You are on page 1of 9

Applied Energy 162 (2016) 257265

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Influence of the operational conditions on the performance of a


chemisorption chiller driven by hot water between 65 C and 80 C
Rogrio Gomes de Oliveira a,, Daniel Joo Generoso b
a
Federal University of Santa Catarina (UFSC), Ararangu Cmpus, Laboratory of Applied Thermal Science (LABCITEA), R. Pedro Joo Pereira, 150, Araragu, SC 88905-120, Brazil
b
Federal Institute of Santa Catarina (IFSC), Ararangu Cmpus, Av. XV de Novembro, 61, Araragu, SC 88905-112, Brazil

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Design and construction of a


chemisorption chiller driven by hot
water between 65 C and 80 C.
 The influence of the operation
conditions on the COP and cooling
power are presented.
 Heat and mass recovery are effective
to improve COP.
 Minimum chilled water temperature
of 4 C using heat source at 70 C.

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

Article history: We designed, built and experimentally assessed the cooling power and the coefficient of performance
Received 22 June 2015 (COP) of a chemisorption chiller that used a consolidated composite made of NaBr impregnated in
Received in revised form 25 September 2015 expanded graphite as sorbent and ammonia as refrigerant. The influence of the operation conditions
Accepted 6 October 2015
on the COP and the cooling power was calculated through the values of inlet and outlet water tempera-
ture and water mass flow across the chillers heat exchangers. From the experimental results, we con-
cluded that the utilization of mass recovery simultaneously to heat recovery improved the COP and
Keywords:
cooling power up to 11.7% and 15.4%, respectively, when compared to the values obtained when only
Adsorption
Ammonia
heat recovery was employed. Regarding the influence of the heat source temperature, maximum cooling
Chiller power and COP were obtained when that temperature was, respectively, 75 C and 65 C. As for the influ-
Sodium bromide ence of the cycle time, the highest values of COP were obtained with an 88 min cycle, whereas the highest
Solar cooling values of cooling power were obtained at different cycle time, depending on the value of other operation
Waste heat conditions. When the chiller was driven by hot water at 70 C, and the inlet water temperature in the
evaporator was not kept constant, it cooled water from 23 C to 10 C in 90 min with mean cooling power
of 1400 W and COP 0.33.
2015 Elsevier Ltd. All rights reserved.

Corresponding author.
E-mail addresses: rogerio.oliveira@ufsc.br (R.G. de Oliveira), danieljoaogeneroso@gmail.com (D.J. Generoso).

http://dx.doi.org/10.1016/j.apenergy.2015.10.057
0306-2619/ 2015 Elsevier Ltd. All rights reserved.
258 R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265

1. Introduction Other examples of the utilization of NaBr as sorbent in


chemisorption refrigerators can be seem in the work presented
Sorption refrigerator are among the alternatives for cooling by Jiang et al. [19] and by Kiplagat et al. [20]. In the former work,
power production through directly use of thermal energy, instead a refrigerator operating with a resorption process in double stage
of electricity, which is especially attractive if the heat is provided could effectively produce cooling power at 20 C, when the heat
without extra costs. These refrigerators can be of the absorption source was between 70 C and 90 C. In the latter work, NaBr
or adsorption type, and they have been the focus of much impregnated in expanded graphite was used as composite sorbent
research [19], due their capability to produce cooling power in the reactor of a chemisorption chiller, which cooled water 13 C
directly from waste heat or solar energy. However, for these with COP above 0.4 when hot water with temperature between
refrigerators effectively use solar radiation or low grade waste 60 C and 70 C was used as heat source and tap water at 25 C
heat, from a broad range of processes, as main source of energy, was used as heat sink.
the temperature level of the heat source should be as low as Because the prototype studied by the latter authors had only
possible. one reactor, the cooling power halted during the regeneration of
Saha et al. [10] developed a two-stage silica gel adsorption chil- the sorbent, and the machine could not operate as the usual
ler driven with hot water at 55 C that produced 3.6 kW of cooling adsorption chillers, which can produce cooling effect even during
power from cold water at 14 C, with COP of 0.36. The silica- the regeneration period of the sorbent. Moreover, as the machine
gelwater adsorption chiller studied by Xia et al. [11] had similar had only one reactor, it could not operate using heat and mass
COP, but was driven by hot water at 75 C and cooled water from recovery processes. Previous studies with adsorption refrigerators
16 C. A higher COP of 0.43 was achieved by the silica-gel adsorp- using physical adsorbents [21,22] or composite of physical and
tion chiller developed by Chen et al. [12] when the machine was chemical adsorbents [23,24] had shown that heat recovery and
driven by hot water at 73 C and also cooled water from 16 C. refrigerant mass recovery between the reactors are effective ways
Vasta et al. [13] used zeolite as adsorbent in their adsorption chiller to enhance the COP and the cooling power. The importance of
and achieved COP of 0.41 to cool water from 12 C when the heat these recovery processes was also highlighted in a recent study
source was 90 C. [25] related to the use of chemisorption reactors combined with
Both silica-gel and zeolite are adsorbents in which the adsorp- a turbine to provide electricity and cooling power. The authors of
tion process occurs due to intermolecular forces similar to those the latter study showed with their theoretical analysis that
responsible for the condensation of vapors, and hence, this type combined heat and mass recovery between the chemisorption
of process is called physical adsorption [14]. However, there is reactors could enhance the COP by 3538%, and increase the elec-
another type of adsorption, called chemical adsorption or tricity efficiency from 8% to 12%.
chemisorption, in which the adsorption process occurs due to Therefore, to overcome the limitations of the study carried by
valence forces similar to those operating in the formation of Kiplagat et al. [20] and to provide experimental evidence of the
chemical compounds [14]. In practical terms, the thermodynamic heat and mass recovery process influence on machines with chem-
equilibrium is the main feature to distinguish between physical ical adsorption, we design, built and tested a chemisorption chiller
and chemical adsorption, which have respectively, divariant and with two reactors that could operate with or without heat and
monovariant equilibrium. mass recovery processes. Moreover, in the present study, the
The use of physical adsorbents in adsorption chillers is more composite sorbent of NaBr impregnated in expanded graphite
common than the use of chemical adsorbents like metallic salts. was consolidated in blocks, instead of be filled as powder around
The latter type of adsorbent usually has the advantage of higher fins, as done by the latter authors. Because consolidated sorbent
adsorption capacity but the disadvantage of lower adsorption exhibits higher thermal conductivity than powder sorbent, the
rate when compared to physical adsorbents, and relatively less utilization of fins was unnecessary.
research about chemisorption chillers is available in the The COP and cooling power of the chiller was assessed under
literature. different working conditions, which included the recovery
The chemisorption machines studied by Lepinasse et al. [15] processes previously mentioned, under variable cycle time and
and by Bao et al. [16] were refrigerators were the cooling power operation temperatures in the evaporator and in the reactors.
was produced through a resorption process in the reactor placed
inside an insulated box. The purpose of these machines was to 2. Materials and methods
keep a low temperature inside the box. The machine developed
by the former authors kept air inside the box between 3 C The chemisorption chiller presented in this work had two reac-
and 5 C for about 3 h when the ambient temperature was tors, one evaporator and one condenser. Each of these heat
20 C, whereas the machine developed by the latter authors kept exchangers had their temperature, pressure and water mass flow
air at the bottom zone of the cold box below 0 C for 5 h when measured by sensors, which sent periodically, their signal to a data
the ambient temperature was 20 C. In both machines, the heat logger. The scheme of the chiller and the characteristics of the
source to drive the machines was at least 150 C, which was sensors are presented in sub Section 2.1, whereas the preparation
also the heat source temperature used to drive a chemisorption of the composite sorbent is described in sub Section 2.2.
chiller that produced cooling power at 0 C with COP of 0.30 The experiments were divided in two types, whose description
[17]. is shown in sub Section 2.3. In the first type of experiments, we
The above-mentioned machines employed resorption process, compared the performance of the chiller at selected operation
which eliminates the use of condenser and evaporator, because conditions. In the second type of experiment, we assessed the
the cooling power was produced inside a reactor due an endother- minimum temperature of the cooled water produced by this chiller
mic desorption process of ammonia from one of the following type when it was employed to cool water stored in a 125 L tank. The
of salts: PbCl2, NH4Cl or NaBr. mean cooling power and coefficient of performance (COP) were
Experimental results presented in the literature [18] show that calculated along the whole cooling process.
the latter salt is suitable to be used in both resorption refrigerators Finally, in Section 2.4, we explain the calculation procedures
and in conventional sorption chillers in which the cooling power used to obtain the cooling power and COP, together with their
production occurs in the evaporator. respective experimental uncertainties.
R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265 259

Fig. 1. Scheme of the chemisorption chiller.

2.1. Description of the chemisorption chiller and the refrigeration


sorption cycle

The scheme of the chiller is presented in Fig. 1. The chiller


operated in a cycle where while one reactor was being heated
and desorbed refrigerant to the condenser, the other reactor
was being cooled and adsorbed refrigerant from the evaporator.
The refrigerant condensed was kept in an insulated bottle until
it was transferred to the evaporator at predefined periods, defined
according to the operation condition tested, as explained in sub
Section 2.3.
The reactors were tube and shell heat exchangers with the
composite sorbent inside the tubes and water flowing in the shell.
The photo of the reactor without the shell and with the shell is
shown in Fig. 2. Other characteristics of the reactor are shown in
Table 1.
The current chiller design considered the possibility of changing
the adsorbent bed inside the reactor to allow the utilization of a
different adsorbent. For a commercially designed chiller, this pos-
sibility would not be necessary, which would allow the construc-
tion of a reactor with about 22 kg less due the absence of flanges
and pipes lids. These components were employed to allow the
opening of the reactor to access and change the adsorbent. More-
over, the bed thickness, the expanded graphite to salt mass ratio
and the bed density were not optimized, and future experiments
may show that the mass ratio between adsorbent bed and reactor
can increase to allow higher cooling power and COP in the same
machine.
The condenser and the evaporator were also tube and shell
reactors. In the condenser, the refrigerant condensed inside the
tubes and water flowed through the shell, whereas in the evapora- Fig. 2. (a) Top view of the reactor tubes. (b) Side view of reactor tubes. (c) Complete
tor, the refrigerant evaporated in the shell, around the tubes, and reactor with shell.
260 R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265

Table 1 NaBr 5:25NH3 () NaBr  5:25NH3 5:25DHR 1


Constructive characteristics of each reactor.
The preparation of the composite sorbent was presented else-
Heat exchange area (m2) 2.81
where [26], and mainly included the expansion of the expandable
Mass of tubes (kg) 82.5
Total metallic mass (tubes, shell, cover, etc.) (kg) 140 graphite at temperatures between 700 C and 750 C, followed by
Inner diameter of the tubes (m) 0.055 the mixture of the expanded graphite to an appropriate amount
Number of tubes 19 of salt solution with mass concentration around 23%, so that the
Inner diameter of the shell (m) 0.31 slurry formed by the these components had the proportion
Number of baffles 1
between salt and expanded graphite equal to 3:2. The slurry was
Mass of dry composite adsorbent (kg) 9.4
Thickness of the adsorbent layer (m) 8.4  103 then, dried in an electric oven at 130 C for a period between 6
Length of the adsorbent inside the tube (m) 0.81 and 12 h, depending on the amount of slurry to be dried.
Once a batch of slurry was turn into a dried composite powder,
three samples of this powder were put in a moisture analyzer (SHI-
water flowed in multiple passes, through the tubes. The evaporator MADZU, MOC 63u) to have its moisture content identified. In all
was only partially filled with tubes to ensure that even with the batches, the final moisture content of the composite powder was
reduction of refrigerant inside the evaporator during the adsorp- below 1.5%. The composite powder was then, put inside the tubes
tion period, the tubes remained immersed in the refrigerant. of the reactor, and compressed until it reached the bulk density of
Fig. 3 shows the disposition of the tubes inside the evaporator 500 kg/m3.
and the cover used to ensure multiple passes. Other characteristics
of the evaporator are shown in Table 2. 2.3. Experimental procedures
Each reactor was heated and cooled under predefined periods of
time, as described in Section 2.3. During the heating period of one Experiments were conducted at different controlled operation
reactor, the composite sorbent desorbed refrigerant to the con- conditions to investigate how these conditions influence the per-
denser, and the heat consumed by the sorbent and by the reactor formance of the chiller. The controlled conditions were the inlet
was provided by hot water heated in the boiler. At the same time, hot water temperature in the reactor (TIn,Hot), the inlet cold water
the heat released in the condensation process of the refrigerant temperature in the evaporator (TIn,Cold), the cycle time, the use of
was being removed by water coming from a sink, which in this heat recovery and the use of mass recovery. During the heat recov-
work, was tap water from the city grid supply. The water leaving ery process, water flowed from one reactor to the other one, with-
the condenser entered the other reactor to cool it down and to out pass through the boiler. This process lasted 4 min, and each
remove the adsorption heat released during the adsorption cycle had two heat recoveries. The influence of mass recovery in
process. the performance of the chiller was only assessed in the cycles with
In an alternative approach, the reactors could have their heat heat recovery, and both processes occurred simultaneously. During
removed directly from the water coming from the heat sink, but the mass recovery, the reactor that had desorbed refrigerant to the
this would imply in two sinks instead of one, which would increase condenser sent refrigerant mass to the reactor that had adsorbed
the number of valves, connections, pipes, etc. of the system. refrigerant from the evaporator due to the pressure difference
Once the heating and cooling period of the reactors was between them.
completed, there was a lag period where the refrigerant stored in In the cycle without heat recovery, there was a lag period of one
the bottle was transferred to the evaporator. During this period, minute between the heating and cooling period of a reactor. The
the reactors could exchange heat between them or also, exchange lag period was necessary to allow the refrigerant stored in the bot-
refrigerant mass. The influence of these recovery processes in the tle to flow to the evaporator. In the cycles with heat recovery, this
performance of the chiller was investigated, as further explained lag period was not necessary because the refrigerant flowed from
in Section 2.3. the bottle to the evaporator during the heat recovery period. Nev-
ertheless, in all cycles, the heating and the cooling period of the
2.2. Preparation of the composite consolidated sorbent reactors had the same length of time to ensure that while one reac-
tor was receiving heat from the source, the other one was releasing
Each reactor contained a consolidated block of composite sor- heat to the sink.
bent made from the impregnation of NaBr in expanded graphite. Table 3 summarizes the controlled conditions tested and also
NaBr was chosen as the sorbent salt because previous research shows the inlet water temperature in the condenser (TIn,Med,Cn)
[18] indicated that this salt is capable to desorb ammonia at the and in the reactor (TIn,Med,Rc) during the cooling period. Although
temperature of 65 C when the condensation temperature is these temperatures were not controlled because the water came
30 C. The desorption and adsorption of ammonia by NaBr occurs directly from the city grid supply, it is possible to see that the for-
according to the following equation: mer temperature was approximately constant at 28 C in most

Fig. 3. (a) Front view of the tubes inside the evaporator shell. (b) Cover to ensure multiple passes through the tubes.
R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265 261

Table 2 where Dt is the interval of time of each data acquisition by the data
V_ Ev is the
Constructive characteristics of the evaporator.
logger, qW is the water density at the inlet temperature,
Heat exchange area (m2) 1.57 volumetric flow of water through the evaporator, CpW is the specific
Total volume of refrigerant (m3) 1.06  102 heat of water, TOut,Cold is the outlet water temperature from the
Volume of refrigerant above the tubes (m3) 4.89  103
evaporator, ti is the length of time since the beginning of a cycle,
Length of the tube (m) 0.93
Number of tubes 43 n is the number of times that data was acquired by the data logger
during a cycle, mCS is the total mass of composite adsorbent inside
the two reactors and V CS is the total volume of composite adsorbent
experiments, and that the latter temperature had higher variation
inside the two reactors. As the mass and volume of the adsorbent
because it depended on the amount of heat released in the con-
were constant in all experiments, each 150 W of cooling correspond
denser under the different operational conditions.
to 8 W/kg and 4 kW/m3.
The results of the experimental conditions shown in Table 3
Eq. (2) gives the accumulated mean value of the cooling power
presented in this work were obtained after the second consecutive
since the beginning of the cycle, and the values presented in this work
cycle in which the controlled operation conditions remained con-
are those obtained at the end of the cycle, unless stated otherwise.
stant. This procedure was adopted to remove the influence in the
The COP was calculated with Eq. (5) and indicates the amount of
machine operation that the previously tested operation condition
heat that need to be provided to the system to obtain a certain
could have in the current operation condition being tested.
amount of cooling load.
Once the experiments at the conditions shown in Table 3 were
concluded, a different type of experiment were conducted where Q Cool
COP 5
the chiller operated under a 68 min cycle with heat and mass Q Hot
recovery and inlet hot water temperature of 70 C. These experi-
n h
X i
ments were used to identify the minimum temperature that water
stored in a 125 L tank at ambient temperature (23 C) could be
Q Hot V_ Source CpW T In;Hot  T Out;Hot
Dt qW 6
i
i1
chilled, and to assess how the COP and cooling power varied in suc-
cessive cycles when the inlet water temperature to the evaporator n h
X i
varied with the time. Q Cool V_ Ev CpW T In;Cold  T Out;Cold
Dt qW 7
i
i1

2.4. Performance indicators


where V_ Source is the volumetric flow of water from the heat source
and TOut,Hot is the outlet water temperature from the reactor during
The most important performance indicators for adsorption chillers
the heating period.
are the cooling power Q_ Cool and the coefficient of performance TIn and TOut were measured with resistance temperature detec-
(COP). The cooling power was obtained with Eq. (2), and it was used tors (RTD) 1/10 DIN class with uncertainty at 0 C of 0.03 C. The
to calculate the specific cooling power (SCP) and the cooling power uncertainty in the temperature acquisition by the data logger
density (CPD), as shown in Eqs. (3) and (4), respectively. Both SCP
was 0.06 C. V_ Ev was measured with a flowmeter FPR301 (Omega
and CPD are important parameter in the design of adsorption chillers
because the former indicates the amount of cooling power that can be Engineering, Inc.) and V_ Source was measured with a flowmeter
produced per mass unit of adsorbent whereas the latter one indicates FPR204 (Omega Engineering, Inc.), and both have uncertainty of
the minimum volume necessary to accommodate the adsorbent. 3.2  103 L/s. The uncertainty in the flow acquisition by the data
logger was 2.8  105 L/s for the sensor FPR301 and 8.8  104 L/s
Pn h i
i1 V_ Ev CpW T In;Cold  T Out;Cold
Dt qW for the sensor FPR204. The data acquisition was done with a data
Q_ Cool i
2 logger 34772a (Agilent Technologies, Inc.), and occurred every 6 s.
ti The uncertainty in the values of cooling power and COP were
obtained, as recommended in the literature [27], with Eq. (8) when
Q_ Cool the uncertainty of the variables were independent of each other or
SCP 3
mCS with Eq. (9), when they were dependent of each other.
s
 2  2  2
Q_ Cool dP dP dP
CPD 4 DP DV 1 DV 2    DV n 8
V CS
dV 1 dV 2 dV n

Table 3
Operation conditions for the chemisorption chilller.

Heat recovery Mass recovery Cycle time (min) TIn,Cold (C) TIn,Med,Cn (C) TIn,Med,Rc (C) TIn,Hot (C)
Yes No 68 12.1 0.1 27.9 0.1 29.1 0.6 70.0 1.3
Yes Yes 68 12.0 0.2 27.4 0.1 29.1 0.7 69.9 1.8
Yes Yes 78 12.0 0.1 28.0 0.1 29.5 0.7 69.9 1.3
Yes Yes 88 12.0 0.1 28.4 0.1 29.9 0.8 70.0 1.4
Yes No 68 12.1 0.2 27.5 0.1 28.9 0.9 75.1 1.4
Yes Yes 68 12.0 0.2 27.5 0.1 29.0 0.9 75.0 1.4
Yes Yes 68 12.0 0.2 28.4 0.1 29.9 1.4 79.9 1.6
Yes Yes 68 15.1 0.1 28.4 0.1 29.7 0.4 65.0 1.4
Yes Yes 78 15.1 0.1 27.9 0.1 29.5 0.5 65.0 1.1
No No 62 15.1 0.2 28.2 0.1 30.0 0.7 69.9 1.1
Yes No 68 15.1 0.1 29.1 0.1 30.7 0.7 70.0 1.4
Yes Yes 68 15.1 0.2 28.9 0.1 30.5 0.6 70.0 1.3
Yes Yes 78 15.0 0.1 27.6 0.1 29.6 0.7 69.9 1.4
Yes Yes 88 15.1 0.1 27.8 0.1 29.6 0.7 70.0 1.1
Yes No 68 15.2 0.3 28.2 0.1 30.0 0.9 75.1 1.7
Yes Yes 68 15.1 0.2 28.3 0.1 30.0 0.8 75.1 1.8
262 R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265

dP dP dP
DP DV 1 DV 2    DV n 9
dV 1 dV 2 dV n

where DP is the uncertainty in the value of a certain parameter P


calculated through an equation, DV is the uncertainty in the mea-
sured value of a certain variable used to calculate a certain param-
eter P and the subscript n is the number of variables used to
calculate a certain parameter P.

3. Results and discussion

3.1. Influence of heat and mass recovery in COP and cooling power

The variation of the COP and cooling power of the chiller when
Fig. 5. Performance variation with the utilization of mass recovery at different
heat recovery or heat and mass recovery were used is shown in working conditions.
Fig. 4, for the operation conditions of TIn,Hot equal to 70 C and
TIn,Cold equal to 15 C. The reference values were those obtained
when neither heat recovery nor mass recovery were employed.
It is possible to see that the use of heat recovery increases the Fig. 5 presents the performance variation of the chiller at differ-
COP, when compared to that obtained when the machine operated ent operation conditions when both heat and mass recovery were
without heat recovery, but this process decreases the cooling used instead of only heat recovery. In this case, the reference
power. The cooling load absorbed in the evaporator during the values were those obtained with the cycles that employed only
cycle without heat and mass recovery and during the cycle with heat recovery.
heat recovery were practically the same within the experimental At all conditions of TIn,Hot and TIn,Cold, the utilization of mass
uncertainty and equal to respectively, 4.1 0.2 MJ and recovery in the cycle with heat recovery enhanced the cooling
4.0 0.2 MJ, whereas in the cycle with heat and mass recovery, power and the COP. The enhancement in the cooling power
the heat absorbed by the evaporator was equal to 4.3 0.2 MJ. decreased with the increase of TIn,Hot and TIn,Cold, as the operation
Hence, the main reason to the reduction of the cooling power conditions during the heating and cooling periods became more
was the increase of cycle time in almost 10%, as it passed from favorable to complete the reaction shown in Eq. (1). This behavior
62 min in the cycle without heat recovery to 68 min in the cycle can be explained due to the fact that as the value of TIn,Hot and
with heat recovery or with heat and mass recovery. The COP TIn,Cold increase, the amount of ammonia desorbed and adsorbed
increased with the heat recovery because 29% of the heat used also increases, and hence, there is less ammonia available to be
by the reactor could be recovered at this operation condition. transferred from the hot adsorbent bed to the cold adsorbent bed
The utilization of mass recovery simultaneously to the heat recov- during the mass transfer process. As for the COP, the greatest
ery caused further increment in the COP and helped to decrease the contribution of the mass recovery was the reduction of the amount
influence of the longer cycle time in the reduction of the cooling of heat that need to be supplied by the external source, as previous
power. explained, and this reduction was more significant at the
The mass recovery increased the COP due to an increase of the conditions that demanded more sensible heat, as those when
amount of heat absorbed in the evaporator and due to the TIn,Hot was in its highest value.
reduction of the total amount of heat that had to be supplied by
an external source to the reactors. The former phenomenon
occurred because the mass recovery reduced the amount of 3.2. Influence of the heat source and evaporator temperature in COP
ammonia remained in the adsorbent bed after the heating period, and cooling power
and hence, the adsorbent bed could adsorb more refrigerant
from the evaporator, when the cooling period started. The latter The influence of TIn,Hot and TIn,Cold in the cooling power and COP
phenomenon occurred because during the mass recovery, the of the 68 min cycles with heat and mass recovery can be seen in
temperature of the adsorbent that adsorbed ammonia increased Fig. 6. For the values of cooling power shown in Fig. 6a, the specific
due to the reaction heat released in the adsorption process. cooling power (SCP) ranged from 44.8 W/kg to 64.1 W/kg, whereas
the cooling power density (CPD) varied between 22.4 kW/m3 and
32.1 kW/m3.
As expected, the elevation of TIn,Cold increases both the cooling
power and the COP, because the evaporation occurs at higher
temperatures, which benefits the adsorption process. The mean
evaporation temperature of the cycle with TIn,Hot of 75 C and TIn,Cold
of 15 C was 13.3 1.0 C, whereas it was 10.3 0.8 C in the cycle
with TIn,Cold of 12 C and same TIn,Hot. For the length of time chosen in
these experiments, the increase of TIn,Hot reduces the COP because
the increment of the cooling load due to the larger amount of
refrigerant desorbed with the increase of TIn,Hot was smaller than
the increment of the heat consumed due to the higher temperature
achieved by the adsorbent beds and reactors. For example, when
TIn,Cold was 15 C, the cooling load absorbed by the evaporator
increased 19% when TIn,Hot was raised from 65 C to 75 C, but
the heat absorbed by the reactors increased 35%. Nevertheless,
Fig. 4. Performance variation with the utilization of heat and mass recovery. the cooling power also increased because there was an increment
TIn,Hot = 70 C and TIn,Cold = 15 C. of the cooling load in the cycles within this range of TIn,Hot.
R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265 263

Fig. 6. (a) Cooling power at different TIn,Hot and TIn,Cold. (b) COP at different TIn,Hot and TIn,Cold. Cycle time of 68 min with heat and mass recovery.

When TIn,Cold was 12 C and TIn,Hot varied between 70 C and Nevertheless, as most of the heat absorbed by the reactors was
80 C, the COP also decrease with the increase of TIn,Hot, but the sensible heat, the increase of the cycle length enhanced the COP
cooling power first increased when TIn,Hot varied from 70 C to because as the heating period occurred, the inlet and outlet heat
75 C, to then, decreased as TIn,Hot was raised to 80 C. The cooling source fluid temperature approached to each other, and the per-
load absorbed by the evaporator at TIn,Hot of 70 C and 75 C were cent increase of the heat consumed by the reactors was smaller
similar and respectively, 3.78 0.19 MJ and 3.84 0.19 MJ, but at than the percent increase of the heat absorbed by the evaporator.
TIn,Hot of 80 C, it dropped to 3.21 0.19 MJ. A possible explanation As expected, the highest COP and cooling power were obtained
to this fact is that the amount of ammonia adsorbed during the when TIn,Cold was 15 C instead of 12 C. When the chiller was dri-
cooling period at TIn,Cold of 12 C was relatively smaller when ven with TIn,Hot at 70 C in a 78 min cycle, it achieved its maximum
compared to that when TIn,Cold was 15 C. In the former situation, cooling power (1200 60 W) among he conditions presented in
a TIn,Hot of 70 C was enough to desorb most of ammonia within a Table 3. The highest value of COP (0.36 0.03) was also obtained
period of 30 min and hence, the values of cooling load at 70 C with the latter TIn,Hot, but cycle time of 88 min, or with TIn,Hot at
and 75 C were similar, and further increment of TIn,Hot would not 65 C and cycle time of 78 min.
increase significantly the amount of ammonia desorbed and
adsorbed. Moreover, the adsorbed amount during the cooling period 3.4. Performance comparison
of the reactors was not a function of only TIn,Cold but also a function
of TIn,Med,Rc, which was higher and had larger standard deviation in A large number of the adsorption chillers presented in the liter-
the experiment where TIn,Hot was 80 C. A higher value of TIn,Med,Rc ature were tested to operate with heat sources of 80 C or above, as
with larger standard deviation implies in lower adsorbed amount in these temperatures, those chillers may show their best perfor-
during the cooling period of the reactor and for such a reason, mance. Moreover, in those studies, the sink temperature usually
the evaporator absorbed less heat when TIn,Hot was 80 C. varied between 25 C and 35 C, the inlet temperature in the evap-
orator may had varied between 10 C and 20 C, and in some cases,
3.3. Influence of cycle time in the COP and cooling power the evaporation temperature was informed instead of the chilled
water temperature. Hence, a direct comparison between different
Fig. 7 shows the cooling power and COP obtained at cycles with machines is hardly possible, and Table 4 is presented with the
heat and mass recovery, but which had different length of time. purpose to shown how the performance of the chiller discussed
For the values of cooling power shown in Fig. 7a, the SCP ranged in this work stands among the performance of other chillers
from 41.9 W/kg to 57.2 W/kg, whereas the CPD varied between described in the literature.
21.0 kW/m3 and 28.6 kW/m3. From the data presented in Table 4, one can conclude that the
In the cycle with TIn,Hot of 70 C and TIn,Cold of 12 C, the increase chiller presented in this work has COP and SCP compatible to those
of the cycle time reduced the cooling power in up to 9% and obtained by other machines that operate with heat source temper-
increased the COP in 18%. As previously mentioned, under the atures up to 75 C. From this table, it is also possible to realize that
above mentioned temperature conditions, a heating period of even machines with the same type of adsorbent can have a consid-
30 min was enough to desorb most of the ammonia from the erable different performance. As the chiller presented in this work
adsorbent bed, and the decrease in cooling power indicates that was effectively driven by hot water at 65 C with COP similar to
the percent increase of the cycle length was higher than the that obtained by machines driven by hot water at higher tempera-
percent increase of the cooling load absorbed in the evaporator. ture, the utilization of composite sorbent of NaBr impregnated in

Fig. 7. (a) Cooling power at different cycle time. (b) COP at different the cycle time. Three selected operation conditions with heat and mass recovery.
264 R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265

Table 4
Performance comparison with the chiller presented in this work and in previous ones.

Pair TIn,Med,Rc (C) TIn,Hot (C) TIn,Med,Cn (C) TIn,Cold (C) COP SCP Ref.
NaBr/EG-NH3 30 70 28 12 0.32 46.6 This work
NaBr/EG-NH3 30 70 28 15 0.34 64.1 This work
NaBr/EG-NH3 30 65 28 15 0.36 57.6 This work
Silica-gelH2O 28 70 28 10 0.27 52 [28]
Silica-gelH2O 31 73 31 16 0.43 69 [12]
Silica-gelH2O 31 75 31 16 0.36 62.7 [11]
Silica-gelH2O 28 80 28 7 0.33 92 [28]
Silica-gelH2O 28 85 28 12 0.20 38.3 [29]
Silica-gelH2O 28 87 28 12 0.55 150 [30]
Silica-gel/Ca(NO3)2H2O 30 80 30 15 0.23 245 [31]
Silica-gel/LiClH2O 30 87 30 12 0.23 160 [32]
Zeolite AQSOATM FAM Z02-H2O 28 90 28 12 0.41 600 [13]

expanded graphite in different chillers design may lead to the


developed of chillers more suitable to be driven by hot water
heated in flat plate solar collectors or in other sources of hot water
with temperatures limited to less than 80 C.
It is worth mention that when compared to the chillers that
have water as refrigerant and operate under vacuum, the chillers
that operate with pressure above atmospheric, as the one pre-
sented in this work, have the advantage that their performance is
not affected by minor refrigerant leakages. In the machines that
have water as refrigerant, periodically purge must be conducted
to remove the infiltrated air, which reduces the performance of
the machine.
Fig. 9. Cooling power and COP through successive cycles, with variable inlet
3.5. Variation of COP and cooling power at variable evaporator temperature in the evaporator.

temperature

Once the cooling power and COP were assessed at different Table 3. The cooling power and COP in the latter experiment were,
cycle time, TIn,Hot and TIn,Cold, these performance indicators were respectively, 930 50 W and 0.28 0.03, but in the experiment
then, assessed under a condition of constant cycle time and with variable TIn,Cold, these performance indicators were, respec-
TIn,Hot, but variable TIn,Cold. This type of condition occurs when the tively, 1690 60 W and 0.41 0.03. In both experiments, the outlet
chiller is used to cool a certain amount of water from a specified temperature from the evaporator was close to 10.5 C at the end of
temperature to a lower value, instead of remove a certain heat load the cycle, but in the experiment that the inlet temperature to the
from an air-conditioned space. evaporator was kept constant at 12.0 C, the cooling power of the
Fig. 8 shows the variation of the inlet and outlet temperature chiller was 45% smaller and the COP was 32% smaller, when com-
through the evaporator in successive cycles. The inlet water to pared to those figures obtained when the inlet temperature varied.
the evaporator (TIn,Cold) came from the top of a 125 L insulated tank, As the TIn,Cold decreased in the successive cycles, the mean cool-
while the outlet water from the evaporator returned to the bottom ing power and mean COP also decreased, until they reached the
of this tank. The cooling power and COP in the successive cycles values of respectively, 710 40 W and 0.19 0.02, when the outlet
can be seen in Fig. 9. temperature from the evaporator reached 3.8 C at the end of the
From the beginning to the end of the first cycle, the inlet water fifth cycle.
temperature dropped from 23.5 C to 12.3 C, and the outlet
temperature was 1.5 C lower, as in the experiment conducted at 4. Conclusions
similar conditions, but constant TIn,Cold of 12.0 C, as shown in
We assessed the COP and cooling power of a chemisorption
chiller at different operation conditions and from the analysis of
the results we obtained the following findings:

 The heat recovery process increases the COP in about 32%, but it
decreases the cooling power in about 11%, because there is no
cooling production during the period in which the heat recovery
occurs. When mass recovery occurred simultaneously to heat
recovery, the increment of COP reached 44%, and the decrease
in the cooling power was only 5%.
 Mass recovery simultaneously to heat recovery process
improved the COP in 8.711.7% and the cooling power in 5.8
15.4%, when compared to the values obtained in the cycles with
only heat recovery. These results indicate that mass recovery is
appropriate to enhance the performance not only in physical
adsorption processes, as well described in the literature, but
Fig. 8. Inlet and outlet temperatures in the evaporator through successive cycles. also in chemical adsorption processes.
R.G. de Oliveira, D.J. Generoso / Applied Energy 162 (2016) 257265 265

 When the cycle time was 68 min, maximum cooling power and [9] Wang RZ, Oliveira RG. Adsorption refrigeration an efficient way to make good
use of waste heat and solar energy. Progr Energy Combust 2006;32(4):42458.
COP were obtained when the reactors were fed with hot water
[10] Saha BB, Akisawa A, Kashiwagi T. Solar/waste heat driven two-stage
at, respectively, 75 C and 65 C. adsorption chiller: the prototype. Renew Energy 2001;23(1):93101.
 Longer cycle time enhanced the COP in all operation conditions [11] Xia ZZ, Wang RZ, Wang DC, Liu YL, Wu JY, Chen C. Development and
tested, but decreased the cooling power in some operation comparison of two-bed silica gelwater adsorption chillers driven by low-
grade heat source. Int J Therm Sci 2009;48(5):101725.
conditions. [12] Chen CJ, Wang RZ, Xia ZZ, Kiplagat JK, Lu ZS. Study on a compact silica gel
 The chiller with composite NaBr-ammonia working pair when water adsorption chiller without vacuum valves: design and experimental
driven by hot water at 65 C could achieve COP similar to that study. Appl Energy 2010;87(8):267381.
[13] Vasta S, Freni A, Sapienza A, Costa F, Restuccia G. Development and lab-test of
obtained by chillers with other working pairs, but driven with a mobile adsorption air-conditioner. Int J Refrig 2012;35(3):7018.
hot water at higher temperatures. [14] Oliveira RG. Solar powered sorption refrigeration and air conditioning. In:
 When the inlet cold water to the evaporator varied with the Larsen ME, editor. Refrigeration: theory, technology and
applications. Hauppauge: Nova Publishers; 2010. p. 20538.
time, the chiller driven by hot water at 70 C could chill water [15] Lepinasse E, Marion M, Goetz V. Cooling storage with resorption process.
from 23 C to 10 C in about 1 h and half with mean cooling Application to a box temperature control. Appl Therm Eng 2001;21:125163.
power of 1400 W and COP of 0.33 or from the same initial tem- [16] Bao HS, Oliveira RG, Wang RZ, Wang LW. Choice of low temperature salt for a
resorption refrigerator. Ind Eng Chem Res 2010;49(10):4897903.
perature to 4 C in almost 6 h, with a mean COP of 0.19 and [17] Bao HS, Wang RZ, Oliveira RG, Li TX. Resorption system for cold storage and
mean cooling power of 710 W. long-distance refrigeration. Appl Energy 2012;93:47987.
[18] Oliveira RG, Wang RZ, Kiplagat JK, Wang CY. Novel composite sorbent for
resorption systems and for chemisorption air conditioners driven by low
generation temperature. Renew Energy 2009;34(12):275764.
Acknowledgements [19] Jiang L, Wang LW, Luo WL, Wang RZ. Experimental study on working pairs for
two-stage chemisorption freezing cycle. Renew Energy 2015;74:28797.
Prof. Dr. Oliveira thanks CNPq for the productivity in research [20] Kiplagat JK, Wang RZ, Oliveira RG, Li TX, Liang M. Experimental study on the
effects of the operation conditions on the performance of a chemisorption air
grant 306364/2010-7 and for the Universal 14/2012 Grant conditioner powered by low grade heat. Appl Energy 2013;103:57180.
478654/2012-0, FAPESC for the Universal 04/2012 Grant [21] Rezk ARM, Al-Dadah RK. Physical and operating conditions effects on silica gel/
3819/2012, the company Nacional de Grafite LTDA for the expansi- water adsorption chiller performance. App Energy 2012;89(1):1429.
[22] Lu ZS, Wang RZ. Performance improvement by mass-heat recovery of an
ble graphite used in the composite sorbent, and the IFSC for the innovative adsorption air-conditioner driven by 5080 C hot water. Appl
machinery used to manufacture the chiller. Therm Eng 2013;55(12):11320.
[23] Pan QW, Wang RZ, Lu ZS, Wang LW. Experimental investigation of an
adsorption refrigeration prototype with the working pair of composite
References adsorbentammonia. Appl Therm Eng 2014;72(2):27582.
[24] Tso CY, Chan KC, Chao CYH, Wu CL. Experimental performance analysis on an
[1] Sarbu I, Sebarchievici C. Review of solar refrigeration and cooling systems. adsorption cooling system using zeolite 13X/CaCl2 adsorbent with various
Energy Build 2013;67:28697. operation sequences. Int J Heat Mass Transf 2015;85:34355.
[2] Boudehenn F, Demasles H, Wyttenbach J, Jobard X, Cheze D, Papillon P. [25] Lu YJ, Wang YD, Bao HS, Yuan Y, Wang LW, Roskilly AP. Analysis of an optimal
Development of a 5 kW cooling capacity ammoniawater absorption chiller for resorption cogeneration using mass and heat recovery processes. Appl Energy
solar cooling applications. In: Proceeding of the 1st international conference 2015. http://dx.doi.org/10.1016/j.apenergy.2015.01.138.
on solar heating and cooling for buildings and industry, San Francisco, USA; [26] Oliveira RG, Wang RZ, Wang C. Evaluation of the cooling performance of a
2012. consolidated expanded graphitecalcium chloride reactive bed for
[3] Zhai XQ, Wang RZ. A review for absorbtion and adsorbtion solar cooling chemisorption icemaker. Int J Refrig 2007;30(1):10312.
systems in China. Renew Sust Energy Rev 2009;13:152331. [27] Taylor JR. An introduction to error analysis: the study of uncertainties in
[4] Pongtornkulpanich A, Thepa S, Amornkitbamrung M, Butcher C. Experience physical measurements. 2nd ed. Salsalito: University Science Books; 1997. p.
with fully operational solar-driven 10-ton LiBr/H2O single-effect absorption 327.
cooling system in Thailand. Renew Energy 2008;33:9439. [28] Liu YL, Wang RZ, Xia ZZ. Experimental performance of a silica gelwater
[5] Zaltash A, Petrov A, Linkous R, Vineyard E, Goodnack D, Egilegor B. adsorption chiller. Appl Therm Eng 2005;25(23):35975.
Performance evaluation of a 4.5 kW (1.3 refrigeration tons) air-cooled [29] Liu YL, Wang RZ, Xia ZZ. Experimental study on a continuous adsorption water
lithium bromide/water hot-water-fired absorption unit. In: Proceedings of chiller with novel design. Int J Refrig 2005;28(2):21830.
the ASME international mechanical engineering congress and exposition, [30] Grisel RJH, Smeding SF, de Boer R. Waste heat driven silica gel/water
Seattle, USA; 2007. adsorption cooling in trigeneration. Appl Therm Eng 2010;30(89):103946.
[6] Erickson DC. 100 ton absorption chiller/heat pump demonstrates the real cost [31] Freni A, Sapienza A, Glaznev IS, Aristov YI, Restuccia G. Experimental testing of
of saving energy.. ASHRAE Trans 2007;113(Pt 2):904. a lab-scale adsorption chiller using a novel selective water sorbent silica
[7] Kwak HY, Choi CY. Demonstration study of solar absorption cooling system modified by calcium nitrate. Int J Refrig-Rev Int Du Froid 2012;35(3):51824.
with evacuated tubular solar collectors. In: Proceedings of ISES solar world [32] Zheng X, Ge TS, Wang RZ, Hu LM. Performance study of composite silica gels
congress 2007: solar energy and human settlement, vols. IV; 2007. p. 25215. with different pore sizes and different impregnating hygroscopic salts. Chem
[8] Castro J, Oliva A, Perez-Segarra CD, Cadafalch J. Evaluation of a small capacity, Eng Sci 2014;120:19.
hot water driven, air-cooled H2OLiBr absorption machine. HVAC&R Res
2007;13(1):5975.

You might also like