International Journal of Heat and Mass Transfer 176 (2021) 121299

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Analysis of indirect evaporative cooler performance under various heat

and mass exchanger dimensions and flow parameters
Abdalazeem Adam a,b, Dong Han a,∗, Weifeng He a,∗, Majid Amidpour c
a
Energy Conservation Research Group (ECRG), College of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing,
Jiangsu 210016, China
b
Department of Mechanical Engineering, College of Engineering, Karary University, Omdurman 12304, Sudan
c
Department of Mechanical Engineering, K.N. Toosi University of Technology, Energy Systems Division, No. 19, Pardis Street, Molla Sadra Ave., Vanak Sq.,
P.O. Box 19395-1999, Tehran, Iran

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

Article history: This paper presents a study investigating and analyzing the effect of heat and mass exchanger (HMX)
Received 2 December 2020 dimensions and the flow parameters on the indirect evaporative cooler (IEC) system’s performance. For
Revised 10 February 2021
this purpose, a mathematical model based on heat and mass transfer concepts was developed, and the
Accepted 27 March 2021
model was verified using experimental and simulation results from published papers. Simulations were
Available online 25 May 2021
performed based on different operating and structural conditions affecting the outlet air temperature,
Keywords: cooling capacity, coefficient of performance, and wet-bulb efficiency. Results showed that the optimal
Indirect evaporative cooler dimensions that give good efficiency in climates with moderate humidity, the length of the duct should
Numerical simulation be between 0.6 to 1.0 m, the width of the channel between 0.3 to 0.5 m, and the channel gap between
Cooling capacity 0.004 to 0.008 m. It has also been observed that the increasing working air velocity has a positive effect
Heat and mass exchanger on cooling performance, and its velocity should not be less than 1 m/s. As for increasing the product air
velocity, it reduces the heat exchange period, which is undesirable, so it should not exceed 1 m/s. The
prescribed flow rate ratio between product and working air should be 0.3 to 0.5. This analysis provides
desirable operating conditions to achieve high operational efficiency as well as optimal dimensions for
designing such a cooler.
© 2021 Elsevier Ltd. All rights reserved.

1. Introduction cant systems for their high efficiency and low energy consumption
[10].
Indirect evaporative cooling (IEC) technology is one of the most Up to date, many researchers have studied the effect of various
promising energy-saving and heat recovery applications. IEC tech- flow parameters and operating conditions on the performance of
nology operates based on heat and mass transfer between air and IEC systems [11–14]. Riangvilaikul [15] introduced an experimental
cooling water in the wet duct to cool the product air in the dry study on a novel dew point evaporative cooling system to examine
channel [1]. The principle of heat and mass transfer between water outlet air conditions and the system’s efficiency under various in-
and air in IECs is very similar to the cooling tower and humidifier let air conditions. The results showed that the dew point and the
[2,3]. Several kinds of IEC units, such as plate type, irregular type, wet-bulb efficiency vary between (58 - 84)% and (92 - 114)%, re-
and tube type are commonly used in the cooling applications [4,5]. spectively. Chen et al. [16] performed a parametric investigation of
According to flow arrangement, the plate type IEC is divided into IEC regarding condensation from product air. They used the return
three main types: cross-flow, parallel flow, and counter-flow [6,7], air with low humidity and temperature from the cooling space as
and also divided into single, two-stage [8], and multi-stage [9]. The inlet air in the wet duct. The simulation was executed to estimate
single-stage plate types are widely used in energy recovery appli- four evaluation indexes under three operation states, while various
cations such as heating, ventilation and air conditioning, and desic- parameters were analyzed in detail. The simulation results showed
that a decrease in channel gap and an increase of working air ve-
locity have a positive effect on enhancing the system performance.
Wang et al. [17] produced a study on the optimization and entropy
∗
Corresponding authors.
analysis of heat and mass exchanger of the counter-flow IEC work-
E-mail addresses: kdadao9@hotmail.com (A. Adam), handong@nuaa.edu.cn (D. ing on M-Cycle. The results pointed that the suggested average
Han), wfhe@nuaa.edu.cn (W. He). product air velocity should not exceed 1.0 m/s, the ratio between

https://doi.org/10.1016/j.ijheatmasstransfer.2021.121299
0017-9310/© 2021 Elsevier Ltd. All rights reserved.
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

an increase of the product air flow rate. Wan et al. [22] proposed
Nomenclature a new method for investigating the linked heat and mass trans-
fer aspects in IEC with counter-flow arrangements. They imple-
cp Specific heat, J/kg.K mented the computational fluid dynamics (CFD) model to simu-
L Length, m late the heat and mass transfer process. Oh et al. [23] introduced a
Dh Hydraulic diameter, m comparative study on purge configurations for indirect evaporative
m˙ Mass flow rate, kg/s cooling. They performed a mathematical model to understand fur-
h Heat transfer coefficient, W/m2 .K ther the combined heat and mass transfer processes of a counter-
U Overall heat transfer coefficient, W/m2 .K flow IEC in semi-arid regions. The results showed that lower prod-
Q̄ Cooling capacity rate, kJ/kg uct air temperature and higher dew-point effectiveness were ob-
t Temperature, ◦ C tained with longer channel length and larger purge ratio.
k Thermal conductivity, W/m.K Furthermore, the nozzles placement, water flow rate, and heat
u Velocity, m/s and mass exchanger (HME) dimensions can significantly affect the
hv Latent heat of vaporization of water kJ/kg IEC system performance. Al-Zubaydi and Hong [24] introduced
P Theoretical fan power W three suitable designs for water nozzles including, the mixed-
Re Reynolds number mode, internal spraying mode, and the external spraying mode the
Le Lewis factor, dimensionless installation in which the system effectiveness can increase. De An-
Nu Nusselt number tonellis et al. [25] introduced an experimental investigation on a
Pr Prandtl number cross-flow IEC to reduce water use, with a water mass flow rate
between 0.4% and 4%. They investigated the influences of spry noz-
Special characters
zle placement, the change of water flow rate, and working air tem-
ν Kinematic viscosity, m2 /s
perature. The results noted that the performance is somewhat re-
ω Humidity ratio, kg/kg dry air
liant on nozzles number and size, but the water flow rate dramati-
β Mass transfer coefficient, kg/m2 .s
cally affects it. Numerous plate shapes, including micro-baffled, pil-
ρ Density, kg/m3
low, corrugated, and diamond-shaped plates, intended at enhanc-
η Efficiency, %
ing surface wettability have been investigated in [26–28]. The ex-
σ Wettability, factor
perimental and simulation outcomes noted that simple geometries
δ Plate thickness, m
could reach to high wettability. Cui et al. [29] studied the impacts
Subscripts of the channel dimensions and the inlet air condition, including
s Working air stream the air stream ratio and the air velocity on the system perfor-
p Product air stream mance. The results noted that the selected evaporative air cooler
w Water film could achieve higher wet-bulb efficiency with small channel width
x Channel height and low air velocity. Antonellis et al. [30] presented an experimen-
y Channel width tal study of IEC to investigate the impact of different plate geom-
z Channel gap etry on the efficiency and pressure drop. The results revealed that
i Inlet plate geometry affects both surface wettability and the heat trans-
o Outlet fer rate. They concluded that reticulated plate protrusion is an ex-
int Interface cellent design to improve system effectiveness; also, the parallel
wb Wet-bulb water and working airflow order lead to better performance than
the counter flow arrangement.
Superscripts The comprehensive literature survey shows a shortage in re-
s Sensible heat search deals with the effect of HMX geometry on IEC performance.
l Latent heat In an effort to narrow the research gap, this paper aims to de-
velop a numerical model based on the principle of heat and mass
working and product air should be in the range of (0.3 − 0.4), the transfer to evaluate the impact of HMX size and operating param-
passage length in the range between (1 − 1.75) m, and the height eters on the IEC performance. The proposed mathematical model
between (3 − 5) mm. was solved numerically by adopting the 4th order Runge Kutta
On the other hand, some studies focus on the effect of heat ex- method. The IEC system’s performance was assessed under vari-
changer arrangement and flow patterns on the IEC’s effectiveness. ous exchanger geometry and operating conditions, including prod-
Anisimov et al. [18] provided a numerical investigation on heat and uct to working air flow rate ratio, product air inlet parameters, and
mass transfer in IEC, including four airflow patterns. The simula- working air parameters. This analysis provides desirable operating
tion results showed that the increase in efficiency is sensitive to conditions to achieve high operational efficiency as well as optimal
different inlet conditions. Li et al. [19] performed a comparative dimensions for designing such a cooler.
study between the counter-flow and cross-flow IEC for heat recov-
ery utilization. Besides, they addressed the influence of the plate
installation (vertical and horizontal) on the system’s thermal per- 2. Description of the IEC system
formance. The results showed that under constant air flow rate, the
vertical arrangement gives better cooling capacity and lower outlet The indirect evaporative cooling technology operates based on
temperature than the horizontal layout. Khalid et al. [20] presented heat and mass transfer between air and cooling water in the wet
an experimental study to improve the IEC design under low ve- duct to cool the dry channel’s product air. IEC is acknowledged as
locity conditions. They used a cross-flow pattern between dry and a cooling technology fit for various purposes of heating, ventila-
wet channels, which has the advantage of increasing the heat ex- tion, and air conditioning for data centers, sports centers, service
change area. Moshari et al. [21] introduced a study on the effect buildings, markets, cinemas, etc [10]. The IEC system consists of
of different flow arrangements of two-stage IEC on the wet-bulb flat plates that act as a heat and mass exchanger, a water tank,
effectiveness. They concluded that in the counter-flow regenera- water spray nozzles installed in the upper part of the plate, and a
tive IEC, the dimensionless water evaporation rate decreases with pump to circulate water, as shown in Fig. 1.

2
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Fig. 2. IEC control volume.

Fig. 1. Scheme of the counter-flow indirect evaporative cooling system.

3.1. Governing equations of heat and mass transfer

The IEC system’s working principle divides the air stream into
two parts separated by a plate, one of them called product air (pri- In counterflow arrangement, the working air flows upward in
mary, main), and the second one is working air (secondary, ex- the opposite direction with the water film’s direction while the
haust) [13,31]. The recirculated water is diffused on the plate’s tip product air flows downward in the dry channel. Fig. 2 shows the
and flows along the cooling air duct’s wall surfaces as a thin film. control volume in the counterflow arrangement. According to the
The heat is then transferred from product air to the water film in above assumption, the studied IEC’s mathematical model was de-
the wet channel without any addition to its moisture. The work- veloped based on the heat and mass balance equations of the con-
ing air at the interface with the water film is saturated with water trol volume as follows:
vapor at average water temperature. Since the partial pressure of From Fig. 2, we can express the energy balance in the control
water vapor at the interface is higher than the partial pressure of volume of the product air as follows:
the vapor in the air, there is a mass transfer of water vapor associ- ∂ tp
m˙ p c p,a dy = −dQ p (1)
ated with the latent heat transfer due to water evaporation. Simul- ∂y
taneously, there is heat transfer by convection from working air to
where dQ p is the sensible heat transfer from the product air to the
water film due to the temperature difference.
water film, which can be calculated by:
IEC can be used as four possible cooling technologies [32]: (1)
Classic IEC equipment; (2) Classic IEC equipment in dry drive sys- dQ p = U (t p − tw )Lx .dy (2)
tem, which is used when the working air temperature is low at the where U is the overall heat transfer coefficient can calculated from:
inlet; (3) Regenerative IEC equipment operating in the wet system;
(4) As an energy-saving unit compatible with the classic cooling 1
U= (3)
system. The outside air temperature circumscribes the frontiers of 1
hp
+ kδwall + 1
hw
wall
IEC technologies in terms of cooling capacity and humidity. In the
IEC of the counter-flow arrangement with a plate of infinite surface Substituting Eq. (2) into Eq. (1) and rearranging the energy balance
area, if the product air to the working air has an appropriate air- equation yields:
flow-ratio, the air temperature in the plate’s dry side will reach the dt p Lx
wet-bulb temperature of the incoming working air. The plate’s wet = (Up (tw − t p ) ) (4)
dy m˙ p c p,a
side’s working air temperature will be cooled from its inlet dry-
where
bulb temperature to the inlet wet-bulb temperature. But in fact,
1
this high operating efficiency cannot be obtained due to several m˙ p = ρ p u p Lz Lx (5)
attribution facts [33]: There is limited area for heat exchange; No 2
pure counter-flow pattern can be achieved; It is difficult to get a The mass balance in the working air channel can be expressed as:
uniform and even distribution of water on the wetted sides of the
plate. dm˙ w = m˙ s dω (6)
The rate of heat transfer between the water film and working air
3. Mathematical model
can be expressed as:

The IEC system’s mathematical model was developed based on dQ s = hs (tw − ts )σ .Lx dy (7)
the heat and mass balance of separate air streams during steady- Due to the hydrophobic effect, the water membrane does not vir-
state operation. The model was derived in the form of differential tually overlay the entire plate area of the IECs heat exchanger.
equations using the control volume of two flow arrangements. The Hence, the wettability factor σ estimates the ratio of actual wet
following assumptions were used in the model derivation: plate surface area of the IECs heat exchanger.
• Steady-state operation. dQ l = hv β (ωint − ωs )σ .Lx dy (8)
• The system is adiabatic, no heat is transferred from the system
where dQ l is the latent heat transfer and dQ s is the sensible heat
to the surrounding.
transfer.
• The properties of working fluids are uniform at the inlet.
The overall change in the water membrane energy can be de-
• The working air at the interface with water-membrane is satu-
termined by adding Eqs. (2), (7) and (8) as:
rated corresponding to the water temperature.  
• The transfer of the heat and the mass is perpendicular to the U (tw − t p ) + hs (tw − ts )σ +
d Qw = Lx dy (9)
plate. hv β (ωint − ωs )σ

3
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

The change in the water mass flow rate due to the evaporation The heat transfer coefficient of the water membrane can be calcu-
process can be calculated as: lated from:
dm˙ w kw
= −β (ωint − ωs )σ .Lx (10) hw = 1.88 (23)
dy δw
From the energy balance equation in the water film and combining 3.3. Performance evaluation indexes
Eqs. (9) and (10) the change in the water internal energy neglect-
ing the second-order derivative can be expressed as: The following indexes were considered to evaluate the working
 
dtw Lx β (ωint − ωs )(c p,wtw − hν )σ + performance of the selected IECs:
= (11)
dy c p,w m˙ w hs (ts − tw )σ + U (t p − tw ) • The wet-bulb efficiency, which is defined as the ratio of the dif-
ference between inlet and outlet temperature of the product air
The rate of change in the working air humidity ratio in the wet
to the difference between the inlet temperature of the product
control volume can be expressed as:
air and the wet-bulb temperature [11]:
d ωs Lx
= β (ωint − ωs )σ (12) t p,i − t p,o
dy m˙ s ηwb = (24)
t p,i − twb
The change in the working air enthalpy due to the latent and sen-
• The cooling capacity of the system can be calculated as:
sible heat transfer can be determined by adding Eqs. (7) and (8) as:  
Q̄ p = m˙ p c p,a t p,i − t p,o (25)
d hs Lx σ
= [hν β (ωint − ωs ) + hs (tw − ts )] (13) • The coefficient of performance (COP) of the system can be ex-
dy m˙ s
pressed as:
From the energy balance equation in the working air and substi-
tuting equation of humid air enthalpy in Eq. (13), the working air Q̄ p
COP = (26)
energy change can be expressed as: P
dts Lx (tw − ts )σ The power consumption of the pump in is negligible and the
= [β (ωint − ωs )c p,w + hs ] (14) theoretical fan power of the system can be expressed as:
dy c p,a m˙ s
P= p pVp + psVs (27)
where:
1 The friction pressure loss can be calculated as:
m˙ s = ρs u s L z L x (15)
2 ρv2 l
p= f (28)
The initial conditions of the working and product air stream at the 2 Dh
inlet to corresponding channels:
    The fluid friction resistance factor for lamina flow in rectangular
t t t ωs,i channel can be calculated using the following equation [35].
t p = p,i ; tw = w,i ; ts = s,i ; ωs = (16)
y = Ly y = Ly y=0 y=0 96
f = (29)
Re
3.2. Heat and mass transfer coefficients
4. Numerical approach and its validation
The coefficient of the mass transfer is determined from the
Lewis relationship as: The set of the presented governing equations describes the heat
and mass transfer process in every IEC control volume. Conse-
h
Le = (17) quently, the proposed set of differential equations was solved nu-
β .c p merically together with the initial conditions. The 4th order Runge-
where Le is the Lewis factor and assumed as a unity. Kutta method was implemented to carry out the computer sim-
The coefficient of heat transfer by convection in the dry channel ulation. Therefore the computational domain was discretized into
can be obtained from [16]: eighty cell control elements along the airflow direction, which
0.8 gives an accurate solution. To verify the outlet air conditions
0.0238(ν p /υ p ) .P r p .k p
0.3

hp = (18) obtained by computer simulation, the established mathematical

D0h.2 model validated against published experimental data from the lit-
erature.
The coefficient of heat transfer by convection in the dry channel
can be obtained from [34]:
4.1. Model validation against experimental data
36.31(ρνs )0.68
hs = (19)
(L/Dh )0.08 The experimental data published by Riangvilaikul and Kumar
[15] were used to validated the described mathematical model. In
For steady, fully developed laminar flow along a vertical surface, this validation, the exchanger width, length, and channel gap were
the film thickness; flow rate per length, and Reynolds number can 0.08 m, 1.2 m, and 0.005 m, respectively. The same heat and mass
be expressed [34]: exchanger dimensions, airflow velocity, and the same inlet airflow
m˙ w parameters were used to set the developed model to the same op-
Rew = (20)
L μw erating conditions as in the experimental case. This analysis deter-
mined the model’s accuracy in predicting a true renewable indirect
m˙ w evaporative air cooler’s performance. Fig. 3a illustrates the compar-
w = (21)
( + 1 )L
n ison results between the proposed model and the experiment data.
 1/3 The inlet product air stream temperature (Tpi ) was changed from
3uw 25 ◦ C to 45 ◦ C with the difference of 5 ◦ C, while the airflow ve-
δw = (22)
ρw g locity was kept constant. The comparison was made using air with

4
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Fig. 3. Mathematical model validation; (a) Product air stream outlet temperature Tp,o against inlet air temperature Tp,i under constant air stream velocity conditions; (b)
Product air stream outlet temperature Tp,o against inlet air stream velocity Vp,i under constant inlet air temperature and humidity.

Table 1
shows the IECs geometry and simulation parameters.

Geometery Value Units Parameter Value Units

◦
Channel length 0.5 - 1.0 m Product air temperature 30 - 38 C
Channel width 0.35 - 0.55 m Product air relative humidity 30 - 60 %
Channel gap 0.002 - 0.010 m Product air velocity 1-5 m/s
◦
Plate thickness 0.0005 m Working air temperature 22 - 30 C
Number of plates 92 – Working air velocity 1-5 m/s
Working air relative humidity 40 - 60 %

four different degrees of relative humidity including, very dry air 5.1. Effect of HMX dimensions
(ω p,i = 6.9 g/kg), dry air (ω p,i = 11.2 g/kg), moist air (ω p,i = 20.0
g/kg) and very humid air (ω p,i = 26.4 g/kg). It is noticed that the In this analysis, the impacts of three HMX dimensions on the
maximum inconsistency in the product air stream outlet temper- system performance were studied using the operation conditions
ature between the model and the experimental data was 5.31% listed in Table 1. The simulation results, including, product air tem-
when humidity ratio was 20.0 g/kg. Considering the relevant as- perature difference, wet bulb efficiency, cooling capacity, and coef-
sumptions and accuracy of test equipment, the developed model ficient of performance were obtained and analyzed. In the study of
can simulate the IEC system’s heat and mass transfer process. the channel gap effect the product air temperature, velocity, rela-
tive humidity were set as 30 ◦ C, 1 m/s, and 30 % respectively; and
the working air parameters as same as product air except the tem-
4.2. Model validation against simulation data
perature was 24 ◦ C. The channel gap was varied from 2 mm to 10
mm with 2 mm increment. The temperature difference improved
Additionally, the product air stream outlet temperature was val-
from 9.74 ◦ C to 11 ◦ C when the gap increased from 2 mm to 6
idated, applying the plate-type heat and mass exchanger’s numer-
mm, and accordingly, the rest of the indicators improved, as illus-
ical results derived by Anisimov el at. [18]. The selected heat and
trated Table 2. The effect of the channel gap on the product air
mass exchanger was a counter-flow indirect evaporative exchanger
outlet temperature along the channel is shown in Fig. 4a. As can
for an IEC system. The geometry of the exchanger was (0.5 x 0.5
be seen, the air temperature begins to decrease in the first part of
x 0.5) m, the channel gap was 0.004 m, and the thickness of
the channel. The response depends on the channel gap, so when
the plate was 0.0 0 05 m. The product air inlet velocity (Vp,i ) was
the gap was 2 mm there was a rapid decrease in the air temper-
changed from 2 m/s to 6 m/s with 1 m/s increments. The inlet air
ature, and it reached 20 ◦ C at a length of 0.38 m and continues
temperature was constant at 30 ◦ C, and the relative humidity was
steadily for the rest of the channel. Due to the small channel gap,
45%. The maximum conflict between the two models in the third
a small amount of working air is entering the wet duct, which is
case was 0.74% when the velocity was 4 m/s, as shown in Fig. 3b.
not sufficient to result in the evaporative cooling effect because it
According to the model validation versus the selected data, it can
reaches the saturation point instantly. For this reason, we find that
be commented that the introduced model can predict the perfor-
the temperature shown by the black dashed line becomes almost
mance of the IEC system with great approaches.
constant at 20 ◦ C in the second half of the channel. It is clear that
the increase in the channel gap will lead to a further decrease in
5. Results and discussion the air temperature, for example when the gap is between 4 mm
and 6 mm the outlet temperature is less than 19.5 ◦ C along (0.6 -
This section describes the results and discussion of the pro- 0.8) m, and there is no difference significant at the exit. The rea-
posed system. Based on the stated mathematical model, influences son behind this is that increasing the gap increases the air mass
due to the inlet air condition, the airflow velocity, the product-to- flow rate, and this will increase the evaporative cooling in the wet
working airflow ratio, and the dimension of the airflow passages duct and enhance the heat transfer rate between the two channels.
on the system performance are calculated and analyzed. The sim- Fig. 4b shows the effect of the channel gap on wet-bulb efficiency
ulation was executed with the dimension and operating condition along the channel. The maximum efficiency of 0.68 % was obtained
listed in Table 1. when the gap is varied between 6 mm to 8 mm. Increasing the gap

5
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Table 2
Results of HMX dimensions effect on the IEC performance.

Studied object Temperature difference δ T (◦ C) Wet-bulb efficiency ηwb (%) Cooling capacity CC (W ) Coefficient of performance COP (–)

Channel gap (mm)

2 9.740 0.596 280.20 11.16
4 10.64 0.651 612.34 48.59
6 11.08 0.678 956.98 113.47
8 11.18 0.684 1287.0 202.66
10 11.04 0.675 1588.6 311.46
Channel length (m)
0.5 10.10 0.618 726.75 143.90
0.6 10.46 0.640 752.67 149.03
0.7 10.67 0.653 767.87 152.04
0.8 10.80 0.660 776.78 153.81
0.9 10.87 0.665 782.02 154.84
1.0 10.91 0.668 785.08 155.45
Channel width (m)
0.35 10.68 0.653 768.08 108.17
0.40 10.80 0.660 776.87 109.60
0.45 10.87 0.665 782.04 110.48
0.50 10.92 0.667 785.08 111.03
0.55 10.96 0.669 786.87 111.39

Fig. 4. Shows effect of channel gap: (a) illustrate the product air temperature variation Vs the channel length for different gaps; (b) shows the variation of wet-bulb efficiency
Vs the channel length for different gaps.

will result in improving the efficiency but need more longer chan- At the end of the HMX dimensions analysis, all of the operating
nel because of the high amount of airflow rate. Due to that, the conditions were remained constant, including the mass flow rate
gap size of (4 - 6) mm is an adequate size which results in proper and the value of width was varied from 0.35 m to 0.55 m with 0.05
outlet temperature and wet bulb efficiency at a suitable length of m increment. The results showed that increasing the channel width
the channel. had no significant effect except in the case of ex-changers with
In the study of channel length effect, the size was varied from the small channel length. This effect can be seen in the product
0.5 m to 1 m with a step size of 0.1 m, while the other param- air temperature variation along the channel with different channel
eters remained as constant. The results obtained from this analy- width in Fig. 6a. Working air in the wet passage with large width
sis showed that increasing the length leads to an increase in the reaches saturation point faster,as shown in Fig. 6b. The simulation
heat transfer area which enhances the cooling process and also in- results indicated that the humidity ratio reaches a value of 14.28
creases the wet-bulb efficiency and the cooling capacity, as shown g/kg which is close to the saturation value when the channel width
in Table 2. However, the increment in channel length is related to in the range (0.4 - 0.5) m at the acceptable length of the channel.
the relative humidity of the working air, because when the work-
ing air reaches the saturation point, the evaporative cooling in the
wet duct stops and the increase in channel length converts inef- 5.2. Effect of the working air velocity and temperature
fective. This phenomenon can be seen in the variation of the prod-
uct air temperature, which reaches a value close to it minimum In this simulation impact of three working air parameters, in-
value and continues to decline slowly when the length is in the cluding velocity, inlet temperature, and humidity on the system
range between (0.6-0.8) m Fig. 5a. Because the working air in its performance were studied and analyzed. In this study, all the prod-
wet channel becomes nearly saturated with water vapour and can- uct air flow parameter (Vp,i = 1 m/s; Tp,i = 30 ◦ C; p =30%) were
not cool The wet plate away. Also, this can be seen in the change kept constant. The working air velocity was varied from 1 m/s to 5
in the value of the cooling capacity and coefficient of performance m/s, and the temperature changed from 22 ◦ C to 30 ◦ C with 2 ◦ C
in Fig. 5b. The change in the cooling capacity and COP is minimal increments. During the simulation, the working air humidity ration
when the length reaches 0.8 m. As a result, the suitable channel was kept constant ωs,i = 6.9 g/kg dry air. Fig. 7a shows the influ-
length is (0.6 − 0.8) m for climates in which the humidity ratio is ence of working air velocity on the outlet temperature difference
not very high. for various value of Ts,i . The outlet air temperature difference was
increased from 9.5 ◦ C to 12.8 ◦ C when the speed boosted from 1

6
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Fig. 5. Shows effect of channel length: (a) illustrate the product air temperature variation, wet-bulb efficiency, and working air humidity ratio Vs the channel length; (b)
shows the variation of cooling capacity and COP Vs the channel length.

Fig. 6. Shows effect of channel width: (a) illustrate the product air temperature variation Vs the channel length for different width; (b) shows working air humidity ratio Vs
the channel length for different width.

m/s to 5 m/s. Fig. 7b shows the impact of working air velocity on temperature at the inlet was 22 ◦ C and then began to decrease
the wet bulb efficiency for the various value of Ts,i . It can be seen when the working humidity value was raised to 45 %. As can be
that the wet-bulb efficiency enhanced with increasing the velocity seen, increasing working air humidity negatively affects the cooling
and reaches 84% when the velocity 5 m/s. As can be noted that performance of the system. The reason behind this is because the
the working air velocity has a significant effect on the cooling per- working air flows into the wet duct with a high humidity value,
formance. Because increasing the air velocity will raise the mass and this is unable to carry a large amount of water vapor and thus
flow rate, which enhances the evaporative cooling in the wet chan- affects the evaporative cooling process and, accordingly, the heat
nel and results in a significant temperature drop in the dry chan- exchange process between the two channels is affected. This effect
nel. in contrast, increasing the working air inlet temperature has can also be observed in Fig. 8c that shows the cooling capacity val-
a negative effect on cooling performance. The outlet temperature ues against the working air temperatures at different relative hu-
difference was decreased about 0.75 ◦ C when the inlet tempera- midity values. The cooling capacity value decreased from 780.5 W
ture increased from 22 ◦ C to 30 ◦ C at Vs,i =1 m/s. Also, this effect to 601.8 W when increasing the relative humidity from 40 % to 60
can be noticed in the system’s cooling capacity in Fig. 7c. The cool- %. Fig. 8d shows the effect of relative humidity on the COP with
ing capacity was enhanced when velocity was increased, but when respect to the inlet temperature. The value of COP decreased from
the working air inlet temperature increased, and velocity kept con- 77.43 to 59.2 when the relative humidity was raised from 40% to
stant, it decreased. Fig. 7d shows the COP for different working air 60 % at a temperature of 22 ◦ C. Fig. 8b shows the impact of work-
temperatures and velocities. It can be seen that the COP decreases ing air humidity on the wet bulb efficiency for the various value
through an increase in velocity as a result of a decrease in pressure of Ts,i . It can be seen that the wet-bulb efficiency increases with
through the exchanger. increasing the humidity because the wet-bulb temperature in the
inlet was increased.
5.3. Effect of the working air humidity
5.4. Effect of the product air temperature
In the study of the effect of working air humidity, the relative
humidity ranged from 40 % to 60% with step size 5%. The sim- in this analysis, the product air inlet temperature was varied
ulation also included a variety of working air inlet temperature from 30 ◦ C to 38 ◦ C with 2 ◦ C increments. Fig. 9a shows the vari-
and other flow parameters for both airstreams were kept constant. ation of the product air temperature along the dry duct for dif-
Fig. 8a shows the temperature difference of the outlet air with re- ferent inlet air temperatures. It can be seen that the air temper-
spect to the working air inlet temperature. The highest value of ature begins to drop sharply in the first part of the channel and
the outlet air temperature difference was 10.85 ◦ C, which was ob- continues to decrease slowly in the remaining channel part. When
tained when the relative humidity was 40 % and the working air the inlet air temperature has a small value, it can also be ob-

7
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Fig. 7. Shows effect of working air parameters: (a) shows the product air temperature difference Vs Ts,i for different Vs,i ; (b) shows the wet-bulb efficiency Vs Ts,i for different
Vs,i ; (c) shows the cooling capacity Vs Ts,i for different Vs,i ; (d) shows COP Vs Ts,i for different Vs,i .

Fig. 8. Shows effect of working air parameters: (a) shows the product air temperature difference Vs Ts,i for different s,i ; (b) shows the wet-bulb efficiency Vs Ts,i for
different s,i ; (c) shows the cooling capacity Vs Ts,i for different s,i ; (d) shows COP Vs Ts,i for different s,i .

8
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Fig. 9. Shows effect of product air temperature: (a) shows the product air temperature variation Vs channel length for different Tp,i ; (b) shows the COP and wet-bulb
efficiency Vs Tp,i .

Fig. 10. Shows effect of product air velocity: (a) shows the product air temperature variation and cooling capacity Vs Vp,i ; (b) shows the COP and wet-bulb efficiency Vs Vp,i .

served that it drops faster to small degrees with a shorter chan- as shown in Fig. 10b. Therefore, the initial air velocity should not
nel length compared to the air with a high inlet temperature. For exceed 1 m/s to maintain high efficiency.
example, when the inlet air temperature was 30 ◦ C, the tempera-
ture rapidly decreased to 20.28 ◦ C at a channel length of 0.56 m, 5.6. Effect of the mass flow rate ratio
but when the temperature at the entrance was 38 ◦ C, the temper-
ature reached 20.31 ◦ C at a length of 1 m. However, the tempera- The flow rate ratio between product and working air varied
ture at the outlet was not of a big difference. Fig. 9b illustrates the from 1 to 0.33, with a five-point interval. The results obtained
wet bulb efficiency and COP against various product air inlet tem- from this simulation indicated that the outlet air temperature has
perature. it can be noticed that both the efficiency and the COP a lower value at the lower flow rate ratio, as shown in Fig. 11a.
were enhanced. the wet bulb efficiency was increased from 0.7 to As can be seen the lowest outlet temperature of 18.98 ◦ C was ob-
0.784 and the COP raised from 78 to 128 when the temperature tained when the flow rate ratio was 0.33. The lower flow rate ratio
increased from 30 ◦ C to 38 ◦ C. because the high temperature at can be obtained by increasing the working air velocity, and this
the inlet will enhance the heat transfer between the two channel will increase the working air mass, which will enhance the evapo-
which result in enhancing the cooling capacity. rative cooling process in the wet channel. As a result, the temper-
5.5. Effect of the product air velocity ature of the plate will decrease, and the heat transfer between the
two channels will increase. This ratio can also be obtained by using
In contrast to the positive effect obtained from increasing the a larger gap size for the wet channel compared to the dry channel.
speed of the working air, increasing the velocity of the product air Fig. 11b shows the wet-bulb efficiency and cooling capacity versus
has a negative impact on the system’s performance. In this analy- the flow rate ratio. It can be seen the wet-bulb efficiency improved
sis, the air velocity was changed from 1 m/s to 5 m/s with a step by 13% when the flow rate changed from 1 to 0.33, and the cooling
size of 0.5 m/s while all other parameters kept constant. Fig. 10a capacity also improved by 136.5 W. The simulation results showed
shows the difference in product air temperature and cooling ca- that the recommended flow rate ratio should be in the range 0.3
pacity versus the input air velocity. It can be seen that the tem- to 0.5, and it is better to design the wet channel with a bigger
perature difference decreased from 10.26 ◦ C to 7.98 ◦ C when the gap size than the dry channel to obtain the appropriate flow rate.
product air velocity was increased from 1 m/s to 5 m/s. This is be- Hence, reducing the mass flow rate ratio is a useful technique for
cause increasing the speed increases the product air’s mass flow improving the cooling performance of the IEC system.
rate and reduces the resident time in the passage for heat to be
exchanged with the wet channel. The increase in the cooling ca- 6. Comparisons between present and previous studies
pacity resulted from the rise in the product air’s mass flow rate.
This effect also appears in the decrease of the wet-bulb efficiency, Table 3 illustrates the comparison of indirect evaporative cool-
as well as the COP due to the increase in the speed of the product, ing systems’ performance for the current and previous studies. The

9
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Table 3
Comparison of operating parameters and performances of the present and previous studies.

Author information Present study Zhu et al. [12] Riangvilaikul and Kumar[15] Oh et al. [23] Khalid et al. [20]
Method Simulation Simulation Experimental Simulation Experimental
Type of the cooler Indirect evaporative cooler dew point IEC cooler Indirect evaporative cooler Purge configuration IEC Indirect evaporative cooler

Flow arrangement Counter flow Counter flow Counter flow Counter flow cross-flow
Parameters
Channel gap 2-10 mm 5-10 mm 5 mm 5 mm 4 mm
Channel length 0.5-1.0 m 0.5 m 1.2 m 0.6-1.4 m 0.508 m
Channel width 0.35-0.50 m N/A m 0.8 m 0.8 m 25 mm
Inlet air temperature 30-38 ◦ C 22-36 ◦ C 25-45 ◦ C 30 ◦ C 25-45 ◦ C
Inlet air humidity 7.9-12.4 g/kg 5.3-18.8 g/kg 7-26 g/kg 10 g/kg 11.2 g/kg
Intake air velocity 1-5 m/s 0.6-3 m/s 2.4 m/s 2 m/s 0.5-1.1 m/s
Outlet air temperature 18.8-23 ◦ C 20-26 ◦ C 25 ◦ C 19.3-21.5 ◦ C 19-35 ◦ C
Wet bulb effectiveness 62 - 100.05 % 75-90 % 92-114 % 69 % 90-98 %
Coefficient of performance 70 − 160 N/A N/A N/A N/A

Fig. 11. Shows effect of mass flow rate ratio: (a) shows the product air temperature variation Vs channel length for different mass flow rate ratio (mr ); (b) shows the cooling
capacity and wet-bulb efficiency Vs mass flow rate ratio.

present study investigated the effect of HMX dimensions and var- duct should be between (0.6 − 1.0) m, the width of the chan-
ious inlet air conditions on the IEC system performance. When nel between (0.3 − 0.5) m, and the channel gap between (3 − 6)
comparing the current simulation results with a study of Oh et al. mm.
[23], it was found that the outlet temperature range is close to cur- • It was found that increasing the working air velocity had a pos-
rent study. The indirect evaporative cooling, which was studied in itive effect on cooling performance. When the speed increased
[15], also showed a similar feature of cooling; that is, there was from 1 m/s to 5 m/s, the product air temperature difference in-
no significant difference in the outgoing air temperature. However, creased from 9.5 ◦ C to 12.8 ◦ C accordingly, improved cooling
higher efficiency values were observed compared to the results of capacity and efficiency. In contrast to the positive effect pre-
the current study due to the difference in channel length. Khalid sented by the working air velocity, it was found that the in-
et al. [20] used a cross-flow arrangement of heat transfer between crease in the product air velocity has a negative effect on the
dry and wet channels, which has the advantage of increasing the cooling performance and that the recommended speed is not
heat exchange area. The rest of the studies, including this one, used more than 1 m/s.
the counter flow arrangement method to transfer heat between the • From the study presented on the working air parameters effect,
dry and wet ducts. Using counter flow in the wet duct, the heat re- it was found that air with high temperature and high humidity
leased from the water on the wet surface can be high and making gives lower cooling performance compared to medium cold air.
this design more compact, easy to build and operate. Therefore it is better to use the expelled air from the cooling
spaces or to use the air after partial cooling in the dry duct.
7. Conclusions • The outlet temperature decreased from 21 ◦ C to 18.98 ◦ C when
the flow rate ratio is changed from 1 to 0.33 and the efficiency
In this paper, a study is presented to examine and analyze the was increased by 13%, this shows that reducing the mass flow
effect of heat and mass exchanger dimensions and flow parame- rate ratio is an important technique for improving the cooling
ters on IEC system performance. For this purpose, a mathemati- performance of the IEC system. The desired flow rate ratio be-
cal model was developed based on the concepts of heat and mass tween product and working air can be achieved by increasing
transfer. The proposed mathematical model was verified using ex- the velocity of the working air or design the wet channel with
perimental data and numerical simulation results from published a bigger gap size than the dry channel in the case of equal ve-
papers. Simulations were conducted based on the various opera- locities.
tional and structural conditions that influence the outlet air tem-
perature, cooling capacity, coefficient of performance, and wet-bulb
Declaration of Competing Interest
effectiveness. After a detailed analysis of the system, the main re-
sults can be summarized as follows:
The authors declare that they have no known competing finan-
• In order to reach optimal operating conditions that give good cial interests or personal relationships that could have appeared to
efficiency in climates with moderate humidity, the length of the influence the work reported in this paper.

10
A. Adam, D. Han, W. He et al. International Journal of Heat and Mass Transfer 176 (2021) 121299

Acknowledgments [17] L. Wang, C. Zhan, J. Zhang, X. Zhao, Optimization of the counter-flow heat and
mass exchanger for M-cycle indirect evaporative cooling assisted with entropy
analysis, Energy 171 (2019) 1206–1216, doi:10.1016/j.energy.2019.01.099.
The authors acknowledge the financial support from the Natural [18] S. Anisimov, D. Pandelidis, J. Danielewicz, Numerical analysis of selected evap-
Science Fund Project in Jiangsu Province (Grant No. BK20201294), orative exchangers with the Maisotsenko cycle, Energy Convers. Manage. 88
and the Key Research and Development Program of Jiangsu (2014) 426–441, doi:10.1016/j.enconman.2014.08.055.
[19] W.-Y. Li, Y.-C. Li, L.-y. Zeng, J. Lu, Comparative study of vertical and horizontal
Province (Grant No. BE2019088). indirect evaporative cooling heat recovery exchangers, Int. J. Heat Mass Transf.
124 (2018) 1245–1261, doi:10.1016/j.ijheatmasstransfer.2018.04.041.
References [20] O. Khalid, M. Ali, N.A. Sheikh, H.M. Ali, M. Shehryar, Experimental analysis of
an improved Maisotsenko cycle design under low velocity conditions, Appl.
[1] Y. Wan, C. Ren, Z. Wang, Y. Yang, L. Yu, Numerical study and performance cor- Therm. Eng. 95 (2016) 288–295, doi:10.1016/j.applthermaleng.2015.11.030.
relation development on counter-flow indirect evaporative air coolers, Int. J. [21] S. Moshari, G. Heidarinejad, A. Fathipour, Numerical investigation of wet-bulb
Heat Mass Transf. 115 (2017) 826–830, doi:10.1016/j.ijheatmasstransfer.2017.07. effectiveness and water consumption in one-and two-stage indirect evapo-
109. rative coolers, Energy Convers. Manage. 108 (2016) 309–321, doi:10.1016/j.
[2] J. Chen, D. Han, W. He, Y. Liu, J. Gu, Theoretical and experimental analysis of enconman.2015.11.022.
the thermodynamic and economic performance for a packed bed humidifier, [22] Y. Wan, C. Ren, L. Xing, An approach to the analysis of heat and mass transfer
Energy Convers. Manage. 206 (December 2019) (2020) 112497, doi:10.1016/j. characteristics in indirect evaporative cooling with counter flow configurations,
enconman.2020.112497. Int. J. Heat Mass Transf. 108 (2017) 1750–1763, doi:10.1016/j.ijheatmasstransfer.
[3] J. Chen, D. Han, W. He, Thermodynamic balancing of heat and mass transfer 2017.01.019.
process to minimize its entropy generation by mass injection and extraction, [23] S.J. Oh, M.W. Shahzad, M. Burhan, W. Chun, C. Kian Jon, M. KumJa, K.C. Ng,
Int. J. Heat Mass Transf. 161 (2020) 120261, doi:10.1016/j.ijheatmasstransfer. Approaches to energy efficiency in air conditioning: a comparative study on
2020.120261. purge configurations for indirect evaporative cooling, Energy 168 (2019) 505–
[4] P. Xu, X. Ma, T.M. Diallo, X. Zhao, K. Fancey, D. Li, H. Chen, Numerical inves- 515, doi:10.1016/j.energy.2018.11.077.
tigation of the energy performance of a guideless irregular heat and mass ex- [24] A.Y.T. Al-Zubaydi, G. Hong, Experimental study of a novel water-spraying con-
changer with corrugated heat transfer surface for dew point cooling, Energy figuration in indirect evaporative cooling, Appl. Therm. Eng. 151 (November
109 (2016) 803–817, doi:10.1016/j.energy.2016.05.062. 2018) (2019) 283–293, doi:10.1016/j.applthermaleng.2019.02.019.
[5] F. Wang, T. Sun, X. Huang, Y. Chen, H. Yang, Experimental research on a novel [25] S. De Antonellis, C.M. Joppolo, P. Liberati, S. Milani, L. Molinaroli, Experimental
porous ceramic tube type indirect evaporative cooler, Appl. Therm. Eng. 125 analysis of a cross flow indirect evaporative cooling system, Energy Build. 121
(2017) 1191–1199, doi:10.1016/j.applthermaleng.2017.07.111. (2016) 130–138, doi:10.1016/j.enbuild.2016.03.076.
[6] D. Pandelidis, A. Cichoń, A. Pacak, S. Anisimov, P. Dra̧g, Counter-flow indi- [26] K. Siebeneck, W. Popov, T. Stefanak, S. Scholl, Pillow plate heat exchangers -
rect evaporative cooler for heat recovery in the temperate climate, Energy 165 investigation of flow characteristics and wetting behavior at single-flow con-
(2018) 877–894, doi:10.1016/j.energy.2018.09.123. ditions, Chemie Ingenieur Technik 87 (3) (2015) 235–243, doi:10.1002/cite.
[7] D. Pandelidis, A. Cichoń, A. Pacak, S. Anisimov, P. Dra̧g, Performance compar- 20140 0 055.
ison between counter- and cross-flow indirect evaporative coolers for heat [27] H. Ishikawa, S. Ookawara, S. Yoshikawa, A study of wavy falling film flow on
recovery in air conditioning systems in the presence of condensation in the micro-baffled plate, Chem. Eng. Sci. 149 (2016) 104–116, doi:10.1016/j.ces.2016.
product air channels, Int. J. Heat Mass Transf. 130 (2019) 757–777, doi:10.1016/ 04.021.
j.ijheatmasstransfer.2018.10.134. [28] A. Gonda, P. Lancereau, P. Bandelier, L. Luo, Y. Fan, S. Benezech, Water falling
[8] A. Sohani, H. Sayyaadi, Thermal comfort based resources consumption and film evaporation on a corrugated plate, Int. J. Therm. Sci. 81 (1) (2014) 29–37,
economic analysis of a two-stage direct-indirect evaporative cooler with di- doi:10.1016/j.ijthermalsci.2014.02.010.
verse water to electricity tariff conditions, Energy Convers. Manage. 172 (2018) [29] X. Cui, K. Chua, W. Yang, Numerical simulation of a novel energy-efficient dew-
248–264, doi:10.1016/j.enconman.2018.07.008. point evaporative air cooler, Appl. Energy 136 (2014) 979–988, doi:10.1016/j.
[9] X. Cui, X. Yang, Q. Kong, X. Meng, L. Jin, Performance evaluation and compari- apenergy.2014.04.040.
son of multistage indirect evaporative cooling systems in two operation modes, [30] S. De Antonellis, L. Cignatta, C. Facchini, P. Liberati, Effect of heat exchanger
Int. J. Energy Res. 44 (12) (2020) 9298–9308, doi:10.1002/er.4803. plates geometry on performance of an indirect evaporative cooling system,
[10] M.H. Mahmood, M. Sultan, T. Miyazaki, S. Koyama, V.S. Maisotsenko, Overview Appl. Therm. Eng. 173 (March) (2020) 115200, doi:10.1016/j.applthermaleng.
of the Maisotsenko cycle - a way towards dew point evaporative cooling, Re- 2020.115200.
new. Sustain. Energy Rev. 66 (2016) 537–555, doi:10.1016/j.rser.2016.08.022. [31] X. Zhao, J. Li, S. Riffat, Numerical study of a novel counter-flow heat and mass
[11] C. Zhan, Z. Duan, X. Zhao, S. Smith, H. Jin, S. Riffat, Comparative study of the exchanger for dew point evaporative cooling, Appl. Therm. Eng. 28 (14–15)
performance of the M-cycle counter-flow and cross-flow heat exchangers for (2008) 1942–1951, doi:10.1016/j.applthermaleng.20 07.12.0 06.
indirect evaporative cooling - Paving the path toward sustainable cooling of [32] B. Porumb, P. Ungureşan, L.F. Tutunaru, A. Şerban, M. Bălan, A review of indi-
buildings, Energy 36 (12) (2011) 6790–6805, doi:10.1016/j.energy.2011.10.019. rect evaporative cooling operating conditions and performances, Energy Proce-
[12] G. Zhu, T.-T. Chow, C. Lee, Performance analysis of counter-flow regenerative dia 85 (2016) 452–460, doi:10.1016/j.egypro.2015.12.226.
heat and mass exchanger for indirect evaporative cooling based on data-driven [33] Z. Duan, C. Zhan, X. Zhang, M. Mustafa, X. Zhao, B. Alimohammadisagvand,
model, Energy Build. 155 (2017) 503–512, doi:10.1016/j.enbuild.2017.09.053. A. Hasan, Indirect evaporative cooling: past, present and future potentials, Re-
[13] A. Adam, D. Han, W. He, J. Chen, Numerical analysis of cross-flow plate type new. Sustain. Energy Rev. 16 (9) (2012) 6823–6850, doi:10.1016/j.rser.2012.07.
indirect evaporative cooler: modeling and parametric analysis, Appl. Therm. 007.
Eng. 185 (2021) 116379, doi:10.1016/j.applthermaleng.2020.116379. [34] N. Stoitchkov, G. Dimitrov, Effectiveness of crossflow plate heat exchanger for
[14] J. Lin, R. Wang, M. Kumja, T. Bui, K. Chua, Multivariate scaling and dimen- indirect evaporative cooling, Int. J. Refrig. 21 (6) (1998) 463–471, doi:10.1016/
sional analysis of the counter-flow dew point evaporative cooler, Energy Con- S0140-70 07(98)0 0 0 04-8.
vers. Manage. 150 (2017) 172–187, doi:10.1016/j.enconman.2017.08.003. [35] A. Bejan, The thermodynamic design of heat and mass transfer processes
[15] B. Riangvilaikul, S. Kumar, An experimental study of a novel dew point evap- and devices, Int. J. Heat Fluid Flow 8 (4) (1987) 258–276, doi:10.1016/
orative cooling system, Energy Build. 42 (5) (2010) 637–644, doi:10.1016/j. 0142- 727X(87)90062- 2.
enbuild.2009.10.034.
[16] Y. Chen, H. Yang, Y. Luo, Indirect evaporative cooler considering condensation
from primary air: model development and parameter analysis, Build. Environ.
95 (2016) 330–345, doi:10.1016/j.buildenv.2015.09.030.

11