Energy 86 (2015) 423e435

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Stand-alone seawater RO (reverse osmosis) desalination powered

by PV (photovoltaic) and PRO (pressure retarded osmosis)
Wei He, Yang Wang, Mohammad Hasan Shaheed*
School of Engineering and Materials Science, Queen Mary, University of London, London E1 4NS, UK

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

Article history: A novel RO seawater desalination plant powered by PV (Photovoltaic) and PRO (PVROPRO) is proposed
Received 14 October 2014 and the feasibility of two stand-alone schemes, SSRO (salinity-solar powered RO) operation and SRO
Received in revised form (salinity powered RO) operation, are investigated. First, the stand-alone feasibility of the plant is ther-
9 April 2015
modynamically analysed. In doing so, on the basis of mathematical models describing RO, PRO and the
Accepted 11 April 2015
PV array, the stand-alone feasibility is numerically investigated and the feasible operational windows for
Available online 19 May 2015
the two operation schemes, SSRO and SRO, are identified. In addition, the detrimental effects, CP (con-
centration polarization) and RSP (reverse salt permeation) in the mass transfer, on the operational
Keywords:
Stand-alone desalination plant
windows are investigated. Finally, a case study of the proposed PVROPRO plant is developed based on the
Hybrid salinity-solar power generation hourly solar data of Perth Australia in a year. The highest weekly production rate is found to be almost 20
Concentration polarization effect times the rate in PVRO in the same week. Annual production is increased more than nine times
Reverse salt permeation effect compared to the stand-alone PVRO plant. Furthermore, it is found that, due to detrimental effects the
Optimal operation strategy weekly PW (product water) production rate is decreased in the range of 16e20% and the overall annual
reduction is 18.07%.
Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.

1. Introduction countries [6] have many hours of sun shine which also makes solar
energy an attractive source of renewable energy in those areas [7].
Water is one of the most abundant resources on earth. However, Actually, renewable energy powered RO (reverse osmosis)
most of it, about 97%, is saline water in the oceans and the desalination and the optimum operation management have been
remaining 3% is freshwater. The freshwater scarcity is becoming an widely studied in previous literature [8e13]. Among the renewable
increasingly significant problem in many areas around the world energy powered RO desalination applications, stand-alone PVRO
[1]. Desalination has been demonstrated to be a promising and plant has been demonstrated to be feasible both in terms of tech-
viable technology to provide drinking water [2]. But the main issue niques and costs in lab-scale and pilot-scale systems [14e17]. At
impeding the wider use of desalination technologies is the high present, it is considered as a proper solution for small-scale desa-
economic cost involved, especially due to intensive energy con- lination applications in rural areas with high solar insulation [18].
sumption [3]. In addition, current use of conventional fossil fuels as The effectiveness of the stand-alone plant depends on the location,
main power source is increasingly raising concerns over climate geographical conditions, topography of the site, and the capacity of
change and promoting global awareness for carbon dioxide emis- the plant. Bilton et al. presented a generalised methodology to
sion reduction and cleaner energy supplies [4]. In fact, because of evaluate the feasibility of small-scale PVRO systems in challenging
the shortage of freshwater, the opportunity for the use of renew- environment [19]. Their findings indicate that the freshwater cost
able energy to power small seawater and brackish water desali- of PVRO is economically feasible for most remote areas with high
nation has become possible in many small cities and villages on the availability of solar energy [20]. Fraidenraich et al. proposed a
coastal areas and in many of the remote small villages and cities in simple and general theoretical procedure for estimating the SEC
the mainland [5]. Regions like GCC (Gulf Cooperation Council) (specific energy consumption) to evaluate feasibility of a PVRO
plant and validate the methodology with experiments [21]. During
the last decade, with the significant development in optimal de-
* Corresponding author. Tel.: þ44 20 7882 3774.
signs [22,23] and control strategies [24,25] for PVRO plants, the cost
E-mail addresses: [email protected] (W. He), [email protected] (Y. Wang), of freshwater has considerably reduced. However, the sunshine is
[email protected] (M.H. Shaheed). not available at night. In order to prolong the operational hours, and

http://dx.doi.org/10.1016/j.energy.2015.04.046
0360-5442/Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.
424 W. He et al. / Energy 86 (2015) 423e435

to increase the renewable energy supply, alternative power sources effect of CP, and found a reduction in performance [46]. Zwan et al.
need to be identified and integrated to the designs for night time derived a hydrodynamic mass transfer model for a PRO process
operation. Previous investigations have focused on a hybrid system which accounted for the actual size of the membrane sheet and
combining solar and wind energy to power RO desalination pro- concluded that it is feasible to produce 0.5 MW with 1 m3/s of
cesses [26,27]. An estimation of energy and water production freshwater [47]. Kim et al. modified the PRO models in order to
during a large-scale time frame from photovoltaic-wind hybrid incorporate the spatial distribution of concentration and the ve-
system coupled with RO desalination unit was developed based on locity in the two flow channels, and compared four different RO-
the local solar and wind data [28]. Their results demonstrated the PRO hybrid configurations [48]. Recently, Feinberg et al. demon-
appropriation of the proposed hybrid system to produce water strated that the power densities achievable from PRO are well
from brackish water (6 g/L) in southern Tunisia. Novosel et al. below those predicted by extrapolating lab-scale measurement
evaluated the impact of desalination in combination with water with an idealised model and stated that future work should focus
pump storage and penetration of wind and solar energy [29]. on increasing salinity difference and identifying optimum oper-
The analysis of the case study in Jordan demonstrated that ating conditions [49]. Thus, using brine from RO is a natural solu-
the integration of water and renewable energy generation could tion to increase the initial salinity concentration difference.
provide a real benefit to the country water supply, energy security However, so far no research has addressed ICP, ECP and RSP effects
and ecology [29]. Recently, osmotic energy from water with on salinity energy generation in a hybrid RO-PRO plant. Therefore,
different salinities has emerged as a viable alternative. Compared further investigation is needed to address these issues.
with other renewable energy sources, osmotic energy is less peri- Although several previous studies have focused on the salinity
odic and has no significant operational hazards. It therefore has the energy powered RO desalination plant, it is still a challenge to
potential to formulate a hybrid energy system to supplement the develop the stand-alone salinity power driven desalination based
power supply. on the current PRO membrane performance. In fact, to deal with
PRO (pressure retarded osmosis) is one of the most investigated the problems in a stand-alone renewable energy powered system, a
technologies in osmotic energy harvesting [30]. In a PRO plant, hybrid renewable energy source is a promising solution [50e52].
because of the osmotic pressure difference across the membrane, There are a large number of examples of stand-alone renewable
freshwater permeates from the unpressurized feed solution to the energy powered systems at off-grid locations. An experimental
pressurised draw solution. Then, the pressurised permeation is study of hybrid energy generation including PV (Photovoltaic),
expanded in a hydro-turbine to generate electricity [31]. Following wind emulator, battery, and controller was constructed and
the rapid developments made in the area of high performance demonstrated to be capable to operate stand-alone mode and grid-
membrane in the last decade, it has started to be utilised in reality connected mode [53]. A work of the solar PV and the solar/hydro
[32]. In 2009, the world's first PRO salinity power plant was schemes for rural electrification was evaluated and shown to be
launched in Norway with a 4 kW capacity [33]. In prior studies, PRO more reliable and sustainable than the use of a diesel genset [54]. A
process has been investigated not only as an independent power similar application for off-grid rural electrification by hybrid diesel
plant [34e36], but also as an osmotic ERD (energy recovery device) power plant with high-penetration renewable and compressed air
positioned in pre- or post-treatment in hybrid RO-PRO plant energy storage was also found in literature [55]. Moreover, opti-
frameworks [37,38]. Compared with other current ERDs for RO mum design and control of the integrated PV and wind powered RO
desalination plants (e.g. pressure exchangers and hydro turbines), desalination plant was illustrated by a series of simulations to
the second law efficiency of the PRO is relatively high [39]. Feinberg demonstrate applicability and effectiveness [56,57]. Therefore this
et al. carried out a theoretical study on a hybrid RO desalination study focuses on investigating the hybrid solar-salinity energy
plant with two-stage PRO osmotic ERD and indicated significant supply in an RO desalination application to improve the freshwater
recoverable energy from mixing concentrated and dilute wa- production. In fact, the hybrid solar-salinity power generation has
ters [40]. Sharqawy and Banchik derived systematic effectiveness- several advantages: i) the salinity power improves the energy ef-
mass transfer units (ε-MTU) models of the RO and PRO plants ficiency of the solar powered system by the recovery of osmotic
which are significantly helpful for engineers in the design [41,42]. energy during daytime and by prolonging the operational hours
Also, an experimental pilot system was designed and constructed to over night through salinity power harvest; ii) the solar power helps
investigate the reduced RO energy consumption by integrating improve the total water production of the hybrid RO-PRO system by
with PRO [43,44]. The experiments showed that the enhanced providing external power to compensate the lack of commercially
power densities for the RO-PRO system ranged from 1.1 to 2.3 W/m2 high performance PRO membrane in osmotic energy extraction
and indicated that future RO-PRO systems may reduce the specific currently in practice. Therefore, in the hybrid system, the stand-
energy consumption requirements for desalination by ~1 kWh/m3 alone feasibility can be realised by two operations: hybrid power
[44]. A thermodynamic analysis on a stand-alone salinity power source of salinity power from PRO and solar power from PV array
driven RO plant is developed and, for the first time, the feasible during daytime, and only salinity power at night. With the osmotic
conditions of the stand-alone salinity power supply is identified in energy generation, more freshwater can be treated under the
a desalination application [45]. However, as suggested in the available solar radiation. Conversely, at lower RO water recovery,
literature, no research has addressed the potential integration of salinity power generated by PRO is potentially capable to sustain
the salinity energy with other renewable energy sources. Thus, this continuous operation when the sun is unavailable. Such integration
study aims to investigate the integration of salinity power and solar of power supplies ensures that the desalination plant meets the
power and to identify the optimum operations of this hybrid demand for the freshwater production. To this end, a study on the
renewable system in a desalination application. stand-alone RO desalination plant powered by PV and PRO is
Furthermore, the detrimental effects, including ICP (internal developed. First, the hybrid plant is proposed and thermodynam-
concentration polarization), ECP (external concentration polariza- ically analysed using a state-diagram. Following this, the stand-
tion) and RSP (reverse solute permeation), are inevitable in a real alone feasibility of the plant is studied and derived mathemati-
PRO plant. These detrimental effects significantly reduce the per- cally. Based on the models, the performance of the RO plant and the
formance of PRO salinity power generation. Banchik et al. investi- entire hybrid system is evaluated. The feasible operational win-
gated the overall membrane performance and the optimum dows of the two operations are identified and discussed. Finally,
operations of PRO plant based on ε-MTU model by evaluating the with the known hourly solar data available for Perth, Australia, over
 W. He et al. / Energy 86 (2015) 423e435 425

duration of a year, a case study of the proposed hybrid powered RO the osmotic membrane process [58]. In this study, IW (impaired
desalination is presented. water) is selected as the feed solution for the early-stage investi-
gation. In order to overcome the pressure loss along the flow
channel, the IW is pressurised by a BP (boost pump) which is also
2. Stand-alone seawater RO desalination plant powered by driven by the induction motor. Finally, the draw solution including
hybrid system of PV and PRO the permeated water from the PRO feed solution is expanded in a
HT (hydro-turbine) to generate electricity. Both renewable energy
A proposed stand-alone solar-salinity power driven RO desali- generators, the PV array and the PRO plant, are interconnected to an
nation system is illustrated in Fig. 1 and it shows the three main AC bus through DC/DC/AC and AC/DC/AC converters. For simplicity,
sub-systems in this hybrid plant. The hybrid system consists of RO the efficiencies of all the converters and motors are assumed to be
desalination and renewable power generation including solar and 100%.
salinity power; the two parts are closely interacted. The renewable
energy generation supplies the power to the desalination plant and
the brine from the RO desalination is the source of the chemical 2.1. Thermodynamic analysis of the stand-alone salinity-solar
potential for salinity power generation. In operations of SSRO (so- power driven seawater RO plant
lar-salinity powered RO) during daytime, both the PV array and the
PRO plant are working to generate electricity. In contrast, at night in Before further analysis of the hybrid system, some key states of
operations of SRO (salinity powered RO), only the PRO plant is the saline streams are presented in the pressureeflow rate (PeQ)
working. In Fig. 1(b), the detail diagram of the hybrid plant is diagram as illustrated in Fig. 2. As can be seen from the diagram, the
plotted. SW (seawater) is pressurised by a HP (high-pressure pump) pressure loss in the membrane and flow channels is negligible
and a hydraulic ERD before it flows into an RO membrane module. compared with the hydraulic pressure applied on the saline
The implementation of ERD can significantly reduce the exergy streams. So the applied pressure is considered as constant. Also, it is
destruction of the RO plant [12]. The HP is driven by the induction assumed that no fouling or membrane deformation is occurred.
motor. The freshwater is produced from the SW in the RO plant. And because a very small amount of energy is used by the BP
Accordingly, two streams flow out from the RO module: the PW compared to the energy consumed by the HP, in this study, the
(product water) and the CW (concentrated brine water). The CW is energy consumption is only considered as the work of the HP in the
further used to pressurise the SW in the ERD before it flows into the RO sub-system.
PRO plant, and the PW is the product of the hybrid system. In In the PeQ diagram, the energy consumption by the RO plant
another sub-system, the solar power is harvested by solar PV and the energy generation by the PRO plant can be represented by
technology and the salinity power is generated by the PRO plant. the areas illustrated in Fig. 2, i.e. the energy consumed can be
The low concentration streams (secondary wastewater and BW or represented by the area O-0-1-C, the energy recovered by ERD can
their mixture) are the potential candidates for the feed solution of be represented by the area 3-2-C-B and the energy generated by the

Fig. 1. Illustration of two operations in the proposed solar-salinity power driven RO desalination plant is presented in (a) and schematic diagram of the solar-salinity power driven
seawater RO desalination plant is presented in (b).
426 W. He et al. / Energy 86 (2015) 423e435

Fig. 2. Thermodynamic analysis of the hybrid salinity-solar power driven RO desalination plant in hydraulic pressure and flow rate diagram, PeQ diagram. In (a), the salinity cycle of
the SSRO operation is illustrated, and the cycle of the SRO operation is illustrated in (b).

PRO plant can be represented by the area O-5-4-B. These areas are order to operate the hybrid system stand-alone. With negligible
determined by the specific states of the saline streams, namely pressure drop along the membrane channels in both RO and PRO
states 2, 4 and 5 in Fig. 2. Moreover, these states of the saline plants [41,42], the energy generated by PRO, EPRO, can be repre-
streams can be controlled by the operations of the RO and PRO sented by the area O-5-4-B, which is
plants. Other states are usually determined by the local conditions
or/and RO TR (thermodynamic restrictions) [59]. Therefore, this EPRO ¼ SO54B ¼ ðQCW þ DQ ÞDPPRO (2)
‘salinity cycle’ has similar attributes as the classical thermodynamic
cycles by changing states of the salinity concentration [45]. where QCW is the volumetric flow rate of the CW, DQ is the volu-
Two operational strategies in this hybrid power system are metric flow rate of water permeation in the transportation, DPPRO is
hybrid power of salinity and solar and stand-alone salinity power, the hydraulic pressure applied on the CW. In addition, considering
namely SSRO and SRO operation. In the SRO operation, without the the hydraulic energy recovery by the ERD, the energy consumption
solar energy harvesting, the overall energy surplus between the by the RO plant can be expressed by the area O-0-1-2-3-B, as,
generation and consumption by the hybrid system can be repre-
sented as the difference between the areas D-1-2-3 and 0-5-4-D. In
WRO ¼ SO01C  SB32C ¼ QSW DPRO  QCW ðDPRO  DPPRO Þ
contrast, in SSRO operation, the overall energy surplus includes the
electricity generated from the PV array. Therefore, the overall en- (3)
ergy surplus can be represented as
where QSW is the volumetric flow rate of the SW and DPRO is the
SSRO applied hydraulic pressure on the SW. As discussed above, in order
DE ¼ EPV þ EPRO  WRO (1)
DESRO ¼ EPRO  WRO to operate the hybrid plant stand-alone by the SSRO and SRO
schemes, the energy surplus should be non-negative at each
where EPV, EPRO and WRO are the energy generated from the PV operation.
array, energy generated from the PRO plant and the energy In an RO plant, TR is the limiting operation close to the mini-
consumed by the RO-ERD plant, respectively. If the overall energy mum level of applied pressure (i.e. pressure approaching the
surplus is non-negative, theoretically, the hybrid system can be concentrated water osmotic pressure plus frictional pressure los-
operated as stand-alone. Otherwise, the hybrid system needs an ses). With the current generation of high permeable RO membrane,
extra power source to cover the exceeding energy consumption. it is feasible to operate the RO plant over a wide range of water
Therefore in the hybrid system, the stand-alone feasibility is recoveries to the limit of TR [59]. Zhu has developed a systematic
determined by the states of the streams and the availability of the steady-state model of RO operated at TR [59e61] which is used to
solar irradiation. In the SSRO operation, due to the availability of the simulate the performance of the RO plant in this study. In the TR
solar PV power, more freshwater can be separated from the saline operation of an RO plant, the applied pressure on the feed solution
stream as illustrated in Fig. 2(a). Thus, it allows higher applied equals the osmotic pressure of the brine at the outlet of the RO
hydraulic pressure in RO, and the energy consumption in the RO membrane module [59], which can be expressed as,
plant is higher than the energy generated by the PRO plant. In
contrast, without the solar PV power, the stand-alone feasibility of pSW Cvan0 t cSW
DPRO ¼ ¼ (4)
an RO desalination plant with osmotic energy generation by the 1Y 1Y
PRO plant can be realized by operating it at a lower water recovery
ratio [45]. In such an operation, the energy consumption by the RO where pSW is the osmotic pressure of the SW, Cvan't is the osmotic
plant is fully covered by the energy generation of the PRO plant. As pressure coefficient and cSW is the concentration of the SW. It is
illustrated in Fig. 2(b), at the limiting condition of the SRO opera- noted that the van't Hoff law is restricted to use on dilute, ideal
tion, the area 3-D-1-2 equals to the area 0-5-4-D. solutions [41,42]. In the salinity range of 0e70 g/kg, the amended
linear osmotic pressure coefficient 0.7345 bar kg g1 is validated
and the maximum deviation is 6.8% [41].
2.2. RO and PRO membrane process For a constant pressure PRO process (C-PRO), with enough
membrane area available (allowable for full-scale PRO discharge),
In the SSRO operation during daytime, the RO desalination is the applied pressure on the draw solution of the PRO process de-
powered by both the PV array and the PRO plant. According to termines the termination of water permeation. This means that the
Equation (1), the overall energy surplus should be non-negative in water permeation terminates when the net driving force of the
 W. He et al. / Energy 86 (2015) 423e435 427

water permeation between the two sides of the membrane is zero equations for real PRO and have validated these models with
[35], which is experimental results [63]. The steady-state model of the PRO has
been demonstrated successfully considering these detrimental ef-
DPPRO ¼ DpOutlet (5) fects and describing the reduced water flux and inevitable reverse
solute flux. According to the model, reduced water flux, JW, can be
where DpOutlet is the osmotic pressure difference at the outlets. For expressed as
a co-current PRO process, because both the draw and feed solution
flow towards the same direction, only one outlet needs to be ! !
cPRO
D expðJW =kÞ  cPRO
F expðJW S=DÞ
considered. In contrast, for a counter-current PRO process, with a JW ¼ A Cos  DPPRO
different pressure applied, the net driving force at either of the two 1 þ JBW ðexpðJW S=DÞ  expðJW =kÞÞ
outlets may satisfy the condition and terminate the water perme-
ation. Usually, the counter-current scheme performs better than (8)
the co-current scheme due to its high effectiveness [42]. For
simplicity, the co-current PRO process is considered first. where A is the membrane water permeability coefficient, B is the
In this investigation, the feed water desalinated by the RO sys- membrane solute permeability coefficient. cPRO D and cPRO
F are the
tem is assumed as SW with 35 g/kg salinity and its osmotic pressure concentration of the PRO draw and the PRO feed, respectively. D is
can be obtained according to the van't Hoff's law [41,42]. Thus, the bulk diffusion coefficient, k ¼ D/d is the boundary layer mass
without the consideration of the RSP, based on the mass balance, transfer coefficient in which d is the boundary layer thickness, and
the osmotic pressure difference at the outlet of the membrane can S ¼ tst/ε is the support layer structural parameter in which ts is the
be expressed as [35], thickness of the porous layer, t and ε are the tortuosity and porosity
of the support layer of the membrane respectively. In addition, the

   reverse solute flux is
cCW qCW c q
DpOutlet ¼ Cvan0 t cOut Out
CW  cIW ¼ Cvan0 t  IW IW
qCW þ Dq qIW  Dq !
(6) cPRO
D expðJW =kÞ  cPRO
F expðJW S=DÞ
JS ¼ B (9)
1 þ JBW ðexpðJW S=DÞ  expðJW =kÞÞ
where cOut Out
CW is the concentration of the CW at outlet, cIW and cIW are
the concentration of the IW at the inlet and outlet, respectively. qCW
and qIW are the inlet mass flow rate of the CW and the IW. The inlet where JS is the reverse solute flux.
flow rate of draw solution is the flow rate of the CW water from the Three sets of experimental data from literature are selected to
RO system with leakage assumed to be negligible. With 100% salt validate the transport equation of water flux represented in Equa-
rejection RO membrane, the concentration of the CW is determined tion (8). The membrane parameters used in the validation are
by the water recovery that is cCW ¼ cSW/(1  Y). Dq is flow rate of shown in Table 1. Power density, W, is another important mem-
water permeation across the PRO membrane. brane characteristic which is the product of the water flux and the
However, in a real PRO process, it is inevitable that the salt operating pressure. For the validation of water flux, the closely
permeates from the draw side to the feed side and the performance related power density is also compared with the experimental
of the PRO salinity power generation is significantly changed. If the results.
CP effect is considered, the flow rate of the permeation also In the comparisons, two membranes with different properties
changes. When these effects are considered, the mass balance is are selected. With the first membrane, two different salinity gra-
changed to dients are used, in which feed solution are RW (river water) with
concentration 0.9 mM and BW (brackish water) with concentration
 
cCW qCW  DS cIW qIW þ DS 80 mM, respectively. The validations of the PRO model are shown in
DpOutlet ¼ Cvan0 t  (7)
qCW þ DqS qIW  DqS Fig. 3. The results clearly indicate the satisfactory agreement be-
tween the modelling results and the experimental results in all the
where DS is the mass rate of the reverse solute, DqS is flow rate of conditions, including the mixing of different salinity gradients and
water permeation across the PRO membrane considering CP and the utilization of different membrane properties.
RSP effects. Furthermore, due to limited experiments of scale-up PRO plant
The mathematical models and the framework for modelling a available in the literature, it is difficult to verify the modelled
process considering the CP and RSP effects are presented in our process performance by the experimental results. Thus, modelling
previous investigation [62]. Generally, during the mass transfer in a a full-scale PRO process using the validated transport equation can
real PRO, the water permeates across the membrane. On one hand, be used to determine the viability of real systems and inform their
the feed solutes are selectively retained by the active layer and design and operation [64]. In this study, considering the detri-
accumulated in the support layer, resulting ICP. On the other hand, mental effects of CP and RSP in salinity power generation, the flow
the permeated water dilutes the draw solution near another side of rates of water permeation (DqS) and solute permeation (DS) is ob-
membrane active layer and causes ECP. In addition, RSP exists tained by integrating water flux (JW) and solute flux (JS) along the
because of the non-perfect rejection of the current PRO membrane. membrane from the inlet to the outlet until the balance as shown
Previous studies have developed water flux and solute flux Equation (5) is established. Detailed steps and framework of PRO

Table 1
Selected membranes and salinities for validation of the water flux equation.

Ref. number Concentration of the Concentration of the Membrane water permeability Membrane salt permeability Membrane structural
draw solution M feed solution mM coefficient L m2 h1 bar1 coefficient L m2 h1 parameter mm

[72] 1.06 0.9 1.23 ± 0.15 0.28 ± 0.07 149 ± 17

80
[73] 1 0 2.5 ± 0.2 0.9 450
428 W. He et al. / Energy 86 (2015) 423e435

Fig. 3. Validation of the water flux and power density modelled based on the PRO model used in this study. In (a) and (b), the simulation results of water flux and power density are
validated with the experimental results from [72]. And in (c) and (d), the simulation results of water flux and power density are validated with the experimental results from [73].

modelling can be found in [62]. In addition to the validation with process, both the changing flow rates and the concentration are
the available experimental data of transport flux, the model of the relatively close. Thus, along with the theoretical steady-state model
scale-up PRO process is also validated with the published simula- of RO, the overall performance of the hybrid system can be
tion work. Based on the membrane parameters from [64] which are evaluated.
A 3 L m2 h1 bar1, B 0.36 L m2 h1 and S 100 mm, and the salinity
gradients of draw solution 0.6 M and feed solution 0.015 M, the 2.3. Solar PV array
comparison between the two modelling results of scale-up PRO
process are made. Other parameters used in [64] are specific Datasheets of a PV array provide the information of the per-
membrane scale (membrane area per 1 L h1 initial feed solution) formance of PV devices with respect to STC (standard test condi-
0.2 m2 (L h1)1, temperature 298 K, and gas constant tion), namely irradiation 1000 W/m2 with an ambient temperature
8.314 J K1 mol1. The osmotic pressure is estimated based on the of 25  C (298 K). However, practical PV arrays are not always
van't Hoff's law which is the product of number of ion species, operated at STC. The performance of a PV array depends on the
temperature and gas constant. The comparison is shown in Fig. 4 in solar irradiation level and the ambient temperature. In this study,
which the changing flow rate and concentration of the draw and single-diode model of a PV array is used to find the non-linear
feed along the flow channels are plotted in a scale-up PRO process. currentevoltage equation with the parameters from the product
According to [64], the flow rates are converted into the normalized datasheet. In the single-diode PV model, the effect of the series and
value which is divided by the sum of initial flow rate of the draw parallel resistances are considered and it is warranted that the
and feed solution. As seen in Fig. 4, it is clear that the results maximum power of the model matches with the maximum power
simulated based on the modelling framework in this study are in of the real array [65]. The currentevoltage (IeV) characteristics of
agreement with the modelled results from [64]. In the scale-up PRO the single-diode PV cell is given by [66].

Fig. 4. Comparison between the results based on the scale-up PRO model in this study and the results from [64]. The changing flow rate and concentration of the draw and feed
along the flow channel in a scale-up PRO are shown in (a) and (b), respectively. The flow rate is represented by the normalized flow rate which is the flow rate at a given position
divided by the total initial flow rate. The subscript i denotes the draw or the feed solution.
 W. He et al. / Energy 86 (2015) 423e435 429

Table 2
The technical data of solar array Bosch M2453BB.

Short-circuit current [A], ISC 8.7 Nominal output [W], Pmpp 245
Open-circuit voltage [V], VOC 37.7 Voltage/temperature coefficient [V/K], KV 0.1206
Nominal current [A], Impp 8.2 Current/temperature coefficient [A/K], KI 0.0028
Nominal voltage [V], Vmpp 30.1 Number of series cell, NS 60

   
V þ RS I V þ RS I recovery in all the range of the ratio studied. In contrast, in the cases
I ¼ IPV  I0 exp 1  (10)
Vt a RP with real devices which are not 100% efficient, the profile of the SEC
has a minimum among the water recovery ratio. The reason is that
where IPV and I0 are the PV and saturation currents of the array, at very low water recovery, the energy loss due to the inefficient HP
respectively. Vt is the thermal voltage of the array and is given by and ERD is relatively large. And hence, the optimum SEC moves to
Vt ¼ NSkT/q. NS is the number of cell connected in series, k is the the high water recovery ratio. For instance, the solid line repre-
Boltzmann constant (1.3806503  1023 J/K), q is electron charge senting the HP efficiency 85% and the ERD efficiency 95% has the
(1.60217646  1019 C), T is the temperature. RS and RP are the optimum SEC approximately 1.3 in the water recovery around 0.18.
equivalent series and the parallel resistance of the array, respec- Actually, the TR operation is the limiting case when the high
tively. Detailed derivation and numerical algorithm to adjust the permeable membrane or/and sufficient area of the membrane is
IeV mode can be found in [65]. The technical data of solar array used. In practice, the higher SEC is required when the operation of
Bosch M2453BB used in this work is listed in Table 2. the RO plant does not meet the thermodynamic restriction (non-
TR). For instance, based on the ε-MTU model of the RO plant, with
3. Stand-alone salinity-solar power driven RO desalination the membrane permeability 3.61  106 kg m2 kPa1 s1, flow rate
plant of feed solution 1 kg s1, and area 100, 200 and 300 m2, the SEC is
presented in Fig. 5(b). The efficiency of the HP is 90%, and efficiency
3.1. Stand-alone solar PV powered RO desalination plant of the ERD is 98%.
In Fig. 5(b), the results clearly indicate that the profiles of the
Works investigating the SEC in steady-state operation of the RO SEC are different with different areas. For areas 100, 200 and
plant are available in the literature [61,67,68]. But different from the 300 m2, less the area requires, more the SEC in RO desalination. The
RO plants powered by the electric grid in which the power input is optimums of the SEC are shifted towards the lower water recovery
constant, PV powered RO plants are operated along a large variety ratio with the decrease on the membrane area. In addition,
of power supplies subject to the availability of intermittent solar comparing the profile of area 300 m2 with the SEC profile of a TR RO
energy. To deal with the intermittent power input, operation of the plant in Fig. 5(a) with the devices with same efficiency, it is found
RO plant is needed to change for the high effectiveness. The SEC that the profile of area 300 m2 approximately have the same values
performance of the RO plant has been investigated in previous in all the range of water recovery. Consequently, in this case, with
studies [61,68]. The outcomes of the SEC of the RO plant operated at the usage of 300 m2 area, the TR operation of the RO can be
TR with hydraulic energy recovery are shown in Fig. 5(a) in which achieved.
three sets of the efficiencies of devices in the RO plant are consid- Based on the results shown in Fig. 5, with different parameters
ered, including the HP and the ERD. (conditions and usages of devices and membrane), the RO plant
The inefficiency of the devices shifts the monotonically ideal SEC performs differently. But at each fixed condition, a minimum SEC
profile and the inefficiency increase of the devices increases the can be found, thus the RO plant can be operated optimally. To this
SEC. In the ideal case with devices of 100% efficiency, the lowest SEC end, to deal with the variable load from the solar PV array, the flow
occurs and the SEC increases with the increase of the water rate of the feed solution can be adjusted to ensure that the RO plant

Fig. 5. SEC profile of the RO plant. The SEC profiles of the RO operated at the TR condition are presented in (a). The SEC profiles of the RO plant with different membrane area usage
are shown in (b).
430 W. He et al. / Energy 86 (2015) 423e435

is operated most efficiently with respect to energy consumption. the PRO, the net SEC can be obtained. In the salinity energy har-
Thus, in a PVRO plant (operation with TR and non-TR), with certain vesting by the PRO, the CP and RSP cannot be avoided in real ap-
power supply from the solar PV array, the maximum water pro- plications. As a result, with the reduced energy generation in the
duction rate can be expressed as PRO, the net SEC is increased. According to the framework of
simulating the CP and RSP effect in salinity energy generation [62],
EPV the net SEC considering the overall detrimental effects in the PRO
qMAX
P ¼ (11)
SEC MIN plant and the reduced performance can also be obtained. The re-
sults of the optimum net SEC of the RO desalination plant with the
where SECMIN denotes the minimum SEC of the RO plant and qMAX is
P osmotic energy generation are presented in Fig. 6(a). The optimum
the maximum water production under certain solar energy supply,
net SEC of the RO-PRO plant without the CP and RSP is obtained
EPV.
through the maximum salinity energy harvest from the mixture of
Previous studies have investigated the operation of RO plants
CW and IW by applying the optimum hydraulic pressure on the CW
under variable-load and suggested that the RO desalination plant
[35]. The optimum net SEC of the RO-PRO plant considering the CP
can operate successfully under varying flow rate and pressure
and RSP effects is numerically obtained. The parameters used in the
without any technical problems [15,16]. In addition, it has been
simulations of the salinity power harvest by the PRO are: water
pointed out that relatively short time is needed for the transition
permeability 1.74 L m2 h1bar1, salt permeability 0.16 L m2 h1,
from one steady state to another steady state for an RO process and
and structural parameter 307 mm, temperature 298 K, mass transfer
the pseudo steady-state model can be assumed for time steps more
coefficient 138.6 L m2 h1 [63], diffusion coefficient
than 0.25 h [69]. Therefore, it is practical to study the RO plant in
1.49  109 m2 s1 [70], and van't Hoff coefficient 0.7345 bar kg g1
this work with the hourly data of solar irradiation and ambient
[42]. The efficiency of the HP is 90%, efficiency of the ERD is 98% and
temperature. In this work the optimal strategy of the PVRO plant is
efficiency of the HT is 90%.
to control the flow rate of the feed solution (CW) and the pressure,
The results clearly indicate that the optimum net SEC of RO
ensuring that the RO plant is operated at the available minimum
desalination with osmotic energy recovery by the PRO is signifi-
SEC state.
cantly decreased compared to that of the SEC of the RO plant
without osmotic energy recovery. Furthermore, the detrimental
3.2. Dynamics of the hybrid RO-PRO system effects raise the optimum net SEC in all studied RO operations.
Considering the CP and RSP effects, less the osmotic energy is
When the osmotic energy recovery is considered in the RO generated from the PRO plant, more the net SEC is required.
desalination plant, the net SEC (SECnet) that is the overall SEC At the lower range of the water recovery, the negative net SEC
considering the specific energy generation in the PRO plant, can be indicates that the salinity power generated by the PRO is larger
expressed as than the energy consumed by the RO. Therefore, the operations at
RO water recovery ratios with non-positive optimum net SEC are
WRO  EPRO the feasible operations for the SRO operation of the hybrid plant. In
SECnet ¼ (12)
QPW Fig. 6(a), the feasible operational window for the SRO operation is
O-B-E-C and considering the detrimental effects the feasible
where QPW is the volumetric flow rate of the PW. Based on equa- operational window becomes O-A-D-C. Moreover, with the in-
tions of energy consumed by the RO plant and energy generated by crease of the RO water recovery, the optimum net SEC increases as

Fig. 6. SEC profiles of RO desalination with osmotic energy recovery by PRO.

 W. He et al. / Energy 86 (2015) 423e435 431

well. In the range of the water recovery ratio with positive optimum accordingly, causes the lower water production rate at the constant
net SEC, the osmotic energy generated by the PRO cannot fully flow rate of the SW or the higher required flow rate of the SW to
cover the energy consumed by the RO desalination. The power produce a certain volume of the PW.
shortage can be supplemented by the energy from solar PV array in
the proposed hybrid system. Thus, these RO water recovery ratios 3.3. Simulation framework of the stand-alone salinity-solar power
belong to the SSRO operation. Under the constrained range of the driven RO desalination plant
RO water recovery, from 0.1 to 0.5, in Fig. 6(a), the maximum
feasible operational window of the SSRO operation is restricted by A flow chart illustrating the simulation framework is shown in
the window E-F-H-G for the ideal PRO salinity harvest and by the Fig. 7. In order to simulate the PVROPRO hybrid power desalination
window D-F-J-I for the PRO salinity harvest considering the CP and plant, several inputs are needed, including the input parameters of
RSP effects. the system and the environmental data of solar and salinity. The
The upper and lower limits of the feasible operational windows input parameters of the PV array are the efficiency of the PV panel
are determined by different factors in the two operational schemes. which is 15% and the number of the PV panels which is 20. The
In the feasible operational window of the SRO operation, the lower parameters of the PRO membrane are the same as that in Section
limits of both operations with or without the CP and RSP effects are 3.2. The RO plant is assumed to be operated at the TR operation. The
restricted by the lowest RO water recovery ratio. But the upper yearly data of solar irradiation and ambient temperature of Perth,
limits are different due to the performance limiting effects. The Australia is provided by Meteonorm 7 software [71] and used in this
upper limit is significantly decreased when the CP and RSP effects study for a case study of the proposed hybrid power RO desalina-
are considered. Conversely, theoretically the lower limit of the SSRO tion plant. The concentration of SW is 35 g/kg and the concentra-
operation is the upper limit of the SRO operation in both cases. The tion of the IW is assumed to be 0.1 g/kg.
upper limit depends on the specific available solar power per unit In the simulation, firstly the performance of the stand-alone
production water, ePV ¼ EPV/VPW. If the required water recovery by PVRO plant is evaluated. The PVRO plant is optimized to find the
the specific available solar power is lower than the maximum water appropriate flow rate of the SW to ensure the minimum SEC. Then,
recovery ratio studied (0.5), the upper limit of the SSRO operation is with the same devices (e.g. HP pumps) within the PV array and the
the required water recovery ratio by the specific available solar RO plant, the proposed hybrid powered RO desalination plant is
power. Otherwise, with the enough energy from the PV array, the modelled. The same pumps implementation denotes that the
upper limit is restricted by the maximum water recovery 0.5. One maximum flow rate in the PVROPRO plant is restricted by the
specific SSRO operation with a certain specific solar power is maximum flow rate in the PVRO plant. In the simulation, the solar
illustrated in Fig. 6(b). The upper limit of the RO water recovery is PV power generation is assumed to be operated with maximum
less than 0.5, namely state A for the PRO salinity extraction power point tracking that the available maximum power is sup-
considering the CP and RSP effects and state B for the ideal PRO plied to the desalination plant under certain environmental
salinity extraction. condition.
Actually, similar to the PVRO plant, the water is needed to be
optimally produced by the PVROPRO plant. The optimum operation 4. Results and analysis
of the SRO and SSRO schemes are achieved at their upper limits of
the RO water recovery ratio. From the perspective of the optimum A stand-alone RO desalination plant powered by PV array is first
operation, the detrimental effects reduce the upper limit of the RO modelled with predefined parameters. The RO plant is operated at
water recovery ratio in both the SRO and SSRO schemes. In Fig. 6(a), its minimum SEC, and then the maximum water production is
in the window of the SRO operation, the optimum operation of the obtained under certain level of solar PV power based on Equation
RO with ideal PRO power generation is the state E, whilst the op- (11). In addition, with the same scale of flow rate of the SW used in
timum is state D due to the real CP and RSP effects. In the SSRO the PVRO plant, the hybrid solar-salinity powered RO desalination
operational window, under certain specific available solar power plant is modelled. Because the stand-alone feasibility of the hybrid
(ePV), the upper limit is also further restricted by the detrimental solar-salinity powered RO desalination system depends on the
effects as illustrated in Fig. 6(b). The optimum operation of the RO relation between the flow rates of the SW and the IW, another
with ideal PRO power generation is the state B. In contrast, the advantage of the PVROPRO plant is that it can be operated at a
optimum ratio is reduced to state A due to the real CP and RSP larger SW flow rate if the volume of IW is sufficient. In this study,
effects. The resulting lower maximum water recovery ratio, the total dimensionless flow rate which is the ratio of the IW flow

Fig. 7. Illustration of the simulation framework.

432 W. He et al. / Energy 86 (2015) 423e435

rate to the sum of the SW and the IW flow rate is assumed as 0.5 in stand-alone PVRO plant and the stand-alone PVROPRO plant with
modelling both the processes with and without osmotic energy and without the detrimental effects.
generation. The flow rate of the SW is selected as the maximum The results clearly indicate the dependence of the PW flow rate
flow rate of the SW required in the stand-alone PVRO plant. Thus, on the available PV solar power. The water production rate varies
the same flow rate of the IW is required to meet the predefined with respect to different solar power availability in both the PVRO
overall dimensionless flow rate. In both the SSRO and SRO schemes, and PVROPRO plants. However, with the salinity power generation,
the maximum RO water recovery ratio can be obtained by solving the PW water production rate is considerably increased and less
Equation (1) with the limiting zero overall energy surplus. More- fluctuated. Comparing Fig. 8(a) and (b), the hourly production rate
over, the effects of the CP and RSP are investigated with the same of the PVROPRO plant is much larger than that of the PVRO plant. In
flow operation, namely the flow rate of the SW and the IW. the PVROPRO plant, due to the salinity power generation, the
desalination plant can be operated continuously. In the SRO oper-
4.1. Overall optimum production rates of the PW ation, a constant water production is achievable at the upper limit
of the RO water recovery as shown in Fig. 6(a) (state E). In addition,
The objective is to maximize the flow rate of PW in the PVRO during daytime, with the available solar power, more PW can be
and the PVROPRO plants. With hourly PV solar power calculated, treated in the stand-alone desalination plant as illustrated in
both the SSRO and SRO operations are optimised. During the Fig. 6(b) (state B). The relative PW increase is presented in Fig. 8(c)
search, with the step-size of the RO water recovery ratio 0.01 (1%), with respect to weekly PW production. The results show significant
all the operations in the range of RO water recovery (0.1e0.5) are increase in PW production due to incorporation of the salinity
calculated and compared. Then, the hourly optimum RO operation power harvesting technology in the PVRO desalination plant. It is
for the maximum water production is obtained. The results, as observed in Fig. 8(c) that, the PW production rate of the most
presented in Fig. 8, show the increased water production by the improved week is almost 20 times the rate in PVRO plant. On
osmotic power generation and the water reduction due to the CP average, the yearly PW production of the PVROPRO plant is
and RSP effects. These two influences are represented by increased more than nine times the stand-alone PVRO plant.
the relative PW increase and decrease, respectively, which are Furthermore, the profile of the relative PW production increase
defined as shows that the more increase occurs when the PW rate in the PVRO
plant are less, i.e. hour 4000e5000 in Fig. 8(a) and week 20e30 in
qPVROPRO  qPVRO Fig. 8(c). Because there was less irradiation in magnitude and
qþ
PW ¼
PW PW
;
qPVRO
PW
shorter daytime during such periods in the year, the PV solar power
(13) is reduced. Simultaneously, the continuous osmotic power gener-
qPVROPRO  qPVROPRO;CPRSP ation which is less periodic, plays a more important role to sustain
q
PW ¼
PW PW
qPVROPRO
PW
the plant when the solar PV power is relatively lower.
When the overall performance limiting effects in PRO salinity
where qþ 
PW and qPW are the relative PW increase due to osmotic
energy harvest are considered, the salinity power generation de-
energy generation and the relative water decrease due to the CP creases. Thus, the overall PW production of the hybrid desalination
and RSP effects in PRO salinity power harvest, respectively. qPVRO
PW , decreases as well. The weekly relative PW decrease is shown in
qPVROPRO
PW , and qPVROPRO;CPRSP
PW are the flow rates of the PW in the Fig. 8(d). According to the simulations with the membrane studied,

Fig. 8. The stand-alone RO desalination plant. In (a), the hourly water production of the stand-alone PVRO desalination plant are presented. In (b), the hourly water production of
the stand-alone PVROPRO desalination plant is shown. The water production increase due to the osmotic energy generation is presented in (c), and the water production reduction
due to the CP and RSP effects in PRO plant are presented in (d).
 W. He et al. / Energy 86 (2015) 423e435 433

the weekly PW production rates are decreased in the range of be controlled by the hydraulic pressure applied on the SW based on
16e20%. Annually, the overall PW production is reduced by 18.07% the TR condition.
due to the CP and RSP effects. In addition, it is found that the most Conversely, the simulations of the PVROPRO plant are developed
significant reduction also occurs when the solar power is less. In the with the constant flow rate of the SW as presented in Fig. 9(a). All
SSRO operation, the optimum RO water recovery ratio is lower the flow rates during a day are constant. In the PVROPRO plant, the
when the solar power is less. And for the proposed hybrid desali- flow rate of the SW is selected as the maximum in the PVRO plant
nation plant, lower RO water recovery rate causes more energy loss for the entire year. With the osmotic energy generation by the PRO
due to the detrimental effects. As shown in Fig. 6(b), with the in- plant, the RO desalination plant can be operated consecutively over
crease of the RO water recovery rate, the deviation between the two night at a low water recovery, as shown in Fig. 9(b). However, the
dotted lines become less at the same water recovery rate. Therefore, performance limiting effects reduce the optimum water recovery
during the period with less solar irradiation, more reduction due to ratio in both the SSRO and SRO operation. In Fig. 9(b), the optimum
the CP and RSP effect may result in. states discussed in Fig. 6 can be identified in the 24 h operation. It is
noted that the PVROPRO without the CP and RSP effects, operation
4.2. Optimum operations of the stand-alone RO desalination plant of hours 1e5 and hours 21e24, is carried out at the state E shown in
Fig. 6(a), and operation of hours 6e20 is carried out at the state B
The optimum operation of the stand-alone RO desalination shown in Fig. 6(b) subject to different solar power. In addition,
plants is studied, including the stand-alone PVRO plant and the when the CP and RSP effects are considered, operation of hours 1e5
stand-alone PVROPRO plant with and without consideration of the and hours 21e24 is carried out at the state D as shown in Fig. 6(a),
detrimental effects. The optimum flow rate of the SW and the op- and operation of hours 6e20 is carried out at the state A as shown
timum water recovery ratio are shown in Fig. 9. A summer day with in Fig. 6(b).
24 h (from the 25th to 48th hours, as shown in Fig. 8), is considered
to compare the different optimum operation of the stand-alone RO 4.3. Optimum operations of the PRO plant
desalination plant.
The results clearly indicate different optimum operation In the PRO plant, the flow rate of the CW is determined by the
required for the two RO desalination plants. The SW flow rate of the RO water recovery ratio in the RO plant. The flow rate of the IW
PVRO plant clearly indicates the varying optimum flow rate with might be restricted by the local condition of the low concentration
respect to the available solar power. In the range of hours without streams. Because compared to available SW in coastal regions, the
solar power (hour 1e5 and hour 21e24), the flow rate of the SW are availability of low concentration stream is always limited. In this
zero and also zero treated freshwater is produced. In contrast, study, the overall dimensionless flow rate is assumed to be 0.5 that
during daytime, the flow rate is changed in order to ensure the RO the IW flow rate is same to that of the SW. Therefore, in order to
plant operated at its minimum SEC operation as shown in Fig. 9(b). extract the maximum salinity energy from the given volume of the
The optimum water recovery ratio of the PVRO plant is 0.13. It can SW and the IW, the hydraulic pressure applied on the CW should be

Fig. 9. The optimum operations of the RO plant. In (a), the optimum flow rates of the SW are presented. In (b), the optimum RO water recovery ratios are shown.
434 W. He et al. / Energy 86 (2015) 423e435

production of the PVROPRO plant is increased more than nine times

that of the stand-alone PVRO plant; 4) the CP and RSP effects in the
PRO plant reduce the performance of the PVROPRO plant. The
weekly PW production rate is decreased in the range of 16e20% due
to the detrimental effects. Annually, the overall production is
reduced by 18.07%.

References

[1] Kalogirou SA. Seawater desalination using renewable energy sources. Prog
Energy Combust Sci 2005;31:242e81.
[2] Shannon MA, Bohn PW, Elimelech M, Georgiadis JG, Marinas BJ, Mayes AM.
Science and technology for water purification in the coming decades. Nature
Fig. 10. Optimum hydraulic pressure applied on the CW in the PRO plant. 2008;452:301e10.
[3] Fritzmann C, Lo €wenberg J, Wintgens T, Melin T. State-of-the-art of reverse
osmosis desalination. Desalination 2007;216:1e76.
[4] Subramani A, Badruzzaman M, Oppenheimer J, Jacangelo JG. Energy minimi-
optimised. In the ideal PVROPRO plant in which the CP and RSP zation strategies and renewable energy utilization for desalination: a review.
effects are ignored, the maximum osmotic energy extraction can be Water Res 2011;45:1907e20.
[5] Al-Karaghouli A, Renne D, Kazmerski LL. Solar and wind opportunities for
analytically obtained [35]. But for a PRO plant with the CP and RSP water desalination in the Arab regions. Renew Sustain Energy Rev 2009;13:
effects considered, the maximum extractable energy cannot be 2397e407.
easily calculated. It needs simulation by certain step-size of the [6] Doukas H, Patlitzianas KD, Kagiannas AG, Psarras J. Renewable energy sources
and rationale use of energy development in the countries of GCC: myth or
pressure and comparison between the results to search for the reality? Renew Energy 2006;31:755e70.
optimum [62]. With different membrane used (water and salt [7] Mokheimer EMA, Sahin AZ, Al-Sharafi A, Ali AI. Modeling and optimization of
permeability coefficients, structural parameter, and et al.) and flow hybrid windesolar-powered reverse osmosis water desalination system in
Saudi Arabia. Energy Convers Manag 2013;75:86e97.
parameters (flow rates, flow directions et al.), the phenomena and
[8] Xia GH, Sun QX, Cao X, Wang JF, Yu YZ, Wang LS. Thermodynamic analysis and
detrimental effects of the CP and RSP are different. Based on the optimization of a solar-powered transcritical CO2 (carbon dioxide) power
membrane and flow parameters in this study, the results of the cycle for reverse osmosis desalination based on the recovery of cryogenic
energy of LNG (liquefied natural gas). Energy 2014;66:643e53.
optimum operations are shown in Fig. 10 in which both the
[9] Antipova E, Boer D, Cabeza LF, Guille n-Gos
albez G, Jimenez L. Uncovering
PVROPRO plant with and without consideration of the CP and RSP relationships between environmental metrics in the multi-objective optimi-
effects are included. zation of energy systems: a case study of a thermal solar Rankine reverse
The results indicate the optimum operational widow during a osmosis desalination plant. Energy 2013;51:50e60.
[10] Blanco-Marigorta AM, Masi M, Manfrida G. Exergo-environmental analysis of
day subject to the availability of solar power. Higher hydraulic a reverse osmosis desalination plant in Gran Canaria. Energy 2014;76:
pressure is required during daytime for both PVROPRO plants. It is a 223e32.
result of the more concentrated CW caused by the higher water [11] Calise F, d'Accadia MD, Piacentino A. A novel solar trigeneration system
integrating PVT (photovoltaic/thermal collectors) and SW (seawater) desali-
recovery in the RO plant with the PV array in operation. Moreover, nation: dynamic simulation and economic assessment. Energy 2014;67:
the overall detrimental effect on the optimum applied pressure is 129e48.
also studied. The results show that, when the CP and RSP effects are [12] El-Emam RS, Dincer I. Thermodynamic and thermoeconomic analyses of
seawater reverse osmosis desalination plant with energy recovery. Energy
considered, the required optimum pressure is lower in both the 2014;64:154e63.
SSRO and SRO operation, which is a result of the reduced osmotic [13] Ghobeity A, Mitsos A. Optimal time-dependent operation of seawater reverse
energy generation. osmosis. Desalination 2010;263:76e88.
[14] de Carvalho PCM, Riffel DB, Freire C, Montenegro FFD. The Brazilian experi-
ence with a photovoltaic powered reverse osmosis plant. Prog Photovolt Res
Appl 2004;12:373e85.
5. Conclusions [15] Thomson M, Miranda MS, Infield D. A small-scale seawater reverse-osmosis
system with excellent energy efficiency over a wide operating range. Desa-
lination 2003;153:229e36.
An investigation into the development of a novel stand-alone RO
[16] Thomson M, Infield D. A photovoltaic-powered seawater reverse-osmosis
seawater desalination plant powered by a solar PV and a PRO is system without batteries. Desalination 2003;153:1e8.
carried out. Two stand-alone schemes, the SSRO and SRO operation, [17] Joyce A, Loureiro D, Rodrigues C, Castro S. Small reverse osmosis units using
are proposed and investigated using a state-diagram. With the PV systems for water purification in rural places. Desalination 2001;137:
39e44.
mathematical models describing the membrane process of the RO, [18] Aybar HŞ, Akhatov JS, Avezova NR, Halimov AS. Solar powered RO desalina-
the PRO and the solar PV energy harvest, the stand-alone feasibility tion: investigations on pilot project of PV powered RO desalination system.
is studied numerically and both the feasible operational windows Appl Sol Energy 2010;46:275e84.
[19] Bilton AM, Kelley LC, Dubowsky S. Photovoltaic reverse osmosis d Feasibility
of the SSRO and SRO operation are analysed. In addition, the and a pathway to develop technology. Desalin Water Treat 2011;31:24e34.
detrimental effects, the CP and the RSP, are also investigated. [20] Bilton AM, Wiesman R, Arif AFM, Zubair SM, Dubowsky S. On the feasibility of
Finally, with the hourly solar data of Perth, Australia in a year, the community-scale photovoltaic-powered reverse osmosis desalination sys-
tems for remote locations. Renew Energy 2011;36:3246e56.
production rates of the PVROPRO plant during a year is modelled [21] Fraidenraich N, Vilela OC, Lima GA. Specific energy consumption of PV reverse
and the optimum operational widows are identified and discussed. osmosis systems. Exp Theory Prog Photovolt Res Appl 2013;21:612e9.
Based on the results, some conclusions can be drawn: 1) the [22] Poullikkas A. An optimization model for the production of desalinated water
using photovoltaic systems. Desalination 2010;258:100e5.
feasibility of the PVROPRO plant can be realized by the SSRO [23] Kelley LC, Dubowsky S. Thermal control to maximize photovoltaic powered
operation during the daytime and by the SRO operation over night reverse osmosis desalination systems productivity. Desalination 2013;314:
when sun is unavailable; 2) the operational windows are identified 10e9.
[24] Ben Chaabene A, Sellami A. A novel control of a reverse osmosis desalination
in the case of both the SRO and SSRO operation in the PVROPRO
system powered by photovoltaic generator. In: Electrical engineering and
plant, and the upper limit in each operational window is the op- software applications (ICEESA), 2013 international conference; 2013. p. 1e6.
timum operation; 3) the production rate is significantly increased [25] Bialasiewicz JT. Renewable energy systems with photovoltaic power genera-
by the integration of the salinity power generation by the PRO tors: operation and modeling. Ind Electron IEEE Trans 2008;55:2752e8.
[26] Ipsakis D, Voutetakis S, Seferlis P, Stergiopoulos F, Elmasides C. Power man-
plant. The highest weekly production rate of the PVROPRO plant is agement strategies for a stand-alone power system using renewable energy
almost 20 times the rate in PVRO at the same week. Annual sources and hydrogen storage. Int J Hydrog Energy 2009;34:7081e95.
 W. He et al. / Energy 86 (2015) 423e435 435

[27] Bekele G, Tadesse G. Feasibility study of small Hydro/PV/Wind hybrid system [51] Garcia HE, Mohanty A, Lin W-C, Cherry RS. Dynamic analysis of hybrid energy
for off-grid rural electrification in Ethiopia. Appl Energy 2012;97:5e15. systems under flexible operation and variable renewable generation e Part II:
[28] Cherif H, Belhadj J. Large-scale time evaluation for energy estimation of stand- dynamic cost analysis. Energy 2013;52:17e26.
alone hybrid photovoltaic-wind system feeding a reverse osmosis desalina- [52] Ali Ehyaei MR, Ahmadi M. Meeting the electrical energy needs of a residential
tion unit. Energy 2011;36:6058e67. building with a wind-photovoltaic hybrid system. In: In proceedings of the
 
[29] Novosel T, Cosi c B, Kraja
ci
c G, Dui
c N, Puksec T, Mohsen M, et al. The influence 4th world sustain. forum. Sciforum electronic conference series; 2015.
of reverse osmosis desalination in a combination with pump storage on the [53] Dali M, Belhadj J, Roboam X. Hybrid solarewind system with battery storage
penetration of wind and PV energy: a case study for Jordan. Energy 2014;76: operating in grid-connected and standalone mode: control and energy
73e81. management e experimental investigation. Energy 2010;35:2587e95.
[30] Logan BE, Elimelech M. Membrane-based processes for sustainable power [54] Ibrahim H, Youne s R, Basbous T, Ilinca A, Dimitrova M. Optimization of diesel
generation using water. Nature 2012;488:313e9. engine performances for a hybrid windediesel system with compressed air
[31] He W, Wang Y, Shaheed MH. Energy and thermodynamic analysis of power energy storage. Energy 2011;36:3079e91.
generation using a natural salinity gradient based pressure retarded osmosis [55] Abdullah MO, Yung VC, Anyi M, Othman AK, Hamid KB Ab, Tarawe J. Review
process. Desalination 2014;350:86e94. and comparison study of hybrid diesel/solar/hydro/fuel cell energy schemes
[32] Skilhagen SE. Osmotic power d a new, renewable energy source. Desalin for a rural ICT Telecenter. Energy 2010;35:639e46.
Water Treat 2010;15:271e8. [56] Qi W, Liu J, Christofides PD. A distributed control framework for smart grid
[33] Achilli A, Childress AE. Pressure retarded osmosis: from the vision of Sidney development: energy/water system optimal operation and electric grid inte-
Loeb to the first prototype installation d Review. Desalination 2010;261: gration. J Process Control 2011;21:1504e16.
205e11. [57] Qi W, Liu J, Christofides PD. Supervisory predictive control for long-term
[34] Skilhagen SE, Dugstad JE, Aaberg RJ. Osmotic power d power production scheduling of an integrated wind/solar energy generation and water desali-
based on the osmotic pressure difference between waters with varying salt nation system. Control Syst Technol IEEE Trans 2012;20:504e12.
gradients. Desalination 2008;220:476e82. [58] Zhang S, Chung T-S. Minimizing the instant and accumulative effects of salt
[35] Yip NY, Elimelech M. Thermodynamic and energy efficiency analysis of power permeability to sustain ultrahigh osmotic power density. Environ Sci Technol
generation from natural salinity gradients by pressure retarded osmosis. En- 2013;47:10085e92.
viron Sci Technol 2012;46:5230e9. [59] Zhu A, Christofides PD, Cohen Y. Energy consumption optimization of reverse
[36] He W, Wang Y, Shaheed MH. Enhanced energy generation and membrane osmosis membrane water desalination subject to feed salinity fluctuation. Ind
performance by two-stage pressure retarded osmosis (PRO). Desalination Eng Chem Res 2009;48:9581e9.
2015;359:186e99. [60] Zhu A, Christofides PD, Cohen Y. Effect of thermodynamic restriction on en-
[37] Kim YC, Elimelech M. Potential of osmotic power generation by pressure ergy cost optimization of RO membrane water desalination. Ind Eng Chem Res
retarded osmosis using seawater as feed solution: analysis and experiments. 2008;48:6010e21.
J Membr Sci 2013;429:330e7. [61] Zhu A, Christofides PD, Cohen Y. Minimization of energy consumption for a
[38] Altaee A, Zaragoza Guillermo, Sharif Adel. Pressure retarded osmosis for po- two-pass membrane desalination: effect of energy recovery, membrane
wer generation and seawater desalination: performance analysis. Desalina- rejection and retentate recycling. J Membr Sci 2009;339:126e37.
tion 2014;344. [62] He W, Wang Y, Shaheed MH. Modelling of osmotic energy from natural salt
[39] Sharqawy MH, Zubair SM, Lienhard V JH. Second law analysis of reverse gradients due to pressure retarded osmosis: effects of detrimental factors and
osmosis desalination plants: an alternative design using pressure retarded flow schemes. J Membr Sci 2014;471:247e57.
osmosis. Energy 2011;36:6617e26. [63] Yip NY, Tiraferri A, Phillip WA, Schiffman JD, Hoover LA, Kim YC, et al. Thin-
[40] Feinberg BJ, Ramon GZ, Hoek EMV. Thermodynamic analysis of osmotic en- film composite pressure retarded osmosis membranes for sustainable power
ergy recovery at a reverse osmosis desalination plant. Environ Sci Technol generation from salinity gradients. Environ Sci Technol 2011;45:4360e9.
2013;47:2982e9. [64] Straub AP, Lin S, Elimelech M. Module-scale analysis of pressure retarded
[41] Banchik LD, Sharqawy MH, Lienhard V JH. Effectiveness-mass transfer units osmosis: performance limitations and implications for full-scale operation.
(ε-MTU) model of a reverse osmosis membrane mass exchanger. J Membr Sci Environ Sci Technol 2014;48:12435e44.
2014;458:189e98. [65] Villalva MG, Gazoli JR, Filho ER. Comprehensive approach to modeling and
[42] Sharqawy MH, Banchik LD, Lienhard V JH. Effectivenessemass transfer units simulation of photovoltaic arrays. Power Electron IEEE Trans 2009;24:
(εeMTU) model of an ideal pressure retarded osmosis membrane mass 1198e208.
exchanger. J Membr Sci 2013;445:211e9. [66] Villalva MG, Gazoli JR, Filho ER. Modeling and circuit-based simulation of
[43] Prante JL, Ruskowitz JA, Childress AE, Achilli A. RO-PRO desalination: an in- photovoltaic arrays. In: Power electronics conference, 2009. COBEP '09. Bra-
tegrated low-energy approach to seawater desalination. Appl Energy zilian; 2009. p. 1244e54.
2014;120:104e14. [67] Bartman AR, Zhu A, Christofides PD, Cohen Y. Minimizing energy consump-
[44] Achilli A, Prante JL, Hancock NT, Maxwell EB, Childress AE. Experimental re- tion in reverse osmosis membrane desalination using optimization-based
sults from RO-PRO: a next generation system for low-energy desalination. control. J Process Control 2010;20:1261e9.
Environ Sci Technol 2014;48:6437e43. [68] Zhu A, Rahardianto A, Christofides PD, Cohen Y. Reverse osmosis desalination
[45] He W, Wang Y, Sharif A, Shaheed MH. Thermodynamic analysis of a stand- with high permeability membranes d Cost optimization and research needs.
alone reverse osmosis desalination system powered by pressure retarded Desalin Water Treat 2010;15:256e66.
osmosis. Desalination 2014;352:27e37. [69] Sassi KM, Mujtaba IM. Optimal operation of RO system with daily variation of
[46] Banchik LD, Sharqawy MH, Lienhard V JH. Limits of power production due to freshwater demand and seawater temperature. Comput Chem Eng 2013;59:
finite membrane area in pressure retarded osmosis. J Membr Sci 2014;468: 101e10.
81e9. [70] Achilli A, Cath TY, Childress AE. Power generation with pressure retarded
[47] van der Zwan S, Pothof IWM, Blankert B, Bara JI. Feasibility of osmotic power osmosis: an experimental and theoretical investigation. J Membr Sci
from a hydrodynamic analysis at module and plant scale. J Membr Sci 2009;343:42e52.
2012;389:324e33. [71] Meteonorm Software 7 (http://meteonorm.com/), in, Access in September
[48] Kim J, Park M, Snyder SA, Kim JH. Reverse osmosis (RO) and pressure retarded 2014.
osmosis (PRO) hybrid processes: model-based scenario study. Desalination [72] Song X, Liu Z, Sun DD. Energy recovery from concentrated seawater brine by
2013;322:121e30. thin-film nanofiber composite pressure retarded osmosis membranes with
[49] Feinberg BJ, Ramon GZ, Hoek EMV. Scale-up characteristics of membrane- high power density. Energy Environ Sci 2013;6:1199e210.
based salinity-gradient power production. J Membr Sci 2015;476:311e20. [73] Chen SC, Wan CF, Chung T-S. Enhanced fouling by inorganic and organic
[50] Garcia HE, Mohanty A, Lin W-C, Cherry RS. Dynamic analysis of hybrid energy foulants on pressure retarded osmosis (PRO) hollow fiber membranes under
systems under flexible operation and variable renewable generation e Part I: high pressures. J Membr Sci 2015;479:190e203.
dynamic performance analysis. Energy 2013;52:1e16.