Energy Conversion and Management 87 (2014) 199207

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Performance analysis of an absorption power cycle for ocean thermal

energy conversion
Han Yuan a,b, Ning Mei a,, Peilin Zhou b
a
College of Engineering, Ocean University of China, 238 Songling Road, Laoshan District, Qingdao 266100, China
b
Department of Naval Architecture and Marine Engineering, University of Strathclyde, Glasgow G4 0LZ, United Kingdom

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

Article history: An absorption power cycle with two ejectors is proposed for ocean thermal energy conversion. The
Received 4 May 2014 ammoniawater is used as the working uid. The ejectors are driven by vapor and solution from the
Accepted 2 July 2014 sub-generator. Based on the rst and second law, the mathematical model for this cycle is developed
and theoretical analysis is conducted to evaluate the effects of thermodynamic parameters on the perfor-
mance of this cycle. Results show that the absorption temperature is increased by 2.06.5 C by employ-
Keywords: ing the two-stage ejector sub-cycle, which indicates that this proposed cycle can be driven with a lower
Ocean thermal energy conversion (OTEC)
temperature difference. Further, the thermal efciency, net thermal efciency and exergy efciency of
Ammoniawater absorption
Power cycle
this cycle can reach to 4.17%, 3.10% and 39.92% respectively. Besides, the generation pressure, the heating
Ejector source temperature, the solution concentration, and the expansion ratio, as well as the entrainment ratio
Energy efciency of the rst stage ejector have signicant effects on the absorption temperature, the thermal efciency, the
Exergy efciency exergy efciency and the exergy loss of this cycle. In addition, 49.80% of exergy loss in this proposed cycle
occurs in the generators and reheater, followed by the ejectors of 36.12%.
2014 Elsevier Ltd. All rights reserved.

1. Introduction Compared with the open-cycle, the closed-cycle has a higher

Ocean thermal energy is a kind of solar energy absorbed and
stored in the upper layer of the ocean. It is exploited by means of
ocean thermal energy conversion (OTEC) cycle. Such cycle utilize
the top layer seawater (3032 C) as the heating source and deep
seawater at a depth of 1000 m (46 C) as the cooling source to
drive a heat engine cycle and generate power [1]. Due to the rela-
tively small temperature difference between heating and cooling
source, the Carnot cycle efciency for an OTEC cycle is limited at
8% [2]. However, the vast ocean thermal energy reserves still make
it a promising research topic.
Generally, an OTEC cycle can take the form of either open-cycle
or closed-cycle process. The open-cycle employs warm sea water
as working uid. Steam is evaporated in a ash evaporator and
then drives a turbine to produce power. Then the steam is con-
densed and fresh water is obtained as a by-product. The closed-
cycle operates similar to the open-cycle except that a refrigerating
uid is used as the working uid. The refrigerant is evaporated by
warm sea water in the generator and expands in the turbine to
generate power [3].
Corresponding author. Tel./fax: +86 532 66781105.
E-mail address: nmei@ouc.edu.cn (N. Mei).
thermal efciency with the same temperature difference of sea
water. This is due to the use of vacuum equipment in the open-
cycle. On one hand, the entire open-cycle, from ash evaporator
to condenser, operates under a partial vacuum condition. On the
other hand, the non-condensable gas in sea water needs to be
removed. Thus a vacuum compressor is used in the open-cycle
and this increases extra power consumption [4].
But the closed-cycles application in the OTEC cycle is far from
satisfactory. For this reason, current research is mainly focused
on the approaches of improving thermal efciency.
One way to improve the performance of the closed-cycle con-
sists in choosing a proper working uid. Both the pure working
uid (i.e. R134a, R245a, and R601, etc.) and the binary mixture
(i.e. ammoniawater mixture) can be used as the working uid
[5]. The binary mixture was found to be more appropriate and ef-
cient for using in OTEC cycle, enable to reduce the heat transfer
related irreversibility in evaporator and condenser [68]. This is
because at a given pressure, the evaporation or condensation of a
binary mixture occurs over range of temperatures, results in a
lower temperature difference between the heating/cooling source
and the working uid. DiPippo [9] conducted a comparison of
different working uids in a closed-cycle. Results showed that
the ammoniawater based cycle performs better with a low
temperature of heating source (120 C) and the theoretical thermal

http://dx.doi.org/10.1016/j.enconman.2014.07.015
0196-8904/ 2014 Elsevier Ltd. All rights reserved.
200 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207

Nomenclature

Symbols EJ ejector
m mass ow rate, kg s1 net net
h specic enthalpy, kJ kg1 pf primary ow
E exergy, kJ sf secondary ow
I exergy loss, kJ d diffuser
T temperature, C n nozzle
P pressure, MPa m mixing chamber
X solution concentration, kg kg1 s solution mixture
Q heat input, kJ s0 isentropic process
W turbine output, kJ n1 nozzle inlet
u velocity, m s1 n2 nozzle outlet
in input
Subscripts out output
I, II apparatus number ratio ratio of two generators
i = 1, 2. . . state points
0 reference state Greek symbols
G generator l entrainment ratio
T turbine b expansion ratio
R reheater g thermal efciency, %
A absorber m specic volume, m3 kg1
SP solution pump
WP warm seawater pump Acronyms
CP cold seawater pump OTEC ocean thermal energy conversion
WS warm seawater ex exergy

efciency is 3% higher than that of a pure working uid based
cycle.
Another way to make the improvement is to investigate new
thermodynamic cycles. Great efforts were made to develop a novel
binary mixture based cycle and a representative one is the Kalina
cycle developed by Kalina [10]. The Kalina cycle, which employs
the ammoniawater as the working uid, was rst designed to
serve as a bottoming cycle for a combined-cycle energy system
as well as for generating electricity using low-temperature heat
sources. It proved to be an efcient power cycle and to have a
higher thermal efciency than a conventional Rankine cycle by
about 1020% [11]. Furthermore, researchers have proposed a Kali-
na cycle family and they carried out a number of theoretical stud-
ies to demonstrate these cycles performance [1223]. Uehara
cycle, proposed by Uehara and Ikegami [24], is a modied Kali-
na cycle. In this cycle, a part of the vapor is extracted in the turbine
and fed through the heater and absorbed by the liquid in the absor-
ber. This extraction process was added to improve thermal ef-
ciency. Furthermore, Uehara et al. [25] presented numerical
simulation for the Uehara cycle, and results showed that the ther-
mal efciency rises when the extraction mass fraction ratio
increases. The concept of using an ejector in an ammoniawater
based closed-cycle has been proposed recently. Yuan et al. [26]
and Xinguo Li et al. [27] put forward novel types of one-stage ejec-
tor power cycle for OTEC separately. These one-stage ejector power
cycles have showed a higher thermal efciency than traditional
ones. The ejector can lead to depressurization in the turbine outlet
which has signicantly improved the thermal efciency by increas-
ing the turbine power output.
However, little attention has been paid to approaches of
increasing the cooling source seawater temperature for the OTEC
cycle. It is interesting to note that the cooling source seawater tem-
perature is closely related to its depth. When the seawater depth
decreases from 1000 m to 200 m, its temperature draws a rapid
increase [28]. This phenomenon provides a potential of broadening
the scope of the OTEC cycle application: once a shallower depth of
cooling source seawater can be used, both the energy consumption
in the cold water pump and the cost of the expensive cold-water
pipe that is used to pump seawater can be reduced.
In this study, an absorption power cycle with ejectors is pro-
posed. The main objective of this study is to increase the absorp-
tion temperature for the OTEC cycle, which ultimately reduces
the energy consumption in cold water pump by utilizing shallower
depth of seawater as the cooling source. The absorption power
cycle is thermodynamically modelled and assessed with energy
and exergy analyses with a theoretical comparison with the perfor-
mance of the one-stage ejector power cycle proposed in Ref. [26],
including the evaluation of the effects of two-stage ejector sub-
cycle on the absorption temperature and operation parameters.
2. Cycle description
The proposed absorption power cycle consists of two genera-
tors, two separators, two-stage turbines, two ejectors, two heat
exchangers, a reheater, an absorber and a solution pump. This
power cycle employs ammonia water as the working uid and sea-
water as the heating/cooling source.
A two-stage ejector sub-cycle, which is shown in Fig. 1, consists
of a sub-generator, a heat exchanger and two ejectors, is employed
to get a higher absorption temperature in the absorber. A reheater
is introduced between the turbines to improve the thermal ef-
ciency. The generator I performs as the primary generator, provid-
ing ammonia vapor (state point 3) which drives the turbines to
produce power. The generator II performs as the sub-generator,
in which ammonia vapor (state point 7) is generated as the pri-
mary uid of the ejector I. The turbine exhaust vapor (state point
6) is rst ejected as the secondary uid of the ejector I and then
mixes with the ammonia vapor generated from the generator II
before owing into the ejector II as the secondary uid (state point
8). Simultaneously, the weak solutions discharged from the gener-
ators (state point 2, 17) mix together (state point 9) and enter the
ejector II as the primary uid. The weak solution (state point 9) and
the ammonia vapor (state point 8) are preliminarily mixed in the
ejector II and then ow into the absorber to produce the strong
 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207 201

Fig. 1. Schematic diagram of the absorption power cycle.

solution (state point 11). A solution pump is utilized to pump the

strong solution back into the generators. Besides, two heat
exchangers are used to carry out heat transfer between the strong
and the weak solutions.
As is shown in Figs. 2 and 3, the proposed cycle can be divided
into two sub cycles: the primary generator driven cycle and the
sub-generator driven cycle. The former one is identied as the pro-
cesses of 1345681011121 (for vapor) and 12910
11121 (for weak solution), and the latter one is identied as
the processes of 167810111216 (for vapor) and 16179
10111216 (for weak solution). The mass ratio between these
two sub cycles is reected by l I (the entrainment ratio of ejector
I).
The advantage of this proposed power cycle is that it reduces
energy consumption in the cold water pump, which is used to
draw cold water from the deep ocean. The key lies in the increasing
of the absorption pressure in the absorber by introducing the two-
stage ejector sub-cycle. Given a xed ammonia mass concentra- Fig. 3. Pressure-enthalpy diagram of the absorption power cycle.
tion, a higher pressure of the ammoniawater mixture leads to a
higher corresponding saturation temperature. Thus the absorption cooling source for the power cycle. Therefore the energy consump-
temperature is increased. This means shallower depth of cold sea- tion in the cold water pump will be ultimately reduced due to the
water with a relatively higher temperature can be utilized as the shallower depth of cooling source seawater. Also, the cost of the
expensive cold-water pipe that is used to pump seawater can be
reduced.

3. Thermodynamic analysis of the absorption power cycle

3.1. Mathematical model

Based on the rst and second law of thermodynamics, the

mathematical model for this power cycle was established. Both
energy and exergy analyses were conducted to evaluate the theo-
retical performance of this power cycle.
The mass balance equations for both ammonia vapors and
ammoniawater solutions in the cycle are given as follows:
X
Din
out mi 0 1

X
Din
out X i  mi 0 2

Fig. 2. Temperature-ammonia mass concentration diagram of the absorption Each component of this cycle is treated as a control volume and the
power cycle. energy balance equations are given as follows:
202 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207

For the generator I: The performance of the power cycle is evaluated by the thermal
efciency, which is dened as the ratio of the turbine outputs to
Q Gk m3 h3 m2 h2  m1 h1 3 the sum of heat and pump input:

For the generator II: g WT j WT k  W SP =Q Gj Q Gk Q R 22

Q Gk m7 h7 m17 h17  m16 h16 4 Both warm and cold seawater pump power are calculated as
follows:
For the turbines:
W WP mWP DHWP g=gp 23
W TI gT mh3  h4 5
W CP mCP DHCP g=gp 24
W TII gT mh5  h6 6
where mWP, mCP, DHWP, DHWP, gP are mass ow rate, total pump
head difference and pump efciency of the warm and cold seawater
b P3 =P4 P5 =P6 7 piping. The pump head difference of each piping is calculated based
on the equations derived by Ahmadi et al. [32]. In addition, the sea-
For the reheater:
water temperature data at different depth is based on the ocean
Q R m3 h5  h4 8 temperature proles in the coastal area of Taitung county, east cost
of Taiwan [33].
For the heat exchanger I: The net thermal efciency is also calculated, which is given as
follow:
m2 h2  h19 m1 h13  h1 9
For the heat exchanger II: gnet WT j WT k  W SP  W WP  W CP =Q Gj Q Gk Q R 25

m17 h17  h18 m16 h16  h15 10 Exergy analysis was employed to evaluate quantitatively the cause
of thermodynamic imperfection of the power cycle. The exergy of
For the absorber:
each state point is identied as follow:
Q A m10 h10  h11 11 Ei m  hi  h0  T 0 si  s0  26
For the pump: The exergy loss of each component for the absorption power cycle
W P P12  P11 m11 m11 =gp 12 can be found as follow:
X X
Ein  Eout  W out I 27
For the ejector, the simulation is carried out based on the one-
dimensional ow model which was rst introduced by Keenan To be specic, the exergy losses are identied by the following
[29] and then developed by Huang et al. [30] and Dai et al. [31]. equations:
The enthalpy balance equations for the ejectors [27] are given For the generator I:
as follows:
IGj E1 EGj  E2  E3 28
1 lj h8 lj h6 h7 13
EGj 1  T 0 =T Gj Q Gj 29
1 lk h10 lk h8 h9 14
For the generator II:
lj m6 =m7 15 IGk E16 EGk  E7  E17 30

lj m8 =m18 m19 16 EGk 1  T 0 =T Gk Q Gk 31

The ejector consists of three sections: the nozzle section, the mixing For the turbines:
section and the diffuser section Dai et al. [31]. ITj E3  E4  W Tj 32
For the nozzle section:
mpf u2pf;n2 mpf u2pf;n1 IT2 E5  E6  W Tk 33
mpf hpf;n2 mpf hpf;n1 17
2 2 For the reheater:
For the mixing section: IR E4 ER  E5 34
mpf hpf;n2 mpf u2pf;n2 mpf msf umf;m;s 18
ER 1  T 0 =T R Q R 35
For the diffuser section:
For the heat exchanger I:
1 2 
umf;m  u2mf;d;s0 hmf;d;s0  hmf;m 19 IHj E13 E2  E1  E19 36
2
The entrainment ratio can also be derived as: For the heat exchanger II:
q IHk E17 E15  E16  E18 37
l gn gm gd hpf;n1  hpf;n2;s0 =hmf;d;s0  hmf;m  1 20
For the ejector I:
where gn is the nozzle efciency, gm is the mixing efciency, gd is
IEJI E6 E7  E8 38
the diffuser efciency.
The heating input ratio of the two generators is given as follow: For the ejector II:

Q ratio Q GI =Q GII 21 IEJII E8 E9  E10 39

 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207 203

For the absorber:
IA E10  EA  E11
EA T 0 =T A  1Q A
For the pump:
IP E11  E12 W P
The exergy efciency is dened as the ratio of the turbine outputs to
the sum of the exergy inputs:
 X .
gex Ein  I Ein
3.2. Basic assumptions for calculations
To simplify the simulation, the main assumptions for the power
cycle are made as follows:
(1) Each components of the power cycle is in steady state.
(2) The pressure drop and heat losses in this cycle are neglected.
(3) The solution at the absorber outlet is saturated.
(4) The solutions at the generators outlet are saturated.
(5) The temperature of vapor from the separator remains
constant.
(6) The pump and the turbines have a given isentropic ef-
ciency, respectively.
3.3. Solution procedure
A program was developed to analyze the cycle performance and
Table 1 shows the calculation strategy for the absorption power
cycle. In this program, a database was compiled based on the Sulze
Equation [34] to conduct the calculation for the thermodynamic
variables of ammoniawater and ammonia vapor.
4. Results and discussion
4.1. Typical working condition
In order to obtain the details of the thermodynamic perfor-
mance of the proposed cycle, a simulation under a typical working
condition was carried out. Table 2 shows the simulation results of
each state point in the absorption power cycle at a typical working
40
condition. Parameters of the thermodynamic properties, the mass
ow rate and the solution/vapor concentration were obtained. It
41
is noted that the pressure at state point 10 (0.35 MPa) is higher
than that of state point 6 (0.3 MPa), which indicates that by
employing the two-stage ejector sub-cycle the absorption pressure
42 is increased. In addition, the performance of this power cycle with
this typical working condition is shown in Table 3.
The exergy analysis aids in pointing out the irreversibility of the
absorption power cycle and the result is shown in Table 4. It is
43 found that two major exergy losses exist in this cycle. The highest
exergy loss occurs in the generators and the absorber, due to the
temperature difference between the heating/cooling source and
the working uid. Another major exergy loss exists in the ejectors,
and this is because the processes of mixing, shock formation and
friction in the ejectors can lead to the exergy destruction, which
is also found by Khaliq [35].
In order to assess the effect of two-stage ejector sub-cycle on
the absorption temperature and operation parameters, a paramet-
ric analysis was performed and parameters of the generation pres-
sure, the heating source temperature, the solution concentration
and the expansion ratio were selected. The simulation conditions
for the proposed cycle are listed in Table 5.
4.2. Generation pressure
The effect of generation pressure on the absorption temperature
of the proposed cycle is shown in Fig. 4. It is found that given a
xed heating input ratio of the two generators Qratio, the absorption
temperature T11 increases as the generation pressure P11 increases.
The following reason explains this behavior: since b and Qratio are
xed, the increase of P3 can lead to a higher P6 (see Eq. (7)), which
ultimately increases P11; simultaneously X1 is constant, the proper-
ties of the ammoniawater solution determine that T11 increases
when P11 increases. Furthermore, it is found that a higher Qratio
results in a higher absorption temperature. As is shown in this g-
ure, the absorption temperature increases from 7.5 C to 13.1 C
when Qratio = 0.5, which is higher that of from 3.0 C to 8.2 C when
Qratio = 0.1. This is because the increase of Qratio results in an
increase of P11 since P6 is constant (see Eq. (7) and noting that b
and P3 are constant), therefore T11 increases as well.

Table 1
Solution procedure for the absorption power cycle.

1 Compile a database, which is based on the Sulze Equations [34], for ammoniawater and ammonia vapor calculation
2 Fix the values of the inputs: gt, gP, gn, gm, gd, DTG, DTR, DTA, TWS, b, lI, T1, T16, P1, P16, m1, Qratio = QGI/QGII
3 Basic assumptions
The temperature difference exists in the heat-exchange components is taken as 5 C
gn = 0.9, gm = 0.85, gd = 0.85
Ammoniawater solutions at state point 1, 2, 11, 16, 17 are saturated
4 Calculate X1 from P1 and T1; calculate X2 from P2 = P3 = P1 and T2 = T3 = T20  DTG; calculate h1, h2, h3 and s1, s2, s3 from T, P, X in state point 1, 2, 3 and Sulze
Equations; obtain X11 for X11 = X1
5 Calculate m2, m3 from m1 and Eq. (2); calculate QGI from Eq. (3); with the same method, h7, h16, h17 and s7, s16, s17 can be obtained; calculate m7, m17 from lI = m6/
m7 = m1/m16 and Eq. (2); calculate QGII from Eq. (4)
6 Calculate P4 and P6 from the denition of b; evaluate T4 from T3; calculate h4 from T, P, X in state point 4; obtain m4 = m5 = m6 = m3; calculate WTI from Eq. (5)
7 Calculate h5 from T5 = Tws  DTR and P5; calculate QR from Eq. (8)
8 Evaluate T6 from T5, calculate h6 from T6 and P6; calculate WT2 from Eq. (6)
9 Calculate h8 from m8 = m6 + m7 and Eq. (13); calculate P10 = P11 from T11 and X11; calculate P8, P11 = P10 from P7, P6 and lI
10 Obtain P15 = P16, P14 = P1, m14 = m16; obtain P12 = P1, m19 = m2, m18 = m17; calculate m10 from Eq. (1); calculate WP from m10 and Eq. (12)
11 Assume T11 = 273.15 K, calculate X11 from T11 and P11, do T11 = T11 + 0.01, till (X11  X1) < 0.00001, obtain T11; calculate h11 from T11, P11, X11
12 Obtain P13 = P12, calculate h12 from WP and h11; calculate h19 from h13 = h12 and Eq. (9)
13 Obtain h14 = h12, calculate h15 from h14; calculate h18 from Eq. (10); obtain lII from Eq. (16); calculate m10 from Eq. (1), calculate h10 from m10 and Eq. (14)
14 Calculate QA from Eq. (11)
15 Calculate si from T, P, X in each state point and Sulze Equations
16 Calculate Ii from si and Eqs. (26)(43)
17 Calculate g, gnet and gex from Eqs. (18), (25), (43)
204 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207

Table 2
Results of simulation at typical working condition (Th = 30 C, PG = 0.6 MPa, b = 1.42,
Qratio = 0.1).

State T (C) P (MPa) h (kJ kg1) s (kJ kg1 C) m (kg1 s) X

1 22.05 0.60 125.27 1.33 6.36 0.70
2 27.00 0.60 125.70 1.53 5.36 0.64
3 27.00 0.60 1509.96 9.54 1.00 0.99
4 3.67 0.42 1466.73 9.56 1.00 0.99
5 27.00 0.42 1523.08 9.95 1.00 0.99
6 3.67 0.30 1478.03 9.97 1.00 0.99
7 27.00 0.60 1509.95 9.54 0.10 0.99
8 6.34 0.34 1480.94 9.88 1.10 0.99
9 8.85 0.60 228.90 0.93 5.89 0.64
10 8.12 0.35 40.04 1.13 6.99 0.70
11 6.82 0.35 212.59 0.82 6.99 0.70
12 6.82 0.60 212.34 0.82 6.99 0.70
13 6.82 0.60 212.34 0.82 6.36 0.70
14 6.82 0.60 212.34 0.82 0.64 0.70
15 6.82 0.60 212.34 0.82 0.64 0.70
16 22.05 0.60 125.27 1.33 0.64 0.70
17 27.00 0.60 125.70 1.53 0.54 0.64
18 8.85 0.60 228.90 0.93 0.54 0.64 Fig. 4. Comparison between present investigation and Yuan [26] with generation
19 8.85 0.60 228.90 0.93 5.36 0.64 pressure varies.

Table 3
Performance of the absorption power cycle.

Turbine work (kJ s1) 75.02

Generator I heat input (kJ s1) 1632.91
Generator II heat input (kJ s1) 163.29
Reheater heat input (kJ s1) 56.35
Absorber heat output (kJ s1) 1766.78
Thermal efciency (%) 4.17
Net thermal efciency (%) 3.10
Exergy efciency (%) 35.66

Table 4
Exergy losses in each apparatus.

Amount (kJ s1) Percentage (%)

Exergy loss Generator I 385.70 43.22
Generator II 38.92 4.36
Reheater 19.81 2.22
Ejector I 85.58 9.59 Fig. 5. Effect of generation pressure on the thermal efciency.
Ejector II 236.74 26.53
Absorber 107.50 12.05
Pump 1.50 0.17
Turbine 13.24 1.48
Heat exchanger I 3.18 0.36
Heat exchanger II 0.32 0.04

Table 5
Conditions of simulation for the absorption power cycle.

Environment pressure (MPa) 0.1013

Environment temperature (C) 25
Heating source temperature (C) 2833
Generation pressure (MPa) 0.50.625
Strong solution concentration (kg kg1) 0.70.8
Strong solution mass ow rate (kg s1) 1
Pump isentropic efciency (%) 85
Turbine isentropic efciency (%) 85
Expansion ratio 1.41.45
Minimum temperature difference (C)
In the generator 3
In the absorber 3
Fig. 6. Effect of generation pressure on the exergy efciency.

In Fig. 4, a comparison between the proposed cycle and the one- cycle has a relatively higher absorption temperature when the
stage ejector cycle in Ref. [26] is made and it is found that the generation pressure varies. The reason is that by employing
absorption temperature of the one-stage ejector cycle ranges from the two-stage ejector sub-cycle the generation pressure P11 in
1.0 C to 6.2 C. This result indicates that the two-stage ejector the two-stage ejector cycle is higher, which leads to a higher T11.
 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207 205

Fig. 5 shows that the thermal efciency of the proposed cycle

slightly increases as the generation pressure increases. Qualita-
tively this is due to two main reasons. On one hand, the increases
of the generation pressure results in a decreases of the ammonia
vapor mass ow rate, which reduces the thermal efciency. This
is due to X2 increases as P3 increases (since the weak solution at
the generator I outlet is saturated), thus m3 decreases, according
to Eqs. (1) and (2) and noting that m1 is constant. On the other
hand, the increases of the generation pressure leads to an increases
of the turbines output. These two reasons results in the smooth
increase of the thermal efciency. Furthermore, it is found that
the thermal efciency is around 4% when Qratio = 0.1, compared
with approximately 2.8% when Qratio = 0.5, which indicates that
the higher value of Qratio lead to a lower thermal efciency. This
behavior can be explained by noting that m7 is higher with a higher
value of Qratio, which indicates the heat input of the generator II is
higher. According to Eq. (17), a lower thermal efciency is
obtained.
The trend of the exergy efciency is similar with the thermal Fig. 8. Effect of heating source temperature on the exergy efciency.
efciency curves. Fig. 6 shows that the exergy efciency increases
as the generation pressure increases, but it decreases as the value
of Qratio increase, which indicates that the increase of m7 results in
an increase of exergy destruction in the ejectors.

4.3. Heating source temperature

Fig. 7 shows that the thermal efciency increases when the

heating source temperature increases. The following are the rea-
sons for this behavior: on one hand, the increase of heating source
temperature leads to an increase of the ammonia-vapor tempera-
ture T3 and T5, thus h3 and h5 increases; on the other hand, the
weak solution concentration X2 decreases as the heating source
temperature increases, this leads to an increase of the ammonia
vapor mass ow rate m3 (see Eq. (1) and (2)). These two reasons
results in a higher turbine output, therefore the thermal efciency
is enhanced. It is also found that the higher value of Qratio result in a
lower thermal efciency, and the reason for this has been
explained in Fig. 5.
The result in Fig. 8 reveals that the exergy efciency increases as
Fig. 9. Comparison between present investigation and Yuan [26] with strong
the heating source temperature increases, but it decreases as the solution concentration varies.
value of Qratio increase.

4.4. Solution concentration
Fig. 9 shows the effect of strong solution concentration on the
thermal efciency and the absorption temperature. The result
Fig. 7. Effect of heating source temperature on the thermal efciency.
shows that with the increasing of the strong solution concentra-
tion, the thermal efciency increases slightly while the absorption
temperature decreases. The following reasons explain this behav-
ior: when X1 increases m3 decreases (see Eqs. (1) and (2) and not-
ing that m1 and X2 is constant), thus the turbine output increases;
simultaneously X11 increases as X1 increases since X11 = X1, there-
fore T11 decreases (noting that the solution at the absorber outlet
is saturated). The performance of the proposed cycle is compared
with the one-stage ejector power cycle in Ref. [26]. The results
show that this proposed cycle has a relatively higher absorption
temperature but a slightly lower thermal efciency as the strong
solution concentration varies. This is due to the extra heat input
in the generator II which reduces the thermal efciency of the pro-
posed cycle (see Eq. (18)).
The effect of strong and weak solution concentration on the
exergy efciency of the proposed cycle is analyzed in this paper.
In the simulation, the strong solution concentration ranges from
0.7 to 0.8 and the weak solution concentration ranges from 0.6 to
0.7. The result shown in Fig. 10 reveals that the exergy efciency
increases as both the strong and weak solution concentration
increases, but the increase rate for each parameter is different.
The exergy efciency draws a rapid increase with X1 increases at
xed X2 and a smooth increase with X2 increases at xed X1, which
206 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207
Fig. 10. Effect of strong and weak solution concentration on the exergy efciency.
Fig. 11. Effect of strong and weak solution concentration on the total exergy loss.
reects that the value of the strong solution concentration domi-
nates the exergy efciency.
Fig. 11 shows that the total exergy loss decreases as the strong
solution concentration decreases and the weak solution concentra-
tion increases. With a lower concentration difference between the
strong and weak solution, the total exergy loss of the proposed
cycle is lower. Moreover, it is found that the exergy loss experi-
ments a rapid drop when the concentration difference is close to
zero, this is due to the power cycle can hardly work under this
condition.
4.5. Expansion ratio
Fig. 12 shows that the increase of the expansion ratio leads to an
increase of the thermal efciency. Qualitatively when b increases,
the pressure drop across the turbines increases, therefore the tur-
bine performance is enhanced and the thermal efciency is
improved according to Eq. (17). Also it is noted that the absorption
temperature drops as the expansion ratio rises. This behavior can
be explained as follows: when P1 is xed, P6 decreases as b
increases, which ultimately decreases P11, thus T11 decreases
(according to the properties of the ammoniawater solution and
by noting that X11 = X1 is constant).
4.6. Net thermal efciency
The net thermal efciency, which considers the effect of both
warm and cold seawater pump power, is crucial for performance
Fig. 12. Effect of expansion ratio on thermal efciency and absorption temperature.
Fig. 13. Comparison of net thermal efciency between present investigation and
Yuan [26].
evaluation of the proposed cycle. The result is shown in Fig. 13.
Firstly, compared with one-stage ejector cycle, the proposed cycle
has a relatively higher net thermal efciency when Qratio is below
0.3, which indicates the two-stage ejector structure can improve
the cycle performance. The net thermal efciency can reach to
3.1%. However, as Qratio increases, the net thermal efciency of
the proposed cycle decreases. This is because a higher heat input
is required in the generator II. When Qratio is above 0.4, the net
thermal efciency of the proposed cycle is lower than that of
one-stage ejector cycle. Moreover, the net thermal efciency of
both cycles increase as the generation pressure increases. A similar
result is also observed in Fig. 5. Besides, it should be noted that the
net thermal efciency of one-stage ejector cycle is dramatically
reduced as the generation pressure decreases, while the proposed
two-stage ejector cycle performs better under this condition. This
behavior is caused by the following reason: when the one-stage
cycle operates at a lower absorption pressure, a lower absorption
temperature is required. Considering the seawater temperature
at different depth in [33], the depth of cold seawater for cooling
is drastically increased. Thus more pump power is needed for cold
seawater piping and the net thermal efciency is reduced.
5. Conclusions
An absorption power cycle was discussed and a program was
developed to analyze the cycle performance. By carrying out a
 H. Yuan et al. / Energy Conversion and Management 87 (2014) 199207
simulation under a typical working, the thermodynamic perfor-
mance of the proposed cycle was obtained. In addition, the perfor-
mance of the absorption power cycle for a range of generation
pressure, heating source temperature, solution concentration, and
expansion ratio was studied to nd out their effects on the absorp-
tion temperature, thermal efciency, exergy efciency and exergy
loss. The main conclusions from this study can be summarized as
follows:
(1) The utilization of the two-stage ejector sub-cycle in the
proposed cycle can increase the absorption temperature by
2.06.5 C as the heating input ratio of the two generators
Qratio varies.
(2) The thermal efciency of this cycle can reach to 4.17%, and
the net thermal efciency is 3.10%.
(3) The thermal and exergy efciency increases as the genera-
tion pressure, the heating source temperature, the strong
solution concentration and the expansion ratio increases.
(4) The absorption temperature decreases as the generation
pressure, the strong solution concentration and the expan-
sion ratio increases.
(5) Higher value of heating input ratio of the two generators
Qratio leads to a higher absorption temperature but a lower
thermal efciency. When Qratio is above 0.4, its net thermal
efciency is lower than that of one-stage ejector cycle.
(6) Under the typical condition, 49.80% of exergy loss in this
proposed cycle occurs in the generators and reheater, fol-
lowed by the ejectors of 36.12%.
(7) The exergy efciency can reach to 39.92%. The strong solu-
tion concentration dominates the exergy efciency. With a
lower concentration difference between the strong and the
weak solution, the total exergy loss is lower.
Acknowledgments
The authors acknowledge the support provided by the National
Natural Science Foundation of China (NO. 51276174), the National
Key Technology Research and Development Program of the
Ministry of Science and Technology of China (NO. 2012BAC25B02)
and Qingdao Science & Technology Development Program (NO. 12-
1-3-77). H. Yuan is supported by the China Scholarship Council for
a two-year study at University of Strathclyde.
References
[1] Kim NJ, Ng KC, Chun W. Using the condenser efuent from a nuclear power
plant for Ocean Thermal Energy Conversion [OTEC]. Int Commun Heat Mass
Transfer 2009;36:100813.
[2] Yuan H, Mei N, Hu SY, Wang L, Yang S. Experimental investigation on an
ammonia-water based ocean thermal energy conversion system. Appl Therm
Eng 2013;61:32733.
[3] Takahashi PK, Trenka A. Ocean thermal-energy conversion - its promise as a
total resource system. Energy 1992;17:65768.
[4] Semmari H, Stitou D, Mauran S. A novel Carnot-based cycle for ocean thermal
energy conversion. Energy 2012;43:36175.
[5] Becquin G, Lehar M. Two algorithms for the reliable estimation of organic
rankine cycle performance. J Eng Gas Turbines Power-Trans ASME 2012;134.
[6] Wei DH, Lu XS, Lu Z, Gu JM. Performance analysis and optimization of organic
Rankine cycle (ORC) for waste heat recovery. Energy Convers Manage
2007;48:11139.
207
[7] Koroneos C, Rovas D. Exergy analysis of geothermal electricity using the Kalina
cycle. Int J Exergy 2013;12:5469.
[8] Meinel D, Wieland C, Spliethoff H. Effect and comparison of different working
uids on a two-stage organic rankine cycle (ORC) concept. Appl Therm Eng
2014;63:24653.
[9] DiPippo R. Second Law assessment of binary plants generating power from
low-temperature geothermal uids. Geothermics 2004;33:56586.
[10] Kalina AI. Combined-cycle system with novel bottoming cycle. J Eng Gas
Turbines Power-Trans ASME 1984;106:73742.
[11] Singh OK, Kaushik SC. Energy and exergy analysis and optimization of Kalina
cycle coupled with a coal red steam power plant. Appl Therm Eng
2013;51:787800.
[12] Singh OK, Kaushik SC. Thermoeconomic evaluation and optimization of a
BraytonRankineKalina combined triple power cycle. Energy Convers
Manage 2013;71:3242.
[13] Arslan O. Exergoeconomic evaluation of electricity generation by the medium
temperature geothermal resources, using a Kalina cycle: Simav case study. Int
J Therm Sci 2010;49:186673.
[14] Zhang XX, He MG, Zhang Y. A review of research on the Kalina cycle. Renew
Sustain Energy Rev 2012;16:530918.
[15] Fu WC, Zhu JL, Zhang W, Lu ZY. Performance evaluation of Kalina cycle
subsystem on geothermal power generation in the oileld. Appl Therm Eng
2013;54:497506.
[16] Coskun A, Bolatturk A, Kanoglu M. Thermodynamic and economic analysis and
optimization of power cycles for a medium temperature geothermal resource.
Energy Convers Manage 2014;78:3949.
[17] He JC, Liu C, Xu XX, Li YR, Wu SY, Xu JL. Performance research on modied KCS
(Kalina cycle system) 11 without throttle valve. Energy 2014;64:38997.
[18] Larsen U, Nguyen TV, Knudsen T, Haglind F. System analysis and optimisation
of a Kalina split-cycle for waste heat recovery on large marine diesel engines.
Energy 2014;64:48494.
[19] Lolos PA, Rogdakis ED. A Kalina power cycle driven by renewable energy
sources. Energy 2009;34:45764.
[20] Sun FM, Ikegami Y, Jia BJ. A study on Kalina solar system with an auxiliary
superheater. Renewable Energy 2012;41:2109.
[21] Modi A, Haglind F. Performance analysis of a Kalina cycle for a central receiver
solar thermal power plant with direct steam generation. Appl Therm Eng
2014;65:2018.
[22] Nguyen C, Veje CT, Willatzen M, Andersen P. Exergy costing for energy saving
in combined heating and cooling applications. Energy Convers Manage
2014;86:34955.
[23] Walraven D, Laenen B, Dhaeseleer W. Comparison of thermodynamic cycles
for power production from low-temperature geothermal heat sources. Energy
Convers Manage 2013;66:22033.
[24] Uehara H, Ikegami Y. Optimization of a closed-cycle OTEC system. J Solar
Energy Eng-Trans ASME 1990;112:24756.
[25] Noda N, Ikegami Y, Uehara H. Extraction condition of OTEC sing the Uehara
cycle. In: Proceedings of the twelfth (2002) international offshore and polar
engineering conference, vol. 1; 2002. p. 6314.
[26] Yuan H, Mei N, Li Y, Yang S, Hu SY, Han YF. Theoretical and experimental
investigation on a liquid-gas ejector power cycle using ammonia-water. Sci
China-Technol Sci 2013;56:228998.
[27] Li XG, Zhang QL, Li XJ. A Kalina cycle with ejector. Energy 2013;54:2129.
[28] Soto R, Vergara J. Thermal power plant efciency enhancement with Ocean
Thermal Energy Conversion. Appl Therm Eng 2014;62:10512.
[29] Keenan JHNEPLF. An investigation of ejector design by analysis and
experiment. Cambridge, Mass.: Massachusetts Institute of Technology,
Guided Missiles Program; 1948.
[30] Huang BJ, Chang JM, Wang CP, Petrenko VA. A 1-D analysis of ejector
performance. Int J Refrig 1999;22:35464.
[31] Dai Y, Wang J, Gao L. Exergy analysis, parametric analysis and optimization for
a novel combined power and ejector refrigeration cycle. Appl Therm Eng
2009;29:198390.
[32] Ahmadi P, Dincer I, Rosen MA. Energy and exergy analyses of hydrogen
production via solar-boosted ocean thermal energy conversion and PEM
electrolysis. Int J Hydrogen Energy 2013;38:1795805.
[33] Yeh RH, Su TZ, Yang MS. Maximum output of an OTEC power plant. Ocean Eng
2005;32:685700.
[34] Ziegler B, Trepp C. Equation of state for ammonia-water mixtures. Int J Refrig
1984;7:1016.
[35] Khaliq A, Agrawal BK, Kumar R. First and second law investigation of waste
heat based combined power and ejector-absorption refrigeration cycle. Int J
Refrig Revue Internationale Du Froid 2012;35:8897.