Energy and Buildings 123 (2016) 136150

Contents lists available at ScienceDirect

Energy and Buildings

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

Numerical analysis of a desiccant system with cross-ow Maisotsenko

cycle heat and mass exchanger
c
Demis Pandelidis a, , Sergey Anisimov a , William M. Worek b , Pawe Drag
a

Department of Environmental Engineering, Wroclaw University of Technology, 27 Wyspia

nski St., 50-370 Wroclaw, Poland
Stony Brook University, College of Engineering and Applied Sciences, Department of Mechanical Engineering, 127 Engineering Bldg., Stony Brook, NY
11794, USA
c
Faculty of Electronics, Department of Control Systems and Mechatronics, Wroclaw University of Technology, 27 Wyspia
nski St., 50-370 Wroclaw, Poland
b

a r t i c l e

i n f o

Article history:
Received 2 November 2015
Received in revised form 12 March 2016
Accepted 15 April 2016
Available online 25 April 2016
Keywords:
Indirect evaporative cooling
Desiccant system
Maisotsenko cycle

a b s t r a c t
This paper presents a numerical analysis of the desiccant air conditioning system with a cross-ow
Maisotsenko cycle (M-Cycle) heat and mass exchanger (HMX). Presented mathematical models of the
desiccant wheel and HMX are based on the -NTU method. The mathematical model of the desiccant
wheel is also supplemented with algorithm allowing the accurate approximation of the initial conditions
to reduce the calculation time. The novel system was compared with a conventional solution to show its
application potential. The main advantage of the novel system is that it can provide comfort condition
even with less effective dehumidication. Also impact of several operational factors (i.e. inlet air temperature and humidity, outdoor and regeneration airow rate and rotational speed) on the performance of
the system were investigated. The analysis of novel system were performed under assumption that the
desiccant wheel is regenerated with air heated to the relatively low temperature which can be produced
with solar panels in typical moderate climate conditions.
2016 Elsevier B.V. All rights reserved.

1. Introduction
Desiccant air conditioning systems are one of the promising
ways which can improve the performance of traditional air conditioning systems. Such solutions are based on the combination
of different heat exchangers. The typical idea (Fig. 1) includes a
desiccant wheel for dehumidication of the airow, a rotary heat
exchanger for initial cooling of the dehumidied air stream and
a spraying chamber, which cools and humidies the airow with
evaporative cooling. The initial dehumidication allows for effective evaporative cooling which can provide comfort for typical
indoor conditions [14]. The exhaust air is also delivered to the
spraying chamber, where it is cooled and humidied in order to
increase the effectiveness of heat recovery process in the rotary
exchanger. After passing the exchanger, the air is heated with solar
panels or a typical air heater (or both) and it is used for the regeneration of the desiccant wheel.
The main disadvantages of such system are connected with
direct evaporative air cooling: this system adds moisture to the
air, which may result in an uncomfortable indoor conditions. Also

Corresponding author.
E-mail address: demis.pandelidis@pwr.wroc.pl (D. Pandelidis).
http://dx.doi.org/10.1016/j.enbuild.2016.04.039
0378-7788/ 2016 Elsevier B.V. All rights reserved.

the thermal effectiveness of the direct evaporative air cooler is

limited by the wet-bulb temperature of incoming air (in practice
it is impossible to reach the wet-bulb temperature with direct
cooling). The additional disadvantage is the size of the system:
the rotary heat exchanger and spraying chambers require a
lot of volume in the system. For the above mentioned reasons,
researchers are trying to apply new methods in desiccant air
conditioning systems in order to increase their efciency. Montazeri et al. [5] simulated the impact of selected parameters on
evaporative cooling efciency with computational uid dynamics
methods. The selected parameters were: air temperature, air
humidity ratio, airow velocity, water temperature and droplet
size distribution. The results showed that the sensible cooling
capacity of the system can be improve by more than 40% if, for
the given inlet water temperature (35.2 C), the temperature
difference between the inlet air and the inlet water droplets
increases from 0 C to 8 C. Gao et al. [6] analyzed a desiccant
cooling system using a low-grade heat source for regeneration,
equipped with counter-ow indirect evaporative air cooler. Their
study included analysis of impact of the operating parameters
and the number of heat transfer units (NTUs) on the efciency of
rotary heat exchanger and indirect evaporative cooler. The results
indicated that the inlet air temperature should be kept below 35 C,
while the inlet humidity ratio should be lower than 18 g/kg. Chen

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Nomenclature
cd
cg
cp
F
h
G
L
LZ
m
M
q0
qsorp
RH
s
t
V
W
x
X
Y
Z1
Z2

Specic heat capacity of desiccant material [J/(kg K)]

Specic heat capacity of water vapour [J/(kg K)]
Specic heat capacity of moist air [J/(kg K)]
Surface area [m2 ]
Height [m]
Moist air mass ow rate [kg/s]
Length [m]
Streamwise length of the desiccant wheel [m]
Fins temperature coefcient []
Mass [kg]
Latent heat of water evaporation [kJ/kg]
Heat of sorption [kJ/kg]
Relative humidity [%]
Fin pitch dened as a space between the ns [m]
Temperature [ C]
Volumetric airow rate [m3 /h]
Moisture content of desiccant material [kg/kg]
Humidity ratio [kg/kg]
Coordinate along primary air ow direction evaporative cooler [m]
Coordinate along working air ow direction evaporative cooler [m]
Coordinate along the process airow direction desiccant wheel [m]
Coordinate along the regeneration airow direction
desiccant wheel [m]

Special characters
Convective heat transfer coefcient [W/(m2 K)]

Mass transfer coefcient [kg/(m2 s)]

Thickness

 [m]WB 
WB = t1i t1o / t1i t1i
Wet bulb effectiveness [%]
1
Absorption stage duration [s]
2
Desorption stage duration [s]
 1 = 1 /o Non-dimensional absorption stage duration
[s]
 2 = 2 /o Non-dimensional desorption stage duration
[s]
Thermal conductivity [W/(m K)]


Surface wettability factor,  (0.01.0) [s]

Time [s]
o = (1 + 2 ) Time of one revolution of the wheel [s]
 o = ( 1 +  2 ) = 1.0 Non-dimensional time of one revolution of the wheel [s]
Non dimensional coordinates
i
Wheel revolution counter (i = (1. . .n)) []
Stage
counter
(j = 1 absorption
stage,
j
j = 2 desorption stage)
 []

Lewis factor Le = / cp []
Le
NTU
Number of transfer units: airow []
Number of transfer units: desiccant material []
NTU*
X
X = X/L relative X coordinate []
Y
Y = Y/L relative Y coordinate []
Z
Z = Z/ LZ relative Z coordinate []

d
DW
n
g
h
HMX
i
o
p
reg
work
WB


137

Desiccant material
Desiccant wheel
Fin/referenced to n structure
Water vapour
Referenced to the height of the channel
Heat and mass exchanger
Inlet
Outlet
Plate surface
Regeneration airow (desiccant wheel)
Referenced to the working section of heat exchanger
Wet-bulb temperature
Condition at the air/plate interface temperature

et al. [7] analyzed a desiccant material made by combining

silica gel with polyacrylic acid and sodium polyacrylate in order
to determine the most effective mixing ratio. They established
that the novel desiccant has a 41% higher sorption capacity than
silica gel. Subsequently, tests were performed to assess the dehumidication performance of the material at varying air velocities,
regeneration temperatures, and inlet temperatures and humidity.
The results established the optimal mixing proportion to be 10:1:1.
Zhu and Chen [8] used a linear-regression method for a simple
prediction model of a desiccant air conditioning system. The study
included the simulation and analysis of impact of selected factors
on the performance of a marine air-conditioning system It was
established that the dynamic correction factor allows for higher
accuracy of the mathematical model. Gao et al. [9] experimentally
analyzed an integrated liquid-desiccant indirect evaporative
air-cooling system with the M-cycle. The results showed that the
dehumidication process in the rst stage of the cycle has direct
impact on the cooling capacity in the second stage, when the inlet
parameters of the airow or desiccant are changed. The energy
balances obtained were in the range of 20% for all experimental conditions. Jani et al. [10] studied a hybrid air-conditioning
system made of a solid desiccant and mechanical compression
device in TRNSYS environment. The system was tested on a room
with cooling load equal 1.8 kW from March to September. The
obtained results conrmed that presented system allows for a high
efciency in hot and humid climate conditions.
It can be clearly seen that many researchers have discovered
new methods to improve and analyze the desiccant air conditioning
systems. The main problem connected with the typical desiccant
systems is that they require relatively high temperatures for the
regeneration of the desiccant wheel. The best solution is to regenerate the dehumidier using the heat from the solar panels, but
they can only generate such high temperatures in very hot cli-

Subscripts
1
Primary airow in evaporative cooler/process airow in desiccant wheel
2
Working airow in evaporative cooler/regeneration
airow in desiccant wheel

Fig. 1. Scheme of the typical SDEC system.

138

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. 2. Analyzed M-Cycle desiccant system. (a) The scheme of the system. (b) The Maisotsenko cycle cross-ow heat and mass exchanger. (c) The scheme of the desiccant
wheel.

mates. Therefore, traditional desiccant systems are often limited

to hot and humid climates (in hot and dry climates dehumidication of the airow is redundant). In other climatic conditions, they
can effectively operate when there is a source of the waste heat
which can be used for heating of the airow. In order to make such
systems suitable for moderate climate conditions (such as central
Europe and signicant part of the United States), a new component needs to be added to the system which allows for effective
operation under lower regeneration temperatures. One of such
components is the cross-ow Maisotsenko cycle heat and mass
exchanger (HMX) instead of the spray chambers and the rotary heat
exchanger (Fig. 2(a)). This idea has not yet investigated numerically
in the existing studiesthe published data concentrated on the
regenerative counter-ow unit in the desiccant systems [1113].
A regenerative HMX allows obtaining low outlet temperatures, but
due to the unfavorable airow scheme it usually obtains lower cooling capacity than the cross-ow unit [1416]. Many authors claim
that the most effective solution of all evaporative coolers in terms
of both cooling capacity and outlet air temperature is cross-ow MCycle air cooler (More information about cross-ow M-Cycle HMX
can be found in cited literature positions [1418]). It is important
to study its performance in the desiccant systems, especially under
conditions when desiccant wheel is regenerated with relatively low
temperature, because it may have impact on global popularization

of the desiccant systems. For this reason, this study will focus on
the numerical analysis of the desiccant system regenerated with
air with lower temperature levels with the cross-ow Maisotsenko
cycle heat exchanger. The main purpose is to show the levels of the
outlet air temperature which can be obtained with such systems
under different operational conditions.
Two main elements of the desiccant system will be analyzed:
the cross-ow M-cycle HMX (Fig. 2(b)) and the desiccant wheel
(Fig. 2(c)). The HMX is made of a porous material coated with
polyethylene from the one side to create the dry and wet channels.
It is also equipped with ns (Fig. 2(b)). The air delivered to the unit
is divided into the two streams: the primary airow (delivered to
the occupants) and the working airow (delivered to the wet channel and later discharged to the atmosphere). The exhaust airow
is sensibly heated with the solar panels or the standard air heater
(Fig. 1(a)). The working airow is indirectly pre-cooled before it
enters the wet channels, which allows obtaining high effectiveness
(Fig. 1(b)).
2. Methods
The presented desiccant system will be analyzed using mathematical models based on the modied NTU method. The scheme
of the system along with initial conditions is presented in Fig. 3(a).

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

139

- The energy conservation balance:



t1
= NTU 1 t  p1 t1
X


1



fin1

sfin1



2
tanh mfin1 hfin1
mfin1 sfin1
(1)

where NTU1 = 1 F1 / G1 cp .
For the working air stream:
- The energy conservation balance:

Z


2 =1




fin2
2hfin2
t2

t  p2 t2 +
t  fin2 t2 dZ 2
= NTU 2 1

s
s
Y

fin2

fin2

Z 2 =0



(2)

where NTU2 = 1 F2 / G2 cp .
- The mass conservation balance:

x2
= NTU 2
Y

Z
2 =1
 

1 fin2  (x x ) + 2hfin2  (x x )dZ (3)
1
p2
2
2
2
p2
fin2
fin2
Le2
sfin2
sfin2
Z 2 =0

The overall energy balance for the exchanger structure

 W  t
1
1
W2

t2
+
+
X
Y

qo
cp

x2
=0
Y
2

(4)

The mathematical model is supplemented with the initial

conditions, establishing the parameters at the entrance to the
appropriate channels.
For the primary airow at the entrance of the exchange
Fig. 3. Assumptions for mathematical model. (a) General scheme of the system
with initial conditions. (b) Initial conditions: heat and mass exchanger. (c) Initial
conditions: desiccant wheel.

The ambient air enters the desiccant wheel, where it is dehumidied. After that, the air is delivered to the HMX, where it is separated
into the two air streams (the primary and working airows which
are marked 1 and 2, respectively). After passing the HMX, the primary airow is delivered to the conditioned space. After it leaves
the exchanger, the working airow (2o ) is discharged to the atmosphere (Fig. 3(a)).
The scheme of the HMX in this paper is presented in Fig. 3(b),
while the scheme of the model of the desiccant wheel is presented
in Fig. 3(c). The basic assumptions for the mathematical model of
the cross-ow M-Cycle unit, along with the detail conversion of the
heat transfer equations were presented in [17]. Therefore they will
be omitted in this paper. Only the nal forms of the equations will
be presented.
The nal forms of the cross-ow M-Cycle HMX model equations
are written below. The model is based on the mass and energy balances for the primary and working airow. It should be mentioned
that the mathematical model presented in this article describes the
product part of the exchanger (Fig. 3(b)). Equations describing the
initial part of the unit are identical, except the additional algorithm
describing the process of air streams mixing. The processes occurring in the initial part of the unit were described by authors in detail
in their previous studies [17,18].
For the primary air stream:

HMX = t DW
t1 = t1i
1o

HMX = xDW = const

x1 = x1i
1o

| X = 0.0
;


Y = lwork . . .1.0
Y

| X = (0.0. . .1.0)


Y = lwork . . .1.0

(5)

For the working air ow at the entrance to the dry channel of the
exchanger
HMX = t DW
t3 = t3i
1o

HMX = xDW = const

x3 = x3i
1o

| X = 0.0
;


Y = 0.0. . .lwork
Y

| X = (0.0. . .1.0)


Y = 0.0. . .lwork

(6)

For the working airow at the entrance of the wet channel in

units product part of the exchanger

= t4o

t2i
| X = (0.0...1.0)
Y = lwork
Y

= x4o

x2i
| X = (0.0...1.0) ;
Y = lwork
Y

| X = (0.0...1.0)
Y = lwork
Y

| X = (0.0...1.0)
Y = lwork
Y

(7)
The governing equations for the mathematical model of the
rotary dehumidier are presented below.
The energy conservation balance for the process airow:

t1
= NTU1 (td1 t1 )
Z 1

(8)

140

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

where NTU1 = 1 F1 / G1 cp1 .

The mass conservation balance for the process airow:

x1
1
= NTU1
(x x1 )
Le1 d1
Z 1

Desorption side:

Z 2
t2 ,

Wd1
= NTU1


cd
cp Le

(x1 xd1 )

(10)

td2

Z 2
,


1  sorp
q
+ cg1 (t1 td1 )
cd1 1

 = td2 ,
Z 2
Z 1
,

or

t2
= NTU2 (td2 t2 ) 1 +
Z 2

cg2
cp2

cg
cp Le

(td2 t2 )

(13)

Z 1
td1 ,


(xd2 x2 )

(14)

where NTU2 = 2 F2 / G2 cp2 .

The mass conservation balance for the regeneration airow:

x2
1
= NTU2
(xd2 x2 )

Le
Z2
2

(15)

The mass conservation balance for desiccant matrix in the

regeneration sector:

Wd2
= NTU2


cd
cp Le

(x2 xd2 )

(16)

where NTU2 = 2 F2 0 / (Md2 cd2 ).

The energy conservation balance for desiccant matrix in the
regeneration sector:

td2
Wd2 qsorp
= NTU2 (t2 td2 ) +
cd


or

td2
= NTU2 (t2 td2 ) +


qsorp
cp Le

(17)
2


(x2 xd2 )

(18)

The mathematical model is supplemented with initial, switching and boundary conditions.
Initial conditions:

Z 1
td1 ,

 

= tdo Z 1

Z 1
Wd1 ,

| Z 1 = (0.0. . .1.0) ;

 

= Wdo Z 1

| Z 1 = (0.0. . .1.0)

 = 0

(19)

 = 0

Boundary conditions:
Absorption side:

Z 1
t1 ,

= t DW = const

Z 1
x1 ,

1i

| Z 1 = 0.0
 = ((i 1) . . . ( 1 + (i 1)))

= xDW = const
1i

| Z 1 = 0.0
 = ((i 1) . . . ( 1 + (i 1)))

(20)

Wd1



 = Wd2 ,
Z 2
Z 1
,

Z 1 = (0.0. . .1.0)

| Z 2 = (0.0. . .1.0) .

Z 1 = 1 Z 2

Z 1 = 1 Z 2

 = i

 = i

(23)

Conditions of the quasi-steady mode realization (Cyclical steady

state conditions, which are independent of the stating conditions)

Z 1 = (0.0. . .1.0)

 =  1 + (i 1)

| Z 2 = (0.0. . .1.0) ;

Regeneration part
The energy conservation balance for the regeneration airow:

t2
x2
= NTU2 (td2 t2 ) +
Z 2
Z 2



 = Wd1 ,
Z 1

Z 2 = 1 Z 1

Z 1 = (0.0. . .1.0)

(12)

(21)

| Z 2 = (0.0. . .1.0) ; (22)

 =  1 + (i 1)

 sorp

td1
1
= NTU1 (t1 td1 ) +NTU1
(x1 xd1 ) q1 + cg1 (t1 td1 )
cp1 Le1


Z 2
,

Z 2 = 1 Z 1

td1

Wd2

| Z 2 = (0.0. . .1.0) ;

(11)

or

2i

 = (( 1 + (i 1)) . . .i)

Z 1 = (0.0. . .1.0)

= 1 F1 0 / (Md1 cd1 ).
where
The energy conservation balance for desiccant matrix in the
process sector:

Wd1


= xDW = const
| Z 2 = 0.0



 = td1 ,
Z 1

NTU1

Switching conditions:

td1
= NTU1 (t1 td1 ) +


Z 2
x2 ,

2i

 = (( 1 + (i 1)) . . .i)

= t DW = const
| Z 2 = 0.0

(9)

The mass conservation balance for desiccant matrix in the process sector:

Z 1
= td1 ,
|

Z 1 = (0.0. . .1.0)

Z 1
Wd1 ,
;

Z 1
= Wd1 ,
|

Z 1 = (0.0. . .1.0)

 = i

 = i

=i+1


=i+1


(24)

The humidity ratio of air in equilibrium with the desiccant

matrix is dened as a function of matrix temperature and moisture
content of desiccant matrix [11,19]
xd = xd (td , Wd )

(25)

Simulation of multi-cycle heat and mass transfer process in a

rotary desiccant wheel consisted of a dehumidication process followed by a regeneration process in a cyclic manner. This is much
more complicated task in comparison with a single simulation cycle
[20]. In the case of traditional way of simulation, it is necessary to
calculate the rst phase (dehumidication), then the second one
(regeneration), and then repeat such sequence of calculations for
the next cycle. Computations should be repeated for a large number of intermediate cycles in order to approach quasi steady-state
conditions. Such method of simulation requires a signicant time
of computing, which is an obstacle in the research process. Consequently, rst and foremost, it is necessary to minimize the time of
approaching a cyclic steady-state mode. One of the ways to solve
this problem is to nd a rational approximation of the initial conditions. The effectiveness of this approach is based on the fact, that
a cyclic steady-state mode does not depend on the initial conditions and is a result of an asymptotic process. Preliminary series
of numerical simulations allowed to establish an optimal distribution of the initial parameters, based on an asymptotic assumptions
of the investigated heat and mass transfer processes realized in
the rotor of desiccant wheel at a matrix rotation frequency higher
than critical value (turns per minute, no > 7 min1 ) [21]. In this case,
the temperature and moisture content of desiccant material within
each cross-section remains practically substantially constant during the whole cycle (the psychrometric representation of processes
occurring under critical operational conditions in the desiccant
wheel is visible in Fig. 4). This fact allows the heat and mass transfer
in the matrix of desiccant wheel to be described by the set of ordinary differential equations of heat and mass conservation balances.

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

141

The energy conservation balance for desiccant matrix

NTU1 (td t1 ) + NTU1 (td t1 )


+

W2
W1

or
dt1

dZ
+


NTU2 (td t2 ) +

 W  dt
2
2
dZ

W1

dZ

cp

(x1 xd ) + NTU1
1

NTU2



 W  dx  qsorp 
2
2
W1

W2
W1

dx1
dZ

cg
cp Le

qsorp
cp


+
2

cg
cp

qsorp
cp Le

(xd x1 ) +
1

(31)

(xd x2 ) = 0
2


+

qsorp
cp Le

cg
cp


(td t1 )
1

(td t2 ) = 0

(31)

The mass conservation balance for desiccant matrix:

NTU1

1
(x x1 ) +
Le1 d

or
dx1

dZ

 W  dx
2
2
W1

dZ

W 
2

W1

NTU2

1
(x x2 ) = 0
Le2 d

=0

(32)

(33)

Simplied energy conservation balance for desiccant matrix

dt1

dZ

 W  dt
2
2
W1

dx1
+
dZ
dZ

cg
cp

(t1 t2 ) 0

(31*)

The moisture content of desiccant matrix in equilibrium with

the moist air is dened as a function temperature and humidity
ratio of air [22]
Wd = Wd (td , xd )

Fig. 4. Psychrometric representation of process occurring in desiccant wheel under:

(a) very fast rotational speed (b) very slow rotational speed.

The governing equations for the mathematical model of desiccant

rotary wheel at rotation frequency higher than critical value are
presented below.
The energy conservation balance for the process airow:
dt1
= NTU1 (td t1 )
dZ

(26)

(34)

The mathematical model given by Eqs. (26) (34) is supplemented by boundary conditions Eqs. (20) and (21) for absorption
and desorption process. It should be emphasized, that in this case
the developed model, Eqs. (26) (34), is considered to estimate a
rational initial approximation of cyclic steady-state mode. Taking
into account, that obtained proles of the temperature and moisture content of desiccant matrix characterize the distribution of
average values of these parameters in a quasi-steady mode, it is
expedient to move initial conditions (19) into another time layer
corresponding to the mid-stage of the dehumidication phase

Z 1
td1 ,

 

= tdo Z 1

| Z 1 = (0.0. . .1.0) ;
 =  1 /2

where NTU1 = 1 F1 / G1 cp1 .

The mass conservation balance for the process airow:
dx1
1
= NTU1
(x x1 )
Le1 d
dZ

(27)

Z 1
Wd1 ,

 

= Wdo Z 1

| Z 1 = (0.0. . .1.0)

(35)

 =  1 /2

Both models are implemented in a multi-module computer

simulation program. A four-dimensional computational numerical code using a modied Runge-Kutta method was implemented
using the Wolfram Mathematica environment. The validation of
presented mathematical model is presented in Appendix A.

The energy conservation balance for the regeneration airow:

dx2
dt2
= NTU2 (td t2 ) +
dZ
dZ
or

dt2
= NTU2 (td t2 ) 1 +
dZ

cg2
cp2

cg
cp Le

3. Results and discussion


(td t2 )

(28)


(xd x2 )

(29)

where NTU2 = 2 F2 / G2 cp2 .

The mass conservation balance for the regeneration airow:
dx2
1
= NTU2
(x x2 )
Le2 d
dZ

(30)

The analyzed system is characterized by the parameters listed

in Table 1. It is important to mention, that the analysis of methods of heat production for the regeneration of the desiccant wheel
is not the subject of this paper. However, the whole analysis will
be presented for rather low values of regeneration air temperature
(5560 C), which can be produced by solar panels under typical
summer conditions in moderate climate (such as central Europe),
or any other low temperature heat source. There are also several
additional assumptions for the analyzed system: the temperature
and humidity ratio difference between the supply and exhaust air
in conditioned room is 7 C and 1 g/kg respectively. The assumed
temperature increase (caused by the compensation of the heat

142

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Table 1
Properties of analyzed cooling system components.
Parameters
Desiccant wheel matrix
Specic heat capacity (silica gel), J/(kg K)
Channel type
Channel height, mm
Channel width, mm
Wall thickness, mm
Channel length, m

750
Sinusoidal
1.24
2.21
0.1
0.1

Rotor construction
Rotor diameter, m
Percent of rotor surface used for regeneration airow, %
Rotational speed, turns per hour

0.45
25
8

M-Cycle heat exchanger

Length, m
Width, m
Height, m
Channel height, mm
Channel width, mm
Working to product airow ratio

0.5
0.5
0.5
3
25
1.0

Rotary heat exchanger

Temperature effectiveness, %

75

Direct evaporative air coolers in traditionalsystem

WB
Wet bulb effectiveness, WB = (t1i t1o ) / t1i t1i
,%

80

loads) between supply and indoor air is 6 C, also a 1 C temperature increase between indoor air and the exhaust air is assumed.
Therefore, the safe supply air temperature, which allows keeping the indoor air temperature within thermal comfort conditions
at about 25. . .25.5 C is 19. . .19.5 C. The source of heat will not
be analyzed in this paper, but all of the studies will be performed
under consideration that the desiccant wheel is regenerated using
air with relatively low temperature, in compare to the temperatures typically used for analysis of the desiccant systems, which
can be obtained using solar panels in most of the climatic conditions. Due to this fact the regeneration costs along with the COP of
the system is not analyzed in this paper (the low regeneration air
temperature simulates the assumption that solar power covers the
required power for regeneration of the desiccant wheel).
3.1. Comparison between traditional system and the novel
solution
The general comparison between the novel system and the traditional systems is presented on the psychrometric chart shown in
Fig. 5. The simulation was performed for a typical middle Europe
summer conditions: ambient air temperature of 30 C and a humidity ratio of 11 g/kg (RH = 41%). The systems are combined under
assumption that they have to provide the appropriate supply air
temperature (approximately 19 C or lower). The minimal assumed
regeneration air temperature is 55 C, if the system is not able to
provide appropriate temperature, the regeneration air temperature
is increased until appropriate supply air temperature is obtained.
For the purpose of this comparison, a typical wet-bulb effectiveness
of direct evaporative air coolers (80% [23]) and a typical temperature effectiveness of the rotary heat exchanger (75%) are assumed.
It can be seen that there are signicant differences in the air treatment in both analyzed systems. The supply air temperatures are
similar (18.9 C for the system with M-Cycle and 19.1 C for the
traditional system: Fig. 5(a) and (b)). The main advantage of the
Maisotsenko cycle heat and mass exchanger is that it does not add
moisture to the supply airow, which gives two benets: the air in
the conditioned room (marked R in Fig. 5(a) and (b)) lies in the thermal comfort zone, whereas in the case of the traditional system it is
too humid (indoor air RH 5065% in favor of the M-Cycle system).

Fig. 5. Comparison between traditional and novel desiccant air conditioning systems. (a) Traditional system. (b) System with the M-Cycle HMX. (c) System with the
M-Cycle HMX and rotary heat exchanger.

It should also be noted that the exhaust air, which is later used for
regeneration of the desiccant matrix, is substantially dryer in the
case of the novel system (humidity ratio 10.315 g/kg in favor of
the novel system). This allows the use of lower temperature regeneration air in comparison with the traditional system. In the case
of the M-Cycle system the regeneration air temperature, required
to dehumidify the ambient air by approximately 2 g/kg, is 55 C,
whereas for the traditional system it is 63 C. Moreover, when using
the traditional system, it is not always possible to meet indoor air
comfort conditions in terms of relative humidity. In order to provide
thermal comfort conditions it is necessary to increase regeneration
air temperature (about 70 C), which signicantly limits its application to only hot regions of the world. Such temperatures can be
easily obtained when the conditioned object has a source of the
waste heat with relatively high temperature, but in the case of only
solar panels using it is very difcult to reach such temperature level
of regeneration air for typical middle Europe climatic conditions
[22].

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. 6. Impact of inlet ambient air temperature on the system performance.

Another important advantage of the M-Cycle unit is that it can

provide a low supply air temperature using only one device, instead
of three (one HMX instead of two direct evaporative air coolers and
a rotary heat exchanger). Moreover, the heat and mass exchanger
consumes a smaller amount of water, less than half that of the
direct evaporative air cooler [24]. The main disadvantage of the
Maisotsenko cycle exchanger is that it requires a part of the dehumidied air to be delivered to the wet working channels and later
extracted from the system. Therefore, in order to provide the same
amount of air as the traditional system it would require a larger
desiccant wheel, which might increase the cost of the system. However, the savings connected with the reduced number of the system
components should overcome this cost disadvantage.
The system with the Maisotsenko cycle HMX can be also
equipped with the rotary heat exchanger (Fig. 5(c)). It can be seen,
that this conguration gives considerably lower supply and indoor
air temperatures: 16.9 C and 22 C respectively (Fig. 5(c)). Unfortunately, this is connected with higher investment costs and the fact,
that an additional device brings a higher risk of potential malfunction. Moreover, such solution requires additional energy to operate
the rotary exchanger. On the other hand, the ability to provide lower
supply air temperature is an important advantage: in many cases
the supply air temperature is critical in terms of choosing the air
handling unit. Depending on the type of conditioned space and the
investor requirements, this solution may be considered as an alternative to the system equipped only with the HMX. This system will
be also analyzed in this paper under assumption of the constant
temperature effectiveness of the rotary heat exchanger (Table 1) in
order to show the theoretical capabilities of different systems with
the M-Cycle HMX.
3.2. Inuence of selected operational parameters on the system
performance
3.2.1. Impact of inlet ambient air temperature on the system
performance
The impact of the inlet ambient air temperature on the performance of the two M-Cycle systems (with and without the rotary
heat exchanger) is presented in Fig. 6. The numerical simulations
were performed with assumptions of constant exhaust air temperature equal to 26 C and the 1 g/kg humidity ratio increase between
supply and exhaust airow. The ambient airow is equal 600 m3 /h,
the regeneration airow is equal 150 m3 /h. It can be seen, that
increasing ambient air temperature results in higher supply air
temperatures. This is caused by the fact, that the dehumidication of ambient air at higher temperature is realized in desiccant
matrix with lower efciency. The moist air results in decreasing
effectiveness of indirect evaporative cooling through the Maisot-

143

Fig. 7. Impact of inlet ambient air humidity ratio on the system performance.

senko cycle. The system without rotary heat exchanger is more

sensitive to the inlet air temperature changes: for the ambient air
temperature varying from 25 to 40 C, the supply air temperature
provided by this system increases by 3 C, whereas in the case of
the system equipped with rotary exchanger, the change in supply
air temperature is only 1.5 C. This is caused by the fact, that the
heat exchange in the rotary unit under summer conditions is less
dependent on the humidity changes of the inlet airow than the
evaporative air cooling unit. The main disadvantage of the rotary
exchanger is that it is limited by the exhaust air temperature and
it cannot be used as the main cooling source.
The system without a rotary exchanger requires additional heating of the regeneration airow when the outdoor air temperatures
are higher than 33 C: for higher air temperatures, this system is
not able to provide the supply air with a temperature equal 19.5 C
or lower. The positive aspect is that when such temperatures occur,
there is usually enough solar energy to provide higher regeneration
air temperatures with solar panels.
3.2.2. Impact of ambient air humidity ratio on the system
performance
The impact of the inlet ambient air humidity ratio on the performance of the two analyzed systems is presented in Fig. 7. The
simulations were performed under the assumption of constant
exhaust air temperature equal 26 C and the 1 g/kg humidity ratio
increase between supply and exhaust airow. The ambient airow
is equal 600 m3 /h, the regeneration airow is equal 150 m3 /h. It
can be seen that the higher humidity ratio of the airow has negative impact on the system performance, however, the impact of
higher humidity ratio is not as negative as the higher inlet air temperature. This is due to the fact that air with higher humidity is
dehumidied at higher efciency in the desiccant wheel (however, the air with lower inlet humidity always achieves lower outlet
humidity). The more effective dehumidication partly reduces the
negative impact of higher humidity ratio, so the evaporative cooling process provides lower outlet air temperatures. The system
without rotary exchanger requires additional heating of the regeneration airow when the incoming air has higher humidity ratio
than 12 g/kg (Fig. 7) to keep the thermal comfort (maximal temperature in human occupation zone is assumed as 25.5 C). Therefore,
the supply air temperature has to be colder or equal to 19.5 C
when temperature increase between supply air and human occupation zone is assumed as 6 C. It can be seen that when incoming
air has higher humidity ratio than 12 g/kg the supply air temperature is higher than 19.5 C, therefore desiccant wheel requires more
power for regeneration in order to provide air with the appropriate
temperature. The system with the rotary exchanger doesnt require
additional heating in the analyzed case.

144

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. 8. Impact of regeneration air temperature on the system performance.

3.2.3. Impact of regeneration air temperature on the system

performance
The impact of the regeneration air temperature on the analyzed
systems is presented in Fig. 8. The simulations were performed
under assumptions of constant exhaust air temperature equal 26 C
and the 1 g/kg humidity ratio increase between supply and exhaust
airow. The ambient airow is equal 600 m3 /h, the regeneration
airow is equal 150 m3 /h. Fig. 8 clearly shows that higher regeneration air temperature has a positive impact on the effectiveness of
both systems. Also, it can be seen, that under presented conditions,
even low regeneration air temperature (45 C) enables the supply
air temperature to satisfy comfort conditions. This is caused by the
unique features of the Maisotsenko cycle: the efciency of the heat
and mass exchanger which is much higher than efciency of typical direct and indirect evaporative air coolers [24,25] This results in
lower supply air temperatures even under unfavorable conditions,
when incoming airow is relatively hot and moist.
It can be also seen, that the system with the rotary heat
exchanger is more sensitive to the variation of regeneration air
temperature. This is caused by the fact, that the ambient airow
is more intensively heated, when the regeneration air temperature
is higher. This results in a higher temperature difference between
the outdoor and exhaust airow, which results in more effective
pre-cooling in the rotary exchanger before it enters the HMX.

3.2.4. Impact of ambient and regeneration airow rate on the

system performance
The impact of ambient and regeneration airow rate on the analyzed systems is shown in Fig. 9. The simulations were performed
under assumptions of constant exhaust air temperature equal to
26 C and the 1 g/kg humidity ratio increase between supply and
exhaust airow. It can be seen that the increased regeneration airow improves the effectiveness of the system. The negative effect
of the increased outdoor air stream follows from the fact that it
decreases its number of heat transfer units (NTU) when the air is
passing through the desiccant wheel and the HMX. However, both
the systems are still able to provide satisfying supply air temperature in all of the analyzed cases. It is important to mention that
increased ambient airow results in higher cooling capacity (the
drop in supply air temperature is compensated by increasing the
mass airow rate), therefore in some cases it may be reasonable to
provide higher supply airow rate at cost of its temperature.
The positive impact of increased regeneration airow rate
follows from its increased specic heat capacity, which allows
removing the moisture from the desiccant wheel with higher effectiveness. However, the higher regeneration airow requires more
energy to be heated and it causes a higher pressure drop. Therefore,
its value should be kept at the minimal level, which allows effective
dehumidication of the ambient air. In some cases, it may be reasonable though to increase the regeneration airow rate and heat
it to the lower temperature level instead of keeping the smaller
airow, which has to be heated to the higher temperature.
3.2.5. Impact of desiccant wheel rotational speed on the system
performance
The impact of desiccant wheel rotational speed on the performance of analyzed systems is presented in Fig. 10. The simulations
were performed under assumptions of constant exhaust air temperature equal to 26 C and the 1 g/kg humidity ratio increase
between supply and exhaust airow. The ambient airow is equal
600 m3 /h and the regeneration airow is equal 150 m3 /h. It can be
observed that the increased rotational speed of the desiccant wheel
may have both positive and negative impact on the efciency of
the system. When the dehumidier turns very slowly (23 revolutions per hour) the effectiveness of the system is minimal. The
same effect occurs when the wheel turns very fast (15 revolutions
per hour and higher). The negative impact of very slow rotational
speed follows from the fact that wheel remains in the ambient air
section for too long- after a certain time it needs to be regenerated,
otherwise its potential of mass transfer is too low to dehumidify
the airow effectively. The negative impact of fast rotational speed

Fig. 9. Impact of airow rates on the system performance. (a) Ambient airow rate (Vreg = 150 m3 /h). (b) Regeneration airow rate (V1 = 600 m3 /h).

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. 10. Impact of desiccant wheel rotational speed on the system performance.

follows from the fact that very fast rotational speed results in a
high efciency of heat transfer but the effectiveness of desiccant
matrix regeneration signicantly reduces. It can be seen, that the
optimal rotational speed for the analyzed case varies between 6
and 10 revolutions per hour.
3.2.6. Impact of humidity ratio change between supply and
exhaust airow on the system performance
In the typical air conditioning applications, the larger part of
the heating loads in the conditioned spaces are sensible heat loads.
However, in some cases the latent heat loads may also become
signicant (especially in crowded rooms). Therefore the humidity
ratio of the exhaust airow is usually different than of the supply airow. It is important to study how much does it affect the
performance of analyzed systems. Analysis is based on assuming
moisture loads which, combined with previously assumed temperature increase, signicantly change the room process vector (which
is calculated as the ratio between total head loads- sensible+ latent
in [W] to moisture loads in [kg/s] [26]). It can also be obtained
graphically on the psychrometric chart by drawing a line from the
supply air point to the exhaust air point with assumed temperature and humidity ratio increase. In our case it changes from over
50 000 kJ/kg (sometimes marked as innity which is represented
as a straight vertical line [26]) to about 6500 kJ/kg. The room process vector values higher than 50 000 represent spaces with no
or with negligibly small latent heat loads. Values of room process
vector on the level of 6500 kJ/kg represent rooms with very high
latent heat loads (such vectors occur in very crowded rooms or in
places like swimming pools [26]). The results of the numerical simulation are shown in Fig. 11. The simulations were performed with
assumptions of constant exhaust air temperature equal 26 C. The
ambient airow is equal 600 m3 /h and the regeneration airow is
equal 150 m3 /h. It can be seen that increased humidity ratio of the
regeneration airow does not have a signicant impact on the efciency on the systems considered here. This follows from the fact
that due to the use of the M-Cycle HMX, the airow delivered to the
conditioned space remains dry (the same humidity ratio as on the
desiccant wheel outlet). Therefore, even after assimilation of latent
heat loads in the room, the regeneration airow is still relatively dry
in compare with the traditional desiccant system (because exhaust
air is not humidied in the spray chamber). This allows for satisfying regeneration of the desiccant material. For the above mentioned
reasons, the desiccant air conditioning system with the cross-ow
M-Cycle HMX allows for operation on heat generated by the solar
energy instead of relying on the external heat source in more world
regions than the traditional solution.

145

Fig. 11. Impact of humidity ratio change between supply and exhaust airow on
the system performance.

4. Conclusions
In this paper the numerical analysis of the desiccant system
equipped with the cross-ow Maisotsenko cycle indirect evaporative air cooler are presented. The study was performed with the
assumption of relatively low temperature of the desiccant wheel
regeneration airow, which can be obtained using solar energy in
moderate climate conditions, typical of central Europe.
The results of the numerical simulation indicated that the novel
system allows for higher temperature effectiveness than the traditional solution even when the desiccant wheel is regenerated
with lower temperature. The additional advantage of the M-Cycle
exchanger is that it does not add moisture to the airow, which
allows providing more comfort to the conditioned spaces.
The impact of the several operational parameters on the system
performance was investigated. It was established that under very
hot and humid inlet conditions the system may require additional
heat for the regeneration of the desiccant wheel in order to provide
satisfying indoor conditions. The results also showed that the system is sensitive on the ambient and regeneration airow rate and
rotational speed of the desiccant wheel and it is less sensitive on
the humidity ratio change between supply and exhaust airow.
Acknowledgement
Corresponding author received nancial support for his research
from National Science Center Poland under the PhD scholarship
based on the decision no. DEC-2015/161/T1/ST8/00484.
Appendix A. : Validation of mathematical models
Maisotsenko cycle heat and mass exchanger
Mathematical model describing the M-Cycle heat and mass
exchanger was validated against data collected by authors in Coolerado Corporation facilities in Denver, Colorado, USA. The details of
the experiment including the description of the testing bench and
error analysis was presented by authors in Ref. [25]. For this reason
in this Appendix the short description will be presented.
The tested HMX was inside a test chamber (Fig. A1(a)) in order to
minimize the heat exchange with the surroundings. Test chamber
was also equipped with a fan and an air lter (F1 and AF, Fig. A1(a)
and (b)). The airow was delivered from outside the chamber, with
an air handling unit equipped with a fan, a heater fan and an air
humidier (F2, H and VH in Fig. A1(a)). The air is exhausted from

146

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. A1. Measurement station at Coolerado Corporation facilities. (a) Scheme of the testing bench. (b) Photograph of the test chamber.

the chamber by additional fans (SF, EF in Fig. A1(a)). The supply and
working airows were measured in points SI and EI. The dry and wet
bulb temperatures were measured in points DB and WB in Fig. A1(a)
and (b). The measured inlet airow rate varied from 150 m3 /h to
800 m3 /h, the inlet ambient air temperature varied from 18 to 45 C,
inlet relative humidity varied from 20 to 70%.
The results of the validation are visible in Fig. A2. The discrepancies between the primary air temperature obtained by model
and the experiment are at most 3% (the highest differences are
equal to 0.5 C: Fig. A2). The correlation between the experimental
and the simulation results is equal 0.996 for predicting supply air
temperature (Fig. A2(c)) and 0.997 for predicting the outlet working air temperature. It can be seen that both the experimental and
simulation results showed good agreement. For the details of the
experimental analysis and comparison between more experimental and numerical results see ref. [25].

Desiccant wheel
The mathematical model describing the desiccant wheel was
validated against existing experimental data collected by Kowalski [22] at testing bench at Wroclaw University of Technology,
Wroclaw, Poland. The measurement station used by Kowalski was
equipped with basic air treatment devices: air lters, heaters and
humidiers for both the outdoor and regeneration airow (Fig. A3).
The parameters of the desiccant wheel used in the tests are established in Table A1.

Table A1
Parameters of the desiccant wheel used by Kowalski [22].
Desiccant wheel parameters
Specic heat capacity (silica gel), J/(kg K)
Channel type
Channel height, mm
Channel width, mm
Wall thickness, mm
Channel length, m
Wheel diameter, m
Percent of wheel used for regeneration airow, %

750
Sinusoidal
1.24
2.21
0.371
0.1
0.45
25

The inlet ambient air temperature in Kowalskis test varied

from 18.2 to 34.2 C, the regeneration air temperature varied from
30.5 to 101.0 C, the ambient air humidity ratio changed from 4.112.2 g/kg, the regeneration air humidity ratio varied from 4.4 to
15.9 g/kg, the outdoor airow rate varied from 358 to 650 m3 /h, the
regeneration airow rate varied from 80 to 215 m3 /h. The parameters measured included the air temperature, relative humidity,
mass ow rate and the pressure drop on the exchanger.
The comparison between mathematical model and the experimental results is presented in Fig. A4. The comparison includes
the drop in humidity ratio of the outdoor airow for one variable parameter and the other remaining unchanged. The analyzed
parameters include: variable outdoor and regeneration air temperature (Fig. A4(a) and (b)), variable ambient and regeneration air
humidity ratio (Fig. A4(c) and (d)) and variable ambient and regeneration airow rate (Fig. A4(e) and (f)). The discrepancies between
the experimental and simulation results are at most 0.4 g/kg, the

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. A2. Validation of mathematical model describing the M-Cycle HMX. (a) Outlet primary air temperature (constant air stream velocity). (b) Outlet primary air stream temperature t1o (constant inlet air temperature and
humidity). (c) Correlation between numerical model and experiment: outlet primary air temperature. (d) Correlation between numerical model and experiment: outlet working air temperature.

147

148
D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

Fig. A3. Measurement station used by Kowalski at Wroclaw University of Technology [22]. (a) Scheme of the testing bench. (b) Photograph of the tested desiccant wheel. (c) Photograph of the testing bench.

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

149

Fig. A4. Validation of mathematical model describing the desiccant wheel against data collected by Kowalski [22]. (a) Ambient air humidity ratio drop at variable inlet
ambient air temperature. (b) Ambient air humidity ratio drop at variable inlet regeneration air temperature. (c) Ambient air humidity ratio drop at variable inlet ambient
air humidity ratio. (d) Ambient air humidity ratio drop at variable inlet regeneration air humidity ratio. (e) Ambient air humidity ratio drop at variable inlet ambient airow
rate. (f) Ambient air humidity ratio drop at variable inlet regeneration airow rate.

correlation coefcient R2 is equal 0.971. It can be seen that the

model and the experimental results show satisfactory agreement
and trend, therefore the model can be used for numerical analysis
of the desiccant wheel.
References
[1] A. Hasan, Going below the wet-bulb temperature by indirect evaporative
cooling: analysis using a modied NTU method, Appl. Energy 89 (2012)
237245.
[2] B. Riangvilaikul, S. Kumar, An experimental study of a novel dew point
evaporative cooling system, Energy Build. 42 (2010) 637644.

[3] Z. Duan, Z. Changhong, X. Zhang, M. Mustafa, B. Alimohammadisagvand, A.

Hasan, X. Zhao, Indirect evaporative cooling: past, present and future
potentials, Renew. Sustain. Energy Rev. 16 (2012) 68236850.
[4] M. Farmahini-Farahani, H. Heidarinejad, Increasing effectiveness of
evaporative cooling by pre-cooling using nocturnally stored water, Appl.
Therm. Eng. 38 (2012) 117123.
[5] H. Montazeri, B. Blocken, J.L.M. Hensen, CFD analysis of the impact of physical
parameters on evaporative cooling by a mist spray system, Appl. Therm. Eng.
75 (2015) 608622.
[6] W. Gao, W. Worek, V. Konduru, K. Adensin, Numerical study on performance
of a desiccant cooling system with indirect evaporative cooler, Energy Build.
86 (2015) 1624.
[7] C. Chen, C. Hsu, C. Chen, S. Chen, Silica gel polymer composite desiccants for
air conditioning systems, Energy Build. 101 (2015) 122132.

150

D. Pandelidis et al. / Energy and Buildings 123 (2016) 136150

[8] J. Zhu, W. Chen, A novel multivariate linear prediction model for the marine
rotary desiccant air-conditioning by adding a dynamic correction factor, Appl.
Therm. Eng. 78 (2015) 101109.
[9] W.Z. Gao, Y.P. Cheng, A.G. Jiang, T. Liu, K. Anderson, Experimental
investigation on integrated liquid desiccant Indirect evaporative air cooling
system utilizing the Maisotesenko-cycle, Appl. Therm. Eng. 88 (2015)
288296, http://dx.doi.org/10.1016/j.applthermaleng.2014.08.066.
[10] D.B. Jani, M. Mishra, P.K. Sahoo, Performance studies of hybrid solid
desiccantvapor compression air-conditioning system for hot and humid
climates, Energy Build. 102 (2015) 284292.
[11] W. Gaoa, W. Worek, V. Konduru, K. Adensin, Numerical study on performance
of a desiccant cooling system with indirect evaporative cooler, Energy Build.
86 (2015) 1624.
[12] T. Miyazaki, A. Akisawa, I. Nikai, The cooling performance of a building
integrated evaporative cooling system driven by solar energy, Energy Build.
43 (2011) 22112218.
[13] T. Miyazaki, I. Nikai, A. Akisawa, Simulation analysis of open cycle adsorption
air conditioning system numeral modeling of a xed bed dehumidication
unit and the Maisotsenko cycle cooling unit, Int. J. Energy Clean Environ. 12
(2011) 341354.
[14] S. Anisimov, D. Pandelidis, J. Danielewicz, Numerical analysis of selected
evaporative exchangers with the Maisotsenko cycle, Energy Convers. Manage.
88 (2014) 426441.
[15] S. Anisimov, D. Pandelidis, Theoretical study of the basic cycles for indirect
evaporative air cooling, Int. J. Heat Mass Transfer 84 (2015) 974989.
[16] D. Pandelidis, S. Anisimov, W.M. Worek, Comparison study of the
counter-ow regenerative evaporative heat exchangers with numerical
methods, Appl. Therm. Eng. 84 (2015) 211224.

[17] S. Anisimov, D. Pandelidis, Numerical study of the Maisotsenko cycle heat and
mass exchanger, Int. J. Heat Mass Transfer 75 (2014) (2014) 7596.
[18] D. Pandelidis, S. Anisimov, Numerical analysis of the heat and mass transfer
processes in selected M-Cycle heat exchangers for the dew point evaporative
cooling, Energy Convers. Manage. 90 (2015) 6283.
[19] M.N. Golubovic, H.D.M. Hettiarachchi, W.M. Worek, Evaluation of rotary
dehumidier performance with and without heated purge, Int. Commun.
Heat Mass Transfer 34 (2007) 785795.
[20] P. Bareschino, G. Diglio, F. Pepe, G. Angrisani, C. Roselli, M. Sasso, Modelling of
a rotary desiccant wheel: numerical validation of a variable properties model,
Appl. Therm. Eng. 78 (2015) 640648.
[21] V.N. Bogoslovskij, M.J. Poz, Fundamentals of Physics of Heating, Ventilation
and Air-conditioning Units Design, Strojizdat, Moscow, 1983 (in Russian).
[22] P. Kowalski, Heat and Mass Transfer in Rotary Dehumidier for Solar Air
Conditioning Systems, A Thesis submitted to the Wroclaw University of
Technology for the degrees of Doctor of Philosophy, 2007.
[23] Y.M. Xuana, F. Xiao, X.F. Niu, X. Huang, S.W. Wang, Research and application
of evaporative cooling in China: a review (I)Research, Renew. Sustain.
Energy Rev. 16 (2012) 35353546.
[24] L. Gillan, Maisotsenko cycle for cooling process, Clean Air 9 (2008) 118.
[25] S. Anisimov, D. Pandelidis, A. Jedlikowski, V. Polushkin, Performance
investigation of a M (Maisotsenko)-cycle cross-ow heat exchanger used for
indirect evaporative cooling, Energy 76 (2014) 593606.
[26] J. Ferencowicz, Ventilation and Air Conditioning, Arkady, Warsaw, 1980 (in
Polish).