Desalination 454 (2019) 20–37

Contents lists available at ScienceDirect

Desalination
journal homepage: www.elsevier.com/locate/desal

Energy, environmental and economic assessment of a polygeneration system T

of local desalination and CCHP
Mohammad Esrafiliana, Rouhollah Ahmadib,
⁎

a
School of Mechanical Engineering, Iran University of Science & Technology, Islamic Republic of Iran
b
School of New Technologies, Iran University of Science & Technology, Islamic Republic of Iran

ARTICLE INFO ABSTRACT

Keywords: In this investigation, a polygeneration scheme integrating local desalination units with CCHP systems, providing
Desalination cold, heat and power for Kish twin towers as well as providing a portion of the fresh water demand of the island,
Sea water reverse osmosis (SWRO) is proposed and evaluated in terms of energy, environment, and economy aspects. The proposed system is
Multi-effect distillation (MED) compared to the conventional CCHP and desalination system adopting the separate production system as the
Combined cooling, heating and power (CCHP)
reference. Here, two types of desalination technologies, namely multi-effect distillation (MED) and reverse os-
Multi-optimization
Economic analysis
mosis (RO) processes, are used to provide desalinated water. A genetic algorithm is implemented to determine
the optimal operating parameters. It results in that in the polygeneration system the yearly average power
generation efficiency, the annual primary energy saving ratio (APESR) and the annual total cost saving ratio
(ATCSR) of the whole system increase by 3.45%, 9.73%, and 6.49%, respectively, compared to the conventional
system. Using the MED process can increase the ATCSR from 9.8% to 16.0%, depending on the fresh water
market price. Moreover, the carbon dioxide emission (CDE) decreases by 1460.5 tons each year, compared to the
reference. It is also indicated that 93.87% of the MED input heat can be recovered for domestic hot water (DHW)
provision.

1. Introduction combined cooling, heating and power (CCHP) systems, also named
trigeneration systems, providing electrical energy as well as heat and
Water desalination has been considered as an effective solution to cold mainly through recovering the waste heat produced by prime
address the water scarcity problem, especially in the Middle East, where movers such as internal combustion engines (ICE). These systems are
almost 58% of the world's desalination capacity is installed [1]. Con- proven to be environmentally friendly and energy-saving [7]. CCHP
ventional desalination processes can be divided into two major types: systems are conditionally economic, and their performance is strongly
thermal desalination and membrane desalination. The conventional dependent on the energy demand profiles [8]. They have dissatisfactory
thermal desalination technologies, such as the multi-effect distillation performance under partial-load working conditions due to the unstable
(MED) processes, are considered as highly energy-intensive plants, and thermal and electrical demand profiles [9]. In this regard, Li et al. [7]
they could be economical if they utilize the waste heat rejected by in- evaluated the performance of three CCHP systems used for hotels, of-
dustrial processes [2,3]. The reverse osmosis (RO) process as a mem- fices and residential buildings. The results showed that the hotels
brane-based technology is the most promising seawater desalination achieved the best energy performance because of their relatively more
technology owing to its comparatively lower energy consumption and stable electrical demand profile.
its high salt rejection [4]. Combination of this process and high-effi- Both membrane and thermal desalination processes can be coupled
ciency power generation units can lead to a significant reduction in the to CCHP systems in order to provide fresh water as another product.
primary energy consumption and environmental emissions [5,6]. Such systems are called polygeneration systems and considered to be
Nowadays, the growth in the world population and living standards more energy-efficient [10]. A limited number of researchers have
has led to a substantial increase in fossil fuels consumption, greenhouse evaluated the performance of such polygeneration systems. Uche et al.
gas emissions, and energy costs. These factors have invoked the ne- [10] proposed several configurations for polygeneration systems for use
cessity of adopting more efficient methods for energy conversion and in an agro-food factory. The systems provided energy and desalted
generation. One approach to overcome these issues is to develop water for the building demands by a combined cycle coupled with

⁎
Corresponding author at: Narmak, Tehran 1684613114, Islamic Republic of Iran.
E-mail address: [email protected] (R. Ahmadi).

https://doi.org/10.1016/j.desal.2018.12.004
Received 30 June 2018; Received in revised form 17 November 2018; Accepted 10 December 2018
0011-9164/ © 2018 Elsevier B.V. All rights reserved.
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Nomenclature Q1 sensible thermal energy used in the first effect of MED, kW

r membrane recovery factor
Abbreviations Sm membrane active area, m2
T temperature, °C
APESR annual primary energy saving ratio, % Tac,out exhaust gases temperature at the outlet of absorption
ATC annual total cost, $ chiller, °C
ATCSR annual total cost saving ratio, % V total volumetric flow rate, m3/day
BHX brine heat exchanger W power, kW
CDE carbon dioxide emission, tons CO2 X salt mass concentration, kg of salt/kg of water
CDESR carbon dioxide emission saving ratio, %
COP coefficient of performance Greek symbol
CSHX cooling sea water heat exchanger
DHX distillate heat exchanger α high-temperature portion of the exhaust gases recoverable
FC feed condenser heat, %
FE fuel energy, kW Δ difference
FF flow factor η thermal or electrical efficiency, %
GA genetic algorithm π osmotic pressure, Pa
GE gas engine ρ sea water density, kg/m3
HPP high-pressure pump
HRC heat recovery condenser Subscripts
ICE internal combustion engine
MED multi-effect distillation 0 ambient condition
PAF plant availability factor, % ac absorption chiller
PEC primary energy consumption, kWh ann annual value
PESR primary energy saving ratio, % b brine
PHX preheating heat exchanger c cooling
PL partial load, % ca capital
PLF plant load factor, % cond condenser
PT Pelton turbine d distillate
RO reverse osmosis e effects
SEC specific energy consumption, kWh/m3 elc electric chiller
SWIP sea water intake and pretreatment ex exhaust gases
SWRO sea water reverse osmosis f feedwater
TCF temperature correction factor gb gas boiler
grid power grid
Symbols h heating
hrw heat recovery water
Aref reference membrane pure water permeability coefficient, hx heat exchanger
kg/m2·s·Pa i i-th MED effect
B brine flow rate, kg/s imp imported from the grid
BE benefit, $ in inlet
Bm membrane salt permeability coefficient, kg/m2·s j system type
C total cost, $ jw jacket water
c unit cost, $ m membrane
Caac absorption chiller capacity, kW min minimum value
cerng carbon dioxide emission for natural gas, kg CO2/kWh n nominal value
cmem membrane unit cost, $ oil lubricating oil
Cp specific heat, kJ/kg·k op operating
CRF recovery factor out outlet
D distillate mass flow rate, kg/s p permeate
EL electricity, kW pe personnel
F feedwater mass flow rate, kg/s pr preheating water
h specific enthalpy, kJ/kg pump high-pressure pump or SWIP pump
hfg evaporation enthalpy, kJ/kg pv pressure vessel
i interest rate, % pw pure water
J mass flux, kg/s·m2 rec recovery
m mass flow rate, kg/s ref reference system
n life time, year s salt
N number sw sea water
P pressure, Pa tot total
PC purchased equipment cost, $ w wall
Q thermal energy, kW

21
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

heaters, absorption chillers as well as desalination units including MED In water-stressed islands, such as Kish Island located in the south of
and RO plants. Their results showed that such combination con- Iran, fresh water is obtained mainly from local RO desalination units,
siderably improved the energy performance of the whole system. Co- which are energy-intensive processes. Moreover, in order to reduce the
lella et al. [11] proposed a polygeneration scheme providing cooling, islands' pollution and energy consumption, CCHP systems are drawing
heating, power and desalted water simultaneously in a reliable way for great attention. However, these systems often suffer from working at
a tourist sector. They used a natural gas engine and a reverse osmosis partial-load in a dissatisfactory performance due to users' fluctuating
system as the prime mover and the desalination unit, respectively. The demand profiles. Integrating the local desalination units with CCHP
system achieved a payback period and an internal rate of return of systems could be an effective solution to reduce the required primary
16 years and 6.8%, respectively. Rubio et al. [12] showed that such energy consumption (PEC) for desalination and improve CCHP part-
integrations are always profitable from the energy standpoint, but they load efficiency by making the electrical demand profile more stable.
might not achieve the economic merit. Xu et al. [13] proposed a novel However, in the authors' knowledge, no researcher has focused on such
CCHP system integrated with an MED desalination unit. Their system integrations, which are quite different from those systems investigated
produced fresh water in the spring and the autumn when the amount of in the literature. The researchers previously focused on providing en-
cooling and heating demand was at the lowest level during the year. ergy and fresh water for only a limited number of buildings. Using such
Consequently, the MED system had to work in partial-load condition in small-scale desalination processes results in higher specific energy
the summer and winter. Their results showed that the annual energy consumption and is not cost-effective [22]. Furthermore, the integra-
saving ratio of the integrated system increased by 12.6%, compared to tion of these desalination units with CCHP systems has no substantial
conventional CCHP systems without desalination. Rubio-Maya et al. impact on the electrical demand profile due to their lower energy
[14] presented a two-step optimization procedure for a similar poly- consumption, and consequently, such combinations can achieve no
generation unit. They created a superstructure for selection of the ap- considerable improvement on the CCHP system performance.
propriate plant components and size from both economy and energy In this paper, an integrated CCHP system with desalination tech-
aspects. The superstructure regarded various sets of components in- nologies for use in Kish Island is proposed, and the effects of the
cluding internal combustion engines (ICE), gas turbines, micro-gas combination on the performance of both CCHP system and the desali-
turbines, fuel cells and Stirling engines as the prime movers as well as nation plant are investigated. The proposed system provides energy for
RO and MED processes as the desalination units. The optimization re- Kish twin residential towers and covers regional freshwater demand.
sulted in an ICE as the optimal prime mover and an MED system as the Annual total cost (ATC), primary energy consumption and carbon di-
optimal desalination unit. The desalination unit capacity was restricted oxide emission (CDE) are calculated and compared with those of con-
only to the resort water demand. The MED unit utilized the heat which ventional non-integrated CCHP and desalination system, using the se-
was surplus to the building thermal energy demand. parate production system as the reference. The integrated system
Several works in literature have investigated renewable energy- includes a natural gas engine as the prime mover, hybrid cooling de-
based polygeneration systems. Rubio-Maya et al. [15] presented a vices, and an RO desalination unit. In addition, for further utilization of
systematic procedure to select the proper equipment for a poly- the waste heat, a preheated sensible heat-driven multi-effect distillation
generation system fueled by natural gas, solar energy, and gasified (MED) unit is adopted in this system, using the recoverable heat from
biomass. According to their assumptions, using natural gas led to higher the jacket water and lubricating oil as well as the low-temperature part
economic profitability. However, higher primary energy and carbon of the exhaust gases. Moreover, the users' heating demand is provided
dioxide saving were achieved by utilizing renewables. Various config- through recovering the heat rejected by the MED system. A genetic
urations of integrated solar and geothermal energy to produce fresh algorithm (GA) is implemented to find the optimal characteristics of the
water, cooling, and thermal energy, as well as hot water, were proposed systems. In summary, this work addresses the following gaps, which can
and investigated in [16–20]. They produced desalted water by the MED be found in the previously published studies:
desalination process, using geothermal energy and surplus rejected heat
from solar thermal collectors. Ghorbani et al. [21] developed and • The integration of CCHP systems and local desalination units has not
analyzed a structure in order to cover energy and freshwater demands been previously reported.
of a residential complex. They utilized thermal energy provided by solar • Although there are some studies on the performance of the systems
collectors to produce desalted water through an MED system. providing cold, heat, power, and fresh water simultaneously, no

Fig. 1. Energy flow of the separate production system (S-System).

22
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

direct comparison has been made among them and the conventional and the desalination process. In addition, a preheated sensible heat-
CCHP and local desalination systems. driven MED is used to produce distilled water through recovering both
• The effects of integrating CCHP systems and desalination processes the mid-temperature and low-temperature waste heat. The rejected
on the performance of each individual system have not been in- heat by the MED is recovered for DHW or space heating demands. As
vestigated. the amount of distilled water produced by the MED is surplus to the
• The researchers [13–21] mostly focused on utilizing the waste heat regional water demand, it is considered to be sold to the local water
for the MED system and the users' heating demand, separately. In market. Other parts of the I-system are the same as those of the N-
this circumstances, the MED system has to work mostly at partial- system.
load particularly when the users' heating demand is high. Fig. 4 depicts the schematic diagram of the MED system with the
preheaters and the heat exchangers. The mid-temperature heat from the
GE is utilized to produce the vapor in the first effect, and the low-
2. System description temperature heat is used to preheat the feedwater through the pre-
heating heat exchangers (PHX). There are two ways to recover the
The energy flow diagram of the separate production system (S- waste heat rejected by the MED: the first one is from the latent heat of
system), conventional non-integrated CCHP and water desalination the vapor produced in the last effect, and the second one is from the
system (N-system), as well as the integrated system (I-system) are il- sensible heat of the brine and distilled water streams. Part of the last-
lustrated in Figs. 1–3, respectively. Two kinds of demands are con- effect vapor enters the feed condenser (FC) to preheat the feedwater,
sidered in this paper. The first one is related to building demands, in- and the rest enters another condenser, named heat-recovery condenser
cluding electricity, cooling, and heating energy. The second one is the (HRC), to provide space heating or DHW. In order to boost the amount
regional freshwater demand, which is provided by the RO desalination of the heat which can be recovered from the MED unit, the thermal
unit in all three systems. energy of the brine and distilled water streams is used to preheat
In the S-system, the cooling and heating demands are provided by feedwater and heat-recovery water through the brine heat exchanger
an electric chiller and a natural gas boiler, respectively. Furthermore, (BHX) and distillate heat exchanger (DHX), respectively. A cooling
all the electrical energy consumed for lighting, electrical equipment, seawater heat exchanger (CSHX) is also used to remove the surplus heat
electric chiller and water desalination unit is imported from the grid. from the heat-recovery water, when the building heating demand is
The N-system includes a CCHP unit composed of a natural gas en- low. In order to reduce the loading of CSHX, heat-recovery water can
gine (GE) as the prime mover, a natural gas boiler, and a hybrid cooling bypass the DHX through a three-way valve.
system. The recoverable thermal energy from the GE is utilized with
respect to its temperature. The high-temperature part (above 160 °C) of
3. Mathematical modeling
the exhaust gases is used to provide cooling energy through a double-
effect Li-Br/water exhaust absorption chiller, which is shown to be
3.1. The individual system
more favorable than single effect one in such CCHP systems [8]. The
mid-temperature thermal energy (80–160 °C) of jacket water and the
As mentioned above, the individual system of cooling, heating,
exhaust gases as well as the low-temperature heat (below 80 °C) from
power and desalted water production is considered as the reference
the lubricating oil are used to provide heating energy for the building
system. In this system, all the required electrical energy, ELimp, ref, is
through some heat exchangers. Whenever the available heat is not
imported from the grid.
enough to meet the energy demands, an electric chiller and a natural
gas boiler are assisted to cover the deficiency of the cooling and heating ELimp, ref = ELbuilding + ELRO + ELelc (1)
energy, respectively. In this system, the local RO plant is powered by
the grid, and consequently, the GE provides electricity only for the Here, ELbuilding is the building electrical energy demand including
building demands. The deficiency of electricity is purchased from the the electricity required for lighting and electrical equipment. ELRO and
grid. According to the current energy policy in Iran, the possibility of ELelc are the electricity consumed by the RO unit and the electric chiller,
selling surplus electricity produced by fossil fuel-based systems is pro- respectively. The electric chiller serves all the cooling demand, Qc, and
vided only for few years [23]; therefore, for the sake of generality, its power consumption can be obtained as follows:
selling electricity is not considered in this paper. Qc
Regarding the I-system, the RO and CCHP units are integrated, and ELelc =
COPelc (2)
as a result, the GE can provide electricity for both the building demands

Fig. 2. The energy flow of the conventional non-integrated CCHP and desalination system (N-system).

23
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Fig. 3. The energy flow of the proposed integrated system (I-system).

Fig. 4. Schematic diagram of the MED system with the preheaters and heat-recovery heat exchangers.

All the heating demand is covered by a natural gas boiler and its fuel GE and the thermal efficiency of the exhaust gases, jacket water and
energy consumption, FEgb, ref, can be calculated as: lubricating oil as well as the exhaust gases temperature are presented in
Table 1. These equations have been obtained by curve fitting through
Qh
FEgb, ref = the typical performance characteristics data presented in Ref. [24–26]
gb (3) for commercial natural gas engines over a nominal power range from
100 kW to 9 MW. It is assumed that only the electrical efficiency varies
with the engine partial-load, and thermal efficiencies as well as the
3.2. CCHP system
exhaust gases temperature remain constant at a definite nominal power.
The partial-load percentage of the GE can be expressed as:
In the CCHP systems, a GE is used to produce electrical energy. Its
fuel energy consumption, FEGE, can be estimated as: WGE
PL = ×%
WGE Wn (11)
FEGE =
el (4)
where Wn is the nominal power of the GE. As mentioned earlier, the
where, WGE is the power produced by the GE and ηel is the electrical possibility of selling surplus power to the grid is not considered in this
efficiency. The relations used to calculate the electrical efficiency of the paper; therefore, the operation strategy of the CCHP systems is

Table 1
The relations used for the gas engine modeling.
Description Equation

Electrical efficiency ηel = ηel, n(−4.814PL−0.7211 + 1.168) (5)

Nominal electrical efficiency el, n = 3.658 ln(Wn) + 14.029 (6)
Exhaust heat thermal efficiency 2 (7)
( 2e 5) Wn + 0.0321 Wn + 15.545 Wn < 1067.5KW
th, ex = 0.183
80.584 Wn Wn > 1067.5KW
Jacket water thermal efficiency = 253.85Wn
0.428 (8)
th, jw
Lubricating oil thermal efficiency th, oil = 1.3949 ln(Wn) 5.6212 (9)
Exhaust gases temperature Tex = 1214.8Wn
0.14 (10)

All the efficiencies are in terms of lower heating value (LHV) of natural gas.

24
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

considered to be the following electric load. The amount of the power Qh rec , MED QMED
FEgb, CCHP =
produced by the GE can be calculated as: gb (20)
WGE = min (Wn , ELCCHP ) (12) where ηrec,MED is the heat recovery efficiency of the MED unit, which
will be discussed in Section 3.4 for the I-system and is considered to be
where ELCCHP is the amount of electricity provided by the CCHP system
and can be expressed as follows: 1 for the N-system. QMED is the waste heat that is utilized in the MED
system to produce distillate water and can be expressed as:
For the N-system:
ELCCHP = ELbuilding + ELelc (13)
QMED = (1 ) Qex + Qjw + Qoil (21)

For the I-system:

3.3. RO desalination process
ELCCHP = ELbuilding + ELelc + ELRO + ELMED (14)
The schematic diagram of the RO system is shown in Fig. 5. The
where ELMED is the electricity consumed by the MED system.
main parts of this system are seawater intake and pretreatment (SWIP)
The electricity imported from the grid, ELimp, CCHP, to cover all the
pump, pretreatment unit, RO elements, the high-pressure pump (HPP)
electrical demand of the CCHP systems can be obtained as:
providing the required net driving pressure, Pf, and a Pelton turbine
ELimp, CCHP = ELCCHP WGE (15) recovering the energy of the brine stream. The relations used for
mathematical modeling of each component are shown in Table 2. The
A double effect exhaust gases absorption chiller is used to provide required net driving pressure and both the permeate and concentrate
building cooling demand by recovering the high-temperature waste TDS can be obtained by solving these equations simultaneously for all
heat from the exhaust gases. The cooling energy produced by the ab- the RO elements, based on the feedwater and membrane properties as
sorption chiller, Qc, ac, can be calculated as: well as the plant design and operating parameters. In this paper, the
Qc , ac = Qex COPac (16) following assumptions are imposed for the simulation [27]:

where Qex indicates all the recoverable thermal energy from the exhaust
• Steady state operation.
gases and α is the ratio of the high-temperature heat, which can be used
• Constant water and salt permeability coefficients.
in the absorption chiller, to all the thermal energy which can be re-
• The seawater contains only NaCl salt.
covered from the exhaust gases.
• Negligible seawater salt mass comparing to water mass.
Tex Tac , out
= The power consumption of the HPP, ELHPP, and the SWIP pump,
Tex Tmin (17)
ELSWIP, can be calculated as [28]:
where Tex indicates the exhaust gases temperature exiting the GE,
mf , tot (Pf P0 )
which can be estimated using Eq. (10). Tac, out and Tmin are the exhaust ELHPP =
gases temperature at the exiting of the absorption chiller and the final pump (34)
stack, respectively.
mf , tot PSWIP
The recoverable waste heat from the exhaust gases (Qex), jacket ELSWIP =
water (Qjw), and lubricating oil (Qoil) can be calculated based on the pump (35)
respective thermal efficiencies:
where ηpump is the isentropic efficiency of HPP or SWIP pump, P0 is the
Qj = th, j FEGE , j = ex , jw , oil (18) ambient pressure, mf , tot is the total feedwater mass flow rate, ρ is the
density of seawater and PSWIP is the pressure provided by the SWIP
When the cooling energy produced by the absorption chiller is not pump. The amount of power recovered by the Pelton turbine,ELPT, can
enough to satisfy all the cooling demand, the shortage is covered by the be calculated as [29]:
electric chiller. Its electricity consumption can be expressed as:
mb, tot
Qc Qc, ac ELPT = PT (Pb P0)
ELelc = (36)
COPelc (19)
where ηPT is the isentropic efficiency of the Pelton turbine, and Pb is the
The building heating demand is provided by recovering waste heat pressure of the brine flow entering the Pelton turbine. Thus, the re-
from the GE and MED in the N-system and I-system, respectively. quired electrical energy for the RO desalination unit can be calculated
Whenever the waste heat is insufficient to meet all the heating demand, as follow:
the natural gas boiler is activated. The fuel consumption of the
ELRO = ELHPP + ELSWIP ELPT (37)
boiler,FEgb, CCHP, can be calculated as follows:

Fig. 5. Schematic diagram of the RO unit.

25
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Table 2
The equations used for modeling of the RO elements [27,36,46].
Description Equation

Membrane permeate mass flow rate mp = (Jw + Js ) Sm (22)

Permeate mass flux through the membrane (Fick's law) Jpw = Aref. FF. TCF(ΔP − Δπ) (23)
Salt mass flux through the membrane Js = Bm(Xw − Xp) (24)
Pressure difference across the membrane Pf (25)
P = Pf Pp
2
Osmotic pressure Δπ = 284.23. Tρ(Xw − Xp) (26)
Temperature correction factor (27)
22000 1 1
exp , Tf > 25
8.314 298 Tf + 298
TCF =
25000 1 1
exp , Tf 25
8.314 298 Tf + 298
Feedwater pressure drop mf + m c 1.7 (28)
Pf = 9.5 × 108
2
Concentration polarization factor Xf Xc (29)
Xw Xp = Xp exp(0.7 r )
2
The membrane recovery factor mp (30)
r=
mf
Water mass balance across the membrane mf = mp + mc (31)
Salt mass balance across the membrane mf Xf = mp Xp + mc Xc (32)
Permeate salinity Js = Jw ⋅ Xp (33)

The RO specific energy consumption can be defined as follows [30]: respective effect. Q1 used in Eq. (41) indicates the amount of heat re-
covered from the jacket water and exhaust gases and can be calculated
ELRO
SECRO = 24 as:
Vp (38)
Q1 = Qjw + (1 ) Qex (44)
where Vp is the total volumetric flow rate of the permeate.
Here in this study, water is considered as the preheating fluid and its
3.4. MED desalination process mass flow rate, mpr , is calculated based on the thermal energy re-
covered from the lubricating oil.
A mathematical model for the MED system is presented in this
Qoil
section. The following assumptions are considered for the simulation mpr =
cp, pw (Tpr ,1 Tpr , N + 1 ) (45)
[2,31]:
where Tpr, 1, Tpr, N+1 are the inlet and outlet temperature of preheating
• The steady state operation. water, respectively.
• Equal temperature difference across the effects. As mentioned previously, the waste heat from the MED unit is re-
• Negligible heat losses to the surrounding. covered to provide heating energy for the building. In order to make a
• Negligible losses due to Non-Equilibrium Allowance (NEA). similarity with the GE thermal efficiency (Eqs. (7)–(9)), that shows the
• The brine salinity is considered to be 70,000 ppm. proportion of heat which can be recovered from the GE, MED heat
• The distillate is salt-free pure water. recovery efficiency, ηrec,MED, is defined as the ratio of the maximum heat
which can be recovered from the MED unit through HRC and DHX to
The mass flow rate of the produced vapor in each effect can be the overall input heat:
calculated by solving the salinity, mass and energy balance equations
listed in Table 3, simultaneously. All the thermodynamic properties of =
Qrec
the salt water such as specific enthalpy, saturated pressure, and specific
rec, MED
QMED (46)
heat are functions of both salinity and temperature. These properties
Qrec is the maximum amount of heat which can be recovered from
are calculated based on several relations presented in Ref. [32, 33], and
the MED unit. This can be obtained by writing the energy balance for
are also available in Ref. [34] as Matlab functions. As the PHXs are
the whole MED unit:
liquid-liquid heat exchangers, where any evaporation is not allowed
[35], the design intrinsically constrains the temperature of feedwater Qrec = QMED + Ftot cp, sw Tsw (Btot cp, b Tb, out + Dtot cp, pw Td, out ) (47)
entering each effect, Tf, i, up to the saturated temperature of the
where Ftot, Btot and Dtot are the total mass flow rate of the feedwater,
Table 3 brine, and distillate, respectively. Tb, out and Td, out are the brine and
The salinity, mass and energy balance equations for the MED simulation distillate temperature after the heat exchangers, respectively, and can
[31,36]. be estimated using the minimum temperature difference of the heat
Description Equation exchangers, ΔThx, as follows:

Water mass balance Fi = Bi + Di, i = 1, …, N (39) Tb, out = Tsw + Thx (48)
Salt mas balance Fi ⋅ Xf, i = Bi ⋅ Xb, i, i = 1, …, N (40)
First effect energy Q1 + F 1 hf ,1 B1 hb,1 (41) Td, out = Thr + Thx (49)
D1 =
balance hd,1
2 to N effects energy Di 1 hfg, i 1 + Fi hf , i Bi hb, i (42) For the sake of simplicity, cp, b and cp, pw are approximately replaced
Di = , i = 2, …, N
balance hd , i
PHX energy balance Fi cp, sw (Tf , i Tcond ) (43)
by cp, sw. By applying the water mass balance to the whole system and
Tpr , i + 1 = Tpr , i
mpr cp, pw
, i = 1, …, N substituting Eq. (47) into Eq. (46), the following simplified equation
can be obtained:

26
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Dtot (Thrw, in Tsw ) + Ftot Thx produce power. Therefore, the carbon dioxide emissions for all the
1 cp, sw
rec, MED
QMED (50) systems are expressed based on the carbon dioxide emission factor of
the natural gas, cerng, as follows:
Thrw, in is the temperature of heat-recovering water entering the
system. In the case that the space heating load is negligible, Eq. (50) can CDEj = PECj × cerng , j = S system, N system, I system (59)
be more simplified by considering the equality between the inlet tem-
perature of the heat-recovering water and seawater: 4.3. Economic performance
Ftot cp, sw Thx
1
rec, MED
QMED (51) The annual total cost saving ratio (ATCSR) is selected to evaluate
the economic performance. It is defined as the ratio of the annual total
The recovery ratio of the MED process, RMED, is defined as the ratio cost saved in the N-system or I-system, compared with the reference
of the distillate flow rate to the overall amount of feedwater flow rate as system, to the annual total cost of the reference system, and can be
follows [36]: illustrated as follows:
Dtot Xf ATCref ATCj
RMED = =1 ATCSRj = , j = N system, I system
Ftot Xb (52) ATCref (60)
where Xf and Xb are the salt concentration of the feedwater and brine,
The ATC is the annual total cost of the systems, which consists of the
respectively.
annual total capital cost, Cca, ann, annual total operating cost, Cop, ann,
The electricity consumption of the MED process, ELMED, can be es-
and the annual total benefits, BEann, achieved from selling surplus water
timated as follows:
produced by the MED unit. The ATC can be calculated for each system
SECMED Vd as follows:
ELMED =
24 (53)
ATC = Cca, ann + Cop, ann BEann (61)
Vd and SECMED are the daily distillate production rate and specific
energy consumption of the MED process, respectively. The electricity Cca, ann = Cca × CRF (62)
consumption is dominated by the pumping power, which can be esti- where Cca represents the total capital cost of the system, and CRF is the
mated as low as 0.7 to 1.2 kWh per each cubic meter of produced water recovery factor which can be calculated based on the plant life, n, and
for efficient systems [35]. In this paper, the SECMED is approximately the interest rate, i [37].
considered to be 1.0 kWh/m3.
i (1 + i)n
CRF =
4. Performance parameters (1 + i) n 1 (63)

4.1. Energy performance 5. Optimization methodology

In order to assess the energy performance of the system, the primary 5.1. The objective function and decision variables
energy saving ratio (PESR) is adopted as the index. It is introduced as
the ratio of the primary energy saved in the N-system or I-system, In order to identify the optimum design and operating parameters,
compared with the reference system, to the primary energy consumed minimization of the ATC is considered as the objective function for all
by the reference system. This can be expressed as follows for each the systems.
system considered in this study:
The objective function = min(ATC ) (64)
PECref PECj
PESRj = × %, j = N system, I system The RO plant recovery ratio, the number of pressure vessels and the
PECref (54)
number of elements per pressure vessel are the variables that are op-
ELimp, ref timized for the S-system. As the conventional RO plant adopted in the S-
PECref = + FEgb, ref
system and N-system are the same, the nominal power of the GE is the
grid (55)
only variable which is optimized in the N-system. In the I-system, as the
(ELimp, CCHP + ELRO ) RO, MED and CCHP units are integrated, all the aforementioned para-
PECN system = + FEGE + FEgb, CCHP meters as well as the number of MED effects are considered to be op-
grid (56)
timized, simultaneously. The decision variables as well as the con-
ELimp, CCHP straints are shown in Table 4.
PECI system = + FEGE + FEgb, CCHP
grid (57)
5.2. Optimization algorithm

4.2. Environmental performance A genetic algorithm (GA) is adopted here to perform the optimiza-
tion. Genetic algorithms are stochastic, iterative, global search methods
Generally, the global warming is considered as an important en- and developed based on Darwin's evolution theory. These algorithms
vironmental concern. It is directly related to the carbon dioxide emis-
sion. Hence, in this paper, the carbon dioxide emission saving ratio Table 4
(CDESR) is adopted to assess the environmental performance. It is de- Decision variables and their constraints [27,41,47].
fined as the ratio of the carbon dioxide emission saved in the N-system
Decision variable Unit Bound
or I-system, compared with the reference system, to the carbon dioxide
emission of the reference system. Gas engine nominal power, Wn kW [100, 3000]
Number of pressure vessels, NPV – [10, 200]
CDEref CDEj
CDESRj = × %, j = N system, I system Number of membranes per each PV, Nm – [2, 8]
CDEref (58) RO plant recovery ratio % [10, 60]
Number of MED effects, Ne – [2, 6]
In Iran, the conventional power plants mostly utilize natural gas to

27
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Table 5 are very robust as they avoid local optimal points and try to find global
Tuning parameters in the optimization program. optimum solutions. Moreover, they are used for various kinds of pro-
GA factor Value blems. However, they often require more computational time than
other algorithms. In order to solve the problem based on a given po-
Population size 50 pulation, pairs of individuals are chosen based on their fitness to re-
Max. generation size 300
produce subsequent population that ensures the survival of fittest in-
Residual 10−5
Crossover percentage 0.8
dividuals. This process is repeated for a number of generations, and
Mutation percentage 0.01 gradually improves the fitness. Finally, optimization stopping criteria
Selection method Tournament are met to approach the ideal solutions. The main characteristics of the
Tournament size 2 GA method used in the optimization process are expressed in Table 5.
The flow chart of the procedure used for the I-system optimization is
presented in the Fig. 6. The optimization methodology can be described
in the following steps:

1) The initial values of the decision variables are generated, and the
parameters required for the modeling is used as the input.
2) The RO process is modeled by solving the relations presented in
Section 3.3.
3) The relations presented for the CCHP and MED modeling are solved
together using the electricity consumption of the RO unit calculated
from Eq. (37) in the previous step.
4) The economic, environmental and energy performance of the whole
system is evaluated.
5) If termination criteria are satisfied, the optimization procedure is
stopped and the results are reported.
6) Otherwise, new population is generated and new data vector is used
for the first stage. The procedure is repeated until the termination
criteria are met.

6. Case study

Kish twin towers located in Kish island in the south of Iran (latitude
26°N, longitude 54°E) are considered as a case study to assess the
proposed system. The overall view of the towers is shown in Fig. 7. Each
tower includes two 24-story and one 18-story wings. Each floor consists
of four apartments with an average area of almost 110 m2. The 3D
model of the buildings is designed using SketchUp software (Fig. 7b)
and imported into Openstudio software in order to calculate the
building cooling and electricity demands. The cooling demand is cal-
culated based on the weather data gathered in Ref. [38] and is assumed
to be available only from April to October (Fig. 8). The daily electricity
demand profile (Fig. 9) is obtained based on the library data of Open-
studio software and is assumed to be constant throughout the year. In
this paper, the space heating demand is neglected due to the hot climate
of the island, and consequently, only DHW load is considered as the
Fig. 6. The flow chart of the procedure used for the I-system optimization. building heating demand.
DHW load is calculated based on the hot water consumption profile

Fig. 7. a) The real overall view of the towers b) the 3D model.

28
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Fig. 8. The hourly buildings cooling load throughout the year.

Fig. 9. Daily electricity demand of the buildings.

Fig. 10. Daily hot water consumption per occupation.

(Fig. 10) obtained from the water consumption data reported in [39].
The tap water temperature can be estimated as a function of ambient desalination plant is considered to cover one-third of the overall water
temperature, T0, as follows [40]: demand of the island population. The input parameters and relations
for economic and environmental simulation are listed in Tables 7–9.
Ttap = 0.0084T02 + 0.3182T0 + 11.403 (65)
7. Results and discussion
The hourly required thermal energy for DHW provision is shown in
Fig. 11. 7.1. Validation and optimization results for the RO process
Modeling and optimization are performed in Matlab software. The
parameters used for thermodynamic modeling are shown in Table 6. In order to validate the model used for the RO simulation, the re-
The regional water demand is estimated based on the average water sults obtained for each membrane of an RO pressure vessel in the I-
consumption per person, 225 L/day [41], and the island population, system are compared with the corresponding values achieved from
approximately 39,953 according to [42]. In this paper, the RO Water Application Value Engine (WAVE) application developed by the

29
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Fig. 11. Hourly DHW load of the buildings throughout the year.

Table 6 Table 7
Input parameters for the thermodynamic simulation. Input parameters for economic and environmental analysis.
Parameters Unit Value Parameter Unit Value

CCHP & SP systems [8,48,49] Electricity cost [40] ($/kWh) 0.1

Absorption chiller coefficient of performance, COPac – 1.41 Fuel energy cost [40] ($/kWh) 0.02
Electric chiller coefficient of performance, COPelc – 4.5 Distilled water market price [45] ($/m3) 2
°
DHW temperature, TDHW C 50 System lifetime, n year 20
°
Exhaust gases temperature at the outlet of absorption C 160 Interest rate, i % 5
chiller, Tac,out Natural gas carbon dioxide emission factor, cerng [54] kg CO2/kWh 0.201
Gas boiler efficiency, ηgb % 85
Grid efficiency, ηgrid % 33
°
Minimum allowable temperature of exhaust gases, C 130
Dow Chemical Company. The validation results are presented in
Tmin
Table 10.
Reverse osmosis The optimum performance parameters of the RO desalination units
Membrane properties, SW30HR-380 [27,50,51]
Active surface area, Sm m2 35.3
are summarized in Table 11. It can be seen that the calculated average
Maximum recovery ratio % 13 permeate flux for the conventional RO plant and integrated RO plant is
Max feed pressure bar 68.9 13.0 LMH and 14.0 LMH, respectively. These values are in the operating
Max feedwater volumetric flow rate m3/day 337.96 flux range of 13–14 LMH reported for typical SWRO desalination plants
Minimum brine volumetric flow rate m3/day 87.76
[43]. Moreover, the specific power consumption values calculated in
Salt permeability coefficient, B kg/m2·s 2.3 × 10−5
Water permeability coefficient, Aref kg/m2·s·Pa 2.7 × 10−9 this paper, are also in the range of 3.5 to 4.5 kWh/m3 presented for
Plant characteristics [6,52,53] conventional SWRO desalination plants [44].
Flow factor, FF – 0.85 According to Table 11, the optimization leads to a smaller number
Intake and pretreatment pump pressure, PSWIP bar 5 of membranes and a higher specific energy consumption in the I-
Maximum allowable permeate TDS ppm 400
Number of stages – 1
system, compared with in the N-system. The reason can be explained by
Pelton turbine efficiency, ηPT % 85 the fact that in the I-system, the GE provides the majority of the elec-
Plant load factor, PLF % 90 tricity required by the RO unit, which costs less than in the N-system, in
Regional fresh water demand, Vp m3/day 2996 which the RO electricity consumption is only provided by the grid.
HPP and SWIP pump isentropic efficiency, ηpump % 80 Naturally, in the I-system, as the minimization of the costs is considered
MED [31,47] as the objective function, the optimal parameters are achieved by fo-
First effect temperature, T1 °
C 70 cusing more on the reducing the number of membranes in comparison
°
Preheating water inlet temperature, Tpr,1 C 75 with in the N-system.
°
Preheating water outlet temperature, Tpr,N+1 C 50
°
Minimum allowable temperature difference in heat C 3
exchangers, ΔPhx 7.2. Validation and optimization results for the MED process
Plant availability factor, PAF % 98
°
Condenser temperature, Tc C 50 In order to validate the MED model implemented in this paper, the
Salt concentration of the rejected brine, Xb ppm 70,000
Number of personnel, Npe – 3
main performance parameters including the water production rate and
cooling water flow rate obtained from this model are compared to the
General
corresponding values given in Ref. [31] for three different sensible
Sea water salt concentration, Xf ppm 44,000
Sea water temperature, Tf °
C 25 heat-driven MED systems without preheaters. As shown in Table 12, the
Sea water density, ρ kg/m3 1000 maximum error is less than 1%, which shows a good consistency among
the results.
The main performance parameters of the MED process including the

30
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Table 8
The capital and operating costs (excluding utility costs) for economic simulation.
Description Capital cost Operating cost excluding utility cost Ref.

−0.412a
Absorption chiller 3575.5(Caac) (kW) 0.00013 (kWh) [8]
Electric chiller 153 (kW) 0.0015 (kWh) [8]
Heat exchanger 30 (kW) 0.003 (kWh) [55]
Gas fired boiler 47 (kW) 0.00034 (kWh) [8]
Gen set packageb 5137.8(Wn
0.187
) c (kW) ( 0.002 ln(Wn) + 0.0348)(kWh) [26]
RO unit 6.3 ⋅ PCtotd 27.375 ⋅ PLF ⋅ Vp + 0.0148 ⋅ PCtot + 0.2 ⋅ cm ⋅ Nme [28,53]
MED unit 3018.8Vd0.9795f 6000Npe + 25.5 ⋅ PAF ⋅ Vd + 0.01 ⋅ Cc, MEDg [2,35,47]

a
The relation is obtained by curve fitting through Broad X absorption chillers prices provided by Broad corporation [49].
b
The fitting formulas for capital and operating costs are obtained by curve fitting through vendor-supplied data which is gathered in [26] for typical CHP plants
based on natural gas engines over a 100 kW to 9 MW size range.
c
The capital cost includes gen set package price, cost of labor and material, engineering and fees, project contingency, project financing as well as project and
construction management costs.
d
The capital cost consists of indirect costs, including Freight & insurance rate during construction, construction overhead, contingency rate and owner's cost rate,
as well as direct costs, including purchased equipment cost, site development and installation costs. All these costs are presented as 6.32 times the total purchased
equipment cost, PCtot.
e
The first term indicates annual labor cost and annual chemical cost of product. The second term includes annual cost of the membrane and spare parts
replacement as well as insurance cost for sustainable operating. The third term presents annual membrane replacement cost.
f
This cost function can be used for the capital cost of the conventional MED plants with a capacity up to 10,000 m3/day [47].
g
The first term indicates annual labor cost. The second one includes annual chemical cost for water pretreatment and potabilisation as well as annual maintenance
cost. The latter presents annual insurance cost.

Table 9
Purchased equipment costs for RO plant [52,53,56].
Description Equation

Total purchased equipment cost PCtot = PCSWIP + PCHPP + PCPT + PCm

Seawater intake and pretreatment 4/5
PCSWIP = 996 V p
High-pressure pump log10 (PCHPP ) = 3.3892 + 0.0536log10 (WHPP ) + 0.1538[log10 (WHPP )]2
Pelton turbine log10 (PCPT ) = 2.2476 + 1.4965log10 (WPT ) 0.1618[log10 (WPT )]2
Membranes and PVs PCm = cmNm + cPVNPV
PV unit pricea cPV = 166.5Nm + 1576
Membrane unit price cm = 1200

a
The relation is obtained by curve fitting through prices of Codeline pressure vessels available in Ref. [57].

Table 10
The validation results of the presented RO model.
Element Feed pressure (bar) Permeate TDS (ppm) Recovery ratio (%)

Model WAVE Error (%) Model WAVE Error (%) Model WAVE Error (%)

1 68.87 67.78 1.61 181 165 9.93 8.04 8.40 4.25

2 68.55 67.43 1.65 226 208 8.79 7.57 7.80 2.91
3 68.26 67.12 1.70 286 265 7.76 6.96 7.05 1.24
4 68.01 66.84 1.75 364 341 6.80 6.24 6.20 0.69
5 67.78 66.59 1.79 469 442 6.00 5.46 5.32 2.59
6 67.58 66.36 1.84 607 575 5.49 4.66 4.47 4.26
7 67.40 66.14 1.90 788 748 5.33 3.90 3.70 5.41
Plant 68.87 67.78 1.61 348 320 8.75 35.8 35.8 –

Table 11 daily water production rate, the number of effects and the MED heat
RO performance parameters. recovery efficiency are summarized in Table 13. It can be observed that
Parameters Unit Value 93.87% of the overall input heat can be recovered again for DHW
provision.
Conventional RO Integrated RO Fig. 12 shows the influence of the seawater salinity and the number
of effects on the MED heat recovery efficiency. It can be seen that as the
Number of elements per PV – 8 7
Number of PVs, NPV – 34 36 number of effects rises in a certain sea water salinity, the MED heat
Number of membranes, Nm 272 252 recovery efficiency decreases. The reason is that the increase in the
Plant recovery ratio % 37.8 35.78 number of effects leads to a higher distillate flow rate and as a result, a
Feed pressure, Pf bar 68.9 68.9 higher amount of feedwater is needed; according to Eq. (51), this in-
Permeate TDS ppm 384 348
crease in the feedwater flow rate results in a smaller MED heat recovery
Average permeate flux LMH 13.0 14.0
Specific energy consumption, SECRO – 4.13 4.26 efficiency. It can also be observed that in a definite number of effects,
Electrical energy consumption, ELRO kW 515.0 532.4 an increase in the seawater salinity reduces the MED heat recovery
efficiency. It is because of the fact that according to Eq. (52), as the salt

31
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Table 12
The validation results of the presented MED model.
Parameter N = 4, Q1 = 5392 kW N = 5, Q1 = 8569 kW N = 6, Q1 = 11,788 kW

Model Ref. [31] Error (%) Model Ref. [31] Error (%) Model Ref. [31] Error (%)

3
Water production rate, m /day 700 707 0.99 1333 1341 0.60 2081 2103 0.99
Cooling water flow, kg/s 116.9 117.1 0.17 174.1 173.4 0.40 220.1 219.9 0.09

Table 13 Table 16
MED performance parameters. Main performance parameters of the power generation unit.
Parameters Unit Value Parameters N-system I-system

3
Daily distilled water production rate, Vd m /day 82.39 Yearly average working electrical efficiency (%) 35.23 38.68
MED heat recovery efficiency, ηrec, MED % 93.87 Nominal electrical efficiency, ηel·n (%) 37.85 39.90
Number of effects, Ne – 6 Yearly average partial load percentage (%) 73.54 85.15

7.3. CCHP system optimization results

The optimal capacity of the main equipment used in the CCHP

systems including the GE, the absorption chiller, and the electric chiller
is shown in Table 14. It can be observed that the optimal nominal
power of the GE increases from 674 kW in the conventional CCHP
system to 1178 kW in the I-system as, the GE can provide more elec-
tricity to power the RO plant in the I-system. Furthermore, it is clear
that in the I-system, the capacity of the absorption chiller is 811 kW,
which is 215 kW higher than that in the N-system. It can be expressed
by the fact that using the GE with larger capacity increases the amount
of rejected thermal energy and thus, more cooling energy can be pro-
duced by recovering the waste heat in the absorption chiller. Conse-
quently, it leads to a reduction in the capacity of the electric chiller. It
should be also mentioned that no gas boiler is needed in both CCHP
systems since all the heating demand required for the DHW provision is
covered through recovering the waste heat from the GE.

7.4. Energy performance analysis

Table 15 shows the APESRs of both N-system and I-system. It is clear

that adopting conventional CCHP systems decreases the primary energy
consumption by 12.70%, while integrating this technology with the
local desalination plant can lead to an additional 9.73% reduction in
Fig. 12. The influence of the feedwater salinity and the number of effects on the the primary energy consumption. This improvement can be mainly
MED heat recovery efficiency. explained by the rises achieved in the average of the GE electrical ef-
ficiency, the amount of waste heat recovered for useful purposes and
Table 14 the amount of electricity produced by the CCHP unit which reduces the
Rated parameters of the equipment used in CCHP systems. electricity purchased from the power grid in the I-system. These will be
Equipment Unit N-system I-system
discussed in the following sections in more detail.

Nominal power of the gas engine, Wn kW 674 1178

7.4.1. Power generation efficiency
Absorption chiller capacity, Caac kW 592 811
Electric chiller capacity kW 1802 1583
The yearly average of the electrical efficiency and partial-load
percentage, as well as the nominal electrical efficiency of the GE are
shown in Table 16. It can be seen that the year-round average of the GE
Table 15 working efficiency increases from 35.23% in the N-system to 38.68% in
Annual primary energy saving ratios of the systems. the I-system. It can be explained by two facts. Firstly, as mentioned in
Section 7.3, the capacity of the GE used in the I-system is 504 kW higher
Parameters N-system I-system
than in the N-system, leading to a 2.05% rise in the nominal electrical
Annual primary energy saving ratio, APESR (%) 12.70 22.43 efficiency of the GE in the I-system according to Eq. (6). Secondly,
according to Table 16, the yearly average percentage of the partial-load
rises from 73.54% in the N-system to 85.15% in the I-system. This
concentration in the feedwater rises, the recovery factor falls and a improvement can be seen more clearly from Fig. 13 showing the hourly
higher amount of feedwater flow rate is, thus, needed to provide the partial-load percentages in two typical days with and without cooling
same amount of distillate. It can be concluded that using an MED load. The prime mover in the N-system has to work at a lower partial-
process with a high number of effects in a cold climate, where the load whenever the electricity demand of the buildings is low, which can
heating demand is high, could impose a considerable extra heating load be as low as 27.42% during the night, while the GE follows a more
on the gas boiler, which might not be economical and energy-efficient. stable electric demand profile in the I-system, enhancing the partial-

32
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Fig. 13. The hourly partial-load percentages of the gas engine for two typical days: (a) without cooling load (b) with cooling load.

Fig. 14. Monthly average percentage of the GE working partial-load. Fig. 15. The annual recoverable waste heat produced by the gas engine and its
usage. The percentages show the proportion of the useful heat to the produced
waste heat in each system.
load percentage. It is because of the integration of the RO unit to the
CCHP system, which provides the possibility of powering the RO unit
consumed for the electric chiller in the I-system is lower than that in the
by the GE.
conventional CCHP system. Moreover, it can be seen that in the I-
Fig. 14 illustrates the monthly average percentage of the GE partial-
system, the useful thermal energy can increase up to 85.57% of the
load. It can be seen that the engine partial-load is higher during the
overall available heat, which is 32.58% higher than that in the N-
months with the cooling demand, and it is as high as 90.89% in the
system, mainly because of using the MED process. It can also be ob-
hottest month. This is due to the fact that in these months, more elec-
served that the heat used for DHW provision is the same in both I-
tricity is produced by the GE for the electric chiller.
system and N-system. It is because although a MED process is used in
the I-system utilizing the GE waste heat, all the heating demand of the
7.4.2. Thermal energy production buildings can be satisfied by recovering the rejected heat from the MED.
Fig. 15 shows the annual recoverable heat from the GE, as well as
the annual thermal energy used for different purposes. The percentages
at the top of the bars show the proportion of the useful heat to the total 7.4.3. Electricity production
available heat which can be recovered. It is clear that the amount of Fig. 16 depicts the amount of the electricity imported from the grid
recoverable heat in the I-system is higher than that in the N-system and produced by the GE. The percentages shown at the top of the bars
because of the larger capacity of the GE. As a result, more thermal indicate the decrease in the overall electricity demand in the N- and I-
energy can be recovered in the absorption chiller to provide cooling system, compared to that in the S-system. It can be observed that im-
energy (see the blue part in Fig. 15); consequently, the primary energy plementing CCHP systems makes a 5.54% and 6.52% reduction in the

33
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Fig. 18. The carbon dioxide emission for each system. The percentages present
Fig. 16. The annual overall electricity consumption in the systems. The per- the carbon dioxide emission saving ratios.
centages show the reduction in the total electricity demand, compared to the S-
system.
system and I-system reduces the carbon dioxide emissions by 12.70%
and 22.43%, respectively. The I-system produces 5050 tons CO2 each
year, which is 1460.5 tons lower than that in the S-system less
1469.5 tons CO2 per year, compared to the I-system.
As the grid power and the electricity produced by the GE are both
provided through consuming natural gas, the annual CDESRs and
APESRs have the same values in each system. Therefore, the reasons are
the same as those discussed in Section 7.4.

7.6. Economic performance analysis

The annual capital and operating costs of each system are shown in
Fig. 19. It can be observed that the capital and operating costs of the RO
unit in all systems are almost the same. The reason is that the optimum
Fig. 17. The electricity consumption for different purposes in the I-system. characteristics of the RO plants in the conventional systems and I-
system are quite similar. Furthermore, as can be seen, the capital cost of
the cooling devices follows the order of I-system > N-system > S-
overall electricity demand in the N-system and I-system, respectively. It system because the capital cost of the absorption chiller is higher than
is because of adopting the absorption chillers in the CCHP systems, that of the electric chiller, which serves all the cooling demand in the S-
utilizing the waste heat rather than the electricity consumed by the system. As can be seen, adopting the prime mover in the CCHP systems
electric chiller. Additionally, more than 90% of the electricity demand adds an additional capital and maintenance costs to the N- and I-
in the I-system is provided by the GE, which is significantly higher than system. It should be mentioned that in the I-system, although the ca-
that in the N-system. It is because of the possibility of powering the RO pacity and the amount of yearly electricity production of the GE in-
plant by the CCHP unit in the I-system. This can lead to a significant crease by 75% and 102%, compared to those in N-system, respectively,
reduction in the primary energy consumption as the power generation its capital and maintenance costs rise by 57% and 92%, respectively. It
efficiency of the CCHP systems is considerably higher than that of the can be explained by the fact that according to the relations used for the
power grid. capital and maintenance costs of the gen set package showed in Table 8,
As shown in Fig. 17, in the I-system, the electricity consumed for using the GE with higher capacity in the I-system reduces the unit price
desalination, the buildings electricity demand and the electric chiller of both capital and maintenance costs by 9.8% and 5.13%, respectively.
are 48.71%, 43.05% and 8.24% of the overall electricity demand, re- It can also be seen that the total operating cost decreases in the I-system
spectively. mainly through reducing energy charge cost by buying less electricity
from the grid. Overall, according to Table 17, adopting the conven-
7.5. Environmental performance analysis tional CCHP system leads to a 4.88% annual total cost saving ratio and
the combination of this system with local desalination units makes a
Fig. 18 illustrates the CDEs of the S-, N- and I-system as well as the further 6.49% increase.
annual CDESRs of the N- and I-system. The largest amount of carbon Fig. 20 gives more detailed information on the ATCs of the different
dioxide is emitted when the S-system is used. In this system, units and equipment used in the I-system. It is clear that the ATC of the
6244 tonnes CO2 is produced for providing electricity and 266 tonnes RO unit is the highest, at 44.11%, followed by the ATCs of the fuel
for providing domestic hot water. It can be seen that adopting the N- consumption, the GE, imported electricity, cooling equipment, MED

34
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

Fig. 19. Annual capital cost and annual operating cost of the systems.

Table 17
Annual total cost saving ratios of the systems.
Parameters N-system I-system

Annual total cost saving ratio, ATCSR (%) 4.88 11.37

Fig. 20. Annual total cost percentages of the units and equipment adopted in
the I-system.

and heat exchangers, which account for 26.87%, 18.56%, 4.45%,

2.40%, 2.40% and 1.20% of the total capital cost of the I-system, re- Fig. 21. The impact of the water market price on the economic performance of
spectively, the I-system.
As various market prices of fresh water are reported in the literature
[2,6,45] for the Gulf region, the influence of the water price on the
economic performance of the I-system has been investigated here.
the CCHP system is coupled to the local desalination unit. As far as
Fig. 21 represents the impact of the water market price on the ATCSR of
the annual total cost is considered as the objective function, the
the I-system. As can be seen, the prices vary from 1.05$, the interna-
optimization results in an RO desalination unit with higher specific
tional price, to 5$, the local distilled water price, and the ATCSR varies
energy consumption and fewer membrane numbers for the in-
from 9.8% to 16.0% with changing the water price.
tegrated system, compared to the conventional units.
2. Using the gas engine with the higher capacity in the I-system leads
8. Conclusion
to a 2.05% increase in the nominal electric efficiency. Also, it makes
a 9.91% and 5.13% reduction in the capital and maintenance unit
In this paper, the optimization and performance analysis of an in- costs of the gen set package. Furthermore, as more waste heat is
tegrated CCHP system with both RO and MED desalination units based
available in the I-system than in the conventional system, more
on the energy performance, environmental impact and economic per- cooling energy can be produced through the absorption chiller.
formance, are performed. The results are compared with those of the
Consequently, the absorption chiller capacity increases from 592 kW
conventional CCHP and local desalination system using the separate in the N-system to 811 kW in the I-system.
production system as the reference. Then, the proposed model is ap-
3. Powering the RO desalination unit by the CCHP system leads to a
plied for use in Kish island as a case study. The following conclusions more stable electric demand profile, compared to that of the con-
are obtained:
ventional CCHP system. It improves the average gas engine partial-
load from 73.54% in the conventional system up to 85.15% in the
1. The optimum capacity of the gas engine increases by 504 kW, when
integrated system.

35
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

4. Adopting an MED process utilizing the heat recovered from the producing energy and water, Energy Convers. Manag. 115 (2016) 200–220.
jacket water, lubricating oil and surplus exhaust gases increases the [18] F. Calise, M.D. d'Accadia, A. Macaluso, L. Vanoli, A. Piacentino, A novel solar-
geothermal trigeneration system integrating water desalination: design, dynamic
amount of useful thermal energy up to 85.57% of the total produced simulation and economic assessment, Energy 115 (2016) 1533–1547.
heat, which is 32.58% higher than that of the conventional CCHP [19] F. Calise, M.D. d'Accadia, A. Piacentino, Exergetic and exergoeconomic analysis of a
system. renewable polygeneration system and viability study for small isolated commu-
nities, Energy 92 (2015) 290–307.
5. The amount of heat which can be recovered from the MED process [20] F. Calise, M.D. d'Accadia, A. Piacentino, A novel solar trigeneration system in-
to provide building heating demand is mainly dependent on the tegrating PVT (photovoltaic/thermal collectors) and SW (seawater) desalination:
seawater salinity and the MED effects number. In this paper, 93.87% dynamic simulation and economic assessment, Energy 67 (2014) 129–148.
[21] B. Ghorbani, M. Mehrpooya, M. Sadeghzadeh, Developing a tri-generation system of
of the MED input heat can be recovered again for DHW, and all the power, heating, and freshwater (for an industrial town) by using solar flat plate
heating demand of the building can be provided through recovering collectors, multi-stage desalination unit, and Kalina power generation cycle, Energy
the waste heat. However, using the MED system with a high number Convers. Manag. 165 (2018) 113–126 (2018/06/01).
[22] P. Behnam, A. Arefi, M.B. Shafii, Exergetic and thermoeconomic analysis of a tri-
of effects in cold climates is not recommended.
generation system producing electricity, hot water, and fresh water driven by low-
6. More than 90% of the overall electricity required for both building temperature geothermal sources, Energy Convers. Manag. 157 (2018) 266–276
and RO unit can be covered by the gas engine. (2018/02/01).
7. The aforementioned improvements increase the yearly average [23] Iran's Ministry of Energy, Guidance Handbook for Development Microgrid,
Available: http://psp.moe.gov.ir/Psp/media/mojavez/5%20Distribution/
power generation efficiency, the annual primary energy saving ratio Documents/DG/5_2.pdf.
and annual total cost saving ratio up to 38.68%, 22.43%, and [24] Thermoflow Inc, Reciprocating Engines Sets, 22 ed., (2012).
11.37%, respectively, which are 3.45%, 9.73% and 6.49% higher [25] M. Ekwonu, S. Perry, E. Oyedoh, Modelling and simulation of gas engines using
Aspen HYSYS, J. Eng. Sci. Technol. 6 (3) (2013) 1–4.
than those of the conventional system. Moreover, the carbon dioxide [26] U.S. Environmental Protection Agency, Catalog of CHP technologies, Available:
emission falls by 1460.5 tons per year compared to that in the S- https://www.epa.gov/sites/production/files/2015-07/documents/catalog_of_chp_
system. technologies_section_2._technology_characterization_-_reciprocating_internal_
combustion_engines.pdf.
8. As the distilled water produced by the MED system is surplus to the [27] F. Vince, F. Marechal, E. Aoustin, P. Bréant, Multi-objective optimization of RO
regional water demand, the production is considered to be sold to desalination plants, Desalination 222 (1) (2008) 96–118 (2008/03/01).
the water market. This has significant effects on the economic per- [28] K. Park, D.Y. Kim, D.R. Yang, Cost-based feasibility study and sensitivity analysis of
a new draw solution assisted reverse osmosis (DSARO) process for seawater desa-
formance of the system. The annual total cost saving ratio can vary lination, Desalination 422 (2017) 182–193 (2017/11/15).
between 9.8% to 16.0% with changing the water price from 1.05$, [29] V. Eveloy, P. Rodgers, L. Qiu, Integration of an atmospheric solid oxide fuel cell -
international price, to 5$, the local price. gas turbine system with reverse osmosis for distributed seawater desalination in a
process facility, Energy Convers. Manag. 126 (2016) 944–959 (2016/10/15).
[30] M. Laissaoui, P. Palenzuela, M.A. Sharaf Eldean, D. Nehari, D.-C. Alarcón-Padilla,
References Techno-economic analysis of a stand-alone solar desalination plant at variable load
conditions, Appl. Therm. Eng. 133 (2018) 659–670 (2018/03/25).
[1] International Water Summit, Energy efficient desalination, Available: https://www. [31] B. Rahimi, H.T. Chua, Low Grade Heat Driven Multi-effect Distillation and
internationalwatersummit.com/_media/Energy-Efficient-Desalination-2018.pdf. Desalination, Elsevier, 2017.
[2] M. Salimi, M. Amidpour, Modeling, simulation, parametric study and economic [32] K.G. Nayar, M.H. Sharqawy, L.D. Banchik, J.H. Lienhard V, Thermophysical prop-
assessment of reciprocating internal combustion engine integrated with multi-effect erties of seawater: a review and new correlations that include pressure dependence,
desalination unit, Energy Convers. Manag. 138 (2017) 299–311 (2017/04/15). Desalination 390 (2016) 1–24 (2016/07/15).
[3] R. Ahmadi, S.M. Pourfatemi, S. Ghaffari, Exergoeconomic optimization of hybrid [33] M.H. Sharqawy, J.H. Lienhard, S.M. Zubair, Thermophysical properties of seawater:
system of GT, SOFC and MED implementing genetic algorithm, Desalination 411 a review of existing correlations and data, Desalin. Water Treat. 16 (1–3) (2010)
(2017) 76–88 (2017/06/01). 354–380.
[4] S. Loutatidou, B. Chalermthai, P.R. Marpu, H.A. Arafat, Capital cost estimation of [34] Massachusetts Institute of Technology, Thermophysical properties of seawater,
RO plants: GCC countries versus southern Europe, Desalination 347 (2014) Available: http://web.mit.edu/seawater/.
103–111 (2014/08/15). [35] A. Christ, K. Regenauer-Lieb, H.T. Chua, Boosted multi-effect distillation for sen-
[5] V. Eveloy, P. Rodgers, Techno-economic-environmental optimization of a solid sible low-grade heat sources: a comparison with feed pre-heating multi-effect dis-
oxide fuel cell-gas turbine hybrid coupled with small-scale membrane desalination, tillation, Desalination 366 (2015) 32–46 (2015/06/15).
Int. J. Hydrog. Energy 42 (24) (2017) 15828–15850 (2017/06/15). [36] J. Kucera, Desalination: Water From Water, Wiley, 2014.
[6] V. Eveloy, P. Rodgers, A. Al Alili, Multi-objective optimization of a pressurized solid [37] D.G. Newnan, T.G. Eschenbach, J.P. Lavell, Engineering Economic Analysis, 11th
oxide fuel cell–gas turbine hybrid system integrated with seawater reverse osmosis, ed., Oxford University Press, 2012.
Energy 123 (2017) 594–614. [38] Iran Meteorological Organization, Climatic and historical weather data, Available:
[7] M. Li, H. Mu, N. Li, B. Ma, Optimal design and operation strategy for integrated http://irimo.ir/eng/.
evaluation of CCHP (combined cooling heating and power) system, Energy 99 [39] H. Al Moussawi, F. Fardoun, H. Louahlia, 4-E based optimal management of a
(2016) 202–220 (2016/03/15). SOFC-CCHP system model for residential applications, Energy Convers. Manag. 151
[8] L. Wang, J. Lu, W. Wang, J. Ding, Energy, environmental and economic evaluation (2017) 607–629 (2017/11/01).
of the CCHP systems for a remote island in south of China, Appl. Energy 183 (2016) [40] M. Farahnak, M. Farzaneh-Gord, M. Deymi-Dashtebayaz, F. Dashti, Optimal sizing
874–883 (2016/12/01). of power generation unit capacity in ICE-driven CCHP systems for various re-
[9] Y. Zhang, X. Wang, S. Zhuo, Y. Zhang, Pre-feasibility of building cooling heating sidential building sizes, Appl. Energy 158 (2015) 203–219 (2015/11/15).
and power system with thermal energy storage considering energy supply–demand [41] H. Mokhtari, H. Ahmadisedigh, I. Ebrahimi, Comparative 4E analysis for solar de-
mismatch, Appl. Energy 167 (2016) 125–134 (2016/04/01). salinated water production by utilizing organic fluid and water, Desalination 377
[10] J. Uche, L. Serra, A. Sanz, Integration of desalination with cold-heat-power pro- (2016) 108–122 (2016/01/01).
duction in the agro-food industry, Desalination 166 (2004) 379–391. [42] Statistical Center of Iran, Population and Housing Censuses, Available: https://
[11] F. Colella, J. Uche, N. Dejo, Desalted water provided by a poligeneration scheme for www.amar.org.ir/english.
the tourist sector, Desalination 205 (1) (2007) 279–297 (2007/02/05). [43] A. Efraty, R.N. Barak, Z. Gal, Closed circuit desalination series no-2: new affordable
[12] C. Rubio, J. Uche, N. Dejo, Optimization of desalted water production in a poli- technology for sea water desalination of low energy and high flux using short
generation scheme for the tourist sector, Desalination 223 (1) (2008) 464–475 modules without need of energy recovery, Desalin. Water Treat. 42 (1–3) (2012)
(2008/03/01). 189–196.
[13] J. Xu, J. Sui, B. Li, M. Yang, Research, development and the prospect of combined [44] J. Kim, S. Hong, A novel single-pass reverse osmosis configuration for high-purity
cooling, heating, and power systems, Energy 35 (11) (2010) 4361–4367 (2010/ water production and low energy consumption in seawater desalination,
11/01). Desalination 429 (2018) 142–154 (2018/03/01).
[14] C. Rubio-Maya, J. Uche, A. Martínez, Sequential optimization of a polygeneration [45] V. Eveloy, P. Rodgers, L. Qiu, Hybrid gas turbine–organic Rankine cycle for sea-
plant, Energy Convers. Manag. 52 (8–9) (2011) 2861–2869. water desalination by reverse osmosis in a hydrocarbon production facility, Energy
[15] C. Rubio-Maya, J. Uche-Marcuello, A. Martínez-Gracia, A.A. Bayod-Rújula, Design Convers. Manag. 106 (2015) 1134–1148.
optimization of a polygeneration plant fuelled by natural gas and renewable energy [46] J. Kucera, Reverse Osmosis: Industrial Processes and Applications, Wiley, 2015.
sources, Appl. Energy 88 (2) (2011) 449–457 (2011/02/01). [47] A. Christ, B. Rahimi, K. Regenauer-Lieb, H.T. Chua, Techno-economic analysis of
[16] F. Calise, A. Macaluso, A. Piacentino, L. Vanoli, A novel hybrid polygeneration geothermal desalination using Hot Sedimentary Aquifers: a pre-feasibility study for
system supplying energy and desalinated water by renewable sources in Pantelleria Western Australia, Desalination 404 (2017) 167–181 (2017/02/17).
Island, Energy 137 (2017) 1086–1106 (2017/10/15). [48] U.S. Department of Energy, Absorption chillers for CHP systems, Available: https://
[17] F. Calise, M.D. d'Accadia, A. Macaluso, A. Piacentino, L. Vanoli, Exergetic and ex- www.energy.gov/sites/prod/files/2017/06/f35/CHP-Absorption%20Chiller-
ergoeconomic analysis of a novel hybrid solar–geothermal polygeneration system compliant.pdf.
[49] BROAD X, Absorption Chiller Model Selection & Design Manual, (2011).

36
M. Esrafilian, R. Ahmadi Desalination 454 (2019) 20–37

[50] Water Application Value Engine (WAVE), Design Warnings, 1 ed., (2017). https://www.umweltbundesamt.de/sites/default/files/medien/1968/
[51] The DOW Chemical Company, DOW FILMTEC™ SW30HR–380 Element, Available: publikationen/co2_emission_factors_for_fossil_fuels_correction.pdf.
http://www.dow.com/webapps/include/GetDoc.aspx?filepath=liquidseps/pdfs/ [55] J. Zhang, S. Cao, L. Yu, Y. Zhou, Comparison of combined cooling, heating and
noreg/609-00390.pdf. power (CCHP) systems with different cooling modes based on energetic, environ-
[52] Y. Du, L. Xie, J. Liu, Y. Wang, Y. Xu, S. Wang, Multi-objective optimization of re- mental and economic criteria, Energy Convers. Manag. 160 (2018) 60–73 (2018/
verse osmosis networks by lexicographic optimization and augmented epsilon 03/15).
constraint method, Desalination 333 (1) (2014) 66–81 (2014/01/15). [56] WaterAnywhere, SW30HR-380, Available: https://www.wateranywhere.com/8-
[53] R.S. El-Emam, I. Dincer, Thermodynamic and thermoeconomic analyses of seawater x40-6-000-gpd-99-7-rejection-filmtec-seawater-membrane/.
reverse osmosis desalination plant with energy recovery, Energy 64 (2014) [57] Forever Pure Place, Codeline Membrane Vessels, Available: https://www.
154–163. foreverpureplace.com/Codeline-RO-UF-NF-membrane-vessels-s/1914.htm.
[54] German Environment Agency, CO2 Emission Factors for Fossil Fuels, Available:

37