Applied Thermal Engineering 31 (2011) 50e58

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Central air conditioning based on adsorption and solar energy

Antonio Pralon Ferreira Leite a, Francisco Antonio Belo b, Moacir Machado Martins c,
Douglas Bressan Riffel d, *
a
Federal University of ABC, Brazil
b
Solar Energy Laboratory, Federal University of Paraiba, 58051-900, João Pessoa, PB, Brazil
c
Federal Institute of Pernambuco, Av. Professor Luis Freire, 500, Cidade Universitária, 50.740-540, Recife-PE, Brazil
d
Department of Mechanical Engineering, Federal University of Sergipe e UFS, Av. Marechal Rondon S/N, 49100-000 Aracaju-SE, Brazil

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

Article history: This paper presents the characterization and the pre-dimensioning of an adsorption chiller as part of
Received 18 February 2010 a 20 kW air conditioning central unit for cooling a set of rooms that comprises an area of 110 m2. The
Accepted 10 August 2010 system is basically made up of a cold water storage tank supplied by an activated carbonemethanol
Available online 21 August 2010
adsorption chiller, a hot water storage tank, fed by solar energy and natural gas, and a fan-coil. During an
acclimatization of 8 h (9e17 h), the following parameters were obtained for dimensioning the cooling
Keywords:
system: 504 kg of activated carbon, 180 L of methanol, 7000 L of hot water, 10,300 L of cold water with its
Adsorption chiller
temperature varying in the fan-coil from 1
C to 14
C. Considering the mean value of the total daily
Activated carbonemethanol
Thermal storage
irradiation in João Pessoa (7
80 S, 34
500 WG), and a cover of regenerating heat supplied by solar energy
Solar energy equivalent to 70%, the adsorption chiller’s expected coefficient of performance (COP) was found to be
around 0.6.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction a refrigerating fluid, calling for a number of electromechanical

Solar cooling is an interesting application of solar energy because,
the stranger the insolation, the greater needs for cooling. As a rule,
the systems requiring thermal energy as their main power input for
the production of frigorific effect are most dependable on sorption
processes. The use of solar energy for cooling applications e a device
based on the liquid sorption cycle e was reported more than forty
years ago by Chinnappa [1].
The energy needed for a sorption cooling system is essentially
thermal, and may come from different sources, such as: process
heat, residual vapor or solar energy. One of the main advantages of
such a system in relation to the conventional vapor compression
system lies on its minimum use of electric energy and on its low
cost of maintenance.
Because of its functioning simplicity, the adsorption technology
applicable to refrigerating systems differs significantly from that of
absorption. In adsorption there occurs an interaction between
a solid and a fluid e the transportation of the latter being
a dependable thermal gradient, for it does not require the use of
pumps. In the case of absorption, a solution interacts with
* Corresponding author.
E-mail address: dougbr@ufs.br (D.B. Riffel).
devices so as to move both the solution and the fluid.
Moreover, the adsorption cycle depends on fewer electrome-
chanical components (heat exchangers, valves). Conversely,
adsorption refrigeration has exhibited performance coefficients
lower than those obtained by liquid absorption. In the field of solar
refrigeration by adsorption, various kinds of solid pairs have been
considered; zeoliteewater [2,3] and silica gelewater [4,5] are used
for cold storage, whereas activated carbonemethanol [6,7] and
activated carboneammonia [8,9] are used for ice production.
A comparative study between the liquid, chemical reaction and
adsorption cooling systems was published by Meunier [10]. In all
these systems, the mechanical energy consumption is kept to
a minimum or null. This represents a great advantage over the
conventional vapor compression systems, especially in countries
like Brazil whose energy consumption depends heavily on hydro-
electric power.
Solar energy is abundant over large areas in Northeastern Brazil
(annual average being over 6 kW/m2 day), securing, as a result,
a great potential supply for use in helio-thermal conversion
systems applicable for both heating and cooling purposes. Many
farm and fishing products are lost in this region, mainly because of
unsatisfactory refrigeration, or otherwise, these products have their
final value decreased for lack of economical frigorific preservation.
In the last two decades, absorption chillers, running on natural
gas, have been introduced in markets all over the developed world

1359-4311/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.applthermaleng.2010.08.010
 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58 51

e in the US, Japan, France, Italy, Spain and more recently in 2.1. DubinineAstakhov equation
Portugal. As to adsorption systems e in spite of several studies that
demonstrate their technical feasibilities e, there exist only exper- To describe adsorption in microporous materials with poly-
imental units, though stand considerable chances of becoming modal distributions, Dubinin and Astakhov [14] have proposed an
economically viable, especially where thermal comfort applications isotherm that is a log-linear form of DubinineRudishkevich equa-
are concerned [11]. tion, which can be expressed as follows:
This paper presents some of the adsorption fundamentals and
also the energy equations for adsorbers. Then it puts forwards the

a ¼ Wo rl ðTÞexp  D½TlnðPs =PÞn (1)
characterization and the pre-dimensioning of a 20 kW air
conditioning unit based on both the adsorption process and solar where a is the concentration (adsorbed mass per unit of adsorbent
energy. Cold water is produced during night-time in an activated mass), Wo is the maximum adsorption capacity (volume of adsor-
carbonemethanol adsorption chiller and then stored into a tank bate/mass of adsorbent), rl is the specific mass of the liquid
in order to supply cooling air during day-time. Solar energy is adsorbate, D is the “coefficient of affinity” and n is a characteristic
used for regenerating the adsorbent medium by means of highly parameter that depends on the adsorbenteadsorbate pair.
efficient flat-plate collectors. Thermal energy is stored in a hot- This equation has a wide field of application, and it is particu-
water tank so as to feed the chiller at night. The central unit was larly suitable for strongly activated carbon with high pores
created to provide thermal comfort to four adjoined laboratory heterogeneousness. According to Passos [15], for regenerating
rooms, whose total area is 110 m2. temperature ranging from 20 to 100
C, and for concentrations,
raging from 71 to 286 g/kg of adsorbent, a fitted curve, obtained
2. Adsorption fundamentals from methanol adsorption experimental data, showed a residual
error of 2.2%. These results demonstrate that this state equation is
Adsorption constitutes a solid sorption process by which the most suitable for a number of engineering applications of low-
binding forces between fluid molecules (adsorbate) and the solid grade sources, especially those concerning solar energy.
medium (adsorbent) derive from an electrostatic origin or from
dispersionerepulsion forces. It is an exothermic process as a result 2.2. Isosteric heat of adsorption
of the gaseliquid phase change. The energy liberated during
adsorption is called isosteric heat, and it depends on the nature of The energy released during the adsorption process, i.e. the
the adsorbenteadsorbate pair. isosteric heat (Qst) can be derived from the Gibbs’ isotherm, being
A number of state equations e known as isostherms of adsorption e reduced to a constant adsorbed mass function called isoster, which
are proposed. These functions correlate the temperature T, the pres- is given as follows
sure P, and the concentration of the adsorbed phase a, so that f(T, P,
a) ¼ 0. The main isotherms of adsorption are: (a) Henry’s law, valid  
vlnP qst
only for weak concentrations; (b) Langmuir’s approach, which takes ¼  2 (2)
vT a RT
up adsorption on monomolecular layers where there is a dynamic
equilibrium between the phases; (c) Gibbs’ theory, which is based on For the saturating condition (P ¼ Ps), the qst term in Eq. (2) is
the perfect gas equation, where the adsorbate is treated as a micro- replaced by the latent heat L. Then,
scopic and bi-dimensional form; and (d) Eucken and Polany’s theory,
the so-called potential adsorption theory, in which it is assumed that  
vlnPs
within the space around each solid, it is possible to find some iso- L ¼ RT 2 (3)
vT a
potential surfaces restricting the adsorbate that is adsorbed at pres-
sures (or concentrations) lower than those corresponding to the The DubinineAstakhov equation becomes
potential value, being such isopotential surfaces specific for a given
solid surface [12]. vlnP vlnPs n a o
Dubinin and Radushkevich have proposed the micropore volume ¼ þ lnðPs =PÞ T 1 þ ½TlnðPs =PÞn (4)
vT vT nD
filling theory, which is related to Eucken-Polanyi potential theory.
The DubinineRudushkevich isotherm describes adsorption as with
a single type of uniform pores that is similar to Langmuir-like local
vlnðrl =aÞ
isotherms in adsorption on energetically heterogeneous solids. This a ¼ Wo (5)
theory was later expanded by Stoeckli, allowing it to describe vT
adsorption on energetically heterogeneous solids with a contin- where a is the coefficient of thermal expansion of the liquid
uous distribution of pore sizes. Leite [13] puts forward a detailed adsorbate. Integrating Eq. (4) and, after some algebraic operations,
analysis of the thermodynamics of adsorption and its different the isosteric heat of adsorption can be given as
isotherms.
In general, all microporous materials are adsorbent and are  
aRT
characterized by their high porosity. Their structures have pores qst ¼ L þ RTlnðPs =PÞ þ ½TlnðPs =PÞð1nÞ (6)
nD
with diameters smaller than 20  A. The most common absorbents
are silica gel, activated carbon, zeolite and alumines. Zeoliteewater Being the latent heat, the first term of Eq. (6), the other terms do
and activated carbonemethanol are one of the most adsor- correspond to the energy that most specifically concerns the
benteadsorbate pairs utilized in refrigeration systems. Methanol is adsorption binding forces. For temperature around 5
C, the value
easily desorbed from activated carbon when heated, whereas in of L is about 1200 kJ/kg. According to Srivastava and Eames [16], the
zeolite, the water is sustained for much longer. Thus, the activated- isosteric heat of adsorption for activated carbonemethanol ranges
methanol pair is best adapted for operating cycles with small from 1800 to 2000 kJ/kg. So, the energy resulting from the binding
evaporating temperature variations (up to 40
C), whereas the forces in adsorption corresponds to something from 33% to 40% of
adsorption cycles with zeoliteewater pair, a larger range evapo- the isosteric heat; the rest would come as a result of capillary
rating temperature (70
C, or more) is needed. condensation inside the micropores.
52 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58
2.3. Kinetics of adsorption
In most cases, adsorption in microporous materials is mainly
controlled by the diffusion that takes place inside porous structure,
for on the surface of the grains, diffusion happens too fast. In the
case of materials with bidisperse structure, such as activated
carbon, two diffusion mechanisms may be found: a gaseous phase
diffusion through transport pores (mesoporous and macroporous
diffusion) and an adsorbed phase diffusion in micropores. The
relative importance of these mechanisms on the global diffusion
effect greatly depends on pressure. According to Dubinin and
Erashko [17], for pressures lower than 10 hPa, mesoporous and
macroporous diffusions are seen to be most predominant, whereas
for pressures greater than 10 hPa, microporous diffusion tends to
control the mass transfer process.
Thomas and Glauckauf, cited by Sakoda and Suzuki [18],
proposed an approach based on two important hypotheses: the
temperature of the grain is uniform, and the concentration on the
solid surface is equivalent to an equilibrium concentration. These
hypotheses have established, for modeling the mass transfer
resistance of the adsorption kinetics, the following linear equation:
va 15Di 
¼ aeq  a
vt 2
rg
where Di is the diffusion coefficient, aeq is the equilibrium
concentration (given by an isotherm) and rg is the average radius of
the grain.
In both mesopores and macropores, different mechanisms
contribute simultaneously to the diffusion process, the relative
influence of which does depend on the pores dimensions. Among
them, four main mechanisms are identified: superficial diffusion Ds,
molecular diffusion Dm, Knudsen’s diffusion DK, and Poiseuille’s
diffusion DP. For an overall diffusion analysis, an effective diffusion
coefficient De is normally considered, so that De ¼ f(Ds, Dm, DK, DP).
Notwithstanding the importance of the kinetics theory for
modeling adsorption processes, the practice demonstrates that the
mass transport resistance can be neglected in systems under
moderate thermal power. For methanol adsorption in activated
carbon, Kariogas and Meunier [19] have shown that the mass
resistance is negligible when the incident energy is lower than
50 W/kg of adsorbent. In other words, for modeling adsorption
processes related to low-grade energy sources, such as the case of
solar energy, it will not be necessary to consider diffusion through
the porous medium. Moreover, for many engineering applications,
it may be taken into account the existence of an instantaneous
equilibrium between the adsorbed and the gaseous phases. This
can be represented by means of an isotherm, such as the Dubi-
nineAstakhov equation.
3. Properties of the adsorbent and the adsorbate
For cooling purposes, the adsorbent should have high adsorptive
capacity at ambient temperature and low pressures, maintaining
small capacity of adsorption at high temperatures and pressures. The
required temperature at the evaporator depends on the adsorptive
capacity under small pressures; it is the property that allows the
adsorbent, at a given temperature, to retain vapors from a fluid at
lower temperature. On the other hand, the more intense this property
is, the higher the level temperature for regenerating the adsorbent.
The adsorbate is the working or refrigerant fluid, whose choice
depends on the evaporator temperature. It must possess high latent
heat of vaporization and small molecular dimensions to secure an
easy adsorption. Since the chiller will operate under temperatures
below 0
C, methanol seems to be a good adsorbate because:
- it can evaporate at a temperature largely below 0
C (its
melting point is e 94
C);
- its enthalpy of vaporization is high (w1200 kJ/kg, at e 5
C);
- its molecule is small enogh (4  104 mm) to be easily adsorbed
into micropores, with a diameter smaller than 2  103 mm;
- its normal boiling point (w65
C) is much higher the room
temperatures;
- its working pressure is always lower than the atmospheric one,
which means a safety factor in case of leakage.
The activated carbon has been selected as the adsorbent
material because of its affinity with methanol; the thermo physical
properties are those of the AC-35, produced by CECA (France).
These data and those concerning the thermodynamic equilibrium
e the isosters e of the AC-35/methanol pair are well known [20].
The activated carbon AC-35 is obtained from pinewood by heating
it to 950
C in the presence of water vapor. It has a significant
micropores volume, the empty spaces of which correspond to 78%
of the total volume. The adsorbent bed is made up of cylindrical
grains of 2 mm diameter and is 3 mm long in average. The
maximum adsorption capacity of the activated carbonemethanol
pair is about 0.3 kg of absorbate/kg of absorbent. According to
experimental studies, the regenerating temperature of the acti-
(7)
vated carbon AC-35 at this temperature ranges from 65
C to
100
C [21].
4. Functioning principle and the ideal thermodynamic cycle
The adsorption refrigeration cycle consists of two well-defined
stages: one is described as the adsorber cooling, with its conse-
quent adsorption process, when the evaporation of the working
fluid (the adsorbate) takes place. The other stage consists of the
solid medium (the adsorbent) regeneration, when the adsorbate is
desorbed and condensed (Fig. 1a). The ideal thermodynamic cycle
can be represented by two isosters (isolines with constant adsorbed
phase concentration, a) and two intercalated isobars, as shown in
Fig. 1b.
The adsorber cooling corresponds to the isosteric process 1e2,
depending on the ambient conditions. This process continues until
the adsorber pressure reaches its minimum value (point 2), when it
becomes equal to the evaporator pressure. At this point, the
adsorption process starts and prolongs until its temperature rea-
ches the minimum value (point 3). Then, the adsorber is heated,
corresponding to another isosteric process (3e4), until its pressure
reaches a maximum value (point 4). Desorption starts at this point,
and goes on until the adsorber temperature reaches its maximum
value (point 1), completing, in this way, the cycle.
5. Energy equations
The energy equations related to the adsorber, which will be
given next, correspond to a multi-tubular system, whose outer
surface exchanges heat with the water coming from the hot storage
tank or from the water supply network, depending on the stage of
the cycle. The adsorbent occupies an annular space delimited by the
wall of the tube and a metal screen through which the adsorbate
flows.
For the heat transfer in the adsorbent medium, the following
assumptions have been considered: (a) the pressure is uniform and
constant; (b) the heat conduction in the radial direction is negli-
gible; (c) the adsorbenteadsorbate pair is treated as a continuous
medium in relation to thermal conduction; (d) the convective
effects and pressure drops are negligible. The heat transfer in the
adsorber component can be expressed separately by two
 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58 53

Fig. 1. Functioning principle (a); network of isosters and theoretical adsorption refrigeration cycle (b).

equations: one for the adsorbent, and the other for the wall of the  
tubes. In the case of the former, we have da dlnP qst dT
¼ b  2 (11)
dt dt RT dt
vT va
½r1 ðCp1 þ aCp2 Þ ¼ kV2 T þ qst r1 (8) with
vt vt
 
with Ps n1
b ¼ anDT n ln (12)
P
1 v2 T v2 T
V2 T ¼ 2 2 þ 2 (9) The dlnP/dt term depends on the process that occurs in the
r vq vz
adsorber. In the case of an isosteric process, this can be given by
and
ZZ
    d
va va dT va dlnP aðT; PÞ2prdqdz ¼ 0 (13)
¼ þ (10) dt
vt vT lnP dt vlnP T dt
Applying this condition in Eq. (10), we obtain
where Cp is the specific heat (indices 1 and 2 refer to the adsorbent
and the adsorbate, respectively), r, the specific mass and k, the RR qst
dlnP bða; T; PÞ 2 rdqdz
conductivity of the adsorbent. Eq. (10) represents the kinetics of ¼ RR RT (14)
adsorption. After some algebraic operations, we obtain dt bða; T; PÞrdqdz

Fig. 2. Schematic layout of the air conditioning unit.

54 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58

Fig. 3. Scheme of the solar collector.

When adsorption or desorption process takes place, the system
pressure becomes equal to the saturating pressure, i.e. P ¼ Ps. Then,
the energy equation for the adsorbent can be written as
! exchangers, 1 tank of hot water (replenished by solar energy and
 
br1 q2st vT 1 v2 T v2 T dlnP
r1 ðCp1 þ aCp2 Þ þ ¼ k 2 2 þ 2 þ br1 qst
RT 2 vt r vq vz dt
For the heat transfer at the wall of the tubes, the energy equa-
tion is
vTt
mt Cpt ¼ N pL½hdi ðTt  TÞ þ hsh do ðTt  Tw Þ
vt
where N is the number of tubes, L is the tube length, di and do are
the inlet and the outlet diameters, respectively, Tt is the mean tube
temperature, Tw is the mean water temperature, At is the tube
superficial area (the external surface in contact with the water), h is
the conductance at the interface tube/adsorbent, and hsh is heat
coefficients on the external surfaces of the tubes (on the shell side).
For the water flowing between the tubes and the shell, the
energy equation is written as
vTw
mw Cpw ¼ S½hsh do ðTw  Tt Þ
vt
where S is the effective surface through which the water flows (at
the center of the shell); it will be given in Section 9.1.
For modeling the heat and mass transfer processes inside the
adsorber, we need to solve the coupled Eqs. (1), (6), (15), (16) and
(17). To solve this system of equations, all flow and porous
medium properties must be known, especially k, h and hsh.
According to Guilleminot et al. [22], for the AC-35 activated
carbon: k ¼ 0.19 W/mK and h ¼ 16.5 W/m2 K. The methodology for
obtaining hsh is given in Section 9.1.
6. Description and operation of the adsorption chiller
A schematic representation of the central air conditioning unit is
shown in Fig. 2. The system is formed by 3 basic components:
a water-cooling unit (the chiller), a tank of cold water, and an
airewater heat exchanger (the fan-coil). The adsorption chiller is
made up of the following devices: 2 (water-adsorber) heat
gas combustion), 2 air condensers, 1 evaporator, and accessories,
such as valves and circulation pumps. The adsorbers are multi-
tubular and their configuration is the same as that for the tubes of
(15) a shell-and-tube heat exchanger.
The adsorbers (I and II), working by turns, are responsible for
cooling the water in the thermal storage tank. They form, with the
storage tanks plus the solar and gas heater, the so-called primary
cooling system, or the adsorption chiller. The conditioned air is
(16)
obtained by the exchange of heat between the stored cold water and
the entering air through the fan-coil. When the adsorber I receives
methanol vapor from the evaporator, placed at the bottom of the
storage tank, at the same time the adsorber II goes through the
regeneration mode, which is characterized by the desorption process
with the consequent condensation of methanol in an air condenser.
To release the heat of adsorption (from the adsorber I), a secondary
water circuit, including a sky radiator and a storage tank, is used. A
hydraulic network feeds this circuit. The water is cooled at night-time
by irradiative exchanges with the sky, and it circulates around the
tubes to improve the adsorption process. The regeneration (of the
(17) adsorber II) is produced by the water heated by highly efficient solar
collectors and by a gas heater. The water flows in the adsorbers will
be managed by electrovalves (shown in Fig. 2) and micro controllers.
The entering air/cold water exchanger (the fan-coil) is the so-
called secondary cooling system, and is responsible for moving the
necessary amount of cooled air so as to provide thermal comfort
conditions inside the rooms.
Taking into account the fact that the adsorption chiller operation
is predominantly based on thermal energy consumption, the main
reason for using water thermal storage lies on environmental
conditions. So, the use of thermo storage is justified by two view
points: the availability of natural gas to provide, in a complemen-
tary way, the hot water for regenerating the adsorbent, and the
night ambient conditions to improve the kinetics of adsorption.
Besides, the choice of cold water, instead of ice thermal storage, was

Table 1 Table 2
Secondary cooling system parameters. Primary cooling system parameters.

Parameter Value Parameter Value

Total plant area of the rooms 110 m2 Chiller operation time 10 h
Entering air temperature 14
C Cycle time for each adsorber 2h
Return air temperature 26
C Cold water temperature variation 20
C
Water temperature variation 13
C Total solar radiation 21.3 MJ/m2 day
Mean cold water flow 0.37 kg/s Temperature increase in the collectors 35
C
Effective cooling power 20 kW Cooling power 24 kW
 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58
Table 3
Parameters calculated for the primary cooling system.
Parameter
Activated carbon mass in each adsorber
Methanol mass in each adsorber
Cold storage tank volume
Solar collectors total area
Temperature increase in the collectors
Hot storage tank volume
due to the fact that, for the latter, a very low evaporation temper-
atures are required (less than 10
C), being necessary more
expensive heat exchangers than the ones used for liquid storage.
7. Solar technology
The solar collectors are flat and static. The outer surface of the
collectors is painted nonselective matte black, and it is irradiated on
both faces, as originally proposed by Goetzberger et al. [23]. The
absorbing surface is covered on both sides by transparent insu-
lation material e the TIM cover e which is a honeycomb structure
in polycarbonate, bordered by two glass plates. The TIM cover
reduces considerably the losses from convection and radiation,
with a minimum attenuation of solar incidence. The reflectors,
lying underneath the tubes, have a semi cylindrical profile, and they
are made of polished aluminium. A schematic outline of the solar
collector is shown in Fig. 3.
Results from numerical simulations of a collector using TIM
cover and thermal insulation at the bottom, have shown that the
overall heat loss coefficient is about half of that for the same
collector using a single glass cover [24]. These calculations were
carried out by using meteorological data from João Pessoa (7
80 S,
34
500 WG). As result from an experimental study [25], the overall
heat loss coefficient in upper and lower TIM covers varied from 0.54
to 1.90 W/m2 K, respectively.
With the proposed solar technology, we expect to obtain, at the
outlet of the collector, water temperatures between 100
C and
110
C. The experimental results obtained by Leite et al. [26], shows
that this kind of collector can reach a daily thermal efficiency of
about 40%.
8. Pre-dimensioning
Each one of the four rooms to be kept cool has an area of 27.5 m2
and is 2.7 m high, with an occupancy period from 9 a.m. to 5 p.m.
Fig. 4. Schematic view of the water-adsorber heat exchanger.
55
The effective cooling power, which is provided by the secondary
system, was taken as a design parameter. It was based on the
Value nominal output from commercial air conditioning equipments for
252 kg climatic condition such as that found in João Pessoa, which is about
72 kg 67 W/m3.
10,300 L Based on the isosters of the adsorbenteadsorbate pair, the
120 m2
maximum concentration variation of the adsorbed phase was fixed
35
C
7000 L as 0.29 kg of methanol per kg of activated carbon. The solar irra-
diation was taken as the mean total daily radiation for the hottest
six months in João Pessoa. The total collector area was established
for covering 70% of the supplied regenerating heat. A gas heater
supplies the rest. At the end of the chiller operation period, the
temperature of the water in the hot storage tank was considered to
be at 75
C, and, at the sunrise, the entering water in the collectors
was fixed at 55
C. The global heat loss between the primary and
the secondary cooling systems was considered as being 17%.
The parameters used for pre-dimensioning the main compo-
nents of the air conditioning central unit e adsorbers, storage
tanks, solar collectors and gas heater e are shown in Table 1, for the
secondary cooling system, and in Table 2, for the primary cooling
system. As to the available thermal energy for regenerating the
adsorber, we have considered a water volume at a constant and
uniform temperature of 105
C.
The parameters for the primary cooling system (shown in
Table 3) were obtained by establishing the heat balance for each
component. To calculate these parameters, the remainder mass
inside the water-adsorber heat exchanger (between shell-and-
tubes) was neglected. The thermal resistance and the inertia of the
heat exchangers, such as those of the storage tanks and the gas
heater were also neglected.
Based on the parameters shown in Table 3, the solar collectors
will supply to the storage tank hot water at a temperature of 90
C.
The additional energy to provide the required regenerating
temperature of the adsorber will be supplied by the gas heater.
9. The adsorber thermal parameters design
As described in the previous section, the adsorber is a compo-
nent that includes a shell-and-tubes exchanger or, more precisely,
a liquid-adsorbent heat exchanger. Shell-and-tube heat exchangers
have been widely used in process industries. To create a reliable and
efficient design of this kind of exchanger, a good understanding of
both the flow and the temperature field is required. This implies
that one should know the total pressure drop and the overall heat
56 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58

Fig. 5. Dimensions and arrangement of the tube bank.

transfer coefficients. Such information can be obtained through The total heat (Q) for regenerating each one of the adsorbers is
either experimental analysis [27e29], or through simple calcula- the sum of the total isosteric heat (Qst ¼ mmeqst) and the total
tion methods based on empirical correlations [30,31]. In the sensitive heat (Qad) of their components (activated
present case, for calculating the global heat transfer coefficient carbon þ methanol þ tubes). Taking the adsorbate mass value as
between the liquid (water) and the adsorbent (activated carbon), (mme ¼ 72 kg) from Table 3, qst ¼ 1900 kJ/kg (from Ref. [16]), and
we used the Tinker’s method modified by Fraas and Ozisik [32]. Qad ¼ 34,400 kJ, we obtain Q ¼ 171,200 kJ. We have noticed that the
The choice of the geometric configuration of the heat exchanger thermal energy necessary to heat the adsorber components
was based on practical aspects, such as the simplicity of the (tubes þ activated carbon þ methanol) in the ambient temperature
manufacturing technology employed, and the required number of up to 90
C (DT ¼ 65 K) corresponds to about 20% of the total input
tubes, which depends on the amount of adsorbent they can hold. regenerating heat. For the water flow through the shell, we take
Thus, the chosen exchanger was that of a single pass, with a square DT ¼ 30 K, and, considering Dt ¼ 3600 s (half of one cycle), we have
arranged tube bank. A schematic view of the water-adsorbent heat m ¼ 0.38 kg/s and G ¼ 14.6 kg/m2 s.
exchanger is shown in Fig. 4. Segmental baffles were installed on The estimated water flow rate during the regeneration period of
the shell in order to enhance turbulence in the water flow around each cycle corresponds to a consumption of hot water around
the tubes. Details concerning the adsorber tubes and the way they 1400 L. So, a greater part (w70%) of the total volume of hot water
have been arranged are shown in Fig. 5. for regenerating the five adsorption cycles will be provided by solar
The method proposed by Tinker is based on the current analysis, energy, which corresponds to a mean power of 28 kW. Taking into
in which the flow is divided by a series of separated currents, account the natural gas heat combustion available locally
including the ones mentioned below. The currents are: the true (DH ¼ 38,540 kJ/Nm3), for supplying 40% of the adsorber regener-
crossed flow, the current through the space between tubes and ating heat, the daily gas consumption will be of about 11 m3, which
drillings, the current near the shell that does not cross the tube represents a mean power of 12 kW.
bank, and those between the baffles and the shell.
9.2. Heat transfer coefficient
9.1. Heat balance
The heat transfer coefficient on the shell side (hsh) is obtained
The heat transferred by circulating water between the shell and from a diagram (Fig. 7), as function of Reynolds’ number for
the tubes (Q) is given by designing conditions (Reh) and an adimensional number (Yþ),
correlating the thermal properties of the liquid to Nusselt number
Q ¼ mCp DT ¼ GSCp DT (18) (Nu).
Reynolds number can be written as
where DT is the temperature variation of the water, m is the mass
flow rate, Cp is the specific heat at constant pressure, G is the mass Gdo Fh
flow rate per square unit, and S is written as Reh ¼ (20)
12 M

S ¼ lðs  do Þ CNt (19) where m is the dynamic viscosity, Fh and M are geometric param-
eters given by a table and an equation, in the same diagram.
where l is the length of the tube between two baffles, s is the
distance center-to-center of tubes in the same array, and C and t are Table 4
constants. According to Scovill, cited by Fraas and Ozisik [32], for Parameters of the water-adsorbent heat exchanger.
square arrangements and number of tubes (N) between 20 and
Parameter Value
2000, C ¼ 1.4 and t ¼ 0.47.
Tube inside diameter (di) 76 mm
Table 4 shows physical parameters of the heat exchanger used to
Tube outside diameter (do) 78 mm
calculate thermal parameters based on Tinker’s method. Distance between tubes 10 mm
Taking into account a 40% baffle cutting depth (Fig. 6) and Shell inside diameter (Di) 700 mm
a number of baffles equal to 10, then l ¼ 0.298 m. Using the values Length of the tubes (L) 3.0 m
shown in Table 3, and introducing n ¼ 48, do ¼ 0.078 m and Number of tubes (N) 48
Adsorbent mass per length unit 1.75 kg/m
s ¼ 0.088 m in Eq. (18), we obtain S ¼ 0.026 m2.
 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58 57

Fig. 6. Schematic view of the baffles.

The adimensional number Yþ is power to the heat for regenerating the adsorber plus the input
energy for pumping cold and hot water and air (through the fan-
 
Nu 1=3 m 0:14 coil). For the adsorption chiller, or the primary cooling system,
Yþ ¼ Pr (21)
12 mw taking into account just the input thermal energy (40 kW), we
expect a COP equal to 0.6, and, for the secondary cooling system,
where Pr is the Prandtl number and mw is the dynamic viscosity at a COP of 0.5.
the tube wall temperature. The Nusselt number is

hsh do 10. Final remarks

Nu ¼
k
where k is the thermal conductivity of the liquid water.
By introducing into the diagram the geometric data of the
exchanger and the water thermal properties, we obtain Reh ¼ 130
and Yþ z 5, and, finally, hsh ¼ 600 W/m2.
9.3. Coefficient of performance (COP)
COP is the parameter to evaluate the efficiency of an air condi-
tioning system. It is defined as the ratio of the effective cooling
(22)
In this paper, we have specified the main dimensions and
characteristics of a central air conditioning unit based on adsorp-
tion and solar energy. The energy equations for modeling the
adsorption chiller are presented. Design thermal parameters of the
primary and secondary cooling systems are determined. The water-
adsorber heat exchanger has been analysed in full detail; the
external heat transfer coefficient e on the shell side e has been
calculated for designing conditions by means of a simple method.
For a decrease of 105
Ce75
C in the water temperature through
the water/adsorber exchanger, the volume flow rate was about

Fig. 7. Shell side heat transfer characteristics for a multi-tubular shell-and-tube heat exchanger with a square arrangement of tubes; Di is the inside shell diameter, Yþ ¼ (Nu/12)
Pr1/3(m/mw)0.14 (Adapted from [32]).
58 A.P. Ferreira Leite et al. / Applied Thermal Engineering 31 (2011) 50e58
1400 kg/h, and the heat transfer coefficient was around
600 W/m2 K. With the help of the given equations, we can simulate
the behaviour of the adsorber as a function of different water inlet
temperatures and different cycle periods. The effective coefficient
of performance is expected to be 0.5, which is relatively low, in
relation to that of conventional vapor compression systems.
Nevertheless, it is important to consider the null cost of the main
energy source e the solar energy e and the low cost of the available
natural gas.
Acknowledgements
The authors gratefully acknowledge the Brazilian agency,
National Council of Scientific and Technological Development e
CNPq, for the financial support provided to this work through
Research Project Grant No. 504229/2004-4.
References
[1] J.C.V. Chinnappa, Performance of an intermittent refrigerator operated by
a flat-plate collector, Solar Energy 6 (1962) 143e150.
[2] M.R. Chaya, M. Horn, R. Spinoza, A.P.F. Leite, Evaluation of a zeolithe-water
solar adsorption refrigerator, in: Proceedings of Solar World Congress
(ISES’2003), Göteborg, Sweden, 2003.
[3] Ph. Grenier, J.J. Guilleminot, F. Meunier, M. Pons, Solar powered solid
adsorption cold store, J. Solar Energy Eng. 110 (1988) 192e197.
[4] C. Hildebrand, Ph. Dind, M. Pons, F. Buchter, A new solar powered adsorption
refrigerator with high performance, Solar Energy 77 (2004) 311e318.
[5] W.-S. Chang, C.-C. Wang, C.-C. Shieh, Design and performance of a solar-
powered heating and cooling system using silica gel/water adsorption chiller,
Appl. Thermal Eng. 29 (2009) 2100e2105.
[6] A. Boubakri, J.J. Guilleminot, F. Meunier, Adsorptive solar powered ice maker:
experiments and model, Solar Energy 69 (2000) 249e263.
[7] F. Lemini, J. Buret-Bahraoui, M. Pons, F. Meunier, Performance simulation of an
adsorptive solar refrigerator: 1.Comparison of performances for two types of
activated carbon, Int. J. Refrigeration 5 (1992) 159e167.
[8] Z. Tamainot-Telto, R.E. Critoph, Adsorption refrigerator using monolithic
carbon-ammonia pair, Int. J. Refrigeration 20 (1997) 146e155.
[9] A.E. Fadar, A. Mimet, A. Azzabakh, M. Pérez-García, J. Castaing, Study of a new
solar adsorption refrigerator powered by a parabolic trough collector, Appl.
Thermal Eng. 29 (2009) 1267e1270.
[10] F. Meunier, Solid sorption heat powered cycles for cooling and heat pumping
applications, Appl. Thermal Eng. 18 (1998) 715e729.
[11] H.-M. Henning, Solar assisted air conditioning of buildings e an overview,
Appl. Thermal Eng. 27 (2007) 1734e1749.
[12] M. Polanyi, The potential theory of adsorption, Science 141 (1963)
1010e1013.
[13] A.P.F. Leite, Thermodynamic analysis and modelling of an adsorption-cycle
system for refrigeration from low-grade energy sources, J. Braz. Soc. Mech. Sci.
20 (1998) 301e324.
[14] M.M. Dubinin, V.A. Astakhov, Description of adsorption equilibria of vapors on
zeolites over wide ranges of temperature and pressure, Adv. Chem. Ser. 102
(1971) 69.
[15] E.F. Passos, Etude des couples charbon actif-methanol et de leur application à
la refrigeration solaire. PhD thesis, Ecole Polytechnique Fédérale de Lausanne,
Swiss, 1986.
[16] N.C. Srivastava, I.W. Eames, A review of adsorbents in solid-vapor adsorption
heat pump systems, Appl. Thermal Eng. 18 (1998) 707e714.
[17] M.M. Dubinin, I.T. Erashko, Kinetics of physical adsorption by carbonaceous
adsorbents of biporous structure, Carbon 13 (1975) 193e200.
[18] A. Sakoda, M. Suzuki, Fundamental study on solar powered cooling systems,
J. Chem. Eng. 17 (1984) 52e57.
[19] M. Kariogas, F. Meunier, Etude de la relaxation d’un réacteur à zéolite pendant
la désorption éclair provoquée par un échelon de presión, Chem. Eng. J. 32
(1986) 171e192.
[20] M. Pons, Ph. Grenier, A phenomenological adsorption equilibrium law
extracted from experimental and theoretical considerations applied to the
activated carbon and methanol pairs, Carbon 24 (5) (1986) 615e625.
[21] A.P.F. Leite, M.B. Grilo, R.R.D. Andrade, F.A. Belo, F. Meunier, Experimental
evaluation of a multi-tubular adsorber operating with activated, Adsorption
11 (2005) 543e548.
[22] J.J. Guilleminot, F. Meunier, J. Pakleza, Heat and mass transfer in a non-
isothermal fixed bed solid adsorbent reactor: a uniform pressure/non-uniform
temperature case, Int. J. Heat Mass Transfer 30 (8) (1987) 1595e1606.
[23] A. Goetzberger, J. Dengler, M. Rommel, J. Göttsche, V.A. Wittwer, A new
transparently insulated, bifacially irradiated solar flat plate collector, Solar
Energy 49 (1992) 403e411.
[24] A.P.F. Leite, M. Daguenet, Performance of a new solid adsorption ice maker
with solar energy regeneration, Energy Conversion Manage. 41 (2000)
1625e1647.
[25] A.P.F. Leite, M.B. Grilo, R.R.D. Andrade, F.A. Belo, Dimensioning, thermal
analysis and experimental heat loss coefficients of an adsorptive solar ice-
maker, Renew. Energy 29 (2004) 1643e1663.
[26] A.P.F. Leite, M.B. Grilo, R.R.D. Andrade, F.A. Belo, F. Meunier, An improved multi-
tubular solar collector applied to adsorption refrigeration, in: Proceedings of ISES
Solar World Congress, Orlando, Florida, USA, Aug 6e12, 2005.
[27] K.J. Bell, Final Report of the Cooperative Research Program on Shell-and-tube
Heat Exchangers. Bulletin No 5, University of Delaware Engineering Experi-
mental Station, Newark, Delaware, USA, 1963.
[28] H. Halle, J.M. Cheoweth, M.W. Wambsganss, Shell side water flow pressure
drop distributions measurements in an industrial-sized test heat exchanger,
in: Proc. 22th ASME Heat Transfer Conf. and Exhibition, USA, 1984, 39e48.
[29] L.S. Pletcher, M.J. Andrews, Technical/market assessment heat exchanger
technology for users of natural gas. GRI Report GRI-94/0248, 1994.
[30] T. Tinker, Shell-side characteristics of shell-and-tube heat exchangers:
a simplified rating system for commercial heat exchangers, Trans. ASME
(1958) 36e52.
[31] J.W. Pale, J. Taborek, Solution of shell side flow pressure drop and heat
transfer by stream analysis method, Chem. Eng. Symposia Ser. 65 (93) (1969)
53e63.
[32] A.P. Fraas, M.N. Ozisik, Heat Exchanger Design. John Wiley & Sons, New York,
1965, pp. 146e157.