Sunteți pe pagina 1din 15

International Journal of Heat and Mass Transfer 81 (2015) 207–221

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer


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

Performance study of the Maisotsenko Cycle heat exchangers in different


air-conditioning applications
Demis Pandelidis a,⇑, Sergey Anisimov a, William M. Worek b
a
Department of Environmental Engineering, Wroclaw University of Technology, 27 Wyspiański st., 50-370 Wroclaw, Poland
b
Stony Brook University, College of Engineering and Applied Sciences, Department of Mechanical Engineering, 127 Engineering Bldg., Stony Brook, NY 11794, USA

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

Article history: This paper investigates a mathematical simulation of the heat and mass transfer in the two different
Received 12 August 2014 Maisotsenko Cycle (M-Cycle) heat and mass exchangers used for the indirect evaporative cooling in
Received in revised form 13 October 2014 different air-conditioning systems. A two-dimensional heat and mass transfer model is developed to per-
Accepted 13 October 2014
form the thermal calculations of the indirect evaporative cooling process, thus quantifying the overall
heat exchangers’ performance. The mathematical model was validated against the experimental data.
Numerical simulations reveal many unique features of the considered units, enabling an accurate predic-
Keywords:
tion of their performance. Results of the model allow for comparison of the two types of heat exchangers
Maisotsenko Cycle
M-Cycle
in different applications for air conditioning systems in order to obtain optimal efficiency.
Indirect evaporative cooling Ó 2014 Elsevier Ltd. All rights reserved.
Mathematical model
Cross-flow recuperator

1. Introduction humid. Therefore, new methods and technologies are needed for
cooling of buildings.
In recent years, the increase in summer temperatures, improved One of the best solutions to this limitation is the sub-wet bulb
insulation of the buildings, and a growth of indoor facilities have temperature evaporative cooling. There are several studies on
led to an increased requirement for air conditioning in buildings. achieving sub-wet bulb temperatures by the evaporative cooling
Conventional mechanical vapor–compression air-conditioning and some innovative ideas do exist. Stoichkov [1] presented a
systems consume a large amount of the electrical energy that is mathematical model describing a cross-flow heat exchanger with
largely dependent upon a fossil fuel. This mode of air conditioning a flowing down water film. In this study the wet-bulb temperature
is, therefore, neither sustainable nor environmentally-friendly. Due of an ambient air was not reached at the exit. Ren and Yang [2]
to the increasing need for air conditioning and the growing interest developed an analytical model for the coupled heat and mass
in energy savings, seeking ways to reduce fossil fuel consumption transfer processes in an indirect evaporative cooling with paral-
and to increase usage of the renewable energy during air- lel/counter-flow configurations. Maclaine-Cross and Banks [3]
conditioning process in building sector is a matter of great referred to that for regenerative evaporative cooling the process
importance. can approach the dew point temperature of the ambient air if
Evaporative air cooling is an alternative to the conventional appropriate mass flows and cooler geometry are chosen. Hasan
vapor–compression systems to meet above mentioned economic, [4,5] numerically simulated various configurations of indirect
environmental, and regulatory challenges. Direct evaporative cool- evaporative coolers. The results showed that the performance of
ing is the process of evaporating liquid water into the surrounding the system could be improved by manipulating the air flow by
air and causing its temperature to decease. A typical direct branching the working air from the product air, which is indirectly
evaporative cooler uses a fan to draw in outside air through a pre-cooled. Zhao et al. [6] presented a numerical study of the coun-
pad-wetting media and circulates the cool air through the building. ter-flow heat and mass exchanger for the dew point evaporative
Theoretically, the ultimate temperature for the direct evaporative cooling purpose. They proposed a range of design conditions and
cooling process is the ambient air wet-bulb temperature, however flow rates to improve the cooler performance. Wang [7] studied
this temperature is not easy reached and the resulting air stream is the effect of the wettability (the surface wettability factor is the
parameter used to estimate the effect of an incomplete wetting)
of aluminum plates to the cooling performance of the indirect
⇑ Corresponding author.
evaporative systems. A dynamic contact analyzer was applied to
E-mail address: demis.pandelidis@pwr.wroc.pl (D. Pandelidis).

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.10.033
0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
208 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

Nomenclature

cp specific heat capacity of moist air [J/(kg K)] Nu Nusselt number [–]
d hydraulic diameter [m] Pr Prandtl number [–]
F surface area [m2] Re Reynolds number [–]
h specific enthalpy of the moist air [kJ/kg] X X ¼ X=l – relative X coordinate [–]
H height [m] Y Y ¼ Y=l – relative Y coordinate [–]
G moist air mass flow rate [kg/s]
L, l streamwise length of cooler [m] Subscripts
M water vapor mass transfer rate [kg/s] 1 main (primary) air flow
NTU number of transfer units, NTU ¼ aF=ðGcp Þ [–] 2 working (working) air flow in the wet channels (product
q heat flux [W/m2] part of exchanger)
qo latent heat water [kJ/kg] 3 working (working) air flow in the dry channels (pre-cooling
Q rate of heat transfer [W] part of exchanger)
b
Q specific cooling capacity per cubic meter of the heat 4 working (working) air flow in the wet channels (pre-cooling
exchanger’s structure [kW/m3] part of exchanger)
RH relative humidity [%] cond heat transfer by thermal conduction
t temperature [°C] Icond referenced to the first-order boundary conditions
t average temperature [°C] IIcond referenced to the second-order boundary conditions
v air stream velocity [m/s] g water vapor
W heat capacity rate of the fluid [W/K] h referenced to the height of the channel
x humidity ratio [kg/kg] heat heat transfer
X coordinate along primary air flow direction [m] i inlet
Y coordinate along working air flow direction in the wet l latent heat flow
channels [m] mass mass transfer
o output
Special characters p plate surface
a convective heat transfer coefficient [W/(m2 K)] plt channel plate
b mass transfer coefficient [kg/(m2 s)] product referenced to the product section of heat exchanger
d thickness [m] s sensible heat flow
k thermal conductivity [W/(m K)] w water film
e thermal effectiveness [%] work referenced to the working section of heat exchanger
r surface wettability factor, r 2 (0.0–1.0) [–] WB wet-bulb temperature
X air streamwise in the dry channel
Non dimensional coordinates Y air streamwise in the wet channel
Le Lewis factor Le ¼ a=ðbcp Þ [–]  referenced to the elementary plate surface
NTU number of transfer units NTU ¼ aF=ðGcp Þ [–] 0 condition at the air/plate interface temperature

quantitatively measure the advancing and receding contact angles through the dry channel the primary air flow is cooled without
and the water-retention capacity of different aluminum surfaces. increasing the moisture content (process 1i–1o). At the outlet of
Riangvilaikul and Kumar [8] carried out experimental studies on the dry channel a part of the primary flow (working air flow) is
a dew point evaporative cooling exchanger. Their results indicated delivered to the wet channel, where it realizes the evaporative
that the wet-bulb effectiveness achieved by the exchanger was 92– cooling process on the base of direct air–water contact (process
114%. Zhou [9] carried out a study to optimize the design of the 2i–2o in Fig. 1(b)). The remaining portion of the primary flow
water distributor to improve the water distribution uniformity in (product air flow) is delivered to the conditioned space.
the indirect evaporative air coolers. This research outlined several The currently produced HMX has a unique design to maximize
available water distribution modes applicable to these types of the efficiency of the direct and indirect stages of cooling process.
devices. To enhance the cooling performance of the typical evapo- Fig. 1(c) illustrates air-flow arrangement in the HMX produced
rative exchangers, a novel thermodynamic cycle, known as the by the Coolerado Corporation [10].
M-Cycle [10–18], was proposed by Professor Valeriy Maisotsenko The working mechanism of the HMX under the M-Cycle is
as the new approach of making and operating the heat and mass described as follows. Part of the surface on the dry side is used
exchanger (HMX). This was claimed to enable harnessing extra for the primary air flow (Stream 1 in Fig. 1), while the rest is used
amount of energy from the ambient using a dedicated flat plate, for the working air flow (Stream 2 in Fig. 1). The primary and the
cross-flow and perforated heat exchanger. working air streams are guided to flow over the dry side along par-
According to the producer, the maximal water consumption for allel flow channels. The working air stream is delivered to the dry
the unit containing six HMXes is 14 kg/h [13,17]. The water needed channels first, in order to be sensibly precooled before it is split
to provide 1 kW of cooling is only 0.001 kg (for one HMX, at inlet into multiple streams that are directed into the wetted section.
conditions ti = 30 °C and RHi = 40% and primary and working air There are regularly distributed holes in the channel where the
flow rate equal 330 m3/h [17]). working air is retained and each of these allows a certain percent-
In the typical indirect evaporative air coolers (Fig. 1(a)) the age of the air stream to pass through the wet channels. The work-
working air flow is delivered directly to the wet channel, while ing air stream is then gradually delivered to the wet channel
the primary air flow is delivered to the dry channels. The principal (Stream 3 in Fig. 1) as it flows along the dry side, forming an even
idea of the Maisotsenko Cycle is to indirectly pre-cool the working distribution of the air flow over the wetted surface. The pre-cooled
airflow before it is delivered to the wet channel (Fig. 1(b)). Passing working air, delivered to the wet channel, flows over the wet
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 209

Fig. 1. The principal idea of the M-Cycle and the novel heat and mass exchanger which utilizes the cycle: (a) principal operation of the typical indirect air cooler on the
psychrometric chart; (b) principal operation of the air cooler on the base of the M-Cycle on the psychrometric chart; (c) currently produced M-Cycle HMX: 1 – primary air
flow; 2 – working air flow (dry channel); 3 – working air flow (wet channel).

surface arranged at right angles to the dry-side channels, absorbing ambient conditions. Zhan et al. [12] conducted numerical analyses
heat from the working air flow in the dry channels and the primary of the M-Cycle including a cross-flow heat exchanger to obtain
air stream. Owing to its pre-cooling effect, the working air in the thermal performance. A finite-element method was used with
wet side has a lower temperature and therefore, it is able to assim- experimental data for comparison purposes. Miyazaki [15] ana-
ilate more heat from the primary flow. As a result, the wet-bulb lyzed an M-Cycle for an integrated air-cooling system driven by
effectiveness of the novel M-Cycle unit is significantly higher than the solar energy for air conditioning. As shown above, the interest
the effectiveness of the traditional cross-flow exchangers (see Refs. in the M-Cycle is growing and the new cycle has a potential to
[11,12]). From the literature, it can be found that the M-Cycle has replace a significant part of the mechanical compression system
been investigated by several research groups. Gillan [10] presented load. However, no studies have been conducted that focus on the
the study of an applied M-Cycle unit for the commercial purposes. comparative analysis of the different applications of the M-Cycle
Worek et al. [14] showed the possibility of using the M-Cycle HMX heat and mass exchanger in different air-conditioning systems. In
in desiccant-indirect evaporative air-cooling system for various this regard, this study focuses on the comparison of two types of
210 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

the M-Cycle cross-flow HMXs for different applications in air-con- require different operational conditions for the presented HMXs,
ditioning systems. with different temperatures and relative humidities of the primary
The HMXs considered are shown schematically in Fig. 2. The and the working air-flow entering each exchanger. It is essential to
first device (HMX1 – Fig. 2(a)) is currently produced by Coolerado establish the conditions for which it is more reasonable to use one
Corporation, where the working and the primary air flows in the HMX instead of another. In this paper such an evaluation will be
dry channels are flowing parallel to each other. The second unit conducted using a numerical model.
(HMX2 – Fig. 2(b)) is the modification of the first cross-flow
HMX, where entrance to the dry channel for the working and the
product air flow are placed on the opposite sides of the HMX, 2. Methods
which allows for implementation in different configurations in
air-conditioning systems. Both units can be used as an individual In this paper an analytical model based on the modified e–NTU
cooling coils, which are the only source of cooling power in the sys- method is developed to analyze the performance of two M-Cycle
tem. However, in many air conditioning systems, especially in the HMXs. In the e–NTU-model, the air stream in matrix passages is
public buildings and offices, the individual rooms typically use fan considered as a gaseous fluid flow with constant temperature,
coil units to provide individual comfort for the occupants velocity and mass transfer potential (humidity ratio of air flow)
(Fig. 3(a)–(d)). The exhaust air for the systems with the original in the direction normal to the plate surfaces, which equal to bulk
HMX is removed to the outside or it is partly recirculated to the average values [11,16–18]. The e–NTU method was used to
primary air flow (Fig. 3(a)). The modified HMX allows for the sys- describe indirect evaporative air coolers with satisfactory agree-
tem to be implemented in the supply–exhaust flow stream in the ment with the experimental results presented by Hasan [5] and
air-handling unit (Fig. 3(b)). Therefore it can operate on the Miyazaki [15].
exhaust air, which is colder than the ambient air. The system pre- The authors developed the original e–NTU-model describing
sented in Fig. 3(b) can be replaced with a system having a heat coupled heat and mass transfer in considered M-Cycle heat
pump, presented in Fig. 3(c). The operation base in this case is exchangers. Fig. 4 presents the assumed air flow distribution inside
the same as in supply–exhaust system with the fan coil units. In considered devices, with basic information about the inlet condi-
more humid climates, there is a possibility to use the original tions. From the standpoint of the heat and mass transfer processes
HMX with the desiccant wheel (Fig. 3(d)), which dries the air flow occurring in the channels of considered M-Cycle HMXs can be
and increases its temperature [14]. The arrangements presented divided into two main sections [11]. The first one is a typical

Fig. 2. Analyzed HMXs: (a) original cross-flow HMX (HMX1); (b) modified cross-flow HMX (HMX2).

Fig. 3. Analyzed heat exchangers in different applications for air conditioning systems: (a) HMX1 in air conditioning system with fan coils; (b) HMX2 in air conditioning
system with fan coils; (c) HMX2 in air conditioning system with heat pump; (d) HMX1 in air conditioning system with desiccant wheel.
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 211

Fig. 4. Air flow distribution in considered HMXs: (a) HMX 1; (b) HMX 2.

cross-flow indirect evaporative cooler (this portion is responsible  The passage walls are impervious to mass transfer.
for cooling the primary air flow), and the second one, the task of  The temperature of the water film, the sensible heat transfer
which is pre-cooling the working air flow and formation effective coefficient a and the Lewis factor depend on the operating con-
temperature distribution of the working air flow at the entrance ditions [11].
to the wet channels of the first section of the heat exchanger
(Fig. 4). In this regard, depending on the analyzed section, the 2.1. Mathematical model equations
working air stream in both devices will be considered as three dif-
ferent flows 2, 3, 4 (Fig. 4). In the first section, the pre-cooled work- In this section, the equations describing the product part of con-
ing airflow 2 is exchanging heat with the primary airflow 1, like in sidered HMXs will be presented, equations describing the initial
an ordinary cross-flow indirect evaporative cooler but with non- portion of the exchanger are analogous, therefore they will be
uniform initial distribution of the working air flow’s temperature omitted. The only difference in mathematical description of pro-
and humidity ratio at the entrance to the channel 2. The second cesses occurring in considered parts is the fact, that equations
and third portions are responsible for pre-cooling the working describing the initial part additionally consider an algorithm that
air. The above-mentioned portion of the working air stream will describes air streams mixing (which was presented by authors in
be marked by quantitatively equal 2, 3 and 4, where 2 is the part [11]) and must include the variation of the air Stream 3 and 4,
which exchanges heat with product air 1, 3 is pre-cooling air flow due to the continuous mixing of those streams in the wet channel.
which does not contact water film (working air in the dry channel) The equations describing the product part are identical for HMX1
and 4 is pre-cooling air flow which contacts the water film. The and 2, where the difference lies in the different initial conditions.
balance equations for air flow 3 consider only sensible heat trans- According to the above assumptions the following heat and
fer, while set of equations for air flow 4 describes combined heat mass balance equations can be written for the air streams passing
and mass transfer. Stream 3 and 4 are mixed in the wet channel through control volume in the product part of the HMX (see Fig. 4
(Fig. 4). and 5)
The proposed mathematical model for considered HMXs was
based on the following assumptions:  For the air stream in the dry channel.
o The energy conservation balance considering only sensible
 Heat losses to the surroundings are negligible. heat transfer on the plate surface (see Fig. 5(a) and (b)) is
 Steady state operation. given as:
 Driving force of mass transfer is a gradient of moisture content
@t 1
(vapor’s partial pressure). G_ 1 cp1 dX ¼ dQ_ s1 ð1Þ
@X
 Airflow is an ideal, incompressible gas mixture of dry air and
water vapor. where G_ 1 ¼ ðdY=LY ÞG1 .
 Longitudinal molecular diffusion of water vapor in air and lon-  For the air stream in the wet channel.
gitudinal heat conduction along the wall as well as inside the o The energy conservation balance (taking into account the
fluids in the direction of air flow are negligible. specific heat capacity of the evaporated moisture).
 Kinetic properties of air flow and water film are constant and At the entrance part of the wet passage (see Fig. 5(a) and
equal to bulk average values. (d))
 Consumed water rate corresponds to sufficient evaporation and @t 2
keeping up the material of plates in hygroscopic saturated con- G_ 2 cp2 dY ¼ dQ_ s2 þ cg2 ðt0p2  t 2 ÞdM
_2 ð2Þ
@Y
dition. This causes that air flow heat capacity to be much larger
than that of the water film (i.e., W2  Ww. At the exit part of the wet passage (see Fig. 5(b) and (c))
212 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

Fig. 5. Schematic of heat and mass transfer in differential control volumes in the product part of the investigated HMX: (a) at the entrance part of the wet channel (detail view
for the dry channel); (b) at the exit part of the wet channel (detail view for the dry channel); (c) at the exit part of the wet channel (detail view for the wet channel); (d) at the
entrance part of the wet channel (detail view for the wet channel).

 
@t 2 kplt  
G_ 2 cp2 dY ¼ þdQ_ s2 þ cg2 ðt0p2  t 2 ÞdM
_2 ð3Þ tp1  t 0p2  a1 ðt1  tp1 Þ ð7Þ
@Y dplt
where G_ 2 ¼ ðdX=LX ÞG2 . o The overall energy balance for the filling surface results
o The mass conservation balance for the water vapor inside in.
the wet channel (see Fig. 5(c) and (d)) is given as  At the entrance part of the wet passage (see Fig. 5(a) and (d))
@x2 dQ_ s1 þ dQ_ s2 ¼ dQ_ l2 ð8Þ
G_ 2 _2
dY ¼ dM ð4Þ
@Y
 At the exit section of the wet passage (see Fig. 5(b) and (c))
Additionally the M-Cycle mathematical model was supple- we obtain
mented by energy balance equations developed.
o For the airflow/plate surface interface in the dry passage dQ_ s1 ¼ dQ_ s2 þ dQ_ l2 ð9Þ
(see Fig. 5)
  Rewriting Eqs. (8) and (9) using NTUs gives the following rela-
d Q_ cond ¼ dQ_ s1 ð5Þ tionship regardless of the analyzed part of the wet passage
plt
met    
W1 rp q o
which can be converted taking into account the following set of NTU1 ðtp1  t 1 Þ þ NTU2 ðt0p2  t 2 Þ þ NTU2 ðx0  x2 Þ ¼ 0
W2 cp Le 2 p2
energy balance equation
8   ð10Þ
< dQ_ cond ¼ d Q_ cond
plt plt
met ð6Þ As one can see, the heat transfer mechanism in the wet channel is
: t0  ðt 0 Þ  t more complicated than in the dry channel, since both sensible and
p1 p1 met p1
latent heat are exchanged between the working air flow and the
to the following form water film on the plate surface in the wet channel of the
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 213

investigated HMXs (see Eq. (10)). To complete, the set of simulta- The mathematical models are supplemented with:
neous partial differential equations (i.e., (1)–(4), (7) and (10)) the
boundary conditions, establishing initial thermodynamic parameters  The convection heat transfer coefficients in dry and wet chan-
values of exchanged air streams at the entrance to the appropriate nels were obtained by the Nusselt number with uniform heat
channels of the product part of the heat exchangers (see Fig. 3) are flux for both thermal entrance region and fully developed flow.
needed. The air streams are considered to be laminar flow [21,22], due
 For the product air flow at the entrance to the dry channel (both to the low velocity (Re < 2000). The Nusselt number of air can
HMXes) be classified into two regions: the beginning undeveloped
 
t 1  ¼ t1i x1  ¼ x1i ¼ const region (l 6 lb) and the fully-developed region (l > lb). The length
 
X X
  ¼ 0:0   ¼ ð0:0  1:0Þ of the first undeveloped zone can be calculated from Eq. (16):
 ;  ð11Þ
 Y ¼ ðlwork  1:0Þ  Y ¼ ðlwork  1:0Þ
 Y  Y
lb =dh ¼ 0:05RePr ð16Þ
 
For the first zone, the Nusselt number is obtained by using empiri-
 For the working air flow at the entrance to the dry channel cal Eqs. (17) or (18), depending on the boundary conditions (when
(HMX1)
  channel height is considered as channel’s specific dimension)
t 3  ¼ t3i x3  ¼ x3i ¼ const
  [21,22]:
X X
  ¼ 0:0   ¼ ð0:0  1:0Þ
 ;  ð12Þ NuIcond ¼ 1:533ðReh Prðh=LÞÞ
1=3
ð17Þ
 Y ¼ ð0:0  lwork Þ  Y ¼ ð0:0  lwork Þ h
 Y  Y
  NuIIcond ¼ 1:755ðReh Prðh=LÞÞ
1=3
ð18Þ
h

 For the air flow at the entrance to the dry channel (HMX2) For the fully-developed air flow, the Nusselt number is constant. Its

t 3  ¼ t3i x3  ¼ x3i ¼ const value depends on temperature difference profiles between air and
 
X X
  ¼ 1:0   ¼ ð0:0  1:0Þ heat exchange surface (different boundary conditions) and on shape
 ;  ð13Þ of the channel [21,22]. For a rectangular channel:
 Y ¼ ð0:0  lwork Þ  Y ¼ ð0:0  lwork Þ
 Y  Y
  NuIcond ¼ 3:77 ð19Þ
h

 For the working air flow at the entrance to the wet channel in NuIIcond
h ¼ 4:12 ð20Þ
product part (both HMXs)

   
t 2i  ¼ t 4o  x2i  ¼ x4o 
   
X X X X
  ¼ ð0:0  1:0Þ   ¼ ð0:0  1:0Þ   ¼ ð0:0  1:0Þ   ¼ ð0:0  1:0Þ
  work   work ;   work   work ð14Þ
 Y ¼ lY  Y ¼ lY  Y ¼ lY  Y ¼ lY
   
   

 For the working air flow at the entrance to the wet channel in Mass transfer coefficient is designated from well-known Lewis
initial part (both HMXs) relationship [11]:

   
t 4i  ¼ t3  x4i  ¼ x3  ¼ x3i ¼ const
   
X X X 
  ¼ ð0:0  1:0Þ   ¼ ð0:0  1:0Þ 
 ¼ ð0:0  1:0Þ 
 X ¼ ð0:0  1:0Þ
  ;   ð15Þ
 Y ¼ 0:0  Y ¼ 0:0  Y ¼ 0:0  Y ¼ 0:0
   
   

The proposed set of governing equations describing the heat and b ¼ a=ðcp LeÞ ð21Þ
mass transfer process in the working part of the units was addition-
where Lewis factor is considered as the quotient of the Stanton
ally supplemented by algebraic equations, describing air streams
number for heat transfer, to Stanton number for mass transfer
mixing in the wet channels [11,18]. Furthermore, during integration
(Sherwood number) [11]:
of the differential equations, the variable values of the air mass flow
rates and the numbers of transfer units in the dry and wet channels Le ¼ Stheat =Stmass  Nuheat =Numass ð22Þ
(NTU3, NTU4) due to permanent mixing of the air from the dry chan-
Additionally, the e–NTU–models developed are supplemented
nel with the air in the wet channel are taken into account. A computer
by non-linear empirical relationship between the saturation
simulation of the system using a modified Runge–Kutta method,
pressure of water vapor and it is temperature [11].
which has sufficient accuracy and stability based upon experience
with other similar types of problems was used [11,16–20]. The accu-
3. Mathematical model validation (HMX1)
racy of the proposed numerical method was checked by varying the
space domain grid size. Simulations were conducted with decreasing
The detailed description of the experiment and validation of the
grid size, until the values of the local air flow and plate temperatures
model including error analysis with different methods is presented
were unchanged. A convergence test showed that the optimum num-
by authors in Ref. [16]. In this section only short analysis will be
ber of grid nodes and, consequently, the minimum time of computa-
presented.
tion within the given tolerance of accuracy was reached at the step
The model describing the original HMX (HMX1) was validated
size of 0.01 in the direction of X and Y directions.
using test data provided by Coolerado Corporation in Denver, Col-
214 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

orado, USA. Validation was based on comparing the outlet param- Inlet and outlet dry and wet bulb temperatures were measured
eters for the wet and the dry channels. This analysis established with the VWR Traceable thermometers having an accuracy of
the accuracy of the model in predicting the performance of the real, 0.5 °C, WB and DB (i.e., shown in Fig. 6(a)–(c)) were used. Ther-
original indirect evaporative air cooler with the M-Cycle. mometers were pre-calibrated and verified before the tests with
The exchanger analyzed was placed inside a test chamber (see HMP45C sensors (manufacturer: Campbell Scientific) with 0.2 °C
Fig. 6(a)). Inlet air parameters were established by the units placed measure accuracy. The total air flow rate was varied from 150 to
outside the chamber: supply fan (F2 in Fig. 6(a)), an electrical hea- 800 m3/h, the inlet temperature varied from 18 to 45 °C, while
ter (H) and a vapor air humidifier (VH). Inside the chamber, HMX is the inlet relative humidity varied from 20% to 70%.
placed on a working bench. Air is delivered by the main fan (F1 in Modeling results compared with experimental data are pre-
Fig. 6(a) and (b)); the air is filtered before it enters the HMX (AF). sented in Fig. 7(a)–(f). Fig. 7(a) and (b) show product air outlet
Supply and working air flow rates are controlled by the support temperatures obtained by the model and the experiment for vari-
fans (SF, EF) and a set of adjustable dampers on the product and able inlet temperature and velocity, while Fig. 7(c)–(f) show the
the exhaust side of the HMX (SI and EI). The air volume flow rate correlation between the model and the experiment for predicting
is measured using the pressure drop on orifice plate manometers outlet product and working air temperature, working air humidity
(Ma in Fig. 6(a)). Additionally, the inlet air humidity can be ratio and wet bulb effectiveness. Fig. 7(a) shows comparison
increased by returning the exhaust air flow to the chamber, by between the model and the experimental data for constant air flow
switching the damper (D in Fig. 6(a)). Water is delivered to the velocity (constant air flow rate) and variable inlet temperature.
considered HMX directly from the duct (WS in Fig. 6(b)). Comparison was presented for three hypothetical conditions.
All studies were performed under steady-state operational con- These are: dry (x1i = 11.2 g/kg), humid (x1i = 16.5 g/kg) and very
ditions. Each experiment was repeated to ensure consistency and humid (x1i = 25.0 g/kg). Deviation of the primary air stream outlet
repeatability of the measured data. The numerical model was set temperature is at most 3% (i.e., the highest differences in outlet
to the same operating conditions as for the experimental cases: temperatures are up to 0.5 °C). Fig. 7(b) shows the comparison
the same inlet air temperatures and relative humidities and the for constant inlet parameters and variable air flow velocity. The
same air flow rates. The parameters measured included: deviation of the outlet temperature was up to 3 °C. The correlation
between the experimental and the simulation data is equal 0.996
 Dry and wet bulb temperatures of the air flow entering and for predicting outlet product air temperature (Fig. 7(c)). For the
leaving the exchanger (in the dry and wet channels). working air flow, the correlation for output temperature is equal
 Pressure drop in the orifice plates for air volume flow rate 0.998 (Fig. 7(d)). For the working air stream output humidity ratio,
definition. the correlation is equal to 0.997. The maximum discrepancies are

Fig. 6. Testing bench at Coolerado Corporation facilities in Denver, CO, USA: (a) general scheme of the measurement station; (b) visualization of the testing bench inside the
chamber; (c) photo of testing bench.
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 215

Fig. 7. M-Cycle mathematical model validation: (a) average output primary air stream temperature t 1o vs. inlet air temperature t1i under constant air stream velocity
conditions; (b) average output primary air stream temperature t1o vs. air stream velocity v1 under constant inlet air temperature and humidity conditions; (c) comparison
between model and experiment: average output temperature of the primary flow t1o ; (d) comparison between model and experiment: average output temperature of the
working flow t2o ; (e) comparison between model and experiment: average output humidity ratio of the working flow  x2o ; (f) comparison between model and experiment: wet
bulb effectiveness.
216 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

equal 0.3 g/kg (Fig. 7(e)). The correlation for the wet-bulb 4.1. General comparison between considered exchangers
efficiency is equal 0.979 (Fig. 7(f)). The analysis on the basis of
exact differential method indicated that the overall uncertainty Simulations were performed in order to compare the effective-
for the wet bulb effectiveness was within ±4.6% [16]. ness of the considered exchangers for the same inlet conditions (air
Both the experimental and simulation results showed good with the same temperature and relative humidity entering the
agreement. Any differences between the experiment and simu- working and the product channels). The results are presented in
lated results would be due to heat losses, leakage of the air and Fig. 8(a). It can be seen, that for the same inlet conditions HMX1
non-uniform air flow distribution in the channels [13]. The accu- always obtains a lower outlet temperature than HMX2. The specific
racy of the mathematical model prediction is considered to be sim- cooling capacity for the selected inlet air parameters for both units
ilar for HMX2 (see Fig. 4), since the numerical model describing the are presented in Table 2, while the wet-bulb effectiveness is pre-
heat and mass transfer in the channels of HMX2 are analogous to sented in Table 3. It can be assumed that the efficiency of the
ones describing HMX1. HMX2 is significantly lower than the efficiency of the original unit
(HMX1). The differences in output temperatures obtained between
4. Results and discussion the two devices can be up to 2.5 °C, the differences in the specific
cooling capacity are up to 7.5 kW/m3, while the differences in the
The reference operating conditions for the analyzed exchangers wet-bulb effectiveness can be up to 32%.
are presented in Table 1. The considered exchangers will be com- The significant differences in the efficiency between the two
pared for the three hypothetical inlet conditions, representing dif- exchangers seems surprising, since their construction is almost
ferent operational possibilities for the air conditioning system: similar, the only difference is the different air flow direction inside
the dry channel in the initial portion of the exchanger. However,
1. Both exchangers have the same inlet parameters for the work- this small construction difference results in a completely different
ing and the product air channels: t1i = t3i; RH1i = RH3i; (the pur- temperature distribution inside the dry product air channel of the
pose of this study is the general comparison between the two two units, which has a significant impact on their cooling perfor-
exchangers for the same operational conditions). mance. The temperature distribution inside the product part of
2. For HMX1 the working and the product air streams’ inlet tem- the two units is presented in Fig. 8(b) and (c), while the heat flux
perature and humidity are the same, while HMX2 have different distribution inside their primary air channels is presented in
values of the primary and the working air inlet temperature and Fig. 8(d). It can be seen, that HMX1 and HMX2 show different
humidity (comparison between the HMX1 working in the sup- trends in temperature distribution of the product air flow, working
ply air system presented in Fig. 3(a) and HMX2 working in the air flow and the plate surface. In HMX1, the temperature difference
supply–exhaust air system shown in Fig. 3(b) and (c)). In this between the plate surface and the primary air stream is almost
case, the inlet temperature of the working air flow is lower than constant (Fig. 8(b)). At the same time temperature profiles inside
the inlet temperature of the product air flow for HMX2. HMX2 show different trends. The temperature difference
3. HMX1 has hot and dry inlet air conditions (the same for the (t1  t0p1 ) is the highest at the beginning of the dry channel and it
product and the working air flow), while HMX2 has different is the lowest at the final part of the dry channel. The different tem-
values for the primary and the working air inlet temperature perature distributions of the primary air, working air and the plate
and humidity (comparison between HMX1 working in the des- surface are caused by the different arrangement of the initial por-
iccant system presented in Fig. 3(d) and HMX2 working in the tion of the exchangers (see Fig. 4(a) and (b)). The initial part of
supply–exhaust air system visible in Fig. 3(b) and (c)). HMX1 pre-cools the working air flow before entering the product
portion. As a result, the working air temperature is the highest at
Three main parameters (indices) have been selected to study the beginning and lowest at the final portion of the dry channel.
the operational performance of the investigated HMX Temperature distribution of the plate surface in the product chan-
configurations: nels is then similar to the temperature distribution in the counter-
flow exchangers [22]. The temperature difference between the
 Temperature of outlet primary air flow t 1o . plate surface and the primary air stream is almost constant
 The wet-bulb thermal effectiveness, which is defined as the through the entire channel. That is why the effectiveness of the ori-
ratio of the difference between intake and outlet process air ginal M-Cycle cross-flow HMX (HMX1) is significantly higher than
temperature to the difference between intake process air tem- the efficiency of the typical indirect evaporative air coolers. The
perature and its wet bulb temperature [5,23]; initial portion of the modified cross-flow unit (HMX2) creates a
temperature distribution, which is similar to the temperature dis-
t 1i  t 1o tribution inside the parallel-flow exchangers (Fig. 8(c)). This results
eWB ¼ ð23Þ
t 1i  tWB
1i
in the highest heat transfer at the beginning of the dry channel and
the lowest at the end of the dry channel of HMX2 (Fig. 8(d)). The
 The specific cooling capacity per cubic meter of the HMX heat flux inside the dry channel (Fig. 8(d)) of HMX1 is relatively
structure constant through the whole dry channel. In HMX2, however, over
half of the total heat flux transfer occurs in the first 30% of the
^ ¼ Q =V HMX
Q ð24Þ dry channel length (Fig. 8(d)). That is why, for the same inlet air
1
temperatures of the primary and the working air streams, the ori-
where Q 1 ¼ G1 cp1 ðt 1i  t1o Þ is the cooling capacity of the HMX and
ginal unit is characterized by higher efficiency.
VHMX is the volume of the HMX structure, m3.

Table 1
The reference operating conditions for the analyzed exchangers.

Length, m Width, m Channel height, mm Working to primary air ratio, – Primary air stream velocity, m/s
0.5 0.5 3.0 1.0 3.0
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 217

Fig. 8. Numerical simulation results for the same inlet temperature and relative humidity for HMX1 and HMX2: (a) outlet temperatures; (b) temperature distribution inside
the HMX1; (c) temperature distribution inside the HMX2; (d) heat flux distribution inside the primary air channel (both HMXes).

Table 2
Specific cooling capacity of HMX1 and HMX2 for selected inlet conditions. operational conditions in the air conditioning systems. Fig. 9(a)
presents comparison of systems visible in Fig. 3(a)–(c). Here
Inlet parameters b , kW/m3
HMX1: Q b , kW/m3
HMX2: Q HMX1 is using only on ambient air, while in HMX2 the ambient
t1i = 30 °C and RHi = 30% 26.6 21.6 air enters only the product air channels and the working air chan-
t1i = 30 °C and RHi = 40% 22.1 18.1 nels are operating on the exhaust air from the conditioned space.
t1i = 30 °C and RHi = 50% 17.9 14.7
Depending on the type of conditioned space, the exhaust air may
have a different temperature and relative humidity; however the
temperature inside the conditioned space is colder than the ambi-
Table 3 ent temperature, while its relative humidity is usually a little
Wet-bulb effectiveness of HMX1 and HMX2 for selected inlet conditions. higher. Simulations were conducted for an ambient air tempera-
Inlet parameters HMX1: eWB, % HMX2: eWB, % ture equal 30 °C, and the relative humidity of the ambient air
t1i = 30 °C and RHi = 30% 119 97 was varied from 25% to 50%. The exhaust air temperature was
t1i = 30 °C and RHi = 40% 126 103 changed from 21 to 26 °C, with 1 °C increments, while its relative
t1i = 30 °C and RHi = 50% 135 110 humidity was changed from 40% to 60%.
The charts (Fig. 9(a), (c) and (e)) are presented for constant lev-
els of inlet temperatures of the primary air flow (t1i), which are
4.2. Analysis at different operational conditions listed in frames in the upper part of the charts. For HMX1 the inlet
temperatures represent different operation in air conditioning sys-
Another set of simulations was performed in order to compare tem (operation only on ambient air or in a system with a desiccant
the effectiveness of the considered exchangers in the different wheel. HMX2 operates on the ambient air flow in the primary air
218 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

Fig. 9. Simulation results for HMXes operating in different inlet conditions: (a) outlet temperatures HMX1: t1i = 30 °C; RHi = 25–50%; HMX2: t1i = 30 °C; t3i = 21–26 °C;
RHi = 40–60%; (b) heat flux distribution inside the primary air channels; (c) outlet temperatures: HMX1: t1i = 32.5 °C; RHi = 20–50%; HMX2: t1i = 30 °C; t3i = 21–26 °C;
RHi = 40–60%; (d) outlet temperatures: HMX1: t1i = 35 °C; RHi = 20–50%; HMX2: t1i = 30 °C; t3i = 21–26 °C; RHi = 40–60%.

channels and exhaust air from the conditioned spaces in the work- the ambient air. This phenomenon can be explained by analyzing
ing air channels. the heat flux profiles (Fig. 9(b)). Although the heat flux profile for
The X axis in Figs. 9(a), (c) and (d) represents the temperature of the HMX2 operating on the exhaust air is still similar to the profiles
the exhaust air, which enters the channels of the HMX2. It is impor- characteristic for the parallel-flow exchangers, the heat transfer
tant to mention, that the parameters of the exhaust air do not affect rate is higher than the heat transfer rate of the same unit using
the outlet air temperatures obtained by HMX1 (for this unit t1i = t2i), ambient air. The higher value of the heat transfer rate, caused by
therefore values of t1o for this unit are presented as a constant lines. the lower temperature of the working air, results in more efficient
Blue lines represent the outlet air temperatures obtained by the cooling of the primary air. As a result, HMX2 can achieve higher
HMX1 for constant inlet air temperature (t1i = t2i) and different inlet effectiveness than the HMX1 operating on outside air in moderate
relative humidities of the working air flow (listed in the frame on climates (Fig. 9(a)). However, in more dry climatic conditions
the right side of each chart). The black lines represent the outlet (RHi = 35–40%) HMX1 achieves higher efficiency than HMX2 using
temperatures obtained by the HMX2 for constant level of inlet pri- exhaust air (for most of the typical indoor conditions). It can be seen
mary air temperature (t1i), variable inlet working air temperature that the detail analysis of the ambient and indoor air parameters is
(presented on the X axis) and variable inlet working air relative essential to achieve the highest efficiency of the designed air-condi-
humidity (listed in frame on the right side of each chart). tioning system.
It can be seen, that the inlet parameters have significant impact The next set of simulations was performed in order to compare
on the efficiency of considered systems. For moderate climate con- the HMX1 operating in system with the desiccant wheel (see
ditions (RHi = 45–50%) outlet temperatures obtained by HMX1 Fig. 3(d)) with HMX2 operating on the exhaust air (see Fig. 3(b)
were equal 21.8 and 20.8 °C. At the same time, for the typical indoor and (c)). The desiccant wheel dehumidifies the ambient air and it
conditions (t = 24–26 °C; RHi = 50–60%) HMX2 achieved outlet additionally increases its temperature. Depending on the material
temperatures varying from 19.0 °C (t3i = 24 °C and RH3i = 50%) to used for the wheel structure and the range of the dehumidification
21.75 °C (t3i = 26 °C and RH3i = 60%). This shows that HMX2 operat- process, the air temperature may increase or decrease depending
ing on the exhaust air stream at the entrance to the wet channel has on the conditions. For HMX2, the inlet temperature of the primary
more effective performance than the HMX1, which operates only on stream was assumed to be 30 °C, while for HMX1 two types of the
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 219

inlet conditions were selected for this study: dehumidification Table 5


with a small increase in inlet temperature (t1i = t3i = 32.5 °C; RHi = Selected boundary temperature and relative humidity values for
the exhaust air entering the working air channels in HMX2
varied) and dehumidification with a high increase in inlet temper- allowing overcoming the performance of HMX1.
ature (t1i = t3i = 35.0 °C; RHi = varied). The results of the comparison
are shown in Fig. 9(c) and (d), respectively. It can be observed, that HMX1; t1i = 30 °C HMX2; t1i = 30 °C

for the inlet temperature equal to 32.5 °C and RHi 6 25%, HMX1 RHi = 30% RH3i = 40%: t3i < 23 °C
achieves lower outlet temperatures than HMX2 operating on the RH3i = 50%: t3i < 22.2 °C
RHi = 40% RH3i = 50%: t3i < 25.3 °C
exhaust air with the typical indoor conditions (Fig. 9(c)). RH3i = 60%: t3i < 23.5 °C
When the temperature of the inlet air is raised to the higher RHi = 45% RH3i = 50%: t3i < 25.8 °C
level after dehumidification, the HMX1 obtains higher outlet tem- RH3i = 60%: t3i < 25.0 °C
peratures (Fig. 9(d)). In this case, for the inlet temperature equal to HMX1; t1i = 32.5 °C HMX2; t1i = 30 °C
35.0 °C and RHi = 25%, HMX1 achieves higher outlet temperatures RHi = 20% RH3i = 40%: t3i < 22.9 °C
than HMX2 operating using exhaust air with typical indoor condi- RH3i = 45%: t3i < 22.0 °C
RHi = 25% RH3i = 40%: t3i < 24.8 °C
tions. However, for relative humidities equal to 20%, HMX1 shows
RH3i = 50%: t3i < 23.0 °C
better performance than HMX2 for the most of the typical indoor RHi = 30% RH3i = 45%: t3i < 25.6 °C
conditions (Fig. 9(d)). This shows that the inlet air stream for RH3i = 50%: t3i < 24.8 °C
HMX1 has to be dehumidified to very low values of the relative HMX1; t1i = 35.0 °C HMX2; t1i = 30 °C
humidity to overcome the performance of HMX2 operating on RHi = 20% RH3i = 40%: t3i < 24 °C
the exhaust air. RH3i = 50%: t3i < 23.2 °C
Table 4 presents the outlet temperature, specific cooling capac- RHi = 25% RH3i = 40%: t3i < 26.3 °C
RH3i = 50%: t3i < 24.4 °C
ity and the wet bulb effectiveness obtained by HMX1 and HMX2 RHi = 30% RH3i = 50%: t3i < 26.4 °C
for the selected inlet parameters. The analysis of Table 4 illustrates RH3i = 60%: t3i < 24.5 °C
another interesting observation: HMX1 for the inlet conditions of
t1i = 32.5 °C; RHi = 20% obtained lower outlet temperature than
HMX2 (for t1i = 30 °C; t3i = 24 °C and RH3i = 40%), however its seen on the basis of Table 5 and Fig. 9, that in many cases the
wet-bulb effectiveness is 32.5% lower. HMX2 (for t1i = 30 °C; HMX2 can show a better performance than HMX1, especially for
t3i = 24 °C and RH3i = 40%) obtained a lower outlet temperature the colder indoor parameters. This leads to a conclusion, that ‘‘free’’
than HMX1 (for t1i = 35.0 °C and RHi = 20%), however its specific energy available in cold exhaust air should not be wasted by
cooling capacity is 4.4 kW/m3 lower. This shows that the wet-bulb extracting this air to the atmosphere. It also seems reasonable in
effectiveness and the specific cooling capacity are not adequate to some cases to use heat recovery exchangers (or a partially recircu-
describe the performance of the HMXs operating in different lated exhaust air stream) in systems with HMX1 instead of using a
arrangements in the air-conditioning system. The higher specific desiccant wheel. To analyze this aspect, another set of simulations
cooling capacity for the air with higher inlet temperature is caused were performed.
by the higher difference between inlet and outlet product air tem- In this analysis the two types of the heat recovery methods were
peratures ðt 1i  t1o Þ. Greater temperature difference ðt 1i  t1o Þ, for assumed to be placed before HMX1. First a recirculation with
the higher temperature of the inlet air flow causes a very intensive efficiency equal 50% (recovery of the sensible and latent heat:
evaporation process in the wet channels, which improves the effi- Fig. 10(a)) and a cross-flow recuperator with the temperature effec-
ciency of cooling process and results in greater cooling capacity Q ^. tiveness equal to 50% (recovery of the sensible heat only: Fig. 10(b)).
However, the additional cooling capacity can be used only for the The ambient air parameters assumed are an inlet temperature of
reduction of the primary air flow temperature to level before the 30 °C and an inlet relative humidity equal to 45%. The exhaust air
air stream is heated in dehumidification process and thus it does temperature varies from 21 to 26 °C, while its relative humidity
not give any energy benefit. varies from 45% to 55%. The outlet temperatures obtained by
An explanation of the trends in the wet-bulb effectiveness pre- HMX1 with the heat recovery units was compared with outlet
sented in Table 4 is that the wet-bulb effectiveness refers to the temperatures obtained by the same unit in the system with the
wet-bulb temperature of the product air. In case of HMX2, the air desiccant wheel (inlet parameters of t1i = 35.0 °C; RHi = 20–30%).
stream entering the wet channel has a different temperature, rela- The results of the comparison are shown in Fig. 10(c).
tive humidity and therefore different wet-bulb temperature. In this Blue lines in Fig. 10(c) represent the outlet temperatures
case, the efficiency factor based on the primary air’s wet-bulb tem- obtained by the HMX1 in the system with the desiccant wheel
perature is not informative and can lead to an incorrect conclu- for constant level of inlet air-flow temperature (t1i = t2i) and
sions. Therefore it should not be used for the description of different inlet working air-flow relative humidities (listed in the
indirect evaporative exchangers with different primary and work- frame on the right side of each chart). The parameters of the
ing air inlet parameters. exhaust air do not affect the operation of the exchanger in the sys-
The boundary values of the exhaust air parameters at the tem with the desiccant wheel, therefore the results for this system
entrance to the working air channel of HMX2, which allow obtain- are presented as a constant values.
ing higher efficiency than HMX1 are presented in Table 5. It can be The X axis in Fig. 10(c) represents the temperature of the
exhaust air, which enters either heat recovery exchanger or a recir-
Table 4 culation chamber (where it exchanges sensible heat or it is mixed
Outlet temperature, specific cooling capacity and wet bulb effectiveness of HMX1 and
with the ambient air with parameters listed above). The black lines
HMX2.
represent the outlet air-flow temperatures obtained by the HMX1
HMX1: HMX1: HMX2: HMX2: operating with heat recovery units or recirculation system for
t1i = 32.5 °C; t1i = 35.0 °C; t1i = 30 °C; t1i = 30 °C;
constant inlet primary air temperature (t1i), variable exhaust
RHi = 20%; RHi = 20%; t3i = 24 °C; t3i = 24 °C;
RH3i = 40%; RHi = 50%; air temperature (presented on the X axis) and exhaust air
t1o , °C
relative humidity (listed in frame on the right side of each chart).
17.1 18.5 18.17 19.35
b , kW/m3 33.4 35.5 31.1 28.0
The relative humidity of the exhaust air does not affect the heat
Q
eWB, % 114.5 115.0 147.0 132.6
transfer process in heat recovery recuperator, since it is only used
to exchange sensible heat.
220 D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221

Fig. 10. Heat recovery units before the HMX1: (a) system with cross-flow recuperator; (b) system with recirculation; (c) simulation results for the HMX1 operating with heat
recovery unit, recirculation and desiccant wheel.

It can be seen, that performance of the HMX1 operating on air quality, which in some cases makes this solution impossible
partly recirculated air is very close to the performance of the same to apply in air-conditioning systems with higher hygienic require-
unit in the system with the desiccant wheel. At the same time ments. The results obtained in this study show the effectiveness of
HMX1 working with the cross-flow recuperator obtained a rela- the indirect evaporative air-conditioning system strongly depends
tively high outlet temperature (higher by 0.5–2.1 °C than the same on many factors, therefore, several possibilities of operational
exchanger with the recirculation), showing that the recovery of options for the indirect evaporative HMXs should be carefully
latent heat is essential for achieving the highest effectiveness of examined in order to achieve the highest energy efficiency. In sys-
the indirect evaporative exchangers with the M-Cycle. The main tems where the air stream is additionally cooled by a mechanical
disadvantage of the recirculation is the reduction in the supply compression system, a recirculation or a supply–exhaust system
D. Pandelidis et al. / International Journal of Heat and Mass Transfer 81 (2015) 207–221 221

with the HMX2 may be considered. In very humid climates, it may Conflict of interest
be reasonable to use system with the desiccant wheel. In more dry
climates, for typical indoor conditions, a system with HMX1 work- None declared.
ing as the main source for cooling can be applied.
References

5. Summary and conclusions [1] N.J. Stoichkov, G.I. Dimitrov, Effectiveness of cross-flow plate heat exchanger
for indirect evaporative cooling, Int. J. Refrig. 21 (1998) 463–471.
[2] Ren Chengqin, Yang Hongxing, An analytical model for the heat and mass
This study presents the theoretical analyses of the M-Cycle transfer processes in indirect evaporative cooling with parallel/counter flow
cross-flow heat exchangers used for the indirect evaporative cool- configurations, Int. J. Heat Mass Transfer 49 (2006) 617–627.
ing in air-conditioning systems. The units were compared in differ- [3] I.L. Maclaine-Cross, P.J. Banks, A general theory of wet surface heat exchangers
and its application to regenerative evaporative cooling, J. Heat Transfer 103
ent operational conditions, representing potential different (1981) 579–585.
applications in air-conditioning systems. [4] A. Hasan, Indirect evaporative cooling of air to a sub-wet bulb temperature,
Modified e–NTU-model based on the accurate assumptions was Appl. Therm. Eng. 30 (2010) 2460–2468.
[5] A. Hasan, Going below the wet-bulb temperature by indirect evaporative
proposed for the mathematical analysis of the heat and mass trans-
cooling: analysis using a modified e–NTU method, Appl. Energy 89 (2012)
fer processes inside the exchangers. The model was validated 237–245.
against experimental data. [6] X. Zhao, J. Li, S.B. Riffat, Numerical study of a novel counter-flow heat and mass
exchanger for dew point evaporative cooling, Appl. Therm. Eng. 28 (2008)
Using the validated model, the operational parameters of the
1942–1951.
HMXs were varied in order to compare their cooling performance [7] T. Wang, R.L. Reid, Surface wettability effect on an indirect evaporative cooling
in the different types of air-conditioning systems. The basic operat- system, ASHRAE Trans. 102 (Part1) (1996) 427–433.
ing parameters were given in Table 1. The discrepancies in effec- [8] B. Riangvilaikul, S. Kumar, An experimental study of a novel dew point
evaporative cooling system, Energy Build. 42 (2010) 637–644.
tiveness characterizing the considered HMXes were explained on [9] B. Zhou, X. Huang, Y. Di, Effects of water sprays in indirect evaporative coolers
the basis of the unique features of heat and mass transfer phenom- on heat and mass transfer, Build. Energy Environ. (2003).
enon occurring in each of the units. It was established that: [10] L. Gillan, Maisotsenko cycle for cooling process, Clean Air 9 (2008) 1–18.
[11] S. Anisimov, D. Pandelidis, Numerical study of the Maisotsenko cycle heat and
mass exchanger, Int. J. Heat Mass Transfer 2014 (75) (2014) 75–96.
 The energy efficiency of the considered HMXs strongly depends [12] C. Zhan, X. Zhao, S. Smith, S.B. Riffat, Numerical study of a M-Cycle cross-flow
on the inlet air parameters. heat exchanger for indirect evaporative cooling, Build. Environ. 46 (2011) 657–
668.
 HMX1 shows higher efficiency than the HMX2, when both units [13] D. Zube, L. Gillan, Evaluating Coolerado Corportion’s heat-mass
are operating using the same inlet parameters, due to the more exchanger performance through experimental analysis, Int. J. Energy
efficient temperature distribution inside its product portion of Clean Environ. 12 (2-4) (2011) 101–116, http://dx.doi.org/10.1615/
InterJEnerCleanEnv.2012005839. Special Issue on M-Cycle Fundamentals and
HMX1. Applications.
 HMX2 operating on the exhaust air with the typical indoor [14] W.M. Worek, M. Khinkis, D. Kalensky, V. Maisotsenko, Integrated desiccant-
parameters in the working air channels shows higher effective- indirect evaporative cooling system utilizing the Maisotsenko cycle, in:
HT2012, Proceedings of the ASME 2012 Summer Heat Transfer Conference,
ness than HMX1 operating on ambient air in moderate climate
HT2012-58039.
conditions. [15] T. Miyazaki, I. Nikai, A. Akisawa, Simulation analysis of open cycle
 HMX1 operating in the system with the desiccant wheel shows adsorption air conditioning system – numeral modeling of a fixed bed
higher efficiency than HMX2 operating on indoor air, when air dehumidification unit and the Maisotsenko cycle cooling unit, Int. J. Energy
Clean Environ. 12 (2011) (2012) 341–354, http://dx.doi.org/10.1615/
stream is highly dehumidified (RHi = 25% and lower). InterJEnerCleanEnv.2012005977.
 The material used for the desiccant wheel construction should [16] S. Anisimov, D. Pandelidis, A. Jedlikowski, V. Polushkin, Performance
allow for intense dehumidification of the air stream and a rela- investigation of a M (Maisotsenko)-cycle cross-flow heat exchanger used for
indirect evaporative cooling, Energy 76 (2014) 593–606.
tively small temperature increase in order to obtain the lowest [17] S. Anisimov, D. Pandelidis, The analysis of the water consumption by indirect
outlet temperatures. evaporative air coolers, Instal (11) (2013) 50–55. in Polish.
 In systems, where the air stream is additionally cooled by a [18] S. Anisimov, D. Pandelidis, Numerical study of perforated indirect evaporative
air cooler, Int. J. Energy Clean Environ. 12 (2–4) (2011) 239–250, http://
mechanical compression system, it is reasonable to partly recir- dx.doi.org/10.1615/InterJEnerCleanEnv.2013006668 (‘‘Special Issue on M-Cycle
culate the exhaust air stream in systems with HMX1. Fundamentals and Applications’’).
 The wet-bulb effectiveness and the specific cooling capacity are [19] S.V. Patankar, Numerical Heat and Mass Transfer and Fluid Flow, Hemisphere
publishing Corporation, Washington, 1980.
not informative when comparing the indirect evaporative [20] V.N. Bogoslovskij, M.J. Poz, Fundamentals of Physics of Heating, Ventilation
exchangers operating on the different primary and working and Air-conditioning Units Design, Strojizdat, Moscow, 1983 (in Russian).
air inlet parameters. [21] A. Žukauskas, J. Žiugžda, Heat Transfer in Laminar Flow of Fluid, Mintis,
Vilnius, Lithuania, 1969.
[22] R. Shah, D. Sekulic, Fundamentals of Heat Exchanger Design, John Wiley & Sons
The results presented in this paper indicate the attractiveness of Inc, Hoboken, New Jersey, 2003.
the novel evaporative air-cooling and allow extending the poten- [23] B. Riangvilaikul, S. Kumar, Numerical study of a novel dew point evaporative
tial of using these devices in air-conditioning systems. cooling system, Energy Build. 42 (2010) 2241–2250.

S-ar putea să vă placă și