Energy Conversion and Management 148 (2017) 1208–1224

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Evaluation of the ejector refrigeration system with environmentally

friendly working fluids from energy, conventional exergy and advanced
exergy perspectives
Jianyong Chen a,b, Kangda Zhu a, Yisheng Huang a, Ying Chen a,b,⇑, Xianglong Luo a,b
a
School of Material and Energy, Guangdong University of Technology, Guangzhou 510006, China
b
Guangdong Provincial Key Laboratory of Functional Soft Condensed Matter, Guangzhou 510006, China

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

Article history: This paper presents the evaluation of the ejector refrigeration system from three levels: energy analysis,
Received 21 March 2017 conventional exergy analysis and advanced exergy analysis, in light of introducing the thermal conduc-
Received in revised form 8 June 2017 tance and the objective function, the exergy efficiencies, the exergetic rehabilitation ratio and the exer-
Accepted 16 June 2017
getic improvement potential ratio, respectively. Five environmentally friendly working fluids, namely
Available online 3 July 2017
R600, R600a, R601a, R1233zd(E) and R1234ze(E), are used to compare their performance and working
characteristics in the system. At the normal condition, it is found that these candidates have different
Keywords:
orders of highest performance by different analysis, but agree that the ejector has the highest priority
Ejector refrigeration system
Exergy destruction
of improvement. The parametric study shows that these candidates have the same optimum pinch tem-
Working fluid perature differences in the condenser (7 °C and 3 °C) and in the evaporator (3 °C and 2 °C) to minimize the
Performance thermal conductance and the objective function, while they are different in the generator. The ejector
Improvement efficiencies have considerable influence on the system performance, and a 0.1 increase in ejector effi-
ciency could lead to the increase of system exergy efficiency from 1.38% to 10.33%. The pump efficiency
has insignificant influence the system performance, but a 0.1 increase in the pump efficiency results in a
7.37% decrease in exergetic improvement potential ratio of the pump. By comprehensive comparisons,
R1233zd(E) has generally higher system performance than the other four candidates, it is therefore rec-
ommended as the good working fluid for the ejector refrigeration system.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction energy-environment-conscious era. However, it has relatively

Refrigeration and air-conditioning units as well as heat pumps
are dominated by vapor compression refrigeration systems and
consume about 17% electricity world-widely [1]. The electrical
peak load problems during summer and environmental issues have
consequently become more serious than ever before. This raises
voices for seeking approaches of utilizing renewable energy and
low-grade heat to produce cooling. The ejector refrigeration sys-
tem, abbreviated as ERS, is driven by the thermal energy and seems
to be a good alternative. It has merits of simple construction, low
cost, long lifetime, flexible capacity, no-chemical corrosion and
chemical reaction over the sorption refrigeration systems. More-
over, its capability of applying various environmentally-benign
refrigerants makes such system particularly attractive in this
⇑ Corresponding author at: School of Material and Energy, Guangdong University
of Technology, Guangzhou 510006, China.
E-mail address: chenying@gdut.edu.cn (Y. Chen).
low COP, which greatly limits its application in commercial and
wide scales. A number of investigations have been conducted to
get a better understanding of ejector characteristics and to
improve the system performance. Zhu et al. [2] visualized the flow
phenomena inside the ejector. Chen et.al [3] proposed an ejector
model to predict the optimum performance. Chen et al. [4] devel-
oped a 1D model to evaluate the ejector performance at critical and
sub-critical operation.
The working fluid is essential since a proper working fluid can
not only benefits the ejector and system performance, but also
bring less system failures and environmental issues. As new refrig-
erants are merged and developed, they have been extensively
examined and compared in the ERS. Table 1 summarizes the typi-
cal researches [4–13]. It is obvious that (1) different working fluid
performs quite differently and the system COP is generally below
0.9; (2) there are no agreements reached in terms of which work-
ing fluid has the best performance. This is due to the different oper-
ating conditions; (3) many refrigerants are used, including

http://dx.doi.org/10.1016/j.enconman.2017.06.051
0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.
 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1209

Nomenclature

SymbolsCOP
coefficient of performance Subscripts
E_ exergy (kW) CO condenser
EIP exergetic improvement potential ratio d diffuser
ERR exergetic rehabilitation ratio D destruction
F objective function EV evaporator
h specific enthalpy (kJ/kg) EJ ejector
LMTD logarithmic mean temperature difference (°C) F fuel
m _ mass flow rate (kg/s) GE generator
P pressure (kPa) j exergy carrier positions
Q heat load (kW) k the k-th component
s specific entropy (kJ/kg
K) L loss
T temperature (°C) m mixing
DT temperature difference (°C) n nozzle
UA thermal conductance (kW/K) P product
W work (kW) PU pump
TV throttling valve
Greek symbols tot overall
l entrainment ratio 0 reference condition
e exergetic efficiency 1–12 locations in the system
g efficiency e1–e4 locations inside the ejector
Dg efficiency difference

Table 1
Summary of the working fluids compared in the ERS.

Reference Selected working fluids Operating conditions COP Recommendation

Tg (°C) Tc (°C) Te (°C)
Chen et al. [4] R123, R124, R134a, R141b, R142b, R152a, R290, R600, R600a, R717 60–100 16–40 5–15 0.15–1.65 R290
Sun [5] R11, R12, R113, R21, R123, R142b, R134a, R152a, RC318, R500, R718 80, 90 35, 25 5, 5 0.02–0.50 R152a
Selvaraju and Mani [6] R134a, R152a, R290, R600a, R717 62–87 24–36 5 0.05–0.40 134a
Roman and Hernandez [7] R123, R134a, R152a, R290, R600, R600a 70–100 25–35 5–15 0.27–0.84 R290
Chen et al. [8] R134a, R152a, R290, R430A, R600, R245fa, R600a, R1234ze(E), R436B 75–125 27–43 0–16 0.05–0.70 R245fa,
R600
Shestopalov et al. [9] R123, R141b, R142b, R236fa, R245ca, R245fa, R600, R600a 95 32 12 0.41–0.51 R141b
Besagni et al. [10] R134a, R141b, R152a, R290, R600, R600a, R601, R601a 60–180 30–50 5–15 0.03–0.40 R152a
Milazzo and Rocchetti [11] R152a, R245fa, R236ea, R600a, R1233zd(E), R1234ze(E), R134a, 80–180 25–40 5,10 0.02–0.60 R1233zd(E)
R1234yf, R365mfc, R718
Saleh [12] R134a, R227ea, R245ca, R245fa, R236ea, R236fa, R600, R600a 60–110 25–45 5–15 0.06–0.75 R245ca
*
Kasperski and Gil [13] R290, R600, R600a, R601, R601a, R602, R602a, R603, R604 60–200 40 10 0.05–0.32
*
The best COP were obtained by R600a when Tg is at 65–115 °C, R600 at 115–130 °C, R601a at 130–160 °C, R601 at 160–175 °C, and R602a at over 175 °C.

chlorofluorocarbons (CFCs), hydro chlorofluorocarbons (HCFCs)
and hydrofluorocarbon (HFCs) and so on. However, some of them
have already been prohibited or will phase out in a short future.
For example, CFCs and HCFCs have been banned by Montreal pro-
tocol because of the damage of the stratospheric ozone layer
known as ODP (Ozone Depletion Potential). Kyoto protocol was
later proposed to control greenhouse gases including HFCs due to
their global warming impacts. In 2006, EU came out with the
Directive that set the limit on GWP (Global Warming Potential)
value of refrigerant to be used in mobile air conditioning (MAC)
[14]. More recently, EU Regulation No. 517/2014 (known as F-
Gas) imposed a drastic phase-down of HFCs starting from 2015
and the reduction of greenhouse gases emissions by 2030 [15].
Hence the working fluid with zero ODP and low GWP is regulated
and promoted. For this reason, the environmentally-friendly work-
ing fluids are being popularly considered in the ERS in up-to-date
literature. It has to point out that the studies listed in Table 1 are
based on the first law of thermodynamics, called as energy analy-
sis, to comparably investigate the feasibility of working fluids in
the ERS with respect to the energy quantity. However, this method
is insufficient to locate the irreversibility within the components.
The exergy analysis, based on the second law of thermodynam-
ics, takes both quantity and quality of energy into consideration,
and is a powerful tool to identify the location, magnitude and
source of the exergy destruction. It has been widely applied to
the ERS. Pridasawas and Lundqvist [16] performed an exergy anal-
ysis on a solar-driven ERS with R600 and found that the ejector had
the largest exergy destruction, followed by the generator, the con-
denser, the pump, the evaporator and the throttling valve. Similar
result was also found by Dahmani et al. [17] using R134a and Chen
et al. [18] using R245fa as the working fluid, except that the pump
had the least exergy destruction. Alexis [19] concluded that in an
R718 ERS the most significant exergy destruction occurred in the
ejector, then in the condenser, the generator, the evaporator and
the throttling valve. Sadeghi et al. [20] pointed out that the magni-
tude of the exergy destruction in descending order was the gener-
ator, the ejector, the condenser, the evaporator, the pump and the
throttling valve when using R141b in an ERS. The disagreements
are because, on the one hand, the different operating conditions
are employed, on the other hand, the thermodynamic properties
of the working fluid take great responsibility of remarkably
impacting the ERS performance and the ejector behavior.
1210 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224
The above exergy-based method has been coined as conven-
tional exergy analysis and recognized to be useful. However, it can-
not quantify the mutual interdependencies among the system
components, and moreover, it fails to reveal the real improvement
potential since a part of exergy destruction caused by technical and
economic limitations is not considered and specified. To overcome
these shortcomings, the advanced exergy analysis is developed in a
way of further splitting the exergy destruction into several subdi-
vided parts, i.e. endogenous/exogenous parts and/or unavoidable/
avoidable parts. It was firstly proposed by Tsatsaronis et al. [21]
and has received increasing attentions ever since. Morosuk and
Tsatsaronis [22] used seven working fluids in the vapor compres-
sion refrigeration system to demonstrate the effect of different
properties on the results of advanced exergy analysis. Mosaffa
et al. [23] applied this method to an air conditioning system con-
sisting of a latent heat thermal storage unit and vapor compression
refrigeration unit, and found that the exergy destruction within the
former unit was due only to its own irreversibilities. Gong and Bou-
lama [24] discovered that the exergy destruction was mainly
endogenous and avoidable for the condenser and evaporator in
an absorption refrigeration system. New exergetic parameters
based on the advanced exergy analysis were proposed to evaluate
the performance of a district heating system [25]. In the ejector
system, Bai et al. [26] employed this method to an ejector expan-
sion transcritical CO2 refrigeration system, and found that com-
pressor had the highest priority of improvement, followed by the
ejector, the evaporator and gas cooler. An advanced exergy analysis
on an ejector assisted auto-cascade freezer system indicated that
the compressor should be firstly improved, followed by the con-
denser, the evaporator and the ejector, which was different from
the conventional analysis results [27]. Chen et al. [18] concluded
that in an ERS the ejector should be firstly improved, followed by
the condenser and the generator from the advanced exergy analy-
sis, while the conventional method dissimilarly suggested that the
generator and the condenser were the second and the third ones to
improve. The advanced exergy analysis has shown its effectiveness
in revealing more system information. More importantly, it is able
to clarify potential misinterpretations and misleading decisions
inferred from the conventional exergy analysis.
In this paper, five working fluids with zero ODP and low GWP
are selected as the candidates in the ERS with the purposes of (1)
comparing their system performance from three levels, i.e. energy
analysis, conventional exergy analysis and advanced exergy analy-
sis; (2) obtaining the additional information to make a distinction
of the priorities of component improvement among these working
fluids; (3) revealing the system performance and the component
behaviors in terms of several new parameters. It is very helpful
to guide the design and optimization of the ERS when using these
working fluids.
2. System description and working fluids
The ERS consists of a generator (GE), a condenser (CO), an evap-
orator (EV), a vapor ejector (EJ), a pump (PU), and a throttling valve
(TV), as shown in Fig. 1(a) with all processes in T-s diagram shown
in Fig. 1(b). More detailed system descriptions are referred to
[1,18].
R600, R600a, R601a, R1233zd(E) and R1234ze(E) are selected
and their thermodynamic properties are listed in Table 2 [28–
30]. It is seen that they have high critical temperature to adapt
to high heat source temperature. More importantly, they have zero
ODP and rather low GWP smaller than 20, which is a good response
to environmental concern. Noted that R600, R600a, R601a and
R1234ze(E) are flammable. Special attention has to pay on their
flammability in the practical system, which is however not dis-
cussed in present study.
The working fluid in the ERS is categorized into wet fluid, dry
fluid and isotropic fluid according to the slope of the vapor
saturation line in T-s diagram, and dry fluid is more favorable
because no superheat is required [8]. As shown in Fig. 2, the five
candidates are dry fluid due to the negative slope of the vapor
saturation line. This is the other reason why they are selected
in this study. Water is popular in ERS, but it is excluded here
since it is a wet fluid.
3. Mathematical modelling
To establish and simplify the system simulation, some assump-
tions are made as follows:
(1) The system operates at steady state, and the flow is one-
dimensional. The pressure drop and heat loss in the heat
exchangers and connecting pipes are neglected.
(2) The working fluid leaving the heat exchangers is saturated.
The velocities at the inlets and outlet of the ejector are
negligible.
(3) The losses inside the ejector are defined as efficiencies in the
nozzle (gn), the mixing chamber (gm) and diffuser (gd). The
pump has an isentropic efficiency symbolized as gPU.
(4) The system refrigerating capacity is fixed at QEV = 10 kW, i.e.
the system exergy of product is constant.
(5) The system uses water as the external fluid (brine side). The
water temperature at the inlet and outlet of the generator,
the condenser and the evaporator are kept constantly at:
T7 = 100 °C, T8 = 105 °C, T9 = 27 °C, T10 = 32 °C, T11 = 10 °C,
T12 = 15 °C. The reference state is the ambient temperature
(T0) of 25 °C and pressure (P0) of 101.32 kPa.
3.1. Energy modelling
The ejector is the key component in the ERS, and its perfor-
mance is assessed by the entrainment ratio (l), which is defined
as the ratio between the secondary mass flow rate (m _ EV ) and the
primary mass flow rate (m_ GE ). It is written as [3]:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
_
m 2gn
ðh1  he1 Þ  2ðhe4  he3 Þ=ðgm
gd Þ
l ¼ _ EV ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1Þ
mGE 2ðhe4  he3 Þ=ðgm
gd Þ  2ðh6  he2 Þ
where the points e1–e4 represent the ideal processes inside the
ejector as shown in Fig. 1(b). Clealy, l is the function of properties
of the working fluid and ejector component efficiencies gn, gm
and gd.
The coefficient of performance (COP) is the ratio of the refriger-
ating capacity produced in the evaporator (QEV) to the heat sup-
plied to the generator (QGE) and electricity consumed by the
pump (WPU). It is given as:
Q EV m_ EV
ðh6  h5 Þ
COP ¼ ¼
Q GE þ W PU m _ GE
ðh1  h4 Þ þ m _ GE
ðh4  h3 Þ
_ EV
ðh6  h5 Þ
m ðh6  h5 Þ
¼ ¼l
ð2Þ
_ GE
ðh1  h3 Þ
m ðh1  h3 Þ
In heat exchangers, the thermal conductance (UA) is expressed as
the heat load (Q) divided by logarithmic mean temperature differ-
ence (LMTD). For example, UAEV is calculated as:
Q EV
UAEV ¼ ð3Þ
LMTDEV
ðT 5  T 11 Þ  ðT 6  T 12 Þ
LMTDEV ¼ ð4Þ
ln½ðT 5  T 11 Þ=ðT 6  T 12 Þ
UA ¼ UAGE þ UACO þ UAEV ð5Þ
 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1211

Fig. 1. Schematic of (a) the ejector refrigeration system, and (b) its T-s diagram.

Table 2
Properties of the selected working fluids [28–30].

Working Fluids Physical data Safety data Environmental data

R-No Chemical formula M NBP Tc Pc LFL (%) Safety group satm (yr) ODP GWP 100 yr
g/mol °C °C kPa
R600 CH3CH2CH2CH3 58.12 0.50 152.0 3796.0 2.0 A3 0.018 0 20
R600a CH(CH3)2CH3 58.12 11.70 134.7 3640.0 1.6 A3 0.016 0 20
R601a (CH3)2CHCH2CH3 72.15 27.83 187.2 3387.0 1.3 A3 0.009 0 20
R1233zd(E) CF3CH@CHCl 130.50 18.32 165.6 3570.9 – – 0.071 0 <5
R1234ze(E) CHF@CHCF3 114.04 19.0 109.4 3636.4 7.6 – 0.045 0 6

M = molar mass; NBP = normal boiling point; Tc = critical temperature; Pc = critical pressure; LFL = lower flammability limit (% by volume in air); satm = atmospheric life;
GWP = global warming potential.

where QEV, T9 and T12 are fixed, F is proportional to the ratio of UA

and COP that determine, to a large extent, the capital investment
and operational cost, respectively, and F could considered as the
economic indicator of the ERS system [17]. Therefore, it is better
to have a small value of F. These parameters are obtained by the
energy analysis.

3.2. Conventional exergy modelling

The exergy represents the maximum useful work. In the ERS, it

can be expressed as [18]:

E_ j ¼ m
_ j
½ðhj  h0 Þ  T 0
ðsj  s0 Þ ð7Þ
where the subscript j symbolizes the exergy carrier positions 1–12
in Fig. 1, and 0 refers to the reference condition (T0 and P0).
In this study, the exergy is calculated by using the concept
‘‘fuel-product” that is defined by means of exergy dimension. By
using proper definitions for the exergy of fuel and the exergy of
Fig. 2. T-s diagram of the five candidates.
product, the exergy balance is written as following:

E_ F;k ¼ E_ P;k þ E_ D;k ð8Þ

where UAGE is obtained by summing UA of the subcooling and two-
phase areas in the generator. In the same way, UACO is the sum of UA
of the two-phase and de-superheating areas in the condenser. The
total UA in Eq. (5) can be considered as an indication of the system
size [17].
To facilitate the comparison, an objective function F, which is a
non-dimensional parameter introduced by Dahmani et al. [17], is
expressed as:
UA
ðT 9  T 12 Þ
F¼
Q EV
COP
where the subscript k refers to the k-th component in the system,
and it can be the generator (GE), the condenser (CO), the evaporator
(EV), the ejector (EJ), the pump (PU) or the throttling valve (TV). The
exergy of product (E_ P;k ) is the desired result achieved in the k-th
component being considered. The exergy of fuel (E_ F;k ) is the exergy
resources spent to the k-th component, which is conceptual and in a
general sense, not an actual fuel. The term E_ D;k is the exergy destruc-
tion within the k-th component.
At the system level, the exergy balance for overall system
ð6Þ
becomes:
1212 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224

X
E_ F;tot ¼ E_ P;tot þ E_ D;k þ E_ L;tot ð9Þ component, as calculated as Eq. (16). E_ UN;EX
D;k
is the unavoidable
exogenous part, and it cannot be reduced due to technical
where E_ F;tot and E_ P;tot are the exergy of fuel and the exergy of product
limitations in the remaining components. E_ AV;EN is the avoidable
for the overall system. The exergy loss (E_ L;tot ) is the exergy that is
D;k
endogenous part, and it can be reduced through improving the
not further being used in any system. It appears only at the level
efficiency of the k-th component. It should be paid special attention
of the overall system.
to. E_ AV;EX is the avoidable exogenous part, and it can be reduced by a
In the conventional exergy modelling, two parameters are D;k

defined to evaluate the component and system performance: structural improvement of the overall system or by improving the
efficiency of the remaining system components.
E_ In the advanced exergy modelling, several parameters are
ek ¼ _ P;k ð10Þ
EF;k employed as [25]:

E_ P;tot E_
etot ¼ ð11Þ emodified;k ¼ _ P;k
ð17Þ
E_ F;tot EF;k  E_ UN  E_ AV;EX
D;k D;k

where ek and etot are the exergy efficiency of the k-th component
E_
and overall system, respectively. emodified;tot ¼ _ P;tot
ð18Þ
EF;tot  E_ UN _ AV;EX
D;tot  ED;tot
3.3. Advanced exergy modelling
E_ D;k  E_ UN _ AV;EX
D;k  ED;k
The conventional exergy analysis is improved by splitting E_ D;k ERRk ¼ ð19Þ
into endogenous/exogenous parts or unavoidable/avoidable parts. E_ D;tot
Moreover, a combination of the two splitting approaches reveals
more detailed information. E_ UN _ AV;EX
D;k þ ED;k
E_ D;k is not only caused by the considered k-th component EIPk ¼ ð20Þ
E_ F;k
itself, but also related to the remaining ones in the system.
To quantify the component interdependence, E_ D;k can be readily where emodified,k is the modified exergy efficiency of the k-th compo-
split as: nent. emodified,tot is the exergy efficiency of the overall system after
the maximum improvements. The terms ERRk and EIPk are the exer-
E_ D;k ¼ E_ EN _ EX
D;k þ ED;k ð12Þ
getic rehabilitation ratio and exergetic improvement potential ratio,
where E_ EN
D;k is the endogenous part that is associated with its own
respectively.

operational inefficiency. The term E_ EX is the exogenous part that

D;k
is the exergy destruction within k-th component caused by the 3.4. Simulation solution
remaining components. This splitting approach is used to determine
whether the attention should be paid to the considered k-th compo- To quantify all the parts of the exergy destruction, the source
nent or the remaining ones. of E_ D;k in the ERS and their corresponding representative param-
In practical, an improvement is always constrained by the lim- eters are firstly identified. The exergy destruction within the
its. To achieve a more realistic measure of improvement, E_ D;k can heat exchangers is due to heat transfer through the finite tem-
be alternatively split as: perature difference, and is expressed as the pinch temperature
difference, i.e. DTGE, DTCO and DTEV in Fig. 1(b). The source of
E_ D;k ¼ E_ UN _ AV
D;k þ ED;k ð13Þ exergy destruction within the ejector is quite complex, involving
friction, mixing loss, shock loss etc., which are expressed in
E_ D;k
UN
terms of gn, gm and gd in Eq. (1). gPU denotes the source of
E_ UN _
D;k ¼ EP;k
ð _ Þ ð14Þ the exergy destruction within the pump. The exergy destruction
EP;k
within the throttling valve is caused by the reduction in pressure
where E_ UN
D;k is the unavoidable exergy destruction that cannot be
and temperature from the condenser level to the evaporator
level [1].
reduced due to technological limitations (material characteristics,
Secondly, the magnitudes of the parameters that represents the
production costs and manufacturing methods). E_ AV is the avoidable D;k
sources of E_ D;k have to carefully determine. Table 3 summarizes
exergy destruction that can be avoided. Hence, more attentions
their values in the ERS at the normal, ideal and unavoidable condi-
should be paid to reduce E_ AV . Moreover, E_ UN is calculated from Eq.
D;k D;k tions [1,18,24]. Though the temperatures of the external fluid
UN
E_ (water) at the heat exchanger inlet and outlet are fixed, the tem-
(14) in which the term ð E_ D;k Þ is obtained at the unavoidable
P;k
peratures in the heat exchangers for the candidates are not the
condition. same, which can be found in Table A1.
In order to get more insights, these two approaches mentioned Finally, the real cycle, the ideal cycle, the hybrid cycles and
above are combined as: the unavoidable cycle are created to accordingly calculate the
E_ D;k ¼ E_ UN;EN
D;k þ E_ UN;EX
D;k þ E_ AV;EN
D;k þ E_ AV;EX
D;k ð15Þ aforementioned parameters in Eqs. (1)(20). More illustrations
and calculations are found in Ref. [18]. Fig. 3 shows the flow
chart of computational procedure to gain the results of the
E_ D;k
UN

E_ UN;EN
D;k ¼ E_ EN
P;K
ð _ Þ ð16Þ energy, conventional and advanced exergy modelling. In this
EP;k study, the calculating program is written with Fortran Language
and the thermodynamic properties of the selected five working
where E_ UN;EN
D;k is the unavoidable endogenous exergy destruction that fluids and water are taken from the NIST database and
cannot be reduced due to technical limitations in the k-th subroutines [30].
 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1213

Table 3
Parameters used in the normal, ideal and unavoidable conditions.

Component Parameter Real condition Ideal condition Unavoidable condition

Generator (GE) DTGE 7.00 °C 0 0.50 °C
Condenser (CO) DTCO 4.00 °C 0 0.50 °C
Evaporator (EV) DTEV 2.00 °C 0 0.50 °C
Ejector (EJ) gn 0.93 E_ D;EJ ¼ 0 0.98
gm 0.90 0.95
gd 0.90 0.98
Pump (PU) gPU 0.50 1 0.95
Throttling valve (TV) – Isenthalpic Isentropic Isenthalpic

3.5. Modeling validation 4. Results of the ERS at normal operating condition

Validation of the model is carried out in two stages. The first
stage is from the energy analysis. The l and COP from the present
model is compared with the experimental results by Shestopalov
[31] using R245fa. It is found in Table 4 that the calculated l and
COP are in good agreement with those in [31] with the maximum
error of 6.35%. The second stage is from the exergy analysis. The
results obtained in the current study compared with those pre-
sented by Alexis [19] using water. Table 5 shows the comparison
of operating parameter at the typical locations in the ERS which
are corresponding to those in Fig. 1, and Table 6 shows the compar-
ison of the exergy destruction results. The error of E_ D;k and its per-
centage in E_ D;tot is found to be 8.82% maximally, which can be due
to the used software and calculation accuracy of the thermody-
namic properties.
At the normal condition, the performance of the ERS working
with the five candidates is shown in Table 7. A more detailed data
can be found in Table A2.
By using the energy analysis (as shown in column 3–5 of
Table 7), it is obtained that R1233zd(E) has the highest COP and
the lowest UA and F, followed by R600 and R601a, and then by
R600a and R1234ze(E). This is explicitly explained from Eq. (6) that
the smallest value of UA at the numerator is accompanied by the
highest COP at the denominator, which of course makes F the low-
est. Regardless of the working fluid, UACO is larger than the sum of
UAGE and UAEV. This is due the largest QCO and a moderate DTCO in
the condenser at the normal condition.
From the conventional exergy analysis (column 6–8 of Table 7),
for a fixed E_ P;tot = 0.44 kW, a high E_ F;tot leads to a low etot. It is clear

Fig. 3. Flow chart of the computational procedure.

1214 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224

Table 4
Model validation with the experiment l and COP from [31].

Parameters Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8 Case 9
Tg (°C) 90 90 90 95 95 95 100 100 100
Te (°C) 8 8 8 12 12 12 16 16 16
Tc (°C) 34.7 31.7 29 35.6 33 29.7 36.7 33.9 30.8
l [31] 0.355 0.437 0.529 0.449 0.544 0.673 0.587 0.727 0.892
cal 0.357 0.451 0.552 0.456 0.553 0.710 0.566 0.697 0.886
error (%) 0.63 3.13 4.37 1.54 1.63 5.52 3.66 4.18 0.66
COP [31] 0.27 0.334 0.407 0.34 0.413 0.516 0.447 0.556 0.689
cal 0.264 0.336 0.413 0.338 0.412 0.533 0.420 0.521 0.666
error(%) 2.06 0.50 1.58 0.54 0.19 3.28 5.97 6.35 3.30

Cal: calculated value by present model with gn = 0.95, gm = 0.83, gd = 0.88.

Error = (value from the present model – value in [31])/value in [31].

Table 5
Model validation of the operating condition in ERS from [19].

Position T (°C) P (kPa) h (kJ/kg) s (kJ/kg
K) _ (kg/s)

m e (kJ/kg)

[19] Cal [19] Cal [19] Cal [19] Cal [19] Cal [19] Cal
1 158.87 158.87 6.0000 6.0067 2743.80 2756.19 6.7382 6.7589 0.0764 0.0767 770.570 777.730
2 89.71 99.81 0.1008 0.1010 2655.30 2687.08 8.3682 8.4433 0.1198 0.1198 204.290 214.833
3 46.00 46.00 0.1008 0.1010 194.41 192.62 0.6551 0.6517 0.1198 0.1198 4.471 4.462
4 46.02 46.02 6.0000 6.0067 195.01 193.22 0.6551 0.6517 0.0764 0.0767 5.066 5.059
5 6 6 0.0093 0.0094 194.41 192.62 0.6962 0.6910 0.0434 0.0431 7.560 7.046
6 6 6 0.0093 0.0094 2499.50 2511.89 8.9770 8.9993 0.0434 0.0431 129.990 123.363
7 165 165 7 7.0093 699.29 697.24 1.9963 1.9923 0.0949 0.0951 116.170 116.094
8 165 165 7 7.0093 2750.50 2762.81 6.6863 6.7066 0.0949 0.0951 792.500 799.668
9 20 20 – 1 85.54 84.01 0.2990 0.2965 3.5165 3.5465 0 0
10 40 40 – 1 169.38 167.62 0.5756 0.5724 3.5165 3.5465 2.735 2.728
11 12 12 – 1 51.92 50.50 0.1827 0.1806 2.9746 2.9850 0.467 0.466
12 20 20 – 1 85.54 84.01 0.2990 0.2965 2.9746 2.9850 0 0
0 – 20 – 1 – 84.01 – 0.2965 – – – –

Cal: calculated value by the present model with gn = 0.96, gm = 0.91, gd = 0.93, gPU = 1.

Table 6
Model validation of the exergy analysis results from [19].

Comp onent E_ F;k (kW) E_ P;k (kW) E_ D;k (kW) E_ D;k =E_ D;tot

[19] Cal [19] Cal [19] Cal error (%) [19] (%) Cal (%) error (%)
GE 75.635 65.035 69.913 59.245 5.722 5.790 1.19 10.75 10.72 0.29
CO 24.474 25.201 10.152 9.675 14.322 15.525 8.40 26.89 28.74 6.89
EV 0.328 1.390 4.249 5.015 3.921 3.625 7.55 7.36 6.71 8.82
EJ 53.243 43.160 24.474 14.582 28.769 28.578 0.66 54.02 52.91 2.06
TV 0.194 0.212 0.328 0.284 0.522 0.496 4.95 0.98 0.92 6.30
PU 0 0 0 0 – – – 0 0 0
tot – – – – 53.256 54.015 1.42 100 100 0

Cal: calculated value by present model.

Error = (value from the present model – value in [19])/value in [19].

that R1233zd(E) has the highest etot, followed by R601a and R600,
and then by R600a and R1234ze(E). R1233zd(E) is therefore con-
sidered as a good working fluid for the ERS. For all candidates,
E_ D;EJ is the largest, accounting for about a half of E_ D;tot . Thus the
ejector should be firstly improved, followed by the generator and
the condenser. However, the evaporator, the pump and the throt-
tling valve have different priorities of improvement. For instance,
R600 has the fourth improvement of evaporator, then the pump
and the throttling valve. R600a and R1234ze(E) have the improve-
ment order as: the pump, the evaporator and the throttling valve.
R601a and R1233zd(E) follow the improvement order of evapora-
tor, the throttling valve and the pump. Regarding ek, these five can-
didates have the same decreasing order: the throttling valve, the
The advanced exergy analysis gives more detailed insights of
the operating characteristics with a focus on E_ AV;EN
D;k , emodified,k, ERRk
_ AV;EN
and EIPk, as seen in column 9–12 of Table 7. ED;k is an important
parameter and represents the achievable improvements, which
can be used to guide the system enhancement [26]. It is found that
R600, R600a, R601a and R1233zd(E) have the same decreasing
trend of E_ AV;EN : the ejector, the condenser, the generator, the evap-
D;k
orator, the pump and the throttling valve, which indicates the pro-
prieties of improving these components, while R1234ze(E) has a
different order: the ejector, the condenser, the pump, the evapora-
tor, the generator and the throttling valve. Moreover, for all candi-
dates, E_ AV;EN and E_ AV;EN have the largest proportions of E_ AV;EN of
D;EJ D;CO D;tot

generator, the evaporator, the pump, the condenser and the ejec- 54.2%  76.8% and 11.6%  22.0%, respectively. Thus, the major
tor. The lowest eEJ re-emphasizes the highest priority of improving attentions of improvement should be paid to the ejector and the
the ejector. condenser. The throttling valve has zero E_ AV;EN , suggesting the D;TV
 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1215

Table 7
Results of ERS with the selected five candidates at normal condition.

Working fluids Comp-onents Energy analysis Conventional exergy analysis Advanced exergy analysis
COP UA F E_ tot E_ D;k e E_ D;k
AV;EN emodified ERRk EIPk

– kW/K - kW kW % kW % % %
R600 GE – 2.49 – ĖF,tot = 6.04 1.18 79.57 0.07 98.53 1.36 19.24
CO – 6.06 – 0.83 40.39 0.16 77.75 3.19 48.05
EV – 2.51 – ĖP,tot = 0.44 0.17 72.42 0.06 88.61 1.12 18.27
EJ – – – 2.63 30.96 0.83 58.73 16.42 47.28
PU – – – ĖL,tot = 0.56 0.12 51.84 0.03 82.87 0.52 37.44
TV – – – 0.12 84.42 0.00 100.00 0.00 15.58
Tot 0.353 11.06 37.61 5.04 7.25 1.14 20.47 – –
R600a GE – 2.52 – ĖF,tot = 6.44 1.31 78.40 0.07 98.61 1.23 20.50
CO – 6.29 – 0.84 40.81 0.17 77.53 3.10 47.36
EV – 2.51 – ĖP,tot = 0.44 0.17 72.42 0.06 88.61 1.04 18.27
EJ – – – 2.78 30.00 0.99 54.52 18.36 44.98
PU – – – ĖL,tot = 0.58 0.18 51.91 0.04 83.29 0.73 37.68
TV – – – 0.14 82.40 0.00 100.00 0.00 17.60
Tot 0.336 11.32 40.36 5.42 6.80 1.33 18.69 – –
R601a GE – 2.58 ĖF,tot = 6.00 1.17 80.09 0.07 98.46 1.48 18.66
CO – 6.07 0.90 38.90 0.13 81.94 2.52 52.53
EV – 2.51 ĖP,tot = 0.44 0.17 72.42 0.06 88.61 1.13 18.27
EJ – – 2.60 31.68 0.32 78.93 6.44 59.87
PU – – ĖL,tot = 0.57 0.05 51.75 0.01 82.18 0.25 37.04
TV – – 0.11 85.50 0.00 100.00 0.00 14.50
Tot 0.349 11.16 38.38 5.00 7.29 0.59 27.39 – –
R1233zd(E) GE – 2.51 – ĖF,tot = 5.89 1.14 80.13 0.07 98.54 1.39 18.68
CO – 6.03 – 0.83 40.13 0.15 79.17 3.01 49.31
EV – 2.51 – ĖP,tot = 0.44 0.17 72.42 0.06 88.61 1.15 18.27
EJ – – – 2.57 31.19 0.42 73.30 8.68 57.45
PU – – – ĖL,tot = 0.56 0.07 51.79 0.02 82.86 0.33 37.49
TV – – – 0.11 85.54 0.00 100.00 0.00 14.46
Tot 0.358 11.05 37.06 4.89 7.44 0.71 25.62 – –
R1234ze(E) GE – 2.43 – ĖF,tot = 7.05 1.55 75.84 0.05 98.98 0.84 23.38
CO – 6.69 – 0.84 41.94 0.18 76.82 3.06 45.41
EV – 2.51 – ĖP,tot = 0.44 0.17 72.42 0.06 88.61 0.94 18.27
EJ – – – 2.97 28.52 1.19 49.93 19.83 42.88
PU – – – ĖL,tot = 0.61 0.31 52.04 0.07 83.14 1.14 37.41
TV – – – 0.16 79.79 0.00 100.00 0.00 20.21
Tot 0.316 11.63 44.16 6.00 6.21 1.55 16.88 – –

low possibility of enhancing the system performance through the
improvement of the throttling valve itself. Regarding emodified,k, it
is higher than ek that is from the conventional exergy analysis. This
means that when the maximum improvement is realized on a sys-
tem component, etot will be as much as emodified,k [32]. For example,
the etot for R601a (7.29%) can be increased to 27.39% denoted by
[22,26]. E_ D;EJ and E_ D;TV are mainly caused by the ejector and
throttling valve themselves since their E_ EN are larger than E_ EX . This
D;k D;k
means that improvements of these three components are more
effective. On the contrary, the generator, the condenser and the
pump have smaller E_ EN than E_ EX , and they can be improved by
D;k D;k

emodified,tot. What is more, ERREJ and ERRCO have the largest values,
which in another way suggest that the ejector and the condenser
have the top priorities of improvement. These two components
also have the highest values of EIPEJ and EIPCO because their emodi-
fied,EJ and emodified,CO are lowest among all the components.
Fig. 4 shows the grassmann diagrams for energy and conven-
tional energy flow as well as the advanced exergy analysis results
in terms of percentages. The energy analysis leads to the conclu-
sion that the ejector and the throttling valve are free of loss. More-
over, compared to heat loads in the heat exchangers, the pump
power is negligible.
From the conventional exergy analysis, the exergy are mostly
destructed in the ejector, then in the generator and the condenser.
the improvement of other components; (2) splitting the exergy
destruction into unavoidable and avoidable parts suggests that
the five candidates have larger E_ UN than E_ AV within the generator,
D;k D;k
the evaporator and the throttling valve, while the condenser and
the pump have the opposite behaviors. Interestingly, the propor-
tions of E_ UN and E_ AV within the ejector are different for these can-
D;EJ D;EJ
didates, e.g. R600, R600a and R1234ze(E) have larger proportion of
E_ AV than E_ UN , whereas the ejector of R601a and R1233zd(E)
D;EJ D;EJ
behave oppositely. This means that the ejector behavior is more
dependent on the properties of the working fluid; (3) combining
the two spitting approaches demonstrates that for all candidates
the generator is dominated by E_ UN;EX , thus it has low potential of
D;GE

The percentages of E_ D;EV , E_ D;PU and E_ D;TV in E_ D;tot are quite small.
Hence these components are least to consider for improvement.
The advanced exergy analysis focuses more on the individual
components: (1) splitting the exergy destruction into endogenous
and exogenous parts indicates all candidates have the similar pro-
portion of E_ EN and E_ EX . E_ D;EV is entirely from evaporator itself due
D;k D;k
just improving itself and is heavily constrained by technical limited
in the remaining components. The condenser has larger proportion
of E_ AV;EX than E_ AV;EN , so it is more effective to improve the other
D;CO D;CO
components than the condenser. Regardless of the working fluid,
63.5% of the exergy destruction within the evaporator cannot be
reduced due to limitations in itself (E_ UN;EN ). The ejector is generally
D;EV

to 100% of E_ EN
D;EV in order to maintain the fixed refrigerating capacity characterized by E_ AV;EN
D;EJ > E_ AV;EX
D;EJ , which means it is more significant
and overall exergy of product [18], which agrees the findings from to improve the ejector than the remaining components. The pump
1216 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224
Fig. 4. Analytical results in terms of percentages.
has the majority of E_ AV;EX
D;PU , so improving the remaining components
can also largely reduce E_ D;PU . The throttling valve is free of
E_ AV;EN _ AV
D;TV , hence all its ED;TV is from the remaining components and
_ED;TV can only be reduced by improving the remaining components.
5. Parametric evaluation
In this section, the pinch temperature differences in three heat
exchangers (DTGE, DTCO and DTEV) and the ejector efficiencies (gn,
gm and gd) as well as the pump efficiency (gPU) are considered as
the inputs as they are subjected to vary at different normal operat-
ing conditions. Therefore, it is important to carry out a parametric
investigation on how these variables influence the ERS perfor-
mance and the component behaviors, which hereby is expressed
in terms of UA and F at the energy level, ek and etot at the conven-
tional exergy level, ERRk and EIPk at the advanced exergy level. It is
noted that when one of the variables is selected to discuss, the
others are fixed as listed in Table 3.
5.1. Effect of the pinch temperature difference in the generator (DTGE)
Fig. 5 shows the system and generator behaviors versus DTGE.
From the energy analysis, it is obvious that there is an optimum
DTGE (11 °C for R600 and R600a, 12 °C for R601a and R1233zd(E),
10 °C for R1234ze(E)) that minimizes UA. In a general sense, UA
is expected to have the contrary variation with the change of DT
(more precisely LMTD in Eq. (3)) in a heat exchanger. However,
as DTGE increases, the TGE decreases, leading to the decrease of l
and the increase of m _ GE as m
_ EV is constant. Consequently, QGE and
QCO increase and COP reduces. Therefore, the changes of QGE and
LMTDGE cooperatively contribute to the variations of UAGE. While
UACO increases due to the increasing QCO and unchanged LMTDCO,
UAEV keeps constant since QEV and LMTDEV are unchanged. As a
result of these complex combinations, UA, as in Eq. (5), has the
characteristics illustrated in Fig. 5(a). Meanwhile, it is noted that
there also exists an optimum value of DTGE for the minimized F
as well, i.e. 6 °C for R1234ze(E) and 7 °C for the other four candi-
dates. The feature as shown in Fig. 5(b) is due to the variations of
J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1217

Fig. 5. Influence of the pinch temperature difference in the generator.

1218 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224
UA and COP as expressed as Eq. (6). Moreover, R1233zd(E) always
performs the best in terms of UA and F, followed by R600, R601a,
R600a and R1234ze(E).
From the conventional exergy analysis in Fig. 5(c) and (d), eGE
and etot decrease monotonously with the increasing DTGE. For
R600, R600a, R601a and R1233zd(E), an change of one Celsius
degree in DTGE changes eGE and etot by approximately 0.76% and
0.11%, respectively. For R1234ze(E), they are 0.58% and 0.08%. In
this regard, the performance of the five candidates can be
sequenced in descending order as: R1233zd(E), R601a, R600,
R600a and 1234ze(E), which is a little different from that obtained
by the energy analysis.
From the advanced exergy analysis, the larger DTGE is, the more
E_ D;GE can be recovered and the more the generator can be
improved, which is indicated by the increasing ERRGE in Fig. 5(e).
An increase in DTGE gives a linear rise of ERRGE with a sensitivity
of 0.14% for R1234ze(E) and around 0.21% for the other four candi-
dates per Celsius degree. Similarly, EIPGE increases linearly with the
increasing DTGE in Fig. 5(f). It changes 0.45% for R1234ze(E) and
about 0.58% for the other four candidates per Celsius degree. In this
respect, there has the descending order of improvement potentials
as: R1233zd(E), R601a, R600, R600a and R1234ze(E).
5.2. Effect of the pinch temperature difference in the condenser (DTCO)
Fig. 6 shows the system and condenser behaviors versus DTCO.
The behaviors of UA and F in Fig. 6(a) and (b) can be interpreted
similarly as in the discussion of DTGE aforementioned. It is more
interesting to explore that the five candidates have the same opti-
mum DTCO, 7 °C and 3 °C, to gain the minimum UA and F, respec-
tively. From Fig. 6(c) and (d), all candidates have also the similar
variations that one Celsius degree increase in DTCO could cause
average reductions of 3.13% in eCO and 0.48% in etot, which is much
more dramatic than those of DTGE. Fig. 6(e) illustrates that all can-
didates have closely increasing gradient of ERRCO with the increas-
ing DTCO, which is averagely 0.66% per Celsius degree. Fig. 6(f)
indicates that R601a has the highest value and the smoothest
change of EIPCO, followed by R1233zd(E) and R600, then by
R600a and R1234ze(E).
5.3. Effect of the pinch temperature difference in the evaporator (DTEV)
Fig. 7 shows the system and evaporator behaviors versus DTEV.
It is also noted all candidates have the similar optimum DTEV, i.e.
3 °C and 2 °C, that minimize UA and F, respectively, as shown in
Fig. 7(a) and (b). Moreover, they have the same eEV as shown in
Fig. 7(c), and this is derived from the same temperatures and the
fixed QEV = 10 kW in the evaporator. One Celsius degree increase
in DTEV decreases eEV for 3.55%. Speaking of etot in Fig. 7(d), R600,
R600a and R1234ze(E) have the similar decreasing tendency when
increasing DTEV and sensitivity of 0.32% per Celsius degree, while
R601a and R1233zd(E) have the a sensitivity of 0.49% per Celsius
degree of DTEV. Furthermore, it is seen from Fig. 7(e) that for all
candidates the increases of ERREV are quite smooth when DTEV is
below 6 °C. As DTEV is beyond 6 °C, the increases of ERREV for
R601a and R1233zd(E) become much slower than that of the other
three candidates. All candidates have the same EIPEV as well, as
shown in Fig. 7(f), which suggests the EIPEV is solely related to
the operating conditions in the evaporator. One Celsius degree
increase in DTEV results in 0.90% decreasing EIPEV.
5.4. Effect of the efficiencies in the ejector
Fig. 8 shows the system and ejector behaviors versus the ejector
efficiencies in the ejector. It is noted that the x-axis is the general-
ized ejector efficiency difference (DgEJ) between the considered
values and the baseline at normal condition in Table 3, gn, gm
and gd change simultaneously at the same step. This is just to facil-
itate the simulations here. In practical, gn, gm and gd do not neces-
sarily have to vary at the same pace, and they are generally
allocated to different values. From the energy analysis in Fig. 8
(a) and (b), UA and F decrease sharply with the increasing DgEJ.
This can be explained as: when just increasing DgEJ, QGE and QCO
are subjected to decrease without changing the LMTDGE and
LMTDCO, resulting the reducing UA. The increasing DgEJ also bene-
fits COP, F is hence decreased as well. Furthermore, the decreasing
rates of UA and F decrease for all candidates, as indicated by the
concave curves.
The conventional exergy results in Fig. 8(c) and (d) indicate that
eEJ and etot of R600, R600a and R1234ze(E) have approximately lin-
ear increase with the increasing DgEJ, while variations eEJ and etot
for R1233zd(E) and R601a are indicated by the convex curves.
R601a and R1233zd(E) have higher ek and etot than R600, R600a
and R1234ze(E) when DgEJ is smaller than 0.01, it becomes reverse
when DgEJ is beyond 0.01. In addition, the effects of DgEJ on eEJ and
etot are significant. For example, as DgEJ increases from 0.05 to
0.05, eEJ and etot of R1234ze(E) dramatically increase from 6.70%
to 46.62%, and 1.38% to 10.33%, respectively.
The advanced exergy results in Fig. 8(e) show that ERREJ
decreases enormously with the increase of DgEJ. ERREJ has much
larger values than ERRGE, ERRCO and ERREV that are shown Fig. 5
(e), Fig. 6(e) and Fig. 7(e), which implies the importance of improve
the ejector. More interesting features are found in Fig. 8(f) that
R600, R600a and R1234ze(E) have the optimum value of 0.03,
0.03 and 0.02 for DgEJ where EIPEJ has the minimum values.
However, EIPEJ for R601a and R1233zd(E) is characterized by the
monotonous increase with the increasing DgEJ.
5.5. Effect of the pump efficiency (gPU)
Fig. 9 shows the system and pump behaviors versus gPU.
Clearly, the influence of gPU on UA and F is negligible, as illus-
trated in Fig. 9(a) and (b). This is because WPU is insignificant
(for instance, using R1233zd(E) WPU is only 0.37 kW even the
pump has a low gPU of 0.2), and gPU hardly changes the operating
conditions and COP. Moreover, ePU is almost independent on the
working fluid and increases linearly with the increasing gPU, see
Fig. 9(c). Since an increase in gPU causes a little decrease in
WPU, which is good for etot from Eq. (11), leading to the slight
increasing etot, as shown in Fig. 9(d). It is seem from Fig. 9(e) that
ERRPU decreases with the increasing gPU, telling that the potential
of improving the pump gets smaller. Regarding EIPPU, it also has
the features of independence of the working fluid with a roughly
linear trend. A 0.1 increase in gPU causes a 7.37% decrease in
EIPPU, as shown in Fig. 9(f).
Summarily, the energy analysis concludes that the optimum
values of DTGE are different for different working fluids, while
the optimum value of DTCO is the same for the five candidates,
and so is DTEV. The effects of DgEJ on UA and F are considerable,
the effects of gPU are negligible. The conventional exergy analysis
suggests that ek is most impacted by DTEV, follow by DTCO and
DTGE, whereas etot is most impacted by DTCO, follow by DTEV and
DTGE. A slight increase of DgEJ can lead to large improvement of
eEJ and etot. gPU determines ePU and insignificantly impacts on etot.
The advanced exergy analysis indicates that ERR increase with
the increasing DTGE, DTCO and DTEV as well as the decrease DgEJ
and gPU. Moreover, the increase of DTEV and gPU results in the
decreasing EIPEV, which are opposite to these of DTGE and DTCO.
It is very interesting to find that there have values of DgEJ that min-
imizes EIPEJ.
J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1219

Fig. 6. Influence of the pinch temperature difference in the condenser.

1220 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224

Fig. 7. Influence of the pinch temperature difference in the evaporator.

J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1221

Fig. 8. Influence of the ejector efficiencies.

1222 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224

Fig. 9. Influence of the pump efficiency.

 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224 1223

Table A1
Specified operating conditions of ERS with the five candidates at the normal, ideal and unavoidable conditions.

Conditions Working fluids Tg Tc Te gn gm gd gPU

°C °C °C – – – –
Normal Condition R600 94.85 35.54 8.00 0.93 0.90 0.90 0.50
R600a 95.09 35.56 8.00 0.93 0.90 0.90 0.50
R601a 94.71 35.41 8.00 0.93 0.90 0.90 0.50
R1233zd(E) 94.71 35.56 8.00 0.93 0.90 0.90 0.50
R1234ze(E) 95.70 35.77 8.00 0.93 0.90 0.90 0.50
Ideal Condition R600 102.14 31.93 10.00 E_ D;EJ ¼ 0 1.00
R600a 102.44 31.95 10.00 1.00
R601a 101.96 31.89 10.00 1.00
R1233zd(E) 101.96 31.95 10.00 1.00
R1234ze(E) 103.35 32.10 10.00 1.00
Unavoidable Condition R600 101.62 32.21 9.50 0.98 0.95 0.98 0.95
R600a 101.91 32.27 9.50 0.98 0.95 0.98 0.95
R601a 101.45 31.96 9.50 0.98 0.95 0.98 0.95
R1233zd(E) 101.45 32.12 9.50 0.98 0.95 0.98 0.95
R1234ze(E) 102.78 32.55 9.50 0.98 0.95 0.98 0.95

Table A2
Detailed analytical results of ERS with the five candidates.

Working Fluids Compo nents Energy Analysis Conventional exergy Advanced exergy analysis
analysis
Q E_ F;k E_ P;k E_ D;k ĖD,k E_ EN
D;k E_ EX
D;k E_ UN
D;k E_ AV
D;k E_ UN;EN
D;k
E_ UN;EX
D;k E_ AV;EN
D;k
E_ AV;EX
D;k

kW kW kW kW kW kW kW kW kW kW kW kW
R600 GE 28.09 5.80 4.61 1.18 0.27 0.91 0.97 0.21 0.20 0.77 0.07 0.14
CO 38.34 1.39 0.56 0.83 0.33 0.50 0.37 0.46 0.17 0.20 0.16 0.30
EV 10.00 0.60 0.44 0.17 0.17 0.00 0.11 0.06 0.11 0.00 0.06 0.00
EJ 0.00 3.80 1.18 2.63 1.85 0.78 1.35 1.28 1.02 0.33 0.83 0.45
PU 0.25 0.25 0.13 0.12 0.03 0.09 0.01 0.11 0.00 0.01 0.03 0.08
TV 0.00 0.75 0.63 0.12 0.08 0.04 0.09 0.03 0.08 0.01 0.00 0.03
Tot – 6.04 0.44 5.04 2.71 2.33 2.90 2.14 1.57 1.33 1.14 1.00
R600a GE 29.34 6.05 4.75 1.31 0.30 1.01 1.12 0.19 0.23 0.89 0.07 0.12
CO 39.72 1.42 0.58 0.84 0.33 0.51 0.37 0.47 0.16 0.21 0.17 0.30
EV 10.00 0.60 0.44 0.17 0.17 0.00 0.11 0.06 0.11 0.00 0.06 0.00
EJ 0.00 3.97 1.19 2.78 1.94 0.84 1.26 1.52 0.95 0.31 0.99 0.53
PU 0.38 0.38 0.20 0.18 0.04 0.14 0.01 0.17 0.00 0.01 0.04 0.13
TV 0.00 0.77 0.64 0.13 0.08 0.05 0.10 0.03 0.08 0.02 0.00 0.03
Tot – 6.44 0.44 5.42 2.86 2.56 2.97 2.45 1.53 1.44 1.33 1.12
R601a GE 28.56 5.89 4.72 1.17 0.27 0.90 0.93 0.24 0.20 0.73 0.07 0.17
CO 38.67 1.47 0.57 0.90 0.33 0.57 0.45 0.45 0.20 0.25 0.13 0.32
EV 10.00 0.60 0.44 0.17 0.17 0.00 0.11 0.06 0.11 0.00 0.06 0.00
EJ 0.00 3.80 1.20 2.60 1.84 0.76 1.99 0.61 1.52 0.47 0.32 0.29
PU 0.11 0.11 0.06 0.05 0.01 0.04 0.00 0.05 0.00 0.00 0.01 0.04
TV 0.00 0.74 0.63 0.11 0.07 0.04 0.08 0.03 0.07 0.01 0.00 0.03
Tot – 6.01 0.44 5.00 2.69 2.31 3.56 1.44 2.10 1.46 0.59 0.85
R1233zd(E) GE 27.80 5.74 4.60 1.14 0.26 0.88 0.91 0.23 0.19 0.72 0.07 0.16
CO 37.95 1.39 0.56 0.83 0.32 0.51 0.40 0.43 0.18 0.22 0.14 0.29
EV 10.00 0.60 0.44 0.17 0.17 0.00 0.11 0.06 0.11 0.00 0.06 0.00
EJ 0.00 3.73 1.16 2.57 1.77 0.80 1.78 0.79 1.35 0.43 0.42 0.37
PU 0.15 0.15 0.08 0.07 0.02 0.05 0.00 0.07 0.00 0.00 0.02 0.05
TV 0.00 0.74 0.63 0.11 0.07 0.04 0.08 0.03 0.07 0.01 0.00 0.03
Tot – 5.89 0.44 4.89 2.60 2.29 3.28 1.61 1.89 1.39 0.71 0.90
R1234ze(E) GE 31.00 6.40 4.85 1.55 0.33 1.22 1.48 0.07 0.28 1.20 0.05 0.02
CO 41.65 1.45 0.61 0.84 0.35 0.49 0.40 0.44 0.17 0.23 0.18 0.26
EV 10.00 0.60 0.44 0.17 0.17 0.00 0.11 0.06 0.11 0.00 0.06 0.00
EJ 0.00 4.16 1.19 2.97 1.99 0.98 1.07 1.90 0.79 0.28 1.20 0.70
PU 0.65 0.65 0.34 0.31 0.07 0.24 0.02 0.29 0.00 0.02 0.07 0.22
TV 0.00 0.80 0.64 0.16 0.10 0.06 0.12 0.04 0.10 0.02 0.00 0.04
Tot – 7.05 0.44 6.00 3.02 2.98 3.21 2.79 1.47 1.74 1.55 1.24

6. Conclusions
In this study, ERS is investigated through the comprehensive
thermodynamic assessment by using energy analysis, conventional
exergy analysis and advanced exergy analysis. Five working fluids
with zero ODP and low GWP are compared. Their system perfor-
mance and the component behavior are studied. The main findings
are concluded as following:
(1) At the normal condition, the energy analysis shows that the
highest performance in terms of COP, UA and F is ranked as:
R1233zd(E), R600, R601a, R600a and R1234ze(E). However,
the conventional exergy analysis indicates R601a has the
secondary highest performance based on etot and R600 is
the third, while the advanced exergy analysis suggests
R601a have the highest potential performance represented
by emodified,tot.
1224 J. Chen et al. / Energy Conversion and Management 148 (2017) 1208–1224
(2) Regarding the component, the conventional analysis shows
the five candidates have the same highest priority of
improvement, i.e. the ejector, followed by the condenser
and then the generator, but their improvement orders of
the evaporator, the pump and the throttling valve are depen-
dent on the working fluids. The advanced exergy analysis
agrees the ejector should be the first to improve as well,
but disagreeably finds that the condenser should be the sec-
ond. Furthermore, the orders of improving the other four
component for R600, R600a, R601a and R1233zd(E) are the
same, but differs from R1234ze(E).
(3) The parametric study proves the existence of the optimum
values of DTGE, DTCO and DTEV which minimize UA and F.
The influence of DgEJ is more related to the working fluid
properties. The effects of gPU on UA, F and etot are negligible.
(4) It is recommend R1233zd(E) as the good working fluid for
the ERS since it not only has generally high system perfor-
mance, but also is a non-flammable substance.
Acknowledgement
This study is financially supported by Guangdong Special Fund-
ing for Applied Technology R&D (2016B020243010) and Science
and Technology Program of Guangzhou (201704030108). The
authors would like to express their thanks to this sponsorship.
Appendix A.
(see Tables A1 and A2).
References
[1] Chen J. Investigation of Vapor Ejectors in Heat Driven Ejector Refrigeration
Systems. Ph.D Thesis, Royal Institute of Technology-KTH, 2014.
[2] Zhu Y, Wang Z, Yang Y, Jiang P. Flow visualization of supersonic two-phase
transcritical flow of CO2 in an ejector of a refrigeration system. Int J Refrig
2017;74:354–61.
[3] Chen J, Havtun H, Palm B. Investigation of ejectors in refrigeration system:
optimum performance evaluation and ejector area ratios perspectives. Appl
Therm Eng 2014;64(1–2):182–91.
[4] Chen W, Shi C, Zhang S, Chen H, Chong D, Yan J. Theoretical analysis of ejector
refrigeration system performance under overall modes. Appl Energy
2017;185:2074–84.
[5] Sun DW. Comparative study of the performance of an ejector refrigeration
cycle operating with various refrigerants. Energy Convers Manage 1999;40
(8):873–84.
[6] Selvaraju A, Mani A. Analysis of an ejector with environment friendly
refrigerants. Appl Therm Eng 2004;24(5–6):827–38.
[7] Roman R, Hernandez JI. Performance of ejector cooling systems using low
ecological impact refrigerants. Int J Refrig 2011;34(7):1707–16.
[8] Chen J, Havtun H, Palm B. Screening of working fluids for the ejector
refrigeration system. Int J Refrig 2014;47:1–14.
[9] Shestopalov KO, Huang BJ, Petrenko VO, Volovyk OS. Investigation of an
experimental ejector refrigeration machine operating with refrigerant R245fa
at design and off-design working conditions. Part 1. Theoretical analysis. Int J
Refrig 2015;55:201–11.
[10] Besagni G, Mereu R, Leo GD, Inzoli F. A study of working fluids for heat driven
ejector refrigeration using lumped parameter models. Int J Refrig
2015;58:154–71.
[11] Milazzo A, Rocchetti A. Modelling of ejector chillers with steam and other
working fluids. Int J Refrig 2015;57:277–87.
[12] Saleh B. Performance analysis and working fluid selection for ejector
refrigeration cycle. Appl Therm Eng 2016;107:114–24.
[13] Kasperski J, Gil B. Performance estimation of ejector cycles using heavier
hydrocarbon refrigerants. Appl Therm Eng 2014;71(1). 197-03.
[14] European Parliament, Directive 2006/40/EC of the European Parliament and of
the council. Official Journal of the European Union, 2006.
[15] European Parliament and the Council. No 517/2014 of the European
Parliament and of the Council of 16 April 2014 on fluorinated greenhouse
gases and repealing Regulation (EC) No 842/2006. Off J Eur Union 2014.
[16] Pridasawas W, Lundqvist P. An exergy analysis of a solar-driven ejector
refrigeration system. Sol Energy 2004;76(4):369–79.
[17] Dahmani A, Aidoun Z, Galanis N. Optimum design of ejector refrigeration
systems with environmentally benign fluids. Int J Therm Sci 2011;50
(8):1562–72.
[18] Chen J, Havtun H, Palm B. Conventional and advanced exergy analysis of an
ejector refrigeration system. Appl Energy 2015;144:139–51.
[19] Alexis GK. Exergy analysis of ejector-refrigeration cycle using water as
working fluid. Int J Energ Res 2005;29(2):95–105.
[20] Sadeghi M, Mahmoudi SMS, Saray RK. Exergoeconomic analysis and multi-
objective optimization of an ejector refrigeration cycle powered by an internal
combustion (HCCI) engine. Energy Convers Manage 2015;96:403–17.
[21] Tsatsaronis G. Strengths and limitations of exergy analysis. In: Bejan A, Mamut
E, editors. Thermodynamic optimization of complex energy
systems. Dordrecht: Kluwer Academic Publishers; 1999.
[22] Morosuk T, Tsatsaronis G. Advanced exergetic evaluation of refrigeration
machines using different working fluids. Energ 2009;34(12):2248–58.
[23] Mosaffa AH, Farshi LG, Ferreira CAI, Rosen MA. Advanced exergy analysis of an
air conditioning system incorporating thermal energy storage. Energ
2014;77:945–52.
[24] Gong S, Boulama KG. Parametric study of an absorption refrigeration machine
using advanced exergy analysis. Energ 2014;76:453–67.
[25] Yamankaradeniz N. Thermodynamic performance assessments of a district
heating system with geothermal by using advanced exergy analysis. Renew
Energy 2016;85:965–72.
[26] Bai T, Yu J, Yan G. Advanced exergy analyses of an ejector expansion
transcritical CO2 refrigeration system. Energy Convers Manage
2016;126:850–61.
[27] Bai T, Yu J, Yan G. Advanced exergy analysis on a modified auto-cascade freezer
cycle with an ejector. Energ 2016;113:385–98.
[28] Hulse RJ, Basu RS, Singh RR, Thomas RHP. Physical properties of HCFO-1233zd
(E). J Chem Eng Data 2012;57:3581–6.
[29] Calm JM, Hourahan, GC. Physical, safety, and environmental data summary for
current and alternative refrigerants, ID: 915. In: Proceedings of the 23rd
International Congress of Refrigeration, Prague, Czech Republic, 2011.
[30] Lemmon EW, McLinden MO, Huber ML. NIST Standard Reference Database 23:
Reference Fluid Thermodynamic and Transport Properties-REFPROP, Version
9.1. National Institute of Standards and Technology, Standard Reference Data
Program, Gaithersburg, USA, 2013.
[31] Shestopalov KO, Huang BJ, Petrenko VO, Volovyk OS. Investigation of an
experimental ejector refrigeration machine operating with refrigerant R245fa
at design and off-design working conditions. Part 2. Theoretical and
experimental results. Int J Refrig 2015;55:212–23.
[32] Keçebasß A, Gökgedik H. Thermodynamic evaluation of a geothermal power
plant for advanced exergy analysis. Energ 2015;88:746–55.