Energy Conversion and Management 154 (2017) 11–24

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Optimum design of a solar ejector refrigeration system for various operating T

scenarios
⁎
Evangelos Bellos , Christos Tzivanidis
Thermal Department, School of Mechanical Engineering, National Technical University of Athens, Zografou, Heroon Polytechniou 9, 15780 Athens, Greece

A R T I C L E I N F O A B S T R A C T

Keywords: The objective of this work is to examine a solar ejector refrigeration system under various operating scenarios.
Solar refrigeration Evacuated tube collectors of 100 m2 collecting area coupled to a storage tank of 4 m3 are selected for feeding the
Ejector generator of the refrigerator. The system is examined for different evaporator temperatures from −10 °C to 10 °C
ETC and for different heat rejection temperatures from 30 °C to 50 °C. In every case, the system is optimized by
Optimization
selecting the optimum ejector design and the optimum generator temperature. Different refrigerants are ex-
Working fluids
R141b
amined as R123, R245fa, R600a and R134a, while the emphasis is given to the R141b. The innovation of this
work is based on the systematic examination and optimization of a great range of operating scenarios giving a
clear image for the system performance. According to the final results, the R141b is found to be the best can-
didate in all the operating scenarios. More specifically, the maximum system COP is found to be 0.234 when the
system produces refrigeration at 10 °C and rejects heat to the environment at 30 °C. The optimum generator
temperatures are found to be from 114 °C to 157 °C and the cooling capacity is ranged from 1.85 kW up to
23.39 kW. The analysis is performed with a developed model in EES (Engineering Equation Solver) under steady-
state conditions.

1. Introduction design to be optimized [12].

Solar energy exploitation is one of the most promising ways for
facing the numerous threads as the climate change, the increasing en-
ergy demand and the high electricity price [1–3]. Solar energy is able to
be converted either to thermal energy with a solar thermal collector or
directly to electricity with photovoltaic cells, the fact that makes it a
flexible energy source [4,5]. Moreover, the high solar energy potential
in a great variety of courtiers makes it a crucial renewable energy
source for the present and the future.
The utilization of solar energy in refrigeration or cooling applica-
tions gains more and more attention the last years because of the high
compatibility between the solar energy supply and the demand, espe-
cially in the cooling applications [6,7]. Furthermore, solar refrigera-
tion/cooling is a technique which aids the reduction of electricity de-
mands from the grid during the summer period [8,9]. The most usual
solar refrigeration systems include sorption machines as absorption
chillers, adsorption chillers or desiccant wheels. On the other hand,
systems with ejectors are competitive to the previous technologies be-
cause of their lower cost and their lower complexity [10,11]. However,
the ejector refrigerator has usually lower performance compared to
sorption machines and thus a lot of research is conducted in order their
There are numerous studies in the literature which examines and
evaluates solar ejector refrigeration systems. Table 1 summarizes the
most important literature studies about these systems. Totally there are
33 studies in the literature [13–45] and it is obvious that the most usual
solar thermal collectors in these studies are the flat plate collectors
(FPC) [13–26] and the evacuated tube collectors (ETC) [26–40], while
there are only 5 studies about concentrating technologies [41–45].
Moreover, it is obvious that the most recent studies investigate systems
with ETCs and these systems are generally more efficient than the re-
spective with FPC. More specifically, the study of Huang et al. [26]
proved the system COP can reach up to 0.19 with FPC and up to 0.28
with ETC.
Many studies in the literature have examined systems with booster
compressor and/or with an auxiliary heater (usually boiler) in order to
achieve constant heat supply to the generator. The studies with these
modifications are highlighted in table 1, in the comment column. The
studies without extra comment investigate systems driven by solar
energy without other energy input.
The most usual refrigerants according to table 1 are the following:
R134a, R141b, R114, R600a and R123, while the water is used in a
great percentage of the studies. However, the utilization of water is

⁎
Corresponding author.
E-mail address: bellose@central.ntua.gr (E. Bellos).

http://dx.doi.org/10.1016/j.enconman.2017.10.057
Received 20 August 2017; Received in revised form 23 September 2017; Accepted 21 October 2017
Available online 05 November 2017
0196-8904/ © 2017 Elsevier Ltd. All rights reserved.
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

Nomenclature is isentropic
g generator
Acol collecting area, m2 loss heat losses
AT storage tank outer area, m2 low low pressure in the mixing region
cp specific heat capacity, kJ/kg K m mixing region
E exergy flow, kW max maximum
GT solar radiation, W/m2 mf mixing flow
h specific enthalpy, kJ/kg mn motive nozzle
m mass flow rate, kg/s n nozzle
P pressure, kPa o reference
Q heat rate, kW out outlet
T temperature, °C pf primary flow
Tsun sun temperature, K pu pump
UT tank total heat loss coefficient, kW/m2 K s heat source to the generator
(UA)g total heat transfer coefficient in the generator, kW/K sf secondary flow
V velocity, m/s sn suction nozzle
VT storage tank volume, m3 sol solar
W work, kW st storage tank
sun sun
Greek symbols sys system
st,1 storage tank first mixing zone
ΔP pressure drop, kPa st,2 storage tank second mixing zone
η efficiency, – st,3 storage tank third mixing zone
μ entertainment ratio, – u useful
ρ density, kg/m3
Abbreviations
Subscripts and superscripts
COP coefficient of performance
am ambient CPC compound parabolic collector
c condenser EES engineering equation solver
ch chiller ETC evacuated tube collector
col collector FPC flat plate collector
crit critical GWP global warming potential
d diffuser ODP ozone depletion potential
e evaporator of mechanical compression refrigerator PTC parabolic trough collector
ex exergy
in inlet

restricted only to cooling applications because the minimum evaporator
temperature has to be close to 5°C. Thus, the other refrigerants are
suitable for cooling capacity in lower temperature levels.
In the literature studies for solar cooling systems with FPC, the
impact of generator temperature on the system performance has been
extensively examined. Sokolov et al. [14,15] examined solar re-
frigeration systems with R114 as the refrigerant. They found the op-
timum generator temperature close to 110 °C in the system without
booster [14] and lower optimum values up to 100 °C for the system
with booster compressor [15]. Huang et al. [17] found that the op-
timum generator temperature is lower for higher evaporator tempera-
tures. More specifically the proved that for evaporating temperature
equal to 8 °C and to −6 °C, the optimum generator temperatures are
95 °C and 105 °C respectively. Arbel and Sokolov [19] examined a
system with booster and found optimum performance for generator
temperatures in the range of 100–105 °C. Yapici et al. [24] found the
optimum generator temperature equal to 74 °C for evaporator tem-
perature equal to 10 °C.
In a comparative study, Huang et al. [26] found that the use of
higher quality solar collector leads to higher performance. This result is
explained by the higher optimum generator temperature in the systems
with ETC or with improved FPC. These results indicate the need of
using a special collector design for driving the ejector refrigeration
systems. Among the studies with ETCs, the generator temperature op-
timization is a critical issue that has been highlighted. Guo and Shen
[31] found the optimum generator temperature close to 85 °C for a
system with an auxiliary heater. Varga et al. [32] found the optimum
generator temperature for different configuration of the ejector. Chesi
et al. [37] examined the yearly performance of a solar refrigeration
system which can operate in two modes. The first mode is the con-
ventional mechanical compression system, while the second is the
ejector refrigeration system. After a simple sensitivity analysis, they
found the optimum generator temperature close to 65 °C.
The last part of the literature studies investigates systems with
concentrating technologies and ejector refrigeration system. The use of
compound parabolic collectors (CPC) has examined only by Joemann
et al. [41]. On the other hand, more attention has been given to the
parabolic trough collectors (PTC) where there are four literature studies
[42–45]. More specifically, Pollerberg et al. [43–45] have performed
three important studies for the experimental evaluation of a system
with PTC and ejector refrigeration system without auxiliary heater or
booster. In their studies, they found the system COP to be close to 0.20,
a relatively high value compared to the respective studies of Table 1
with only solar energy as the heat source. The main reason for this
result is the high thermal efficiency of the PTCs. However, the use of
concentrating technologies in solar refrigeration systems with ejectors
has not been evolved because of the complexity of these solar collectors
which is associated with the need of continuous tracking system.
The previous literature review gives a clear image of the existing
situation in the literature about the solar ejector refrigeration systems.
It is obvious that the use of ETCs is the most appropriate solar tech-
nology because of their relatively high efficiency and their simplicity in

12
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

the operation. At this paper, a solar refrigeration system with ETC, a separated into the solar field, the storage tank and the refrigerator.
storage tank and a typical ejector refrigerator is investigated. The in- According to the literature review, the utilization of evacuated tube
novation of this work is the optimization of the system by taking into collectors (ETC) is the most usual and appropriate technology for the
account the generator temperature and simultaneously the optimum examined system. The collecting area has been selected at 100 m2 in all
ejector design. The optimum ejector design is achieved by optimizing the study, while the storage tank at 4 m3, a reasonable value for solar
the pressure drop of the secondary flow inside the ejector, something refrigeration systems [47]. The temperature levels in these systems are
that has been also stated by Li et al. [46]. Moreover, the optimization usually reaching close to 150 °C and thus thermal oil is selected in the
procedure is performed for a great range of operation scenarios, as well collector loop and in the storage tank system. Therminol-VP1 is the
as for five different refrigerants which is an extra innovative point. So, selected working fluid which can operate with a safety up to 400 °C
this study comes to present the optimum design of a solar refrigeration [48]. Inside the storage tank, there is thermal oil which stores thermal
system for various operating conditions, for various refrigerants by energy. At the down part of the storage tank, there is relatively cold
optimizing simultaneously the ejector design and the generator tem- thermal oil (Tcol,in) which goes inside the solar field. This quantity
perature. According to our knowledge and the previous detailed lit- absorbs heat from the incident solar energy and its temperature in-
erature review, there is no other study which gives similar results with a creases (Tcol,out). This hot thermal oil enters in the upper part of the
respective optimization procedure. storage tank. On the other side of the storage tank, hot thermal oil of
temperature (Ts,in) leaves the tank and goes to the generator where heat
is given to the refrigerant and the thermal oil temperature decreases to
2. Materials and methods
(Ts,out). This relatively cold quantity returns to the down part of the
storage tank and the thermal oil loop closes.
2.1. The examined configuration
The refrigerator includes a condenser, an ejector, a generator, an
evaporator and a pump. The refrigerant absorbs heat from the hot
At this subsection, the examined solar refrigeration system is de-
thermal oil in the generator and leaves the generator (state point 1) as
scribed with details. Fig. 1 illustrates the examined system which is
saturated vapor. At this study, the superheating in the generator has
been assumed to be negligible, after conducting a sensitivity analysis.
Table 1
Literature studies for the solar ejector refrigeration systems. This mass flow rate (mg) enters in the ejector as primary flow. The
secondary flow (state point 6) comes from the evaporator outlet and it
Study Collector Refrigerant COPsys Comments is also saturated vapor. It is important to state that the state point 1 has
high pressure, the state point 6 has low pressure, while the outlet of the
Korres et al. [13] FPC R11, R12, R22, 0.08 –
R114 ejector (state point 2) has medium pressure. The secondary stream of
Sokolov and FPC R114 0.13 – mass flow rate (me) is mixed with the primary fluid of mass flow rate
Hershgal [14] (mg) and at the outlet of the ejector, the mixture has mass flow rate
Sokolov and FPC R114 0.35 Booster (mc). This stream enters into the condenser where heat is rejected to the
Hershgal [15]
Sun [16] FPC R134a 0.35 Booster
environment. At the outlet of the condenser (the state point 3) is a
Huang et al. [17] FPC R141b 0.12 – saturated liquid which is separated into two streams. The first stream of
Khattab and Barakat FPC Water 0.18 – mass flow rate (mg) goes to the pump where its pressure increases. It is
[18] important to state that the compression in the pump is modeled with an
Arbel and Sokolov FPC R142b, R114 0.21 Booster
isentropic efficiency and its work is extremely low [24]. The low value
[19]
Pridasawas and FPC Butane 0.11 – of this work is explained by the compression of liquid which has high
Lundqvist [20] density. After the pump, the compressed fluid (state point 4) goes to the
Alexis and FPC R134a 0.13 – generator in order to be heated by the heat source. The other part of
Karayiannis mass flow rate (me) goes to the throttling valve where its pressure de-
[21]
Vidal et al. [22] FPC R141b ∼0.20 Auxiliary heater
creases under ideal conditions (the enthalpy is conserved). After the
Colle et al. [23] FPC R141b 0.25 Booster throttling valve, the fluid (state point 5) enters in the evaporator and it
Vidal and Colle [24] FPC R141b ∼0.25 Auxiliary heater absorbs heat. This heat is the refrigeration load and this is the useful
and Booster production of the examined configuration. It is important to state that
Yapici and Akkurt FPC R123 0.13 Auxiliary heater
the main examined refrigerant of this work is the R141b, while the
[25]
Huang et al. [26] FPC, ETC R141b FPC: 0.19 – following refrigerants have also been examined: R123, R245fa, R600a
ETC: 0.28 and R134a. These are usual refrigerants according to Table 1.
Dorantes et al. [27] ETC R142b 0.18 Booster Table 2 includes important information about these refrigerants as
Nguyen et al. [28] ETC Water 0.29 Auxiliary heater the critical pressure, the critical temperature, the Global Warming Po-
Ersoy et al. [29] ETC R123 0.20 –
Pridasawas and ETC R600a 0.23 Auxiliary heater
tential (GWP) and the Ozone Depletion Potential (ODP). It is obvious
Lundqvist [30] that the ODP of the examined refrigerants is approximately zero, while
Guo and Shen [31] ETC R134a 0.48 Auxiliary heater the GWP is low for R123 and R600a.
Varga et al. [32] ETC Water 0.10 Auxiliary heater
Dennis and Garzoli ETC R141b ∼0.15 –
2.2. Mathematical formulation
[33]
Allouche et al. [34] ETC Water ∼0.10 Auxiliary heater
Diaconu [35] ETC Water 0.10 Auxiliary heater 2.2.1. Ejector modeling
Ali [36] ETC Water ∼0.20 – In the present analysis, the examined ejector is depicted in Fig. 2a. It
Chesi et al. [37] ETC R600a ∼0.04 In ejector mode is obvious that there are two inlets and one outlet flow streams. The
Smierciew et al. [38] ETC R600a 0.19 Auxiliary heater
Dennis et al. [39] ETC R245fa 0.24 –
basic parts of the ejector are the nozzle section, the mixing section, the
Tashtoush et al. [40] ETC R134a 0.32 – throat section and the diffuser section. The primary flow goes into the
Joemann et al. [41] CPC Water 0.57 – ejector (state point pf1) and more specifically in the motive nozzle. In
Al-Khalidy [42] PTC R113 0.05 Auxiliary heater this nozzle, the pressure of this flow drops and the pressure at the outlet
Pollerberg et al. [43] PTC Water ∼0.20 –
of the motive nozzle (state point pf2) is extremely low (Pcr). On the
Pollerberg et al. [44] PTC Water ∼0.20 –
Pollerberg et al. [45] PTC Water ∼0.20 – other hand, the secondary flow stream (state point sf1) goes in the
ejector through the suction nozzle. The pressure of this stream goes

13
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

Fig. 1. The examined solar ejector refrigeration system.

Table 2 down (state point sf2) and it takes the value of (Pcr). The two flow
Basic information of the examined refrigerants [49,11]. streams are mixing in the mixing region and after this step, the mixed
flow goes in the throat section (mf). In the present modeling, it is as-
Refrigerants Pcrit (kPa) Tcrit (°C) ODP GWP ASHRAE safety group
sumed that the same pressure level exists in the mixing and the throat
R141b 4249 204.2 0.12 725 A2 sections. At the end of the ejector, there is a diffuser where the pressure
R123 3660 183.7 0.02 77 B1 of the stream is increased up to the outlet value (Pd) and the final outlet
R245fa 3651 154.0 0.00 1030 B1 flow leaves the ejector (state point d). Moreover, Fig. 2b depicts the
R600a 3640 134.7 0.04 3 A3
R134a 4060 101.0 0.00 1430 A1
total thermodynamic cycle in a pressure-enthalpy diagram.
Below the main assumptions of the present modeling are given
[46,50]:

Fig. 2. (a) The examined ejector and (b) The examined thermodynamic cycle in pressure-enthalpy diagram.

14
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

(1) The flow inside the ejector is in steady-state conditions and it is Vpf + μ·Vsf = (1 + μ)·Vmf ,is, (7)
assumed to be a one-dimensional homogeneous equilibrium flow.
However, the mixing procedure is not ideal and the velocity of the
(2) The ejector does not exchange heat with the environment.
mixed flow is found by using the mixing efficiency (ηm). This efficiency
(3) Mixing process in the mixing section of ejector occurs at constant
is the ratio of the real kinematic energy of the mixture to the ideal one
pressure and complies with the conservation of energy and mo-
[50]:
mentum.
2
(4) The effects of frictional and mixing losses in the nozzle, diffuser and Vmf
mixing section are taken into account by using the respective effi- ηm = 2
,
Vmf ,is (8)
ciencies which are assumed to be the constant among the examined
cases. Combining Eqs. (7) and (8), the mixed flow velocity can be written
(5) The velocities of streams at the inlets and at the outlet of the ejector as:
are negligible.
1 μ
(6) The primary and the second flows reach the same pressure at the Vmf = ηm ·⎜⎛ ·Vpf + ·Vsf ⎞⎟,
inlet of the constant area mixing section of ejector, and no mixing ⎝1 + μ 1+μ ⎠ (9)
between the two streams occurs before the inlet of the mixing The energy balance for the mixing process is given below:
section.
2 2
⎛ Vmf ⎞ ⎛ Vpf ⎞ ⎛ Vsf2 ⎞
(1 + μ)·⎜hmf + ⎟ = ⎜hpf + ⎟ + μ·⎜hsf + ,
The mathematical equations for the various parts of the ejector are ⎝ 2 ⎠ ⎝ 2 ⎠ ⎝ 2 ⎟⎠ (10)
given below:

Motive nozzle (primary flow) Diffuser section

In the motive nozzle, the energy is conserved and the inlet velocity In the diffuser section, the energy is conserved and the outlet ve-
at the state point (pf1) is negligible: locity at the state point (d) is negligible:
2
2
Vpf Vmf
hpf 1 = hpf 2 + , hd = hmf + ,
(1) 2 (11)
2
This process is modeled using the diffuser efficiency (ηd):
This process is modeled using the motive nozzle efficiency (ηmn):
hd,is−hmf
hpf 1−hpf 2 ηd = ,
ηmn = , hd−hmf (12)
hpf 1−hpf 2,is (2)
The state point (d,is) is the point with the entropy of the state point
The state point (pf2,is) is the point with the entropy of the state
(mf) and the pressure (Pd). This pressure level is determined according
point (pf1) and the pressure (Plow). This pressure level is assumed to be
to the condenser pressure level.
lower than the secondary flow pressure (Psf). The pressure drop (ΔP)
At this point, it is important to state that usual values for the effi-
has small values (some kPa) [46] and it is depended on the ejector
ciencies of the ejector parts are the following. These are typical values
design:
according to the literature [50,51] and thus they have been selected for
Plow = Psf −ΔP , (3) this study

Nozzle efficiencies: ηn = ηnm = ηns = 90%

Suction nozzle (secondary flow)
Mixing efficiency: ηm = 85%
Diffuser efficiency: ηd = 85%
In the suction nozzle, the energy is conserved and the inlet velocity
at the state point (sf1) is negligible:
The pressure drop (ΔP) is generally small. In ref [46], as well in Ref.
Vsf2 [52], it is clearly stated that this drop is usually up to 50 kPa. At this
hsf 1 = hsf 2 + , study, this parameter is examined parametrically and the optimum
2 (4)
value which maximizes the system performance is selected. The pres-
This process is modeled using the suction nozzle efficiency (ηsn): sure drop is a design parameter and it is associated with the ejector
hsf 1−hsf 2 geometry (for instance the diameter in various sections). So, by de-
ηsn = , termining the optimum pressure drop practically the optimum ejector
hsf 1−hsf 2,is (5)
design is selected. The existence of optimum pressure drop has been
The state point (sf2,is) is the point with the entropy of the state described in Ref. [46] with details.
point (sf1) and the pressure (Plow). This pressure level is assumed to be
lower than the secondary flow pressure (Psf). 2.2.2. Solar field modeling
The available solar energy in the solar field is calculated using the
Mixing section – Throat section total aperture (Acol) and the incident solar irradiation (GT):
Qsol = Acol ·GT , (13)
In the mixing section, the pressure is assumed to be Plow, as it has
been stated in the assumptions. Two are the basic equations which have The selected ETC is the Apricus 30 [53] and this collector is able to
to be given, the momentum and the energy conservation. It is essential passively track the sun and to present more stable thermal behavior
to introduce the entertainment ratio (μ) which is the ratio of the sec- during the daily operation. The thermal efficiency is calculated using
ondary mass flow rate to the primary mass flow rate: the performance equation from the literature [53].
msf Tcol,in−Tam ⎞ (Tcol,in−Tam)2
μ= , ηcol = 0.687−1.505·⎛ ⎜ ⎟ −0.011· ,
mpf (6) ⎝ GT ⎠ GT (14)
The ideal momentum conservation is given as: The useful heat production is calculated using the previous Eqs. (13)

15
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

and (14) as:
Qu = ηcol ·Qsol,
Moreover, it is important to state that the useful heat production can
be written as:
Qu = mcol ·cp·(Tcol,out −Tcol,in),
The mass flow rate is selected to be 2 kg/s, using typical values for
the specific mass flow rate (2 kg/sm2) according to Ref. [54].
2.2.3. Storage tank modeling
The modeling of the examined storage tank is based on the thermal
mixing zones [5,54]. In this modeling, the storage tank is separated into
horizontal zones and the temperature of every zone is assumed to be
uniform. Between the mixing zones, heat and mass are exchanged and
so there is temperature stratification inside the storage tank. The
warmer oil is in the upper part of the tank while the colder oil in the
down part. Totally three mixing zones have been selected in this
modeling [47,55]. Below the energy balances in all the mixing zones
are given. It is essential to state that the first term of Eqs. (17)–(19) is
zero for the analysis in steady-state conditions.
ρ ·VT dTst ,1
· c p· = mcol ·cp·(Tcol,out−Tst ,1) + mh ·cp·(Tst ,2−Tst ,1)
3 dt
−UT ·AT 1 ·(Tst ,1−Tam),
Eq. (23) gives the heat transfer between the heat source (thermal
oil) and the refrigerant in the generator:
(15)
Ts,in−Ts,out
Qg = (UA)g · ,
ln ( Ts,in − Tg
Ts,out − Tg ) (23)
(16) At this work, the total heat transfer coefficient (UA)g is taken equal
to 1.5 kW/K. This value makes the pinch temperature difference in the
generator to be close to 5 K, a usual design value. More specifically, the
system is designed with the outlet temperature of the heat source
(Ts,out) to be about 5 K higher than the generator temperature (Tg).
2.2.4. Refrigerator modeling
At this subsection, the basic equations of the refrigerator modeling
are given. These equations are mainly energy and mass flow rate bal-
ances. It is important to state that the state point 1 is the primary flow
in the ejector, the state point 6 is the secondary flow in the ejector and
the state point 2 is the outlet of the ejector.
The energy input in the generator is given as:
Qg = mg ·(h1−h4 ), (24)
The energy input in the evaporator is given as:
Qe = me ·(h6−h5), (25)
(17) The heat rejection of the condenser to the ambient is calculated as:

dTst ,2 Qc = mc ·(h3−h2), (26)

ρ ·VT
· c p· = mcol ·cp·(Tst ,1−Tst ,2) + mh ·cp·(Tst ,3−Tst ,2)
3 dt The mass flow rate balance is the following:
−UT ·AT 2 ·(Tst ,2−Tam), (18) m c = me + m g , (27)
ρ ·VT dTst ,3 The enthalpy is conserved in the throttling valve:
· c p· = mcol ·cp·(Tst ,2−Tst ,3) + ms ·cp·(Ts,out −Tst ,3)
3 dt h3 = h5, (28)
−UT ·AT 3 ·(Tst ,3−Tam), (19)
The enthalpy after the pump is calculated as:
The inlet temperature in the collector field is equal to the tem-
h4,is−h3
perature level in the lower part of the storage tank: h4 = h3 + ,
ηpu (29)
Tcol,in = Tst ,3, (20)
The state point (4,is) has the entropy of the state point (3) and the
Furthermore, the temperature of the heat source stream to the high pressure of state point (4). Generally, the work of this pump is
generator (Ts,in) is equal to the temperature in the first mixing zone: extremely low because this is a liquid pump. Thus, this quantity is not
Ts,in = Tst ,1, (21) taken into account in the COP calculation [24]. Thus, the COP is de-
fined as:
The mass flow rate to the generator (ms) is selected to be equal to
the total mass flow rate to the solar field (mcol) in order the heat pro- Qe m ·(h −h ) (h −h )
COPch = = e 6 5 = μ· 6 5 ,
duction to be respectful to the heat supply. This assumption creates Qg mg ·(h1−h4 ) (h1−h4 ) (30)
stable heat exchange conditions in the system [5,54]. Moreover, it is
important to state that the outer surfaces of the mixing zones (AT) are
2.2.5. System indexes definitions
described with details in Refs. [5,54].
The main index for the system evaluation is the system Coefficient
At this point, it is important to give the equations for the coupling of
of Performance (COPsys) which is the ratio of the produced refrigeration
the heat source with the generator. The heat input to the generator can
to the indecent solar energy to the solar field.
be calculated as:
Qe
Qg = ṁ s ·cp·(Ts,in−Ts,out ), (22) COPsys = ,
Qsol (31)

Table 3
Validation results with Ref. [57] with water as working fluid.

State points T (°C) P (kPa) h (kJ/kg K) m (kg/s)

Liter. Model Liter. Model Liter. Model Liter. Model

1 158.87 158.87 6.0 6.0 2743 2757 0.0764 0.0758

2 89.71 89.95 0.1008 0.1009 2655 2668 0.1198 0.1189
3 46.0 46.0 0.1008 0.1009 194.4 192.6 0.1198 0.1189
4 46.02 46.02 6.0 6.0 195.0 193.2 0.0764 0.0758
5 6.0 6.0 0.0093 0.0094 194.4 192.6 0.0434 0.0431
6 6.0 6.0 0.0093 0.0094 2500 2512 0.0434 0.0431

Deviation 0.04% 0.39% 0.71% 0.74%

16
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

Table 4 Section 3.2 gives the optimization results for various operating
The system parameters. cases. More specifically, the evaporator or refrigeration temperature is
ranged from −10 °C to 10 °C, while the condenser or heat rejection
Parameter Symbol Value
temperature from 30 °C to 50 °C. For every case, the system is optimized
Collecting area Acol 100 m2 using the “Conjugate Directions Method” or “Powell's method”. This
Storage tank volume VT 4 m3 optimization method is supported by the EES [59]. In the optimization
0.5 W/m2 K
Storage tank thermal loss coefficient UT
procedure, the relative convergence tolerance is selected equal to 10−8
Mass flow rate on the collector loop mcol 2 kg/s
Mass flow rate on the heat source loop ms 2 kg/s and the maximum number of iterations (function calls) is selected to be
Total heat transfer coefficient in the generator (UA)g 1.5 kW/K 5000. The optimization parameters are the generator temperature (Tg)
Nozzle efficiency of the ejector ηn 0.90 and the pressure drop (ΔP), while the goal is to maximize the system
Mixing efficiency of the ejector ηm 0.85 COP (COPsys). The results of this section are presented for the re-
Diffuser efficiency of the ejector ηd 0.85
frigerant R141b.
Pump isentropic efficiency ηpu 0.75
Solar irradiation GT 1000 W/m2 Section 3.3 includes results for different solar irradiation values.
Ambient temperature Tam 25 °C Again the R141b is used and the condenser temperature is constant to
Reference temperature To 298.15 K 40 °C – a typical value. However, the evaporator temperature is ranged
Sun temperature Tsun 5770 K
from −10 °C to 10 °C. The solar irradiation is examined from 500 W/m2
up to 1000 W/m2. For all the illustrated cases of this subsection, an
The system exergetic efficiency is calculated the exergetic output of optimization procedure is conducted in order to maximize the system
the refrigeration (Ee) to the exergy flow of the incident solar irradiation COP. This optimization procedure is the same as it has described for the
(Esol). results of Section 3.2.
Section 3.4 gives results for different working fluids. More specifi-
Ee cally, the following working fluids are examined: R141b, R123, R245fa,
ηex ,sys = ,
Esol (32) R600a and R134a. The condenser temperature is kept equal to 40 °C
The exergy flow of the refrigeration is calculated as [55]: and the solar irradiation intensity equal to 1000 W/m2. The evaporator
temperature is ranged from −10 °C to 10 °C and for every scenario, the
T system COP is maximized by conducting the previous optimization
Ee = Qe ·⎛ o −1⎞,
⎜ ⎟

⎝ Te ⎠ (33) procedure. At the end of Section 3, a discussion of the obtained results

The exergy flow of the incident solar irradiation is estimated using
the Petela formula [56]:
4 T 1 T
Esol = Qsol ·⎡1− · o + ·( o ) 4⎤,
⎢
⎣ 3 Tsun 3 Tsun ⎥ ⎦ (34)
It is important to state that the temperature levels in Eqs. (33) and
(34) are expressed in Kelvin units. Moreover, the reference temperature
(To) is taken equal to 298.15 K, while the sun temperature (Tsun) equal
to 5770 K.
2.3. Ejector modeling validation
The developed model is validated by using literature results from
Refs [57,58]. During this validation procedure, the following effi-
ciencies are used: ηn = 0.70, ηm = 0.80, ηd = 0.80 and isentropic pump
(ηpu = 1) [57,58]. The results of this validation procedure are given in
Table 3. It is obvious that the properties deviation is extremely low
(< 1%). Moreover, the entertainment ratio in Ref. [57] is equal to
0.5680 and the COP 0.5140, while these parameters are calculated at
this modeling 0.5686 and 0.5144 respectively. These results prove the
high accuracy of the developed model and thus this model is adopted as
reliable for the present study.
2.4. Followed methodology
In this paper, the presented system of Fig. 1 is examined and opti-
mized under various operating scenarios. A thermodynamic model is
developed in EES (Engineering Equation Solver) [59] by using the
present equations of Section 2.2. In Section 3, the results of this study
are given with details.
Section 3.1 is devoted to presenting the reason for selecting the
generator temperature and the pressure drop of the secondary ejector
flow as the optimization parameters. More specifically, it is proved that
there are optimum values for the generator temperature and for the
pressure drop, for a typical case with R141b. At this preliminary ana-
lysis, the evaporator temperature is equal to 5 °C and the condenser
temperature equal to 40 °C, typical values for conducting a re-
presentative analysis.
is given in the Section 3.5.
Below, the main parameters of the examined system are summar-
ized in Table 4. These values are the default values of these parameters
and they have kept constant for all the cases. However, it is important
to state that solar irradiation level is examined parametrically in Sec-
tion 3.2. It is noticeable that the outlets of the generator and the eva-
porator are saturated vapor of high and low pressure respectively.
These assumptions are reasonable and they are usual in the literature as
in ref [57]. Moreover, it is important to state that the maximum pres-
sure levels for all the working fluids are set equal to the 90% of the
respective critical pressure [60]. Furthermore, it is important to state
that the parameters have used with the proper units modifications when
they are used in the equations of Section 2.2.
3. Results
3.1. Preliminary analysis
At this subsection, the impact of the optimization parameters on the
system COP is given with details. A typical operating scenario is in-
vestigated in order the emphasis to be given in optimization procedure.
More specifically the scenario of operating with R141b, evaporator
temperature 5 °C and condensing temperature 40 °C is investigated.
Fig. 3 depicts the impact of the pressure drop of the secondary flow
in the ejector on the system COP. These results are obtained for gen-
eration temperature equal to 130 °C. It is clear that there is an optimum
value for this parameter which maximizes the system COP. The op-
timum pressure drop is about 4.5 kPa and this value is the optimum one
for generator temperature equal to 130 °C. The existence of optimum
pressure drop has also been stated in Refs. [46,61–63]. The next step in
this analysis is to show the impact of the generator temperature in the
system COP. The case with pressure drop equal to 4.5 kPa is selected to
be shown in Fig. 4. It is obvious that there is a generator temperature
level close to 130 °C which maximizes the system COP.
At this point, it is interesting to examine the reasons for the existing
of an optimum generator temperature. Higher generator temperature
leads to higher refrigeration performance because of the higher heat
source temperature in the thermodynamic cycle (basic thermodynamic
knowledge). On the other hand, higher generator temperature leads to a

17
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

0.090 optimum generator temperature and pressure drop.

Fig. 7 exhibits the system COP for the optimized cases. It is obvious
0.089
that higher evaporator temperatures lead to higher performance.
0.088 Moreover, lower condenser temperature levels lead to higher system
performance. The interesting point in this figure is the increasing rate of
COPsys

0.087 the curve for higher evaporator temperatures. This result indicates the
utilization of the examined solar refrigeration system in applications
0.086 with higher evaporator temperatures (5–10 °C). This result indicates
that the use of solar ejector refrigeration system is recommended in
0.085
cooling applications. Moreover, it is noticeable that the performances
0.084 for condenser temperature close to 30–35 °C are high, while it is rela-
1 2 3 4 5 6 7 8 9 10 tively low for higher condenser temperatures. This result proves the
P (kPa) high dependence of the system COP with the condenser temperature.
Fig. 8 comes to verify the previous statements by giving a 3-D depiction
Fig. 3. System COP for various pressure drops when the system operates with R141b
of the system COP with the evaporator and the condenser temperatures.
(Te = 5 °C, Tc = 40 °C and Tg = 130 °C).
This depiction shows that the system COP is ranged from 0.018 up to
0.234 and it is maximized for evaporator temperature equal to 10 °C
0.10
and condenser temperature equal to 30 °C.
0.09 The next results are given in Figs. 9 and 10 and they are about the
system exergetic efficiency. Fig. 9 is a figure with parametric curves and
0.08
Fig. 10 is a 3-D depiction of the system exergetic efficiency. It is obvious
0.07 that lower condenser temperatures lead to higher exergetic efficiency.
COPsys

0.06
Higher evaporator temperature leads to higher exergetic efficiency up
to 5 °C and after this point, the exergetic efficiency presents a small
0.05 decrease. This situation is valid for all the condenser temperatures ex-
0.04 cept the 30 °C, where the maximum exergetic efficiency is achieved for
evaporator temperature equal to 10 °C. This result indicates that the
0.03
exergetic efficiency maximization is achieved for an evaporator tem-
80 100 120 140 160 180
Tg (oC) perature close to 5 °C. More specifically, when the condenser tem-
perature is equal to 30 °C, the maximum exergetic efficiency is 1.33%
Fig. 4. System COP for various generator temperatures when the system operates with and it is the global maximum exergetic efficiency among the examined
R141b (Te = 5 °C, Tc = 40 °C and ΔP = 4.5 kPa). scenarios. This small value of the exergetic efficiency is explained by
the low chiller COP and the relatively low operating temperature levels
higher temperature in the thermal oil system and to the solar field, the on the solar collector compared to the huge sun temperature.
fact that reduces the solar thermal efficiency. The result of these reverse At this point, it would be important to explain the reason for the
trends leads to the existence of an optimum generator temperature. maximization of the exergetic efficiency in intermediate evaporator
Fig. 5 comes to verify the above hypothesis by giving the solar thermal temperature for the majority of the condenser temperature levels. Using
efficiency, the chiller COP and the system COP for various generator Eqs. (31)–(34), the system exergetic efficiency can be written as:
temperatures with a constant pressure drop at 4.5 kPa. It is obvious that
4 −1
higher generator temperature increases the chiller COP and decreases T 4 To 1 T
ηex ,sys = COPsys·⎡ o −1⎤·⎡
⎢1− · + ·⎛ o ⎞ ⎤ , ⎜ ⎟

the solar thermal efficiency, while the system COP is maximized in an ⎢

⎣ Te ⎥⎦ ⎣ 3 Tsun 3 ⎝ Tsun ⎠ ⎥
⎦ (35)
intermediate temperature level close to 130 °C.
The previous analysis proved that the couple (Tg = 130 °C and Eq. (35) is a product of three terms. The first one is the system COP
ΔP = 4.5 kPa) is a combination which leads to maximum system COP, which is getting higher for higher evaporator temperatures (see Fig. 7),
close to 0.09. However, the optimization procedures goal is to find the the second term takes lower values for higher evaporator temperatures,
best combination of these parameters which lead to the global max- while the third term remains constant in all the cases. So, the first two
imum of the system COP. Fig. 6 is a 3-D depiction of the system COP terms have a reverse behavior with the increase of the evaporator
with the two examined optimization parameters (Tg and ΔP). It is ob- temperature and this situation leads to the existence of an optimum
vious that the system COP depiction is a surface which presents a
Collector eĸcŝency COPͲcŚŝller COP-system
maximum point. This point is observed for 4.62 kPa pressure drop and
130.5 °C generator temperature. 0.55
The previous analysis clearly indicates that an optimization proce- 0.50
0.45
dure is needed in order to determine the optimum combination of the
0.40
two optimization parameters. Thus, for every operating scenario, this 0.35
Indexes

optimization procedure is applied by using the Conjugate Directions 0.30

Method which is supported by the EES [59]. 0.25
0.20
0.15
3.2. Optimization for various operating scenarios with R141b 0.10
0.05
0.00
Section 3.2 gives the optimum operation results for all the scenarios 80 100 120 140 160 180
with R141b as working fluid. These scenarios are created by examining
Tg (oC)
the evaporator temperature from −10 °C to 10 °C and the condensing
temperature from 30 °C to 50 °C, covering a great range of operating Fig. 5. System COP, chiller COP and collector thermal efficiency for different generator
conditions. For every scenario, an optimization procedure is conducted temperatures when the system operates with R141b (Te = 5 °C, Tc = 40 °C and
ΔP = 4.5 kPa).
using the Conjugate Directions Method in order to determine the

18
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

Tc = 30 °C Tc = 35 °C Tc = 40 °C

0.09 0.014
0.08
0.012
0.07
0.06 0.010
COPsys

ex,sys
0.05
0.008
0.04
0.03 0.006
0.02
0.004
0.01
0.00 0.002
80 90
100 110 8 9 10
120 130
140 150 5 6 7 0.000
160 170 3 4
Tg (oC) 1801 2 P (kPa) -10 -5 0 5 10
Te (oC)
Fig. 6. System COP for various generator temperatures and pressure drops with R141b
(Te = 5 °C and Tc = 40 °C). Fig. 9. System exergetic efficiency for various evaporator and condenser temperatures
with R141b.
Tc = 30 °C Tc = 35 °C Tc = 40 °C Tc = 45 °C Tc = 50 °C

0.25 0.0000-0.0025 0.0025-0.0050 0.0050-0.0075

0.0075-0.0100 0.0100-0.0125 0.0125-0.0150
0.20

0.15
COPsys

0.0150
0.0125
0.10
0.0100

ex,sys
0.0075
0.05
0.0050
30
0.0025
0.00 35
0.0000
-10 -5 0 5 10 40
-10
Te (oC) Tc (oC) 45 -5
0
50 5
Fig. 7. Optimum system COP for various evaporator and condenser temperatures with 10 Te (oC)
R141b.
Fig. 10. 3-D depiction of the optimum exergetic efficiency for various evaporator and
condenser temperatures with R141b.
0.00-0.05 0.05-0.10 0.10-0.15 0.15-0.20 0.20-0.25

values of the generator temperatures. The optimum values are ranged

from 114 °C to 157 °C, high values which validate the selection of ETCs
as the proper collector type in this study. This figure clearly proves that
0.25
higher generator values are needed for higher condenser temperatures
COPsys

0.20 and lower evaporator temperatures. In these cases, the system COP is
0.15 lower and thus higher generator temperature is needed in order the
30 0.10 system to be driven. In other words, the generator temperature has to
35 0.05 be higher in the most difficult operating conditions (high condenser
40 temperature and/or low evaporator temperature). This result indicates
0.00
Tc (oC) 45 -10 the utilization of higher quality collectors in applications with low
-5
50 0 evaporator temperatures and high condenser temperatures.
5
10 Fig. 13 gives the optimum pressure drops for the examined
Te (oC)

Fig. 8. 3-D depiction of the optimum system COP for various evaporator and condenser
temperatures with R141b. 0-5 5-10 10-15 15-20 20-25

evaporator temperature. It is essential to state that for the case of

Tc = 30 °C, the increase of the system COP with the Te is high (great
values of dCOPsys/dTe) and this high increasing rate leads to the max- 25
imization of the exergetic efficiency in a higher evaporator temperature 20
level (at ∼10 °C and not at ∼5 °C as in the other cases).
Qe (kW)

15
Fig. 11 gives the cooling capacity for various operating scenarios.
This parameter is optimized because the optimization of the system 30 10
COP is directly associated with the optimization of the cooling capacity 35 5
(see Eq. (31)). This parameter is ranged from 1.85 kW up to 23.39 kW, 40
0
presenting a great variation. More specifically, this parameter is max- Tc (oC) 45 -10
-5
imized for higher evaporator temperature and lower condenser tem- 50 0
5
perature presenting similar behavior as the system COP. 10 Te (oC)
The next step in this analysis is to present results about the optimum
values of the optimization parameters. Fig. 12 illustrates the optimum Fig. 11. 3-D depiction of the cooling capacity for various evaporator and condenser
temperatures with R141b.

19
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

100-110 110-120 120-130 130-140 140-150 150-160 generator temperature behavior and this adds an extra interest in the
optimization procedure.
At this end of this subsection, Table 5 summarizes the results for the
optimum design of all the examined scenarios. It would be interesting
to give the range of the performance indexes which are given in this
160
table. The solar collector efficiency is ranged from 0.2772 to 0.4408,

OpƟmƵm Tg (oC)
150
the chiller COP is ranged from 0.0667 to 0.5305 and the system COP is
140 ranged from 0.0185 to 0.2339. The system COP is lower than the other
130 parameters because it takes into account both the previous indexes, as
30
120 well as the thermal losses of the storage tank. Lastly, it is important to
35 state that the exergetic efficiency takes low values from 0.0026 to
110
40 0.0133.
100
Tc (oC) 45 -10
-5
50 0
5 3.3. The impact of solar irradiation level on the results
10 Te (oC)

Fig. 12. 3-D depiction of the optimum generator temperature for various evaporator and In this subsection, the impact of the solar irradiation levels on the
condenser temperatures with R141b. collector performance and on the optimum design parameters are in-
vestigated. The solar irradiation levels are examined from 500 W/m2 to
0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 1000 W/m2, covering a great range of possible values. The presented
results have been taken for operating with R141b and for condenser
temperature equal to 40 °C. This condenser temperature is typical and
representative and thus it has been selected in this analysis.
Fig. 14 illustrates the system COP for various irradiation levels and
OpƟmƵm P (kPa)

8
7 evaporator temperature. It is obvious that higher solar irradiation levels
6
lead to higher system performance because of the higher thermal effi-
5
4 ciency of the solar collectors (see Eq. (14)). Similar results are obtained
30
3 for the system exergetic efficiency on Fig. 15. Higher solar irradiation
35 2
levels increase the exergetic efficiency. However, there is an interesting
1
40
0 point at this figure. The maximum exergetic efficiency is achieved for
Tc (oC) 45
-5
-10 evaporating temperature equal to 10 °C when the solar irradiation is up
50 5
0
to 700 W/m2. For higher solar irradiation levels, the optimum eva-
10 Te (oC) porator temperature is equal to 50 °C. This result verifies the results of
Fig. 13. 3-D depiction of the optimum pressure drop for various evaporator and con- Fig. 9, about the optimum evaporating temperature at 5 °C when the
denser temperatures with R141b. condenser temperature is equal to 40 °C. The final conclusion from this
figure is that the solar irradiation level is able to influence on the op-
scenarios. This parameter is ranged from 0.96 kPa to 7.09 kPa. Higher timum evaporating temperature, something interesting and important.
However, the difference between the curves of Te = 5 °C and Te = 10 °C
values are observed for operation with higher evaporator temperatures
and lower condenser temperatures. This behavior is different to the is very small and thus there is no reason for deeper analysis.
The next given parameter is the cooling capacity which is depicted

Table 5
Optimization results for various operating scenarios with R141b.

Te (°C) Tc (°C) ΔP (kPa) Tg (°C) Qe (kW) ηcol COPch COPsys ηex,sys

−10 30 2.29 128.0 7.31 0.3902 0.1873 0.0731 0.0105

−10 35 1.93 135.1 5.33 0.3641 0.1463 0.0533 0.0076
−10 40 1.59 142.2 3.83 0.3366 0.1139 0.0383 0.0055
−10 45 1.25 149.3 2.70 0.3082 0.0877 0.0270 0.0039
−10 50 0.96 156.9 1.85 0.2772 0.0667 0.0185 0.0026
−5 30 3.21 124.3 9.63 0.4038 0.2385 0.0963 0.0116
−5 35 2.78 131.1 7.04 0.3788 0.1857 0.0704 0.0085
−5 40 2.35 137.9 5.11 0.3531 0.1447 0.0511 0.0061
−5 45 1.92 145.0 3.66 0.3254 0.1124 0.0366 0.0044
−5 50 1.49 152.0 2.56 0.2972 0.0863 0.0256 0.0031
0 30 4.30 120.6 12.74 0.4170 0.3056 0.1274 0.0125
0 35 3.86 127.4 9.27 0.3924 0.2363 0.0927 0.0091
0 40 3.35 134.1 6.75 0.3677 0.1836 0.0675 0.0066
0 45 2.82 140.9 4.88 0.3417 0.1428 0.0488 0.0048
0 50 2.30 147.9 3.48 0.3141 0.1107 0.0348 0.0034
5 30 5.66 117.2 17.06 0.4288 0.3979 0.1706 0.0132
5 35 5.19 123.9 12.28 0.4054 0.3028 0.1228 0.0095
5 40 4.62 130.5 8.90 0.3813 0.2335 0.0890 0.0069
5 45 3.98 137.1 6.45 0.3564 0.1811 0.0645 0.0050
5 50 3.38 144.0 4.64 0.3293 0.1410 0.0464 0.0036
10 30 7.09 113.7 23.39 0.4408 0.5305 0.2339 0.0133
10 35 6.71 120.4 16.45 0.4177 0.3938 0.1645 0.0094
10 40 6.16 126.9 11.79 0.3943 0.2992 0.1179 0.0067
10 45 5.51 133.5 8.52 0.3698 0.2304 0.0852 0.0049
10 50 4.68 140.1 6.15 0.3446 0.1785 0.0615 0.0035

20
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

-10 °C -5 °C °C °C °C
R600a present relatively high performance, while R134a leads to low
0.12
performance, the fact that makes it unsuitable for the examined system.
0.10 Fig. 20 gives results about the exergetic performance of the system
with the five working fluids. The performance sequence is the same as
0.08
in Fig. 19. It is interesting to state that all the refrigerants present an
COPsystem

0.06 optimum performance for an evaporating temperature close to 5 °C,

while the R134a curve is not maximized for an intermediate evapor-
0.04
ating temperature.
0.02 Fig. 21 illustrates the cooling capacity for the examined refrigerants.
0.00
The curves are similar to the respective of Fig. 19 about the system
500 600 700 800 900 1000 COP. It would be interesting to give the cooling capacity for the typical
GT (W/m2) evaporator temperature of 5 °C which indicates cooling applications. At
this temperature level, the cooling capacity with R141b, R123, R245fa,
Fig. 14. Optimum system COP for different solar irradiation levels and evaporator tem- R600a and R134a are 8.90 kW, 8.49 kW, 7.87 kW, 7.62 kW and 5.33 kW
peratures with R141b (Tc = 40 °C).
respectively. It is clear that the proper refrigerant selection is critical in
order to achieve adequate cooling capacity.
-10 °C -5 °C °C °C °C
Figs. 22 and 23 illustrate the optimum generator temperature and
0.008
pressure drop for the examined cases of this subsection. Fig. 22 in-
0.007 dicates that higher evaporating temperature leads to lower optimum
generator temperature for R141b, R123 and R245fa, while the optimum
0.006 generator temperature is constant for R600a and R134a. The reason for
ex,sys

0.005
0.004
0.003
0.002
500 600 700 800 900 1000
GT (W/m2)
Fig. 15. System exergetic efficiency for different solar irradiation levels and evaporator
temperatures with R141b (Tc = 40 °C).
in Fig. 16 for all the examined conditions. It is interesting that this
parameter increases with the increase of the solar irradiation level with
a high rate. This result is explained by the increase of the energy input
in the system (Qsol), as well as of the increase in the system COP (see
Fig. 14). Moreover, comparing the slope of the curves between Figs.14
and 16 it is obvious that the cooling capacity is more influenced by the
solar potential than the system COP.
The last figures of this subsection give the optimum values of the
generator temperature (Fig. 17) and of the pressure drop (Fig. 18) for
different values of the solar irradiation. Fig. 17 proves that higher solar
potential leads to higher optimum generator temperatures. Higher solar
irradiation gives the possibility the solar field to work in higher tem-
peratures with an adequate efficiency and thus this result is reasonable.
Furthermore, higher generator temperature is found for lower eva-
porator temperatures.
Fig. 18 shows that higher solar irradiation levels lead to higher
optimum pressure drop. Moreover, higher pressure drops are obtained
for higher evaporator temperatures. Generally, the impact of the solar
irradiation level is not so great on the optimum values of the pressure
drop. This result indicates that the optimum ejector design is not in-
fluenced by the solar irradiation level with a great way
this result is the relatively low critical temperature of R600a and R134a
which creates a limitation in the optimization procedure. At this point,
it is important to remind that the maximum possible pressure level in
the generator is set to be up to the 90% of the respective critical
pressure for safety reasons. So, the optimum generator temperature is
the maximum possible for R600a and R134a for all the evaporator
temperature. This limitation leads to lower performance, especially for
R134a. Moreover, the low critical temperature of R134a is the reason
for the low exergy efficiency with this working fluid.
Fig. 23 gives the optimum pressure drops for the examined cases.
Higher evaporator temperature leads to higher optimum pressure drop.
It is noticeable that generally the higher pressure drops are obtained for
the less efficient refrigerants (R600a and R134a). This result is rea-
sonable for these refrigerants because the other optimization parameter
(Tg) is constrained and equal to the maximum possible value and all the
optimization procedure practically is performed through the pressure
drop. In other words, the pressure drop has to be varied a lot in order
higher system COP to be achieved. For the other refrigerants, both
pressure drop and generator temperature are varied among the cases
and thus the optimization variation is shared.
Table 6 includes the optimum results for the examined refrigerants
and for five evaporator temperatures. It would be interesting to present
the performances of the refrigerants for the evaporating temperature
which maximizes the system COP and it is 10 °C. The solar collector
efficiencies are found 39.41%, 39.35%, 40.01%, 40.12% and 50.15%
for R141b, R123, R245fa, R600a and R134a respectively. The chiller
COP are found 0.2983, 0.2869, 0.2645, 0.2577 and 0.1631, while the
-10 °C -5 °C °C °C °C
14
12
10
Qe (kW)

3.4. Results for different refrigerants 8

6
The next step in the present analysis is the investigation of different
4
working fluids in the system performance. The following results have
taken for condenser temperature equal to 40 °C and solar irradiation 2
equal to 1000 W/m2. Fig. 19 exhibits the system COP for different
0
evaporator temperatures from −10 °C to 10 °C and for the five different 500 600 700 800 900 1000
working fluids. The most efficient working fluid is R141b, while the GT (W/m2)
order of the other continues with R123, R245fa, R600a and R134a. It is
important to state that this order is the same for all the evaporator Fig. 16. Cooling capacity for different solar irradiation levels and evaporator tempera-
tures with R141b (Tc = 40 °C).
temperatures. Moreover, the refrigerants R141b, R123, R245fa and

21
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

-10 °C -5 °C °C °C °C R141b R123 R245fa R600a R134a

150 0.008

0.007
140
0.006
OpƟmƵm Tg (oC)

130 0.005

ex,sys
0.004
120
0.003
110
0.002

100 0.001

0.000
90 -10 -5 0 5 10
500 600 700 800 900 1000 Te (oC)
GT (W/m2)
Fig. 20. System exergetic efficiency for various refrigerants and evaporator temperatures
Fig. 17. Optimum generator temperature for different solar irradiation levels and eva- (Tc = 40 °C).
porator temperatures with R141b (Tc = 40 °C).

R141b R123 R245fa R600a R134a

-10 °C -5 °C °C °C °C 14
7
12
6
10
OpƟmƵm P (kPa)

Qe (kW)
8
4 6

3 4

2 2

1 0
-10 -5 0 5 10
0
Te (oC)
500 600 700 800 900 1000
GT (W/m2) Fig. 21. Cooling capacity for various refrigerants and evaporator temperatures
(Tc = 40 °C).
Fig. 18. Optimum pressure drop for different solar irradiation levels and evaporator
temperatures with R141b (Tc = 40 °C).
R141b R123 R245fa R600a R134a
150
R141b R123 R245fa R600a R134a
0.14 140
0.12
OpƟmƵm Tg (oC)

130
0.10
120
COPsys

0.08

0.06 110

0.04 100

0.02
90
0.00 -10 -5 0 5 10
-10 -5 0 5 10 Te (oC)
Te (oC)
Fig. 22. Optimum generator temperature for various refrigerants and evaporator tem-
Fig. 19. Optimum System COP for various refrigerants and evaporator temperatures peratures (Tc = 40 °C).
(Tc = 40 °C).

optimized for various operating scenarios. The final results seem to be

system COP 0.1179, 0.1129, 0.1058, 0.1034 and 0.0818 for R141b,
R123, R245fa, R600a and R134a respectively. It is obvious that higher
collector efficiencies are obtained for the refrigerants with the lower
system COP. Moreover, the refrigerants with higher chiller COP present
also higher system COP. This result indicates that the chiller COP is the
most important parameter in this analysis because it influences more on
the system performance. The reason for the higher solar thermal effi-
ciency for the working fluids with the lowest system performance is
explained by the operation in lower generator temperatures due to the
limitations of the critical pressure.
3.5. Discussion of the results
At this paper, a solar ejector refrigeration system is examined and
very interesting and they are discussed in this section. Firstly, it is
important to state again that the innovation of this work is the sys-
tematic optimization using two optimization parameters for different
operating scenarios and different refrigerants. This analytical method
aids to extract useful conclusions from the obtained results.
The obtained results indicate that the solar ejector refrigeration
system leads to relatively low system COP up to 0.234, compared to
other solar refrigeration technologies. For example, in other literature
studies with absorption chillers [47], the system COP is up to 0.40.
However, the lower performance of the ejector system is counter-
balanced from the lower capital cost compared to the sorption machine
technologies.
Furthermore, it is important to state that other literature studies
with ejector refrigeration systems driven by ETC without an auxiliary

22
E. Bellos, C. Tzivanidis Energy Conversion and Management 154 (2017) 11–24

R141b R123 R245fa R600a R134a special designed collector, maybe by using a secondary booster re-
25
flector.
At this end of this discussion section, it is important to state that the
20
obtained results indicate the use of ejector in cooling applications
OpƟmƵm P (kPa)

where the evaporator operates in the range from 5 °C to 10 °C.

15 Moreover, the system presents high enhancement in its performance
when it operates with a low condenser temperature, the fact that in-
10 dicates the utilization of water-cooled condenser. Thus, it can be con-
cluded that the ejector systems are ideal technologies for cooling ap-
5 plications with water-cooled condensers.

0 4. Conclusions
-10 -5 0 5 10
Te (oC) The objective of this paper is to investigate the performance of a
Fig. 23. Optimum pressure drop for various refrigerants and evaporator temperatures
solar ejector refrigeration system under different operating scenarios.
(Tc = 40 °C). Evacuated tube collectors with a storage tank are selected to feed the
generator with heat in order the demanded refrigeration to be produced
in the evaporator. The system is examined for evaporator temperatures
heat source, similar COP have been obtained. More specifically, in Refs.
from −10 °C to 10 °C and for condenser temperatures from 30 °C to
[29,33,36], the system COP were found 0.20, 0.15 and 0.20 respec-
50 °C. Five different refrigerants are examined (R141b, R123, R245fa,
tively. This result validates the reliability of the obtained results be-
R600a and R134a), while the emphasis is given in the most efficient
cause they are in accordance with other typical values.
refrigerant which is the R141b. Moreover, the solar irradiation level is
Another important point of this study is associated with the working
examined parametrically. In all the cases, the system is optimized in
fluid investigation. It is found that the most suitable working fluid is
order to maximize the system COP by varying the generator tempera-
R141b, with R123, R245fa, R600a and R134a to follow. An important
ture and the pressure drop in the ejector. The analysis is performed with
parameter for the performance of the working fluids is their critical
a developed model in EES (Engineering Equation Solver). The most
temperature. Table 2 includes the critical temperatures of the five ex-
important conclusions of this study are summarized below:
amined working fluids. It is interesting that higher critical temperature
leads to higher system performance. More specifically, the highest cri-
– There are optimum values for the generator temperature and the
tical temperature is obtained for R141b with 204.2 °C, while the lowest
pressure drop in the ejector which lead to maximum system COP for
for R134a with 101.0 °C. Thus, it is proved that there is an important
all the examined cases. The optimum generator temperature is
correlation between the performance and the critical temperature. The
higher for lower evaporator temperatures and higher condenser
reason for this result is based on the possibility of operation in high-
temperatures. On the other hand, the optimum pressure drop is
temperature levels when the critical temperature is higher.
higher for higher evaporator temperatures and lower condenser
It is essential to state that the results of the present work can be
temperatures.
enhanced if an ETC with higher optical efficiency than the present is
– The system COP is maximized for evaporator temperature equal to
examined. The present solar collector has an optical efficiency close to
10 °C and for condenser temperature equal to 50 °C and it is equal to
68%, a usual value for ETCs which can be increased with a better and

Table 6
Comparison of the refrigerants with Tc = 40 °C.

Te (°C) ΔP (kPa) Tg (°C) Qe (kW) ηcol COPch COPsys ηex,sys

R141b −10 1.59 142.2 3.83 0.3367 0.1139 0.0383 0.0055

−5 2.36 137.9 5.11 0.3533 0.1446 0.0511 0.0061
0 3.36 134.1 6.75 0.3676 0.1836 0.0675 0.0066
5 4.65 130.5 8.90 0.3811 0.2336 0.0890 0.0069
10 6.13 127.0 11.79 0.3941 0.2993 0.1179 0.0067

R123 −10 1.84 141.5 3.62 0.3391 0.1068 0.0362 0.0052

−5 2.74 137.5 4.84 0.3549 0.1364 0.0484 0.0058
0 3.90 133.6 6.42 0.3695 0.1737 0.0642 0.0063
5 5.36 130.0 8.49 0.3831 0.2217 0.0849 0.0066
10 7.03 127.1 11.29 0.3935 0.2869 0.1129 0.0064

R245fa −10 2.50 139.5 3.18 0.3472 0.0916 0.0318 0.0045

−5 3.90 136.0 4.35 0.3605 0.1207 0.0435 0.0052
0 5.71 132.2 5.87 0.3748 0.1567 0.0587 0.0058
5 7.95 128.8 7.87 0.3875 0.2032 0.0787 0.0061
10 10.79 125.3 10.58 0.4001 0.2645 0.1058 0.0060

R600a −10 4.33 125.2 2.80 0.4006 0.0700 0.0280 0.0040

−5 7.38 125.2 4.02 0.4006 0.1004 0.0402 0.0048
0 11.33 125.2 5.58 0.4006 0.1393 0.0558 0.0055
5 14.09 125.2 7.62 0.4006 0.1902 0.0762 0.0059
10 22.13 125.0 10.34 0.4012 0.2577 0.1034 0.0059

R134a −10 0.20 94.9 0.47 0.5015 0.0094 0.0047 0.0007

−5 1.99 94.9 1.67 0.5015 0.0332 0.0167 0.0020
0 5.95 94.9 3.24 0.5015 0.0645 0.0324 0.0032
5 12.17 94.9 5.33 0.5015 0.1062 0.0533 0.0041
10 21.24 94.9 8.18 0.5015 0.1631 0.0818 0.0047

23
E. Bellos, C. Tzivanidis
0.2339 with R141b. The exergetic efficiency at these conditions is
0.0133, while the optimum generator temperature is 144 °C and the
optimum pressure drop 5.64 kPa. Generally, the system presents the
optimum exergetic performance for evaporating temperatures be-
tween 5 °C and 10 °C.
– Higher values of the solar irradiation levels lead to higher system
COP and exergetic efficiency. The increase in the cooling capacity
with solar irradiation is very intense as it is found. Moreover, it is
important to state that the optimum values of the optimization
parameters are different for different solar irradiation levels.
– The investigation of different refrigerants proves that R141b is the
best choice, with R123, R245fa, R600a and R134a to follow. The
lower critical temperature of R600a and especially of R134a is the
most important factor for their lower performance.
– Summarizing the obtained results, the solar ejector technology has
to be applied by using high-quality solar collectors and it is re-
commended for cooling applications with high evaporator tem-
peratures (5 °C ∼10 °C). Moreover, the use of water-cooled con-
denser is important for achieving high performance.
References
[1] Krajačić G, Duić N, Vujanović M, Kılkış Ş, Rosen MA, Al-Nimr MA. Sustainable
development of energy, water and environment systems for future energy tech-
nologies and concepts. Energy Convers Manage 2016;125:1–14.
[2] Loni R, Kasaeian AB, Mahian O, Sahin AZ. Thermodynamic analysis of an organic
rankine cycle using a tubular solar cavity receiver. Energy Convers Manage
2016;127:494–503.
[3] Bellos E, Mathioulakis E, Tzivanidis C, Belessiotis V, Antonopoulos KA.
Experimental and numerical investigation of a linear Fresnel solar collector with flat
plate receiver. Energy Convers Manage 2016;130:44–59.
[4] Qiu Y, He YL, Li P, Du BC. A comprehensive model for analysis of real-time optical
performance of a solar power tower with a multi-tube cavity receiver. Appl Energy
2017;185(1):589–603.
[5] Bellos E, Tzivanidis C, Moschos K, Antonopoulos KA. Energetic and financial eva-
luation of solar assisted heat pump space heating systems. Energy Convers Manage
2016;120:306–19.
[6] Bataineh K, Taamneh Yazan. Review and recent improvements of solar sorption
cooling systems. Energy Build 2016;128:22–37.
[7] Kerme ED, Chafidz A, Agboola OP, Orfi J, Fakeeha AH, Al-Fatesh AS. Energetic and
exergetic analysis of solar-powered lithium bromide-water absorption cooling
system. J Clean Prod 2017;151:60–73.
[8] Mohan G, Kumar U, Pokhrel MK, Martin A. A novel solar thermal polygeneration
system for sustainable production of cooling, clean water and domestic hot water in
United Arab Emirates: dynamic simulation and economic evaluation. Appl Energy
2016;167:173–88.
[9] Bellos E, Tzivanidis C, Symeou C, Antonopoulos KA. Energetic, exergetic and fi-
nancial evaluation of a solar driven absorption chiller – a dynamic approach. Energy
Convers Manage 2017;137:34–48.
[10] Zeyghami M, Goswami DY, Stefanakos E. A review of solar thermo-mechanical
refrigeration and cooling methods. Renew Sustain Energy Rev 2015;51:1428–45.
[11] Chen X, Omer S, Worall M, Riffat S. Recent developments in ejector refrigeration
technologies. Renew Sustain Energy Rev 2013;19:629–51.
[12] Abdulateef JM, Sopian K, Alghoul MA, Sulaiman MY. Review on solar-driven
ejector refrigeration technologies. Renew Sustain Energy Rev 2009;13:1338–49.
[13] Korres CJ, Bardakas EE, Koumoutsos NG. Performance analysis of a solar vapour
thermal compression chiller. Int J Energy Res 1989;13:243–52.
[14] Sokolov M, Hershgal D. Optimal coupling and feasibility of a solar-powered year-
round ejector air conditioner. Sol Energy 1993;50(6):507–16.
[15] Sokolov M, Hershgal D. Solar-powered compression-enhanced ejector air condi-
tioner. Sol Energy 1993;51(3):183–94.
[16] Sun DW. Solar powered combined ejector-vapour compression cycle for air con-
ditioning and refrigeration. Energy Convers Manage 1997;38(5):479–91.
[17] Huang BJ, Chang JM, Petrenko VA, Zhuk KB. A solar ejector cooling system using
refrigerant R141b. Sol Energy 1998;64(4):223–6.
[18] Khattab NM, Barakat MH. Modeling the design and performance characteristics of
solar steam-jet cooling for comfort air conditioning. Sol Energy 2002;73(4):257–67.
[19] Arbel A, Sokolov M. Revisiting solar-powered ejector air conditioner––the greener
the better. Sol Energy 2004;77(1):57–66.
[20] Pridasawas W, Lundqvist P. An exergy analysis of a solar-driven ejector refrigera-
tion system. Sol Energy 2004;76(4):369–79.
[21] Alexis GK, Karayiannis EK. A solar ejector cooling system using refrigerant R134a in
the Athens area. Renewable Energy 2005;30(9):1457–69.
[22] Vidal H, Colle S, dos Santos Pereira G. Modelling and hourly simulation of a solar
ejector cooling system. Appl Therm Eng 2006;26(7):663–72.
[23] Colle S, Pereira GS, Gutiérrez HR Vidal, Moragas R Escobar. On the validity of a
design method for a solar-assisted ejector cooling system. Sol Energy
2009;83(2):139–49.
[24] Vidal H, Colle S. Simulation and economic optimization of a solar assisted combined
ejector–vapor compression cycle for cooling applications. Appl Therm Eng
2010;30(5):478–86.
[25] Yapici P, Akkurt F. Experimental investigation on ejector cooling system
24
Energy Conversion and Management 154 (2017) 11–24
performance at low generator temperatures and a preliminary study on solar en-
ergy. J Mech Sci Technol 2012;26(11):3653–9.
[26] Huang BJ, Petrenko VA, Samofatov IY, Shchetinina NA. Collector selection for solar
ejector cooling system. Sol Energy 2001;71(4):269–74.
[27] Dorantes R, Estrada CA, Pilatowsky I. Mathematical simulation of a solar ejector-
compression refrigeration system. Appl Therm Eng 1996;16(8):669–75.
[28] Nguyen VM, Riffat SB, Doherty PS. Development of a solar-powered passive ejector
cooling system. Appl Therm Eng 2001;21(2):157–68.
[29] Ersoy HK, Yalcin S, Yapici R, Ozgoren M. Performance of a solar ejector cooling-
system in the southern region of Turkey. Appl Energy 2007;84(9):971–83.
[30] Pridasawas W, Lundqvist P. A year-round dynamic simulation of a solar-driven
ejector refrigeration system with iso-butane as a refrigerant. Int J Refrig
2007;30(5):840–50.
[31] Guo J, Shen HG. Modeling solar-driven ejector refrigeration system offering air
conditioning for office buildings. Energy Build 2009;41:175–81.
[32] Varga S, Oliveira AC, Diaconu B. Analysis of a solar-assisted ejector cooling system
for air conditioning. Int J Low-Carbon Technol 2009;4:2–8.
[33] Dennis M, Garzoli K. Use of variable geometry ejector with cold store to achieve
high solar fraction for solar cooling. Int J Refrig 2011;34(7):1626–32.
[34] Allouche Y, Bouden C, Riffat S. A solar-driven ejector refrigeration system for
mediterranean climate: experience improvement and new results performed.
Energy Procedia 2012;18:1115–24.
[35] Diaconu BM. Energy analysis of a solar-assisted ejector cycle air conditioning
system with low temperature thermal energy storage. Renewable Energy
2012;37(1):266–76.
[36] Ali AHH. Solar cooling: experiences and lessons learned with two different systems.
In: International Renewable and Sustainable Energy Conference (IRSEC),
Ouarzazate; 2013. p. 153–60.
[37] Chesi A, Ferrara G, Ferrari L, Tarani F. Analysis of a solar assisted vapour com-
pression cooling system. Renewable Energy 2013;49:48–52.
[38] Śmierciew K, Gagan J, Butrymowicz D, Karwacki J. Experimental investigations of
solar driven ejector air-conditioning system. Energy Build 2014;80:260–7.
[39] Dennis M, Cochrane T, Marina A. A prescription for primary nozzle diameters for
solar driven ejectors. Sol Energy 2015;115:405–12.
[40] Tashtoush B, Alshare A, Al-Rifai S. Hourly dynamic simulation of solar ejector
cooling system using TRNSYS for Jordanian climate. Energy Convers Manage
2015;100:288–99.
[41] Joemann M, Oezcan T, Kauffeld M, Pollerberg C. Process steam and chilled water
production with CPC-collectors, steam jet ejector chiller and latent heat storages.
Energy Procedia 2016;91:767–76.
[42] Al-Khalidy N. Experimental investigation of solar concentrating collectors in a re-
frigerant ejector refrigeration machine. Int J Energy Res 1997;21:1123–31.
[43] Pollerberg C, Ali AHH, Dötsch C. Experimental study on the performance of a solar
driven steam jet ejector chiller. Energy Convers Manage 2008;49(11):3318–25.
[44] Pollerberg C, Heinzel A, Weidner E. Model of a solar driven steam jet ejector chiller
and investigation of its dynamic operational behaviour. Sol Energy
2009;83(5):732–42.
[45] Pollerberg C, Ali AHH, Dötsch C. Solar driven steam jet ejector chiller. Appl Therm
Eng 2009;29(5):1245–52.
[46] Li H, Cao F, Bu X, Wang L, Wang X. Performance characteristics of R1234yf ejector-
expansion refrigeration cycle. Appl Energy 2014;121:96–103.
[47] Bellos E, Tzivanidis C, Antonopoulos KA. Exergetic and energetic comparison of
LiCl-HO and LiBr-HO working pairs in a solar absorption cooling system. Energy
Convers Manage 2016;123:453–61.
[48] https://www.therminol.com/products/Therminol-VP1.
[49] Cizungu K, Mani A, Groll M. Performance comparison of vapour jet refrigeration
system with environment friendly working fluids. Appl Therm Eng
2001;21(5):585–98.
[50] Wang J, Dai Y, Sun Z. A theoretical study on a novel combined power and ejector
refrigeration cycle. Int J Refrig 2009;32(6):1186–94.
[51] Liu F, Groll EA. Study of ejector efficiencies in refrigeration cycles. Appl Therm Eng
2013;52(2):360–70.
[52] Boyaghchi FA, Chavoshi M. Multi-criteria optimization of a micro solar-geothermal
CCHP system applying water/CuO nanofluid based on exergy, exergoeconomic and
exergoenvironmental concepts. Appl Therm Eng 2017;112:660–75.
[53] Shirazi A, Taylor RA, White SD, Morrison GL. A systematic parametric study and
feasibility assessment of solar-assisted single-effect, double-effect, and triple-effect
absorption chillers for heating and cooling applications. Energy Convers Manage
2016;114:258–77.
[54] Duffie JA, Beckman WA. Solar engineering of thermal processes. John Wiley & Sons
INC; 1991.
[55] Bellos E, Tzivanidis C, Antonopoulos KA. Exergetic, energetic and financial eva-
luation of a solar driven absorption cooling system with various collector types.
Appl Therm Eng 2016;102:749–59.
[56] Petela R. Exergy of undiluted thermal radiation. Sol Energy 2003;74(6):469–88.
[57] Alexis GK. Exergy analysis of ejector-refrigeration cycle using water as working
fluid. Int J Energy Res 2005;29:95–105.
[58] Alexis GK, Rogdakis ED. A verification study of steam-ejector refrigeration model.
Appl Therm Eng 2003;23(1):29–36.
[59] F-Chart Software. Engineering Equation Solver (EES); 2015. Available at: http://
www.fchart.com/ees.
[60] Tzivanidis C, Bellos E, Antonopoulos KA. Energetic and financial investigation of a
stand-alone solar-thermal Organic Rankine Cycle power plant. Energy Convers
Manage 2016;126:421–33.
[61] Nehdi E, Kairouani L, Bouzaina M. Performance analysis of the vapour compression
cycle using ejector as an expander. Int J Energy Res 2007;31:364–75.
[62] Sarkar J. Performance characteristics of natural-refrigerants-based ejector expan-
sion refrigeration cycles. Proc IMechE Part A: J Power Energy 2009;223:543–50.
[63] Ersoy HK, Bilir N. Performance characteristics of ejector expander transcritical CO2
refrigeration cycle. Proc IMechE Part A: J Power Energy 2012;226(5):623–35.