You are on page 1of 13

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/270356127

Experimental and numerical analysis of a


cross-flow closed wet cooling tower
Article in Applied Thermal Engineering November 2013
Impact Factor: 2.74 DOI: 10.1016/j.applthermaleng.2013.08.043

CITATIONS

READS

66

3 authors, including:
Xiaohua Liu

Yi Jiang

Tsinghua University

Tsinghua University

89 PUBLICATIONS 1,325 CITATIONS

146 PUBLICATIONS 1,669 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,


letting you access and read them immediately.

SEE PROFILE

Available from: Xiaohua Liu


Retrieved on: 21 June 2016

Applied Thermal Engineering 61 (2013) 678e689

Contents lists available at ScienceDirect

Applied Thermal Engineering


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

Experimental and numerical analysis of a cross-ow closed


wet cooling tower
Jing-Jing Jiang, Xiao-Hua Liu*, Yi Jiang
Department of Building Science, Tsinghua University, Beijing 100084, PR China

h i g h l i g h t s
 A cross-ow closed wet cooling tower (CWCT) is experimentally analyzed.
 Empirical correlations of the heat and mass transfer coefcients are obtained.
 Numerical model of the CWCT is established and validated by experimental data.
 Heat and mass transfer driving forces inside a cross-ow CWCT are more uniform.
 Performance of a cross-ow CWCT is better than parallel/counter-ow patterns.

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 3 January 2013
Accepted 31 August 2013
Available online 10 September 2013

Closed wet cooling tower (CWCT) is an indirect-contact evaporative cooler, in which ambient air, spray
water and process water function together. In this study, a cross-ow CWCT unit based on the plateen
heat exchanger was designed and tested under various conditions in an environmental chamber. The test
results suggest that the heat and mass transfer coefcients and the cooling efciency are remarkably
affected by the temperature of the process water and the ow rates of the air, the spray water and the
process water. Heat and mass transfer coefcients were correlated based on the sensitive parameters.
Two-dimensional steady-state numerical model of the cross-ow CWCT was established and validated
by the experimental data. The numerical analyses revealed that the cross-ow CWCT could breakthrough
the structure limitation of the commonly parallel/counter-ow conguration and obtain more uniform
driving forces, which is benecial for the cooling performance. The ow pattern optimization of the
CWCT shows that air and process water in the opposite direction, spray water and the other uids in the
cross direction is the best ow pattern, which is distinct from the general knowledge of the researches.
2013 Elsevier Ltd. All rights reserved.

Keywords:
Closed wet cooling tower
Experiment
Numerical model
Flow pattern
Cross ow

1. Introduction
Closed wet cooling tower (CWCT) has been adopted in a wide
range of application elds [1], such as refrigeration, airconditioning, manufacturing, power generation, etc. CWCT is an
indirect-contact evaporative cooler mostly based on tubular heat
exchanger structure. Three uids function together in the CWCT,
which are ambient air and spray water owing outside the tubes
and process water running inside the serpentine tubes. The principle of CWCT can be split into evaporative heat and mass transfer
process between the ambient air and the spray water, and heat
transfer process between the spray water and the process water. As
the uid inside the tubes never contact the ambient air, the CWCT

* Corresponding author. Tel.: 86 10 6277 3772; fax: 86 10 6277 0544.


E-mail address: lxh@mail.tsinghua.edu.cn (X.-H. Liu).
1359-4311/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.applthermaleng.2013.08.043

can be used to cool uids other than water and prevent contamination of the airborne dirt and impurities. Furthermore, CWCT
could operate as an air cooling tower by stopping spray water in
severe cold days which makes it possible to run continuously yearround in hospitals, schools, data centers, etc. However, the cost of
CWCT is often higher since tubular heat exchanger needs quantity
of metallic materials [2].
Series of experiments have been conducted for the fundamental
researches of the heat and mass transfer processes in CWCTs. Niitsu
et al. [3] tested the performance of the plain and nned tubes,
including the lm heat transfer coefcient and airewater mass
transfer coefcient. Experimental tests by Heyns and Krger [4]
showed the water-lm heat transfer coefcient was a function of
spray water temperature, spray water and air ow rates, while the
airewater mass transfer coefcient was a function of air and spray
water ow rates. Sarker et al. [5] assessed CWCTs with staggered
arranged bare-type or nned tubes, from the perspectives of

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

679

Fig. 1. The cross-ow CWCT unit: (a) the schematic diagram of the three uids; and (b) the photo from the front view.

Fig. 2. The louver structure of the n.

cooling capacity, wet-bulb efciency and pressure drop. Experimental tests showed that the n-tube CWCT had better thermal
performance although the pressure drop was higher than that of
the bare-tube one. Zheng et al. [6] investigated the thermal
behavior of an oval tube CWCT under different operating conditions. The results showed that the oval tube had a better combined
thermal-hydraulic performance. Some novel CWCTs consisting of
indirect evaporative cooling stage and direct evaporative cooling

stage (or heat transfer stage) were proposed, constructed and


tested by Xia et al. [2] and Heidarinejad et al. [7].
Besides the experimental researches, a number of theoretical
and computational analyses have been conducted aiming to a more
realistic description of the transport phenomena taking place inside a CWCT. Hasan and Sirn [8] presented a computational model
to simulate the performance of the CWCT. The variation of the spray
water temperature was taken into consideration and the saturation
enthalpy was calculated from psychometric relations for moist air.
The coefcients of mass transfer were derived from experimental
data and then implemented in the computational model. Koschenz
[9] presented an analytical model for a CWCT for use with chilled
ceilings, assuming that the spray water temperature kept constant
along the way and the constant temperature was equal to the outlet
process water temperature. However, the accuracy levels of these
assumptions were not quantied with respect to other approaches
or relevant experimental works. Hasan and Gan [10] compared the
cooling performances calculated by the computational model
established by Hasan and the analytical models utilizing the assumptions raised by Koschenz. Gan and Riffat [11] conducted a CFD

Fig. 3. The schematic diagram of the testing conguration.

680

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

Table 1
Accuracies of the measuring instruments.
Parameter

Sensor

Accuracy

Air dry/wet bulb temperatures


Air ow rate
Spray water/process water
temperatures
Spray water ow rate
Process water ow rate

T-type thermocouple
Standard nozzle (GB14294)
T-type thermocouple

0.2 [ C]
1 [%]
0.2 [ C]

Rotameter
Water meter

1.5 [%]
3 [L/h]

method to predict the performance of CWCT according to the


cooling capacity and the pressure drop.
Plenty of works have been done by the researchers concerning
experimental tests and theoretical analyses, which give a good
description of the performance of the CWCT. Moreover, many novel
designs of structures or components have been put forward and
have already reformed the CWCT to a great extent. However, the
ow pattern analysis of the CWCT has barely been involved in
available literature. The majority of the ow pattern of the CWCT
[3e6,8e16] is air owing from the bottom to the top, while the
process water inside the serpentine tubes and the spray water
outside the tubes owing in the opposite direction, which has been
proved the best one-dimensional ow pattern of CWCT by Ren and
Yang [17]. Little work [18,19] has been carried out on the cross-ow
CWCT. The performance of a cross-ow CWCT, with a counter ow
between the air and the process water and two cross ows between
the air and the other two uids, will be analyzed in present study.
Experimental tests will be carried out to investigate the behavior
and inuencing factors of the CWCT. Numerical models of parallel/
counter-ow and cross-ow CWCTs will be established and validated to optimize the ow pattern of the CWCT.
2. Experimental test of the CWCT unit

Fig. 2, to strengthen the heat and mass transfer performance and


the wettability of the ns by disturbing the boundary layer of the
spray water.
The size of the unit is 570 mm (H, the spray water ow
direction)  310 mm (W, width)  180 mm (L, the air ow direction). Its n thickness is 0.127 mm, and the distance between the
ns is 2.2 mm. The inner and external diameters of the tubes are
9.42 mm and 10.02 mm, respectively. The tube bundle consists of 8
rows of steel tubes. The tubes are 0.31 m long and are arranged in a
triangular pattern at a transversal pitch of 25.4 mm. There are 20
tubes per tube row. The external surface of the whole unit is
24.336 m2 (Fm), 0.784 m2 of which is the external surface of the
tubes and 23.552 m2 of which is the surface of the ns. The specic
surface area of the CWCT unit is 790 m2/m3.
2.2. The CWCT testing conguration
The tests for assessing the performance of the CWCT unit were
performed in an environmental chamber. The system conguration
can realize wide range of air, spray water and process water states,
as displayed in Fig. 3. The cooling coil, heater A and humidier can
regulate the air inlet temperature and humidity independently. The
variable frequency fan can control the volume ow rate of the inlet
air. Also, the temperature and ow rate of the process water can be
controlled by Heater B and the water valves.
The environmental chamber provided the measuring instruments for the ow rates and inlet/outlet temperatures of the
air, the spray water and the process water. As listed in Table 1, the
ow rate of the air a was measured by standard nozzles
(GB14294) with the accuracy of 1%. The ow rate of the spray water
s was measured by a rotameter with the range from 60 to 600 L/h
and the accuracy of 1.5%. The ow rate of the process water w was
measured by water meter with the accuracy of 3 L/h. The temperatures of the three uids were measured by T-type thermocouples
with the accuracy of 0.2  C.

2.1. Description of the cross-ow CWCT unit


2.3. Verication of the experimental data
Fig. 1(a) indicates the ow directions of the three uids in the
cross-ow CWCT unit, which are spray water owing from top to
bottom, ambient air owing from front to back through the ns and
process water owing in the serpentine tubes from back to the
front. In other words, the ambient air and process water are in
counter ow and spray water is in cross ow with the other two
uids. As shown in Fig. 1(b), the CWCT unit employs n-tube
structure to expand the heat and mass transfer area. The tubes
and ns of the unit are made of stainless steel to ensure perfect heat
transfer between the spray water and the process water. Furthermore, there are discontinuous louvers on the ns, presented in

In order to study the CWCT unit, a series of experiments were


conducted which intended for nding out the effects of the inlet
parameters on cooling performance. The main parameters are the
ow rate of the three uids and the inlet temperature of the process
water etc. Variable condition analyses were conducted with 11
operating conditions and 46 sets of data. Each set of data was
recorded under approximately steady states, which required all the
temperature points uctuated within 0.2  C for longer than
20 min. The typical experiment data is listed in Table 2. To verify the
reliability of the experiment data, energy balance of the air, spray

Table 2
The experimental data of the CWCT test.
No.

Inlet parameters


1
2
3
4
5
6
7
8
9
10
11

Outlet parameters


Qc kW

0.36
0.35
0.37
0.28
0.32
0.36
0.40
0.45
0.34
0.38
0.40

4.94
5.65
7.26
3.31
4.31
3.49
4.10
3.39
3.22
3.58
3.77

ta,in C

twb,in C

ts,in C

tw,in C

a kg/s

s kg/s

w kg/s

ta,out C

twb,out C

ts,out C

tw,out C

24.8
25.3
27.0
27.2
26.7
27.3
25.1
25.1
26.8
26.9
27.0

20.1
20.9
22.0
21.1
20.6
22.9
20.8
21.1
22.6
22.8
22.6

24.7
25.8
27.9
26.2
25.6
26.1
24.8
24.6
26.2
25.9
25.8

30.3
32.9
36.6
30.2
30.8
30.2
30.4
30.2
30.2
30.2
29.9

0.35
0.35
0.35
0.19
0.27
0.35
0.35
0.35
0.35
0.35
0.35

0.13
0.12
0.12
0.12
0.12
0.12
0.12
0.12
0.06
0.09
0.13

0.32
0.32
0.32
0.32
0.32
0.31
0.26
0.20
0.29
0.30
0.30

24.2
25.3
27.3
26.0
25.2
25.6
24.4
24.2
25.1
25.3
25.3

23.9
25.0
27.0
25.7
24.8
25.2
24.0
23.8
24.6
25.0
25.0

24.4
25.8
27.8
26.0
25.3
25.8
24.5
24.3
25.6
25.5
25.4

26.7
28.8
31.2
27.7
27.6
27.5
26.6
26.1
27.6
27.4
27.0

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

Gained heat of the air (kW)

3. Experimental results and inuencing factors

3.1. Effect of the air ow rate

+ 20

6
5
4

-20 %

3
2
1
0

681

1
2
3
4
5
6
7
Heat loss of process/spray water (kW)

Fig. 4. Energy balance of the CWCT test.

water and process water was adopted. As shown in Fig. 4, the unbalance ratios of the heat gained by the ambient air and the heat
lost by the process water and the spray water are within 20%. The
average absolute unbalance ratio is 7.4%, which means the data are
reliable.
To better describe the cooling processes, some indexes are
introduced, seen in Eqs. (1) and (2). The wet-bulb cooling efciency
[5,13,16] illustrates the distance between the outlet process water
temperature (tw,out) and the ambient wet bulb temperature (twb,in),
which is the limit of the outlet process water temperature. The
cooling capacity Qc [5,12,15] presents the cooling capacity of the
CWCT unit.

In the environmental chamber shown in Fig. 3, the ow rate of


the ambient air is easily conditioned and controlled by regulating
the rotate speed of the fan. The ow rate of the air was at the lowest
rate of 0.19 kg/s and up to the maximum of 0.35 kg/s Fig. 5 displays
that mass transfer coefcient between the spray water and the air
(Km) increased greatly by increasing the air ow rate. While the
heat transfer coefcient between the spray water and the process
water (Kh) was generally constant since the increase of the air ow
rate had little relationship with the heat transfer between the spray
water and the process water. As a result of the strengthening of the
heat and mass transfer performance between the spray water and
the air, and Qc increased with the increase of a. Thus, increasing
air ow rate is a good way to improve the performance of the
CWCT. However, the air ow rate is not the bigger the better if
taking the fan power consumption into consideration. There is an
optimal value of air ow rate depending on the balance of the
CWCT performance and the fan power consumption.
Uncertainty analyses for the experimental results, based on the
accuracies of the measuring instruments introduced in Section 2.2,
were conducted in this study using the method proposed by Kline
and McClintock [20] according to the following expression (Eq. (3)):

Dy



vf
vx1

2

Dx1 2

vf
vx2

2

Dx2 2 /

vf
vxn

2

Dxn 2

1=2
(3)

tw;in  tw;out
tw;in  ta;wb

(1)



_ w tw;in  tw;out
Qc cp;w m

(2)

The uncertainties of , Qc, Km, and Kh were calculated and


expressed in Fig. 5 in the way of error bars. Results show that the Ttype thermocouples are the main sources of errors. Taking Km as an
example, the uncertainties of the air wet-bulb temperature and
spray water temperature account for 81.3% and 16.6% of the total
uncertainty, respectively, while that of the air mass ow rate only

Fig. 5. Effects of the air ow rate (a) on the CWCT: (a) wet-bulb efciency; (b) cooling capacity; (c) Km; and (d) Kh.

682

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

Fig. 6. Effects of the spray water ow rate (s) on the CWCT: (a) wet-bulb efciency; (b) cooling capacity; (c) Km; and (d) Kh.

accounts for 0.8%. Thus improving the temperature measurement


accuracy is the key point of improving the accuracy of the test.
3.2. Effect of the spray water ow rate
The ow rate of the spray water was regulated by changing the
valves in the pipelines. The ow rate of the spray water was from

0.06 kg/s to 0.13 kg/s. When the ow rate of the spray water
increased, wetting degree of the CWCT was improved and heat and
mass transfer area was expanded to a certain extent. Also, the increase of the spray water ow rate would strengthen the heat
transfer process between the spray water and the process water, so
as to take away more heat from the process water. As a result, , Qc,
and Kh increased, as shown in Fig. 6. On the other hand, the

Fig. 7. Effects of the process water ow rate (w) on the CWCT: (a) wet-bulb efciency; (b) cooling capacity; (c) Km; and (d) Kh.

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

683

Fig. 8. Effects of the process water temperature (tw,in) on the CWCT: (a) cooling capacity; (b) spray water inlet temperature; (c) Km; and (d) Kh.

(b)

(a)

Fig. 9. Calculated results of Eqs. (4) and (5): (a) Kh; and (b) Km.

Table 3
Comparison of the experimental parameters in literature [4,6,15,16].
Source

Flow pattern

Ga (kg/m2s)

Gs (kg/m2s)

Gw (kg/m2s)

Kma (kg/m3s)

Kha (kW/m3 K)

a (m2/m3)

Correlations

Heyns [4]

Parallel/counter

0.7e3.6

1.7e4.5

0.5e3.2

42.0e67.2

24

Zheng [6]

Parallel/counter

2.5e5.0

1.2e3.2

2.8e5.3

2.7e5.0

23.9e60.4

31

0.11e0.19

Shim [15]
Faco [16]

Parallel/counter
Parallel/counter

1.2e4.2
0.7e2.4

1.1e3.3
0.3e1.9

0.9e4.8
0.6e1.1

6.6e21.5
1.6e4.3

18.2 31.4
5.5e17.5

33
25

e
0.2e0.65

Present study

Cross

1.3e2.4

1.1e2.3

1.1e1.8

10.3e19.0

30.8e45.0

790

0.28e0.46

0.35 0.3
ts
Kh 470G0.1
a Gs
Km 0.038G0.73
G0.2
a
s
Kh 350.3(1 0.0169ts)G0.59
G1/3
a
s
0.977
Km 0.034Ga
e
Kh 700.3(s/1.39)0.6584
Km 0.1703(a/1.7)0.8099
0.547
Kh 31.79Gs0.238Gw
Km 0.00154t0.471
G0.694
G0.512
w
a
s

684

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

(b)

(a)

Fig. 10. Comparison with literature [4,6]: (a) Kha; and (b) Kma.

improvement of the spray water ow rate enhanced heat and mass


transfer between the ambient air and the spray water, which
resulted in the growth of Km.

by the air and process water states, it rose with the growth of the
process water temperature, as shown in Fig. 8(b).
3.5. Comparison with experimental results from previous studies

3.3. Effect of the process water ow rate


The ow rate of the process water is also a key parameter
inuencing the CWCT performance. It was from 0.20 kg/s to
0.32 kg/s in this set of experiments. Apparently, the rise of the
process water ow rate promoted the heat and mass transfer between the spray water and the process water, therefore Kh
increased. Thus process water could release more heat to the spray
water, which led to the increase of Qc, as shown in Fig. 7(b). Since
the inuence of w to the outlet temperature of the process water
was more remarkable than that of Kh, the cooling effect of the
process water per unit mass was denitely worsened, though the
total cooling capacity was improved. In this way, the outlet temperature of the process water increased and dropped. Since the
process water barely touched the ambient air, Km between the air
and the spray water scarcely changed.
3.4. Effect of the process water temperature
This set of experiment was meant to study the performance of
the CWCT at different process water temperatures. The inlet temperature of the process water was from 30.1  C to 36.6  C. The results showed that Qc and Km increased with the growth of the
process water temperature, seen in Fig. 8(a) and (c), due to the
increase of the heat and mass transfer driving forces between the
three uids. Since the temperature of the spray water was decided

Fig. 11. Control volume schematic of the cross-ow CWCT.

From the sensitivity analyses we could see that Kh is mainly


inuenced by the ow rates of the spray water and the process
water, while Km is dominated by the ow rates of the air and spray
water and the inlet temperature of the process water. Therefore, for
the specic geometry of the tower, the correlation equations for Kh
and Km could be presented as follows:

Kh 31:79G0:238
G0:547
s
w

(4)

0:471 0:694 0:512


Km 0:00154tw;in
Ga
Gs

(5)

where Ga is the air velocity in the minimum ow area, Gs G/do,


Gw w/(H$L) is the process water ow rate per ow area
(1.3 < Ga < 2.4 kg/m2 s; 1.1 < Gs < 2.3 kg/m2 s; 1.1 < Gw < 1.8 kg/
m2 s; 30.1 < Tw,in<36.6  C). Fig. 9 compares the calculated results
based on the correlation equations and the experimental data.
Since good agreement is obtained, the correlation equations can be
used to predict the coefcients.
Table 3 compares the experimental parameters of previous
studies in literature and present study. The ow pattern of parallel/
counter can refer to Fig. 15(a) and the ow pattern of cross can refer
to Fig. 11. As listed in Table 3, the most popular ow pattern is
parallel/counter-ow, while the most common structure is planetube bundles with specic surface area of around 25e35 m2/m3.
In present study, the CWCT is cross-ow with a considerable specic surface area of 790 m2/m3. The correlations of Heyns et al. [4]
shows that Kh has positive correlations with the air and the spray
water ow rates and the spray water temperature, while Km has
positive correlations with the air and the spray water ow rates. Km
is only inuenced by the air ow rate in the studies of Zheng et al.
[6] and Faco et al. [16]. In present study, similar correlations of Kh
and Km were obtained. There are some modications about the
selection of the inuencing factors. Firstly, tw,in, instead of ts, is seen
an inuencing factor because its the source of the heat and mass
transfer process while ts is only an equilibrium parameter of the air
and process water states. Secondly, the process ow rate Gw is
adopted to describe the correlation of Kh since the heat transfer
process actually happens between the spray water and the process
water.
In Fig. 10, Eqs. (4) and (5) are compared to the correlations given
by Heyns et al. [4] and Zheng et al. [6]. Compared to the bare-tube
parallel/counter-ow CWCT, the cross-ow CWCT with plateen
tube structure studied in this paper has much larger volume mass

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

685

Fig. 12. Comparison of the calculated values and the experimental results of the cross-ow CWCT: (a) the variance of the air humidity ratio; and (b) the variance of the temperature
of the process water.

transfer coefcient Kma, due to that the n-tube structure


tremendously extends the contact area, which is exactly the restriction point of the heat and mass transfer between the air and
the spray water. In contract, the performance improvement of the
n-tube structure for waterewater heat transfer process is insignicant. Thats why the volume heat transfer coefcient Kha is no
bigger than the others.
4. Numerical model of the cross-ow CWCT
4.1. Theoretical model
Two-dimensional steady-state model of the cross-ow CWCT, in
which air and process water owing in the opposite direction, spray
water owing in the cross direction with the other two uids, seen
in Fig. 11, will be illustrated in this section. In the CWCT, process
water releases heat to the metal nned tubes while the tubes are
cooled by the spray water. At the same time, heat and mass transfer
takes place between the spray water and the air. The main assumptions for the numerical model are presented as follows
[8,13,17]: 1) Heat and mass transfer processes are at steady state; 2)
The heat and mass exchange between the CWCT unit and the surroundings is negligible; 3) The specic heat of the uids are
assumed to be constant; 4) The spray water lm uniformly covers
all the wall of the tubes and ns, so that Fh equals to Fm and the heat
exchange between the air and process water is negligible; and 5)
The ow of the process water and the air is approximately counter
ow.
Dene the air ow direction as z axis and the spray water ow
direction as x axis. The energy balance equation for the three uids
is:

_ s hs
_ a vha 1 vm
_ w vtw
m
m
 cp;w

0
L
vx
H vz
H vz

(6)

Mass conversion equation of the spray water and the air is given

_s
_ a vda 1 vm
m

0
L vx
H vz

(7)

Heat transfer between the spray water and the process water
driven by the temperature difference between them is shown as:

Table 4
Simulated condition of the cross-ow CWCT.
ta,in  C da,in kg/kg tw,in  C a kg/s s kg/s w kg/s KmFm kg/s KhFh kW/K
0.0136

30.3

(8)

As well known, there is a thin lm of saturated air at the


interface between the spray water and the air. The temperature of
the saturated air is close to that of the spray water. The humidity
ratio of the saturated air is also called the equivalent humidity ratio
of the spray water, which is de. Heat transfer driven by the temperature difference of the saturated air lm and the air ow and
mass transfer driven by the water vapor partial pressure difference
between the two streams take place simultaneously. Thus the mass
transfer equation and the energy balance equation for the air ow
can be expressed by the following equations separately:

vda
Km Fm

de  da
_ aL
vz
m

(9)

vha
K 0 Fm
vda
h
t  ta r
_ aL s
vz
vz
m

(10)

The Lewis factor or Lewis relation Lef could be dened to indicate the relation between the heat and mass transfer in an evaporative process [21e23]. The denition of Lef is as follows:

Lef

K 0h
Km cp;m

(11)

Substitute Eq. (11) into Eq. (10):

i
vha
Km Fm h

Lef $cp;a ts  ta rde  da


_ aL
vz
m
As the enthalpy of the air can be expressed
ha cp;m ta r$da ; Eq. (12) can be transformed into:





vha
Km Fm
1

$Lef he  ha r
 1 de  da
_ aL
Lef
vz
m

(12)
as

(13)

Thus we get all the governing equations of the cross-ow CWCT.


The boundary conditions are shown as follows:

by:

27.2

vtw
Kh Fh
t  ts

_ wL w
vz
cp;w m

0.19

0.12

0.32

0.365

1.17

ta ta;in ;

da da;in ;

ha ha;in ;

z 0

(14)

ts jx0 ts jxH

(15)

tw tw;in ;

(16)

z L

By discretizing the governing equations, the heat and mass


transfer process could be numerically solved. When solving the
model, Lef could be equal to 1 [6,8,14]. The model of the cross-ow
CWCT was validated by the experimental results described in

686

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

Fig. 13. Simulated eld distribution of the cross-ow CWCT: (a) the air wet-bulb temperature; (b) the spray water temperature; (c) the process water temperature; (d) the
temperature difference between the spray water and the process water; and (e) the temperature difference between the spray water and the air (wet-bulb).

Section 3. As shown in Fig. 12, the maximal differences between the


calculated results and the experimental values are within 8.0%, and
the average absolute differences are 3.0% and 4.0% for the variance
of the air humidity ratio and the process water temperature
respectively. On the whole, the calculated parameters by the numerical model agree well with the experimental results, the model
could be used to analyze the heat and mass transfer performance of
the CWCT unit in the following content.
4.2. Typical simulation result of cross-ow CWCT
Since the distribution parameters of the cross-ow CWCT are
two-dimensional, it is difcult to describe it through experimental

results of limited measurement points. Therefore, numerical


modeling results were introduced to investigate the performance of
the CWCT. The boundary conditions from the experimental results
are displayed in Table 4. As seen in Fig. 13, the wet-bulb temperature
of the air increases in the air ow direction, while the temperature of
the process water decreases in the opposite direction. Heat is
transferred from the process water to the air. Without the circulation
of the spray water, the temperature distributions of the air and the
process water should be one-dimensional and the temperature
gradients of the two uids along z should be consistent. Once the
spray water is brought in, the consistency will be disturbed.
To better explain the heat and mass transfer process of crossow CWCT, we could divide it into enough control volumes along

Fig. 14. Simulated temperatures of the x sections: (a) x 0.05H; (b) x 0.5H; and (c) x 0.95H.

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

687

Fig. 15. (a) Parallel/counter-ow CWCT conguration; and (b) simulated temperatures.

z axis and assume that the heat and mass transfer between each
two control volumes could be ignored, which means the performance of the spray water is only affected by the air and the process
water inside the volume. If the inlet temperature of the spray water
is lower than those of the process water and the air (wet-bulb),
spray water will absorb heat from both the uids when falling
along the x axis. As a result, the spray water temperature will go up
until it gains the same heat from the process water as the heat
released to the air. Vice versa, when the inlet temperature of the
spray water is higher than those of the other uids, spray water will
discharge heat to the air and the process water until it transfers the
same quantity of heat from the process water to the air. In this way,
there is an equilibrium temperature of spray water in each control
volume, which is somewhere between the temperatures of the air
and the process water, determined by the heat and mass transfer
ability of the CWCT. As shown in Fig. 13(b), on the bottom of the
CWCT, the spray water temperature barely changes along the way,
which means it already reaches the equilibrium temperature.
Fig. 14 shows the temperatures of the three uids at the sections
of x 0.05H, x 0.5H and x 0.95H, which represent the states of
the spray water from the inlet to the outlet. It can be observed from
Fig. 14(c) that the equilibrium temperature rises from the air inlet to
the process water inlet. Since there is only one sink at the outlet,
spray water of different control volumes with different

Fig. 16. Flow pattern optimization of the CWCT.

temperatures must be mixed to the medium temperature before


going to the inlet. Because of the mixture, the heat and mass
transfer driving forces at the inlet are not uniform, shown in
Fig. 14(a). Fortunately, the spray water of the cross-ow CWCT has
the self-adjust ability to achieve proper equilibrium temperature.
As the simulated mass ow rate of spray water is relatively small in
this article, according to Table 4, its state is easy to be inuenced by
the other two uids and it reaches the equilibrium temperature
very quickly (at about x 0.2H). When x 0.5H, shown in Fig. 14(b),
the three uids have already reached the equilibrium states and
had rather uniform heat and mass transfer driving forces. On the
whole, the heat and mass transfer driving forces are relatively
uniform in the cross-ow CWCT, especially in the lower part.
5. Effect of ow pattern on the performance of the CWCT
For two-ow heat and mass transfer system, scholars have
agreed that counter ow achieves the best performance, followed
by the cross ow and the parallel ow. By analogy, the recommended ow pattern in the CWCT is two counter ows and one
parallel ow between the three uids, achieving as many counter
ows as possible and abandoning the mediocre cross ow. Ren and
Yang [17] studied all the parallel/counter-ow patterns of the
CWCT, nding that the ow pattern shown in Fig. 15(a) achieves the
best cooling performance. Following this conclusion, we simulated
the performance of the model to test the analogy. The boundary
conditions are listed in Table 4.
For the parallel/counter-ow CWCT, the distribution parameters
are one-dimensional. As shown in Fig. 15(b), the process water is
cooled along the way while the air is continuously heated. Since the
inlet and outlet temperatures of the spray water should be the
same, the one-dimensional spray water temperature could not
keep pace with the temperature gradient along x. As a result, the
heat transfer driving force between the process water and the spray
water is not uniform. Neither is the heat and mass transfer driving
force between the air and the spray water. This phenomenon is also
stated by many other researchers [4,10]. Unfortunately, the uneven
driving forces could not be avoided by improving the heat and mass
transfer area or regulating the ow ratios of the three uids. In
other words, the one-dimensional parallel/counter-ow CWCT has
the structural limitation. As mentioned in Section 4.2, the crossow CWCT could achieve uniform heat and mass transfer driving
forces in the most part of the module. In this way, the cross-ow
CWCT turns up the ideal ow pattern, although the two-ow
cross-ow heat and mass transfer performance is not the best.

688

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689

The cooling capacity comparison of the parallel/counter-ow


and cross-ow CWCTs under different KmFm are shown in Fig. 16.
The ow rates of the air, the spray water and the process water are
all 0.5 kg/s, while the other boundary conditions are presented in
Fig. 16. As shown in Fig. 16, the cross-ow CWCT would always have
a larger cooling capacity Qc, and the superiority will be amplied
when the heat and mass transfer coefcient is increased, which
means the cross-ow CWCT has better cooling performance than
the commonly used parallel/counter-ow CWCT.
6. Conclusion
A cross-ow CWCT based on the n-tube structure was
designed and tested in present study. The ow arrangement of the
air and the process water was counter ow, while that of the air and
the spray water was cross ow. Experimental tests were conducted
to investigate the cooling performance and the inuencing factors
on the basis of a good energy balance discrepancy. The main conclusions are:
1) Effect of the process water temperature and ow rates of the
air, spray water and process water on the cooling capacity, wetbulb efciency, heat and mass transfer coefcients were studied. Empirical correlations of the heat and mass transfer coefcients based on the inuencing factors were obtained.
2) Compared to the bare-tube structure in literature, the n-tube
structure tremendously extends the contacting area between
the air and the spray water, thus improves the heat and mass
transfer coefcient. While for the heat transfer coefcient between the spray water and the process water, the n-tube
structure has little impact.
3) Two-dimensional steady-state numerical model of the crossow CWCT was built and validated by the experimental data.
The deviation between the model and the experimental data
was less than 8%, which ensures the accuracy of the model.
4) The numerical results show that the spray water temperature
of the cross-ow CWCT would automatically form a gradient in
the air/process water ow direction to match the temperature
variances of the air and the process water. As a result, the heat
and mass transfer driving forces of the cross-ow CWCT are
fairly uniform, which is benecial for the behavior of the CWCT.
5) The ow pattern optimization of the CWCT shows that the
cooling performance of the cross-ow CWCT is better than that
of the commonly studied parallel/counter-ow CWCT due to
more uniform driving forces. The superiority will be amplied
when heat and mass transfer coefcients are increased.
Acknowledgements
The research described in this paper was supported by National
Natural Science Foundation of China (No. 51138005) and the
foundation for the author of National Excellent Doctoral Dissertation of China (No. 201049).
Nomenclature
a
cp
d
do
Fh
Fm
G
H
h

specic surface area (m2/m3)


specic heat capacity (kJ/kg  C)
humidity ratio (g/kg)
external diameter of the tube (m)
heat transfer area (m2)
mass transfer area (m2)
mass ow rate per ow area (kg/m2s)
height of the CWCT unit (m)
enthalpy (kJ/kg)

Kh
Kh0
Km
L
Lef
_
m
Qc
r
t
W

heat transfer coefcient between spray water and process


water (kW/m2 K)
heat transfer coefcient between spray water and air
(kW/m2 K)
mass transfer coefcient between air and spray water (kg/
m2 s)
thickness of the CWCT unit (m)
Lewis factor (dimensionless)
mass ow rate (kg/s)
cooling capacity (kW)
vaporization latent heat (kJ/kg)
temperature ( C)
width of the CWCT unit (m)

Greek symbols
G
spray water ow per unit breadth (kg/m s)
D
change of or difference between parameters

wet-bulb cooling efciency


Subscripts
a
air
e
air in equilibrium with spray water
in
inlet
m
moist air
out
outlet
s
spray water
w
process water
wb
wet-bulb
References
[1] M. Lemouari, M. Boumaza, A. Kaabi, Experimental analysis of heat and mass
transfer phenomena in a direct contact evaporative cooling tower, Energy
Convers. Manage. 50 (2009) 1610e1617.
[2] Z.Z. Xia, C.J. Chen, R.Z. Wang, Numerical simulation of a closed wet cooling
tower with novel design, Int. J. Heat Mass Transf. 54 (2011) 2367e2374.
[3] Y. Niitsu, K. Naito, T. Anazai, Studies on characteristics and design procedure
of evaporative coolers, J. SHASE. Jpn. 43 (7) (1969) 581e590.
[4] J.A. Heyns, D.G. Krger, Experimental investigation into the thermal-ow
performance characteristics of an evaporative cooler, Appl. Therm. Eng. 30
(2010) 492e498.
[5] M.M.A. Sarker, G.J. Shim, H.S. Lee, C.G. Moon, J.I. Yoon, Enhancement of cooling
capacity in a hybrid closed circuit cooling tower, Appl. Therm. Eng. 29 (2009)
3328e3333.
[6] W.Y. Zheng, D.S. Zhu, J. Song, L.D. Zeng, H.J. Zhou, Experimental and computational analysis of thermal performance of the oval tube closed wet cooling
tower, Appl. Therm. Eng. 35 (2012) 233e239.
[7] G. Heidarinejad, M. Bozorgmehr, S. Dlfani, J. Esmaeelian, Experimental
investigation of two-stage indirect/direct evaporative cooling system in
various climatic conditions, Build. Environ. 44 (2009) 2073e2079.
[8] A. Hasan, K. Sirn, Theoretical and computational analysis of closed wet
cooling tower and its applications in cooling of buildings, Energy Build. 34 (5)
(2002) 477e486.
[9] M. Koschenz, Model for closed circuit evaporative cooling tower, in: Proceedings of the 4th International Conference, International Building Performance Simulation Association, Madison, Wisconsin, USA, 1995.
[10] A. Hasan, G. Gan, Simplication of analytical models and incorporation with
CFD for the performance predication of closed-wet cooling towers, Int. J.
Energy Res. 26 (2002) 1161e1174.
[11] G. Gan, S.B. Riffat, Numerical simulation of closed wet cooling towers for
chilled ceiling systems, Appl. Therm. Eng. 19 (1999) 1279e1296.
[12] S.Y. Yoo, J.H. Kim, K.H. Han, Thermal performance analysis of heat exchanger
for closed wet cooling tower using heat and mass transfer analogy, J. Mech.
Sci. Technol. 24 (4) (2010) 893e898.
[13] P. Stabat, D. Marchio, Simplied model for indirect-contact evaporative
cooling-tower behavior, Appl. Energy 78 (2004) 433e451.
[14] V.D. Papaefthimiou, E.D. Rogdakis, I.P. Koronaki, T.C. Zannis, Thermodynamic
study of the effects of ambient conditions on the thermal performance
characteristics of a closed wet cooling tower, Appl. Therm. Eng. 33e34 (2012)
199e207.
[15] G.J. Shim, M.M.A. Sarker, C.G. Moon, H.S. Lee, J.I. Yoon, Performance characteristics of a closed-circuit cooling tower with multiple paths, Heat Transfer
Eng. 31 (12) (2010) 992e997.
[16] J. Faco, A.C. Oliveira, Thermal behavior of closed wet cooling towers for use
with chilled ceilings, Appl. Therm. Eng. 20 (2000) 1225e1236.

J.-J. Jiang et al. / Applied Thermal Engineering 61 (2013) 678e689


[17] C.Q. Ren, H.X. Yang, An analytical model for the heat and mass transfer processes in indirect evaporative cooling with parallel/counter ow congurations, Int. J. Heat Mass Transf. 49 (2006) 617e627.
[18] J. Yan, N. Liu, X.P. Ouyang, L.X. Zhang, H. Li, Experimental research on crossclosed cooling tower, Fluid Machinery 37 (6) (2009) 53e56 (in Chinese).
[19] C.H. Zhan, X.D. Zhao, Z.Y. Duan, S.B. Riffat, Numerical study on indirect
evaporative cooling performance comparison between counterow and
crossow heat exchangers, Int. J. Low-Carbon Technol. 6 (2011) 100e106.
[20] S.J. Kline, F.A. McClintock, Describing uncertainties in single sample experiments, Mech. Eng. 78 (1953) 3e8.

689

[21] P.F. Li, Y.Y. Li, J.E. Seem, Modelica-based dynamic modeling of a chilled-water
cooling coil, HVAC&R Res. 16 (1) (2010) 35e38.
[22] T. Kusuda, Humidity and Moisture: Measurement and Control in Science and
Industry, Calculation of the Temperature of a Flat-plate Wet Surface under
Adiabatic Conditions with Respect to the Lewis Relation, Reinhold Publishing,
New York, 1965.
[23] J.C. Kloppers, D.G. Krger, The Lewis factor and its inuence on the performance prediction of wet-cooling towers, Int. J. Therm. Sci. 44 (2005)
879e884.

You might also like