
v
a
l
u
e
Fig. 9. Comparison of the heat transfer data with the proposed correlation.
f 17:02Re
0:5636
D
d
o
S
t
_ _
0:3956
f
t
f
s
_ _
0:3728
S
l
S
t
_ _
1:2804
d
o
d
f
_ _
0:1738
: 19
From Fig. 10, one can see the proposed friction factor correlation can predict 75% of the
experimental data within 15% accuracy.
5. Conclusion
This study experimentally investigated the air-side performance of a total of 10 cross ow heat
exchangers having crimped spiral congurations under the dehumidication. The eects of tube
diameter, n spacing, transverse tube pitch are examined. Based on the experimental observa-
tions, the following results are concluded as:
1. The pressure drop of wet surface heat exchanger increases with the mass ow rate of air and the
result is slightly higher or close to that of the dry surface because only water condensate can be
easily drained in the present comparatively large n spacing and individual nned congura-
tion.
2. The heat transfer coecient of the wet surface is slightly lower than that of the dry surface.
3. The eect of the tube diameter on the air-side performance is signicant. A larger tube diameter
not only gives rise to a lower heat transfer coecient but also contribute signicantly to the
increase of pressure drops. This phenomenon is applicable in both dry and wet condition.
4. For the wet surface, the inuence of n height is negligible whereas it has a small eect on the
dry surface.
5. The eect of the n spacing on the heat transfer performance is rather small. However, the
increasing of the n spacing tends to have a lower heat transfer coecient.
6. The tube arrangement plays an important role on the heat transfer coecient. A lower trans-
verse pitch gives a higher heat transfer coecient.
0
0. 2
0. 4
0. 6
0. 8
0.0 0. 2 0.4 0.6 0.8
+15%
-15%
Experimental data
P
r
e
d
i
c
t
e
d
f
v
a
l
u
e
Fig. 10. Comparison of the frictional data with the proposed correlation.
7. Air-side performance in the present study is presented in terms of f and the j factor. The pro-
posed correlations can predict 75% and 95% of the experimental data within 15%.
Acknowledgements
The authors gratefully acknowledge the support provided by the Thailand Research Fund for
carrying out this study. Part of the nancial support provided by the Energy R&D foundation
funding from the Energy Commission of the Ministry of Economic Aairs, Taiwan is also
appreciated.
References
[1] D.E. Briggs, E.H. Young, Convective heat transfer and pressure drop of air owing across triangular pitch banks
of nned tubes, Chem. Eng. Prog. Symp. Ser. 59 (41) (1963) 110.
[2] K.K. Robinson, D.E. Briggs, Pressure drop of air owing across triangular pitch banks of nned tubes, Chem.
Eng. Prog. Symp. Ser. 62 (64) (1966) 177184.
[3] T.J. Rabas, P.W. Eckels, R.A. Sabatino, The eect of n density on the heat transfer and pressure drop
performance of low nned tube banks, Chem. Eng. Commun. 10 (2) (1981) 127147.
[4] A. Nuntaphan, T. Kiatsiriroat, Heat transfer characteristic of cross ow heat exchanger using crimped spiral ns a
case study of staggered arrangement, The 17th Conference of Mechanical Engineering Network of Thailand,
Thailand, 2003.
[5] J.L. Threlkeld, Thermal Environmental Engineering, Prentice-Hall, New York, 1970.
[6] R.J. Myers, The eect of dehumidication on the air-side heat transfer coecient for a nned-tube coil,
M.S. Thesis, University of Minnesota, Minneapolis, 1967.
[7] C.C. Wang, Y.C. Hsieh, Y.T. Lin, Performance of plate nned tube heat exchanger under dehumidifying
conditions, ASME J. Heat Transfer 119 (1997) 109119.
[8] V. Gnielinski, New equation for heat and mass transfer in turbulent pipe and channel ow, Int. Chem. Eng. 16
(1976) 359368.
[9] C.C. Wang, Y.P. Chang, K.Y. Chi, Y.J. Chang, An experimental study of heat transfer and friction characteristics
of typical Louver n and tube heat exchangers, Int. J. Heat Mass Transfer 41 (45) (1998) 817822.
[10] A.H. Elmahdy, Analytical and experimental multi-row, nned-tube heat exchanger performance during cooling
and dehumidication process, Ph.D. thesis, Carleton University, Ottawa, Canada, 1975.
[11] P.W. Eckels, T.J. Rabas, Dehumidication: on the correlation of wet and dry transport process in plate nned-tube
heat exchangers, ASME J. Heat Transfer 109 (1987) 572582.
[12] D.R. Mirth, S. Ramadhyani, Prediction of cooling-coils performance under condensing conditions, Int. J. Heat
Fluid Flow 14 (4) (1993) 391400.
[13] C.C. Wang, J. Lou, Y.T. Lin, C.S. Wei, Flow visualization of annular and delta winlet vortex generators in n-and-
tube heat exchanger application, Int. J. Heat Mass Transfer 45 (2002) 38033815.
[14] F.C. McQuiston, Heat mass and momentum transfer data for ve plate-n tube transfer surface, ASHRAE Trans.
84 (1) (1978) 266293.
[15] F.C. McQuiston, Correlation of heat mass and momentum transport coecients for plate-n tube transfer surfaces
with staggered tubes, ASHRAE Trans. 84 (1) (1978) 294309.