Solar Energy 155 (2017) 130–141

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Experimental evaluation of single stage ejector-absorption cooling cycle

under different design configurations
Azher M. Abed a,b,⇑, K. Sopian a,⇑, M.A. Alghoul c,d,⇑, Hasan Sh. Majadi b, Ali Najah Al-Shamani a,e
a
Solar Energy Research Institute (SERI), Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia
b
Department of Air Conditioning and Refrigeration, Al-Mustaqbal University College, Babylon, Iraq
c
Energy and Building Research Center, Kuwait Institute for Scientific Research, Safat 13109, Kuwait
d
Center of Research Excellence in Renewable Energy (CoRERE), King Fahd University of Petroleum and Minerals (KFUPM), Dhahran 31261, Saudi Arabia
e
Al-Musaib Technical College, Al-Furat-AlAwsat Technical University, 51009 Babylon, Iraq

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

Article history: Experimental performance of a solar assisted single stage absorption cooling system with three different
Received 17 June 2016 design configurations operating on ammonia-water have been conducted. The system was developed and
Received in revised form 31 March 2017 tested under the tropical climate conditions of Malaysia. The prototype cooling capacity was between 3
Accepted 4 June 2017
and 5 kW with 40 evacuated tube collectors sloped at 14° and orientated towards the south. The exper-
imental unit has been designed to operate using three different configurations namely (a) single stage
absorption cycle with single ejector (b) single stage absorption cycle with single ejector and flash tank,
Keywords:
and (c) single stage absorption cycle with dual ejectors and flash tank. Results indicated that the dual
Solar-assisted absorption cycle
Single ejector
ejectors configuration with flash tank has lower generator thermal loads and higher cooling effect com-
Dual ejectors pared to the other two configurations. The solar absorption system with dual ejectors and flash tank has
Flash tank the highest thermal COP between 0.2338 and 0.465. The COP for the basic configuration and single ejector
Thermal COP were between 0.127–0.282 and 0.17–0.362 respectively. The theoretical model for the dual ejector flash
Evacuated tube collector tank cycle was validated by experimental results.
Ammonia-water solution Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction
The ejector offers an attractive solution for the operation pro-
cess of low-grade energy absorption cycle for cooling due to the
simple design, low cost and without moving part (Cizungu et al.,
2005; Wongwises and Disawas, 2005). However, the basic mecha-
nisms that make up the ejector operation are quite intricate due to
interactions during the combination of two fluid streams in sub-
sonic and supersonic conditions. This precludes the fact that the
geometry of the ejector should be designed with extra care to max-
imize the overall efficiency. Numerous theoretical and experimen-
tal studies with ejectors have been conducted to characterize the
cooling system performance (Chen et al., 2014; Dennis and
Garzoli, 2011; Li et al., 2013; Nguyen et al., 2001). (Sankarlal and
Mani, 2006); Sankarlal and Mani (2007) carried out an experimen-
tal study on a vapor ejector refrigeration system operate with
⇑ Corresponding authors at: Solar Energy Research Institute (SERI), Universiti
Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia. Department of Air
Conditioning and Refrigeration, Al-Mustaqbal University College, Babylon, Iraq
(A.M. Abed).
E-mail addresses: azhermuhson@gmail.com (A.M. Abed), ksopian@ukm.edu.my
(K. Sopian), dr.alghoul@gmail.com (M.A. Alghoul).
ammonia. It was found that the entrainment ratio and coefficient
of performance of the system increase with increase in ejector area
ratio. Moreover, the expansion ratio increase with decrease in the
compression ratio.
The application of ejector in solar refrigeration system using
different working fluids were reported by (Abdulateef et al.,
2009; Chen et al., 2013). (Ma et al., 2010) conducted an experimen-
tal study on a solar-assisted ejector cooling system. The primary
flow of the ejector was controlled using a spindle to affect fine-
tuning for ejector operation and realize optimal COP. The influ-
ences of spindle position, boiler temperature, and evaporator tem-
perature upon the system’s performance were assessed as well.
Maximum entrainment ratio (0.34) and COP (0.323) were obtained
from spindle position of 8 mm with a cooling capacity of 2.05 kW.
The COP of the ejector-absorption cooling system is highly reliant
upon the entrainment of ejector (i.e. the ratio of secondary flow
rate to primary flow rate) (Sun, 1997). Entrainment is relative to
the primary flow inlet state, secondary flow inlet state, and mixing
outlet state of the ejector. The three primaries, secondary, and
backpressures are main factors that influence the entrainment
ratio (Aphornratana, 1995; Eames et al., 1995; Huang et al., 1985).

http://dx.doi.org/10.1016/j.solener.2017.06.005
0038-092X/Ó 2017 Elsevier Ltd. All rights reserved.
 A.M. Abed et al. / Solar Energy 155 (2017) 130–141
Nomenclature
At cross sectional area at nozzle throat (m2 )
COP coefficient of performance
COPth thermal coefficient of performance
Cp specific heat (kJ/kg °C)
h enthalpy (kJ/kg)
m_ mass flow rate (kg/s)
P pressure (kPa)
Q thermal load (kW)
T temperature (°C)
v specific volume (m3 = kg)
wp pump work (kW)
x flow entrainment ratio, secondary stream to primary
stream
subscripts
a absorber
c condenser
The principal modifications via the addition of a two-phase
ejector and a liquid–vapor separator (flashing liquid) were studied
in single/multi-stage compression and multi-evaporator cycles
(Aprile et al., 2009) and (Lin et al., 2012). (Kornhauser, 1990) first
studied the ejector expansion refrigeration cycle using R12 as a
refrigerant, which resulted in a COP improvement of
21% over
the standard cycle under standard operating conditions. This is
attributed to the fact that the ejector utilizes kinetic energy gener-
ated by the flash gas to elevate suction pressure from the compres-
sor. (Elbel and Hrnjak, 2008; Xu and Ma, 2010; Shuxue and
Guoyuan, 2011) utilized an ejector in a quasi two-stage compres-
sion heat pump system coupled with vapor injection compressor.
For a multi-evaporator cycle, (Tomasek and Radermacher, 1995)
and (Elakdhar et al., 2007) proposed a compression–ejection
hybrid cycle for domestic refrigeration to decrease energy losses
due to the large temperature gradient between the fresh food sec-
tion and the freezer section. The ejector and separator have been
use to elevate the suction pressure of the compressor and undercut
the compressor pressure ratio. (Xu et al., 2012) analyzed the trans-
critical CO2 heat pump cycle with adjustable ejector. It was conclu-
sively proven that the high-side pressure positively influence the
performance of a system and ignored lower ejector efficiencies,
which falls within 20–30%. Sarkar (2012) reviewed the various
ejector technologies enhanced vapor compressor system configu-
rations, and their performance characteristics have been reviewed
by Sarkar (2012).
This paper is a continuation of the authors’ previous work. The
authors have studied theoretically the potential of enhancing the
COP of absorption cooling cycle, by adding an ejector (Abdulateef
et al., 2009), and then by adding flash tank to the ejector-
absorption cycle (Sirwan et al., 2013a, 2013b). Next, (Abed et al.,
2015b, 2015c) studied the effect of the solution streamlines, heat
recovery and adding a refrigerant heat exchanger (RHE) on the
COP of the ejector-flash tank cycle and enhanced the COP by avoid-
ing the use of flash gas valve or booster in the cycle. This will allow
the ejector to work only under the intermediate pressure of the
flash tank (Abed et al., 2015a). Recently, (Abed et al., 2016) devel-
oped the mathematical model and optimized the performance of
the cycle using two ejectors. The results indicated that the overall
theoretical COPs increment in dual ejector-flash tank system were
11.56%, 12.42%, 13.46% and 14.05% at generator temperature of
80 °C, 85 °C, 90 °C, and 95 °C respectively over the single ejector
flash tank cycle.
131
g generator
amb ambient
hot hot water
coll collector
ch chilled water
cw cooling water
w water
i inlet
o outlet
abs absorber
cond condenser
gen generator
evp evaporator
This paper presents the experimental studies of the perfor-
mance of three different design configuration including the single
stage ejector–absorption system, ejector –flash tank absorption
system and dual ejectors –flash tank absorption system. Further-
more, the comparisons between theoretical and experimental
results have been performed to determine the validity of the theo-
retical model.
2. Theoretical model
The theroretical model has been developed for evaluating the
performance of dual ejector –flash tank cycle by (Abed et al.,
2016). The detailed thermodynamic property equations of
NH3 =H2 O have been obtained using the Engineering Equation Solver
(EES) software (Klein and Alvarado, 2002). The working fluid, which
is a binary mixture of ammonia and water, requires three indepen-
dent parameters to define its state at any location in the system. For
these state points, the additional required parameter is typically
either concentration or quality. The quality at any given state was
chosen as a third independent parameter (wherever appropriate),
and other properties such as enthalpy and specific volume are typ-
ically obtained using these three known independent parameters.
(In some instances, the enthalpy obtained from energy balances is
used as an input to compute quality or concentration, depending
on the specific state point under consideration.) Also, qualities of
0 or 1 as appropriate have been used to obtain concentrations at
two locations in the test rig from measured temperatures and pres-
sures: generator outlet, rectifier vapor outlet. Such state points,
measured temperatures and pressures, and the expected quality
(e.g., saturated liquid or saturated vapor) are used to obtain the
solution concentration. The energy balance equations adopted in
this study were the equations used in the previous work of (Abed
et al., 2016). To evaluate the COP for the cycle, the first law of ther-
modynamics yields the energy balance of each component (each
component can be treated as a control volume with inlet and outlet
streams, and work interactions) of the absorption system as follows:
- Mass balance
X X
m_e¼ m_i ð1Þ
e i
- Pump
132 A.M. Abed et al. / Solar Energy 155 (2017) 130–141

The pump operation on the strong solution flowing from the are the generator, rectifier, three ejectors, condenser, flash tank,
absorber per unit of refrigerant mass can expressed as: evaporator, absorber, solution heat exchanger, pumps and expan-
sion valves. Stainless steel has been used as component material
_ s v s ðPhigh  Plow Þ
wp ¼ m ð2Þ in the system. This will ensure that no reaction will occur between
The entrainment ratio ðxÞ of the ejector can be determined ammonia–water solution and the metal and thus reducing the risk
from the following relationship of corrosion.
 
At
x ¼ f Psecondary ; Tsecondary ; Pprimary ; Tprimary ; Pcond ; ð3Þ 3.1. Design and testing of the absorption system
Ak
Detailed formulations of the above function have been The experimental prototype of the NH3–H2O solar assisted
described in (Abed et al., 2016). absorption cooling cycle is shown in Fig. 2. Three ejectors added
The thermal loads of the components of the combined dual to the experimental test rig to enable the system to work individ-
ejector-flash tank cooling system per unit of refrigerant mass are ually under three modes of operation namely, (a) basic absorption
expressed as follows: cycle with single ejector (b) basic absorption cycle with single ejec-
X X tor and flash tank, and (c) basic absorption cycle with two ejectors
QK ¼ _ e he 
m _ i hi
m ð4Þ and flash tank. In the first configuration, the secondary streamline
of the ejector working only under low pressure of the evaporator,
where Q K is the heat added to component K at temperature T K . and in the second configuration, the secondary streamline of the
COP is used to measure the system performance and is defined ejector utilized the evaporator and the flash tank to work under
as low pressure of the evaporator. In the third configuration,
QE ejector-1 and ejector-2 were working under low pressure of the
COP ¼ ð5Þ evaporator and intermediate pressure of the flash tank respec-
Q G þ wp
tively. In these ejectors, the primary vapor at the high pressure
leaving the generator enters the supersonic nozzle of the ejectors
3. Experiment setup and mixing with secondary streamlines then, passes to the con-
denser. The sub-cooled liquid ammonia then flows to the flash tank
The schematic diagram of the experimental setup is shown in vessel. At the flash tank, as a result of pressure reduction to the
Fig. 1. The experimental system has four loops: hot water (by solar intermediate level, a certain amount of vapor separate from the liq-
thermal collectors), coolant water, chilled water and: ammonia- uid stream and sucked by the ejector. The rest of the liquid ammo-
water. The main components of the absorption cooling system nia enters the evaporator. Inside the evaporator the liquid

Rectifier P2
P1 T1 T2
V-1
Ejector-1
1 F1 V-34
T3
P19 T19 Ejector - 2
P3 P-74
T31 CONDENSER
P-80 GENERATOR
2 V-2
3
T32

F19 P4 T16
Ejector-3
F30 T30
4
T24

P-10
T4
P-81 P-82 T15
V-43

T20
T14
T

PUMP
Butterfly Valve

WATER
STORAGE TANK
11 10 Water supply
P5 T4"

P14
T23

Ammonia supply
P11

V-4
RELIEF T
VALVE
Reducing F4"
T10

F15
F14 valve
Drain P-12
P-13
Ammine
storage tank
FLASH TANK
T7'

P-9
T4
T6

HEAT
T9 F7
EXCHANGER V-46

F5
F12
V-8
T5

12 8
T12

T7
V-42
T7

Expansion valve

Flow Meter P-5

V-16

Vapor Flow Meter P8

ABSORBER EVAPORATOR
P-16
Temperature Sensor
T18

T17

T8

Pressure Gage F21

T22

T26
P21

Butterfly Valve Liquid Refrigerant

Vapor Refrigerant
PUMP P-18
Expansion Valve
T21

COOLING TOWER F17 Strong Solution

T25
Drain WATER TANK
P17 COIL FAN PUMP-1 Weak Solution
Relief valve
T27

V-6 F27 Hot Water

Reducing valve Water Cycle

Check valve PUMP Chilled Water

V-9

Fig. 1. Schematic diagram of multi ejectors-flash tank- absorption cycle.

 A.M. Abed et al. / Solar Energy 155 (2017) 130–141
Ejector -3
Ejector -2
Ejector -1
From the generator
From the evaporator
Fig. 2. Major components of single stage absorption cooling cycle and their locations in the experimental prototype.
ammonia sprays on the tube bundles using full cone nozzles to
make sure that all the tubes inside the shell heat exchanger is wet-
ted by the ammonia droplets. The minimum operating pressure of
the nozzles was 4 bar inside the evaporator. In the evaporator the
liquid refrigerant extracted the heat from the chilled water which
is circulating inside the tubes. The liquid ammonia vaporized
inside the shell and produced the required cooling effect. The
absorption process was carried out inside the absorber. A vertical
multistage circulation pump was used to circulate the ammonia
water solution between the absorber and the generator. The nom-
inal ammonia water solution concentrations at the absorber outlet
varied between 41–53% with mass flow rate between 0.01–
0.029 kg/s.
3.2. Ejector design and geometries
The design of a suitable ejector is the most complicated task in
this work. The ejector was designed based on theoretical model
adopted by (Abed et al., 2016). Three ejectors have three nozzles
with different area ratios and the ejectors main body fabricated
from stainless steel. Fig. 3 shows the characteristic dimensions of
the ejector main body used. Table 1 shows the important geome-
tries for the ejector.
3.3. Solar thermal system
Solar thermal hot water system has been used to provide the
required thermal energy to the absorption cooling cycle. The pri-
mary components of hot water system include the solar collectors,
storage tank, circulating pump, valves and connecting pipes. Forty
evacuated tubes collectors using VARISOL technology [thermomax
heat pipe tube] have been used. The hot water from the generator
was circulated to the storage heat exchanger. The primary loop
pump can provide temperature within (60–98 °C) originated from
the solar collectors. The mass flow rates were measured using
magnetic flow meters. The flow rate in the primary loop has been
varied between (10–45 L/min). The flow rate of the secondary loop
133
To the condenser
Vapor flow meter
From the flash tank Flash tank
Valve
has been varied between (7–18.5 L/min). All temperatures have
been measured by K-type thermocouples having accuracy ±0.4%
between (40 to 375 °C). The pipes were well insulated again
the heat loss to the environment. Fig. 4 shows the photo of the
hot water sytem connceted with the absorption cooling cycle.
3.4. Experimental procedures
The test procedures, precautions and safety issues before, dur-
ing and after experimentation are important since the system con-
tains ammonia and operates at pressures higher than ambient.
Extensive leak tests were performed before charging the system.
All the installation, including connecting pipes was supplied with
thermal insulation, in order to minimize heat losses. Data collec-
tion started from 3 November 2015 until 21 December 2015. Each
experiment was conducted from 8:00 am until 5:00 pm. The mass
flow rates of the hot water subsystem passing through the primary
cycle and secondary cycle, the mass flow rates in the absorption
cycle, the cooling water cycle, and the chilled water cycle were
measured using standard flow meters. The pressures at different
location in the entire unit were measured using pressure gages.
The fluid temperatures at the inlet and outlet of all components
were measured using thermocouple type-K connected to data
acquisition system. Mass flow rate were also varied according to
the design parameters. Table 2 shows the accuracies and toler-
ances of the sensors provided by the manufacture.
Because there are many test days in this work, it is very hard to
present the working conditions of each test day. Table 3 shows the
ranges of the climatic and operational conditions. In fact, the per-
formance variation of the system was mainly affected by the exist-
ing operating conditions during the day of test. The cooling
capacity was design at 5 kW. System operation can be adjust man-
ually to enable different design configuration mode and different
mass flowrates.
To calculate the thermal performance of the cycles, the temper-
ature and flow rates of the hot water, cooling water, and chilled
water cycles that enter and exit the generator, condenser, absorber,
134 A.M. Abed et al. / Solar Energy 155 (2017) 130–141

CHAMBER SECTION CONSTANT AREA SECTION DIFFUSER

29,22 18,3 44,91 73

Throat (At)

D2
Primary Flow

15,5

36°

10°
D1
31

D3

D4

44
35

64
NXP

Secondary Flow

Fig. 3. Dimension of the ejector used in the experimental setup.

Table 1
Summary of the ejector geometrical details.

Ejector D1 De D3 D4 NXP
E1 2.64 4.92 5 19.27 8
E2 2.85 5.13 5.5 19.27 8
E3 2.6 8 8.1 22 6

Fig. 4. Solar thermal system connected with the absorption cooling cycle.

and evaporator were measured. The thermal energy of the main _ cw CP ðTccwi  Tccwo Þ
Q cond ¼ m ð7Þ
components in both cycles was evaluated by using the following
equations: _ cw CP ðTacwi  Tacwo Þ
Q abs ¼ m ð8Þ

_ hotw CP ðTghotw  Tghotwo Þ

Q gen ¼ m ð6Þ _ chw CP ðTchwi  Tchwo Þ
Q evp ¼ m ð9Þ
i
 A.M. Abed et al. / Solar Energy 155 (2017) 130–141 135

Table 2
Specification of the sensors/instruments used in the experimental prototype.

Sensor/instruments Quantity Model Uncertainty

Thermocouples 38 Type K- thermocouple ±0.4% between (40–375) °C
Pressure gages 8 Tecsis (P2325) and (P1778) ± 1% of span
Pressure gages 5 Winters (PFP Premium Stainless Steel Liquid) ±1.5% of full scale
Liquid flow meter 8 BFLZZ25S ±1.5% of full scale
Vapor flow meter 3 LUGB–2310 Liquid ± 1% of reading, gas and steam ± 1.5% of reading
Data acquisition/AD converter 1 ADAM-4018 and 4019 ±0.1% or better

Table 3
Main input parameters and working condition ranges during the different modes of
operation.
Parameter Working
range
Heat source temperature 65–100 °C
Generator pressure 10–18 bar
Evaporator pressure 2–6 bar
Condenser and absorber temperature 25–45 °C
Weak liquid solution mass fraction 25–45%
Cooling water mass flows rate at the absorber and 6–18 L/min
condenser
Chilled water mass flow rate (evaporator) 6–15 L/min
Hot water mass flow rate (generator) 7–18.5 L/min
Solution mass flow rate 1–5 L/min
The consumption of the pumps is usually negligible compared
to the generator capacity (typically less than 1% of the generator
heat input), and is therefore neglected (Chunnanond and
Aphornratana, 2004; Eames et al., 1995). Hence, the thermal COP
is the ratio between the cooling capacity of the system and the
heating power delivered to the system by the solar collectors,
directly or indirectly through the storage vessel.
Q evp
COPth ¼
Q gen
4. Results and discussions
Experimental work carried out to evaluate the performance of
absorption cooling systems under three modes of design configura-
tion. However, the data from a few tests were presented here since
the steady state conditions were difficult to reach. Various practi-
cal problems could not allow testing of absorption cooling system
under a wide range of conditions.
4.1. Comparison of theoretical and experimental results
The theoretical (Abed et al., 2016) and experimental results
were compared for the modified dual ejector –flash tank absorp-
tion cooling system. In order to confirm the agreement between
the theoretical and experimental results, the equilibrium and ther-
modynamic solution properties of ammonia/water mixture were
calculated from the internal library in EES (Klein and Alvarado,
2002), which utilized correlations developed by Ibrahim and
Klein (1993). The correlations used for the solution mixture prop-
erties have less than 5% average deviation from the measured data,
and these correlations are valid for pressures between 0.2 and
11 bar and temperature between 230 and 600 K. The input param-
eters of the theoretical model were obtained from the correspond-
ing experiment at different test days (climatic and operation
conditions) such as temperature, pressure, and flow rates of the
weak solution, vapor, and coolant. Fig. 5 (a-c) illustrate the exper-
imental and theoretical COP, condenser thermal load and evapora-
tor thermal load results respectively at different test days. The
results revealed that relative errors between the experimental
and theoretical results of COP, condenser thermal load and evapo-
rator thermal load are 3.25–11.90%, 4–18% and 3.8–19.6%
respectively.
The increases in the relative errors between theoretical and
experimental results of the condenser and evaporator thermal
loads can be associated with the fact that there are many experi-
mental issues are difficult to control, especially the generator tem-
perature and the efficiency of the generator and the evaporator.
Since the measured experimental values of temperature, pres-
sure, concentration, and flow rates of weak solution, vapor, and
coolant will be input data into the theoretical model; thus, uncer-
tainties in these measurements results in uncertainty in the theo-
retical results. Using the uncertainty values for the various
measurements and their effects on system components were com-
puted using an error propagation method (Taylor and Kuyatt,
1993) and presented in table 4. The uncertainty in the COP calcu-
lated using the input data from the experimental test ranged
between 2.94% to 3.8589% with an average uncertainty of 3.399%.
According to the input data presented in the experimental tests,
and the data obtained by the theoretical model of the same oper-
ating conditions, it can be conclude that the theoretical model
developed is valid. Table 5 lists the operating parameters with their
nominal measured values and uncertainties in measurement as an
ð10Þ
example of one tested days.
However, for further confirmation, a statistical agreement test
was performed to determine whether the theoretical and experi-
mental results are at a confidence level of agreement versus the
test days. The statistical EFF test was proposed as the best overall
measure of agreement (Mayer and Butler, 1993). The test passes if
the EFF values are within [0–1]. If EFF value is negative, there is no
agreement.
Mean Square Error
EFF ¼ 1 
ðVarianceÞexp
The obtained EFF test values for COP, Qevp and Qcon were at 0.80,
0.64, and 0.17 respectively. This means that the evaluation param-
eters of absorption cooling system using two ejectors passed the
EFF test and concluded that the theoretical and experimental
results were at a confidence level of agreement.
4.2. Potential of the proposed absorption cooling system using two
ejectors
The absorption cooling system using two ejectors were com-
pared experimentally with other two design configurations (single
ejector-absorption system and single ejector-flash tank-absorption
system). To achieve reasonable experimental comparison among
the three design configurations, the following steps have been
considered:
1. Data collection of experimental testing for the three experimen-
tal designs took at least 60 days as individual tests under differ-
ent sky conditions.
136 A.M. Abed et al. / Solar Energy 155 (2017) 130–141

Fig. 5. (a–c): Comparison between the experimental and theoretical results of the absorption cooling system using dual ejectors at different tested days (a) COP of the system
(b) Condenser thermal load and (c) Evaporator thermal load.

Table 4 4. Ten (10) individual testing days with almost similar climate
Uncertainties in the Current Experimental Data. conditions of single ejector-flash tank and two ejectors design
Parameter Range Uncertainty Range (%) Average Uncertainty (%) were selected.
Qabs 3–5.1 1.92–5.3 3.6
Qgen 10.34–8.78 1.07–6.27 3.67 Fig. 6 illustrates the experimental COP of different design con-
Qevp 2.8–4.7 3.57145–4.26 3.91 figurations versus several testing days. It can be seen that the dual
Qcond 3–5.6 3.34–3.57 3.46 ejectors-flash tank-absorption system has shown the highest ther-
mal COP value, followed by single ejector-flash tank-absorption
system while the lowest COP value was in the single ejector-
2. During the data collection of the three designs, the days of absorption system.
almost similar ambient temperature (trend and average), simi- The thermal loads of generator and evaporator versus generator
lar solar radiation (trend and average) were selected to compare temperature were compared at different experimental design con-
the experimental performance of these designs. figurations during days of very similar working conditions as
3. Seven (7) individual testing days with almost similar climate shown in Fig. 7. It can be seen that the absorption system with dual
conditions for the three designs were selected. ejectors has the lowest generator thermal load and highest cooling
 A.M. Abed et al. / Solar Energy 155 (2017) 130–141 137

Table 5 ejector-flash tank systems were between 0.127–0.282 and 0.170–

Lists of parameters included in uncertainty analysis of the dual ejectors cycle for one 0.362, respectively.
tested day as an example.
Fig. 9 shows the variation of evaporator temperature with the
Measured Parameter Measured Value ± Uncertainty generator temperature for the three aforementioned configura-
Flow rate tions. The dual ejectors cycle has a lower evaporator temperature
m_ coolant 0.24 ± 0.002 (high cooling effect) than those of the basic cycle with single ejec-
m_ child 0.19 ± 0.0019 tor and the single ejector-flash tank cycle under different generator
m_ hot 0.271 ± 0.002
_ ref
temperatures. As shown in Figs. 9 and 10, a slight time delay
m 0.006 ± 0.00006
Absolute pressure (bar)
between the room and the evaporator temperatures. This time
Pg 11.1 ± 0.01 delay will lead to the cooling becomes less effective. One reason
Pcon 10.6 ± 0.01 for this is that when the chilled water starts to flow through the
Pabs 4.54 ± 0.01 coil, the temperature of the room can exceed the desired limit. It
Absorber Temperatures (°C)
can also be seen that the fluctuation of the room temperature is
Tabs,c,in 30 ± 0.4 °C
Tabs,c,out 33 ± 0.4 °C due to the variation of the room loads from the heat transmission
Tabs,v,in 12.1 ± 0.4 °C from the outside (variable due to changes in ambient air tempera-
Tabs,sol,in 61.5 °C ture). Moreover, the evaporator temperature (Tevp) was affected
Condenser Temperatures (°C)
by the generator temperature and increases gradually as the gener-
Tcon,c,in 32.8 ± 0.4 °C
Tcon,c,out 37.1 ± 0.4 °C
ator temperature decreases. Another problem which may occur is
Tcon,v,in 57.3 ± 0.4 °C the rise in the weak solution temperature that results in rises of
Tcon,ref,out 27.8 ± 0.4 °C the evaporator temperature which affects the low pressure of the
Chilled water Temperatures (°C) system. Therefore, the cooling capacity will decrease with the
Tevp,ch,in 12.9 ± 0.4 °C
decreases of the generator temperature.
Tevp,ch,out 18.6 ± 0.4 °C
Generator Temperatures (°C)
Tgen,h,in 89 ± 0.4 °C
4.3. Hourly performance of solar absorption cooling system
Tgen,h,out 80.3 ± 0.4 °C

Experimental evaluations of different design configuration

effect followed by single ejector-flash tank absorption system and
finally single ejector-absorption system.
When using one ejector was used, the primary flow rate
increases, leading to a decrease in the entrainment ratio. This
decrease has been attributed to the combined effect of the
increased primary mass flow (larger heat input), and the fact that
the primary stream leaves the nozzle slightly under-expanded.
Thus, this resulted in decreased secondary mass flow. In contrast,
the cycle using dual ejectors will operate as a variable ejector,
which can function under a wide range of operating conditions.
Fig. 8 shows the thermal COP as a function of generator temper-
ature for the three experimental design configurations during days
of very similar working conditions. It can be seen that the system
of dual ejectors outperformed the single ejector-absorption and
single ejector-flash tank systems due to the fact that the generator
duty reduced and the cooling effect increased by the employment
of the dual ejector. Experimental results indicate that the thermal
performance of the dual ejectors-flash tank cycle range from 0.234
to 0.465, while the COP of single ejector-absorption and single
including the single stage ejector–absorption system, ejector –flash
tank absorption system and dual ejectors –flash tank absorption
system were carried out. Figs. 11–13 illustrated the hourly varia-
tion of the solar irradiance, generator and evaporator temperatures
for the three different design configurations. The daily solar irradi-
ance increased gradually and reach up to 1000 W/m2, then
decreased to zero in late afternoon. It can be seen from these fig-
ures that the generator temperatures start increasing in the morn-
ing hours following the variation pattern of the solar irradiance.
Fig. 11 shows that in the morning hours the temperature of the
outlet of the vapor raises slowly as the hot water temperature
increases. Therefore, the cooling effect occurred at around 1.18–
2.30 pm. The evaporator temperature (Tevp) is almost dependent
on the generator temperature and increases gradually as the gen-
erator temperature decreases. Therefore, the cooling capacity will
decreases with the decreases of the generator temperature. It can
be seen that the system temperatures were not stable due to the
instability in the hourly hot water temperature from the solar ther-
mal storage system. Moreover, the increase in liquid temperature
at the absorber leading to increases the vapor pressure of the liquid

Fig. 6. Comparison of thermal experimental COP at different design configurations versus several testing days.
138 A.M. Abed et al. / Solar Energy 155 (2017) 130–141

Fig. 7. Generator and evaporator thermal loads versus generator temperature of the three design configurations.

Fig. 8. Experimental COP versus generator temperature of the three design configurations.

solution in the absorber. Thus, the formation of gas bubbles is very
likely to occur, leading to cavitation in the solution pump. In turn,
leads to lack of sufficient strong solution to entire the generator
and consequently the released ammonia vapor in the generator
decreases.
Figs. 12 and 13 showed that the generator temperature
increases steadily, and after nearly 2 h, it stabilizes at a tempera-
ture of
90 °C. The variation of vapor temperature with respect
to time also follows a pattern similar to that of a generator. It
can also be seen that the cooling effect occurred at
11.50 am
and continued until 1.00 pm for the single ejector with flash tank
cycle as shown in Fig. 12. While, the cooling effect (Tevp) of the dual
ejectors cycle was slightly stable and continued for nearly 90 min
as shown in Fig. 13. The vapor temperature first increases to its
maximum, followed by decreasing as the generator temperature
increases, which happens in tandem with the increase in the
evaporator temperature (reduce the cooling effect). As mention
previously, the instability in mass flow rate from the solution
pump due to the cavitation problem resulting in decreased the
concentrations of ammonia to depart from the generator, decrease
the efficiency of generator heat duty, and increases rectification
losses.
5. Conclusions
Experimental studies using three different design configura-
tions of a solar assisted absorption refrigeration system using
 A.M. Abed et al. / Solar Energy 155 (2017) 130–141 139

Fig. 9. Variation of the evaporator temperature and generator temperature for the three different configurations.

Fig. 10. Variation of the Room temperature and generator temperature for the three different configurations.

Fig. 11. Daily variation of the generator and evaporator temperature for the combined ejector –single stage absorption cycle.
140 A.M. Abed et al. / Solar Energy 155 (2017) 130–141

Fig. 12. Daily variation of the generator and evaporator temperature for the combined ejector –flash tank absorption cycle.

Fig. 13. Daily variation of the generator and evaporator temperature for the combined dual ejectors –flash tank absorption cycle.

NH3 H2O as working fluid were conducted. The experimental
results showed significant improvement in the performance of
the dual ejectors –flash tank cycle at different working conditions.
The results based on the mostly clear sky conditions for each
design mode were measured and compared. It was found that
the dual ejectors configuration with the flash tank to be higher
than that of the single ejector and single ejector –flash tank absorp-
tion cooling system. The improvement in the overall COP was due
to lower energy consumption at the generator and this will
improve the quality of the refrigerant that enters the evaporator.
The efficiency of absorption cooling system depends on the tem-
perature inside the generator and hence increases the COP for
absorption refrigeration system.
A number of limitations must be considered in the analysis and
design of absorption systems powered by the low heat source tem-
perature such as solar energy. The first limitation deals with the
problem of reduced efficiency of the absorption cooling cycle dur-
ing start-up and transient conditions. In addition, the absorption
cooling cycle does not reach operating efficiency until after an hour
or more of operation time. Moreover, the system which operates
on and off regularly will drastically reduce the COP compared to
the system in steady state performance. The second limitation is
that the separation process of ammonia inside the generator
depends on the temperature to be formed inside the generator;
therefore, the geometry of distillation column inside the generator
should be designed carefully to ensure that the absorption cycle
works properly under the low heat source. For future work, the
authors will investigate thoroughly the thermal performance of
dual ejectors flash tank-absorption cooling cycle using nano-
refrigerant working fluids under variable power sources namely
electric, and renewable energy sources.
Acknowledgements
The authors would like to sincerely thank the National Univer-
sity of Malaysia for the provision of the grant INOVASI-2014-008 to
support this work.
 A.M. Abed et al. / Solar Energy 155 (2017) 130–141 141

References Klein, S., Alvarado, F., 2002. Engineering equation solver. F-Chart Software, Madison,
WI, p. 1.
Kornhauser, A.A., 1990. The use of an ejector as a refrigerant expander.
Abdulateef, J.M., Sopian, K., Alghoul, M.A., Sulaiman, M.Y., 2009. Review on solar-
Li, X., Li, X., Zhang, Q., 2013. The first and second law analysis on an organic Rankine
driven ejector refrigeration technologies. Renew. Sustain. Energy Rev. 13 (6–7),
cycle with ejector. Sol. Energy 93, 100–108.
1338–1349.
Lin, C., Cai, W., Li, Y., Yan, J., Hu, Y., 2012. The characteristics of pressure recovery in
Abed, A.M., Alghoul, M.A., Al-Shamani, A.N., Sopian, K., 2015a. Evaluating ejector
an adjustable ejector multi-evaporator refrigeration system. Energy 46 (1),
efficiency working under intermediate pressure of flash tank–absorption
148–155.
cooling cycle: Parametric study. Chem. Eng. Process. 95, 222–234.
Mayer, D., Butler, D., 1993. Statistical validation. Ecol. Model. 68 (1–2), 21–32.
Abed, A.M., Alghoul, M.A., Sirawn, R., Al-Shamani, A.N., Sopian, K., 2015b.
Ma, X., Zhang, W., Omer, S.A., Riffat, S.B., 2010. Experimental investigation of a novel
Performance enhancement of ejector–absorption cooling cycle by re-
steam ejector refrigerator suitable for solar energy applications. Appl. Therm.
arrangement of solution streamlines and adding RHE. Appl. Therm. Eng. 77,
Eng. 30 (11–12), 1320–1325.
65–75.
Nguyen, V.M., Riffat, S.B., Doherty, P.S., 2001. Development of a solar-powered
Abed, A.M., Alghoul, M.A., Sopian, K., 2016. Performance evaluation of flash tank-
passive ejector cooling system. Appl. Therm. Eng. 21 (2), 157–168.
absorption cooling cycle using two ejectors. Appl. Therm. Eng. 101, 47–60.
Sankarlal, T., Mani, A., 2006. Experimental studies on an ammonia ejector
Abed, A.M., Sopian, K., Alghoul, M.A., Al Shamani, A.N., Ruslan, M.H., Mat, S., 2015c.
refrigeration system. Int. Commun. Heat Mass Transf. 33 (2), 224–230.
Theoretical study of new combined absorption-ejector refrigeration system. In:
Sankarlal, T., Mani, A., 2007. Experimental investigations on ejector refrigeration
Paper presented at the IOP conference series: materials science and
system with ammonia. Renew. Energy 32 (8), 1403–1413.
engineering.
Sarkar, J., 2012. Ejector enhanced vapor compression refrigeration and heat pump
Aphornratana, S., 1995. Theoretical and experimental investigation of a combined
systems – a review. Renew. Sustain. Energy Rev. 16 (9), 6647–6659.
ejector-absorption refrigerator. University of Sheffield, Department of
Shuxue, X., Guoyuan, M., 2011. Exergy analysis for quasi two-stage compression
Mechanical and Process Engineering.
heat pump system coupled with ejector. Exp. Thermal Fluid Sci. 35 (4), 700–
Aprile, M., Toppi, T., Guerra, M., Motta, M., 2009. Experimental and numerical
705.
analysis of an air-cooled double-lift NH3-H2O absorption refrigeration system.
Sirwan, R., Alghoul, M.A., Sopian, K., Ali, Y., 2013a. Thermodynamic analysis of an
Int. J. Refrig. 2, 1173–1185.
ejector-flash tank-absorption cooling system. Appl. Therm. Eng. 58 (1–2), 85–
Chen, J., Havtun, H., Palm, B., 2014. Investigation of ejectors in refrigeration system:
97.
Optimum performance evaluation and ejector area ratios perspectives. Appl.
Sirwan, R., Alghoul, M.A., Sopian, K., Ali, Y., Abdulateef, J., 2013b. Evaluation of
Therm. Eng. 64 (1–2), 182–191.
adding flash tank to solar combined ejector–absorption refrigeration system.
Chen, X., Omer, S., Worall, M., Riffat, S., 2013. Recent developments in ejector
Sol. Energy 91, 283–296.
refrigeration technologies. Renew. Sustain. Energy Rev. 19, 629–651.
Sun, D.-W., 1997. Solar powered combined ejector-vapour compression cycle for air
Chunnanond, K., Aphornratana, S., 2004. Ejectors: applications in refrigeration
conditioning and refrigeration. Energy Convers. Manage. 38 (5), 479–491.
technology. Renew. Sustain. Energy Rev. 8 (2), 129–155.
Taylor, B.N., Kuyatt, C.E., 1993. Guidelines for evaluating and expressing the
Cizungu, K., Groll, M., Ling, Z.G., 2005. Modelling and optimization of two-phase
uncertainty of NIST measurement results, September (1994). NIST Technical
ejectors for cooling systems. Appl. Therm. Eng. 25 (13), 1979–1994.
Note, 1297.
Dennis, M., Garzoli, K., 2011. Use of variable geometry ejector with cold store to
Tomasek, M.-L., Radermacher, R., 1995. Analysis of a domestic refrigerator cycle
achieve high solar fraction for solar cooling. Int. J. Refrig. 34 (7), 1626–1632.
with an ejector: American society of heating. Refrigerating and Air-Conditioning
Eames, I., Aphornratana, S., Haider, H., 1995. A theoretical and experimental study
Engineers Inc, Atlanta, GA (United States).
of a small-scale steam jet refrigerator. Int. J. Refrig. 18 (6), 378–386.
Wongwises, S., Disawas, S., 2005. Performance of the two-phase ejector expansion
Elakdhar, M., Nehdi, E., Kairouani, L., 2007. Analysis of a compression/ejection cycle
refrigeration cycle. Int. J. Heat Mass Transf. 48 (19–20), 4282–4286.
for domestic refrigeration. Ind. Eng. Chem. Res. 46 (13), 4639–4644.
Xu, S., Ma, G., 2010. Air-source heat pump coupled with economized vapor injection
Elbel, S., Hrnjak, P., 2008. Experimental validation of a prototype ejector designed to
scroll compressor and ejector: Design and experimental research. Sci. China
reduce throttling losses encountered in transcritical R744 system operation. Int.
Technol. Sci. 53 (3), 782–788.
J. Refrig. 31 (3), 411–422.
Xu, X.X., Chen, G.M., Tang, L.M., Zhu, Z.J., 2012. Experimental investigation on
Huang, B., Jiang, C., Hu, F., 1985. Ejector performance characteristics and design
performance of transcritical CO2 heat pump system with ejector under
analysis of jet refrigeration system. J. Eng. Gas Turbines Power 107 (3), 792–
optimum high-side pressure. Energy 44 (1), 870–877.
802.
Ibrahim, O., Klein, S., 1993. Thermodynamic properties of ammonia-water mixtures.
ASHRAE Trans. 99. 1495-1495.