A Hybrid Multi Effect Distillation and Adsorption Cycle 2013

Applied Energy 104 (2013) 810–821

Contents lists available at SciVerse ScienceDirect

Applied Energy

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

A hybrid multi-effect distillation and adsorption cycle

0306-2619/$ - see front matter � 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.apenergy.2012.12.007

⇑ Corresponding author. Tel.: +65 6516 2214; fax: +65 67791459.E-mail address: [email protected] (K.C. Ng).

Kyaw Thu a, Young-Deuk Kim a, Gary Amy a, Won Gee Chun b, Kim Choon Ng c,⇑a Water Desalination and Reuse Center, King Abdullah University of Science and Technology, Thuwal 23955-6900, Saudi Arabiab Department of Nuclear and Energy Engineering, Cheju National University, 66 Jejudaehakno, Jejusi, South Koreac Mechanical Engineering Department, National University of Singapore, 9 Engineering Drive 1, Singapore 117576, Singapore

a r t i c l e i n f o

Article history:Received 23 July 2012Received in revised form 2 December 2012Accepted 3 December 2012Available online 2 January 2013

Keywords:Hybrid desalinationAdsorptionMulti-effect distillationLow-temperature waste heat

a b s t r a c t

This paper describes the development of a simple hybrid desalination system of a Multi-Effect Distillation(MED) and an adsorption (AD) cycle operating at sub-atmospheric pressures and temperatures. Byhybridizing the conventional MED with an AD cycle, there is a symbiotic enhancement of performancesof both cycles. The performance enhancement is attributed to (i) the cascade of adsorbent’s regenerationtemperature and this extended the usage of thermal energy emanating from the brine heater and (ii) thevapor extraction from the last MED stage by AD cycle which provides the effect of lowering saturationtemperatures of all MED stages to the extent of 5 �C, resulting in scavenging of heat leaks into theMED stages from the ambient. The combined effects of the hybrid cycles increase the water productioncapacity of the desalination plant by nearly twofolds.

In this paper, we demonstrate a hybrid cycle by simulating an 8-stage MED cycle which is coupled to anadsorption cycle for direct vapor extraction from the last MED stage. The sorption properties of silica gelis utilized (acting as a mechanical vapor compressor) to reduce the saturation temperatures of MEDstages. The modeling utilizes the adsorption isotherms and kinetics of the adsorbent + adsorbate (sil-ica-gel + water) pair along with the governing equations of mass, energy and concentration. For a 8-stageMED and AD cycles operating at assorted temperatures of 65–90 �C, the results show that the water pro-duction rate increases from 60% to twofolds when compared to the MED alone. The performance ratio(PR) and gain output ratio (GOR) also improve significantly.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Potable water is a necessity for life sustainability but such a re-source, like all other natural resources on Earth, has finite availabil-ity. The accessibility or the balance of supply and demand forpotable water is dependent on factors such as population growth,rapid urbanization, agriculture, climate uncertainties, and techno-logical growth [1]. Furthermore, human’s economic activities andurbanization contaminate the water sources and endangers thesupply of potable or safe water. In many countries, the shortfall be-tween natural water supply and demand can be mitigated by tech-nology-driven desalination processes but such processes areusually energy intensive along with CO2 gas emission and the dis-charge of chemical-laden brine solution. Although the thermody-namic limit for desalting of seawater (35 ppt) is about 0.7 kW h/m3 [2,3], the actual desalting process, whether it is mechanically-or thermally-driven process, would consume two or more foldsof this minimum specific energy consumption. A detailed analysison the energy consumption for the cycle of water supply such as

distribution, end-use, reclamation and disposal of wastewaterhas been reported by Plappally et al. [4].

Commercial desalination methods can be categorized into twomain classes; (i) methods that involved a phase-change mecha-nism and (ii) those processes with liquid separation or simplywithout a phase-change mechanism. All the thermal desalinationsystems fall in the first group while the latter includes mem-brane-based processes such as reverse osmosis (RO). Liquid-saltmolecules separation with membranes needs only to overcomeits osmotic pressures and tend to be more energy efficient as com-pared with the former phase-change methods because it handles afluid of higher density. Recent reports state that seawater reverseosmosis (SWRO) plants with capacities up to 1000 m3/d haveachieved energy consumption below 2.0 kW h/m3 [5,6]. Thermaldesalination systems such as Multi-Stage-Flashing (MSF) and Mul-ti-Effect Distillation (MED) systems are predominant (up to 85% oftotal production capacities) in the Middle East and North Africa(MENA) regions for three reasons: (i) the higher salinity of seawa-ter (up to 45 ppt) in the Gulf and Red Sea lowers the percentage ofwater recovery of membrane processes, (ii) the frequent occur-rence of harmful algae blooms (HABs) have fouled many RO plantsfor weeks to months but has lesser effect to thermally-based MSF

Nomenclature

_m mass flow rateA areaCOP the coefficient of performancecp specific heat capacityDs0 a kinetic constant for the silica gel water systemE activation energy of surface diffusionh enthalpy or heat transfer coefficienthfg latent heatm massn index of the Dubinin–Astakhov isotherm modelP pressureP0 reference pressureq uptake by the adsorbent materialQ the total heat or energyq� the equilibrium uptakeq0 the limiting uptakeQst isosteric heat of adsorptionR gas constantRp average radius of silica gelT temperaturet timeU overall heat transfer coefficient or internal energyv specific volumeV volumeNu Nusselt numberRe Reynolds numberPr Prandtl numberX concentrationq00 heat fluxq densityD diameter

l dynamic viscosityr radiusk thermal conductivityLMTD log mean temperature differencecp,a specific heat capacity of the adsorbed phase

Subscriptsads adsorptiondes desorptionf liquid phase/feedg gaseous phasehw hot waterHX heat exchangerIn inletl liquidOut outletsg silica gelv vaporhw hot watert tubeo outsideb brinee evaporator sidei insideabe adsorbed phases saturationcw cooling waterbh brine heaterbed adsorber/desorber bedcond condenser

K. Thu et al. / Applied Energy 104 (2013) 810–821 811

and MED plants and (iii) The synergy for co-locating power plantsand desalination plants makes cost-effective designs when heavyoil-based fuels are burned.

Although basic thermal desalination systems tend to be energyintensive than RO systems, innovative designs with multiple re-useof latent/condensation heat and cascaded temperature-basedcogeneration concept can be incorporated for efficiency improve-ment. Once the cogeneration concept is introduced, thermal desa-lination systems become more attractive and competitive [7].Thermal desalination systems are coupled with power generationunits where they complement each other since the power genera-tion cycle requires heat sink at cooler temperature to be opera-tional. Thermal desalination system absorbs the rejected heatfrom the upstream systems i.e., the power generating units andproduces potable water. Thus, savings in primary energy consump-tion is realized contrary to RO systems that consumes high-gradeenergy source i.e., electricity which can be utilized for otheractivities.

Among thermal desalination systems, the Multi-Effect Distilla-tion (MED) systems are thermodynamically the most efficient ofall thermal distillation processes [8] but their wide spread usagemay be hindered by corrosion and scaling problems when effortsto raise water production with higher top-brine temperatures(TBTs) [9]. The emergence of advanced tube materials, processtechnology (the implementation of falling film type evaporationdevices) and the use of thermal-vapor compression (TVC) systemshave permitted MED systems to regain considerable market shares[10–14]. The MED-TVC systems operate at sub-atmospheric pres-sures, lowering the fouling and scale problems as well as reducingthe capital and operation costs [15,16].

In MED systems, the performance and cost parameters dependon the ability to design for the number of effects within a giventemperature difference of the TBT to the ambient [17]. Thermo-economic analyses of such systems have been conducted exten-sively using optimization methods and exergy analysis [18–20].Some studies coupled the MED system with an absorption heatpump or refrigeration unit to boost performance of the overall cy-cle [21–26]. El Dessouky [27,28] proposed a MED system to inte-grate with an adsorption vapor compression (ADVC), utilizing thevapor from the last effect of MED for adsorption whilst the regen-erated water vapor from desorption is sent to the first effect of theMED system. It eliminates the condenser of MED system and theperformance ratio improvement is double in this combined cycle[27,28].

In this paper, we propose a simple hybrid desalination systemthat integrates a conventional MED system with an Adsorption(AD) cycle. The AD cycle maintains direct vapor communicationwith the last effect of MED system via the vapor uptake of adsor-bent, establishing a low-pressure environment (hence low satura-tion temperatures) within the MED stages. This implies that thecooling effect from the vapor uptake by the adsorbent of AD cycleis now transferred to all MED stages, lowering their respectivefilm-boiling temperatures of the solution or the feed seawater. Atthe cyclic steady conditions, the last MED stage may reach a tem-perature near to the freezing temperature of solution, rendering anenergy scavenging by heat leaks phenomenon from ambient.Regeneration of the adsorbent, hydrophilic porous silica gel, occursconcomitantly using the hot water that leaves the TBT stage, typi-cally between 55 and 85 �C [29–31], extending the energy extrac-tion from the waste heat sources. The desorbed water vapor is sent

812 K. Thu et al. / Applied Energy 104 (2013) 810–821

via a set of vapor ducts to the condenser where condensate isformed as the distillate: A heat rejection process to the ambientwhich is a necessary process for any thermodynamic cycle. Herethe condensation heat is rejected to the ambient via a coolingtower. A continuous evaporative vapor uptake and condensationof desorbed vapor is achieved by the batch-operation of a multi-bed scheme [32–35].

The sections below present a numerical model for the proposedhybrid AD + MED or in short the ADMED cycle. The governingequations of heat and mass balances are computed along withthe sorption models such as the adsorption isotherms and kineticsof the silica gel–water pair for specific temperature and pressureconditions. The predicted results of the combined cycle are pre-sented in terms of key performance parameters found in any desa-lination plants, namely, the Water Production Rate (WPR), thePerformance Ratio (PR) and the Gain Output Ratio (GOR).

2. Description of the Hybrid ADMED cycle

The schematic diagram of the proposed hybrid ADMED cycle isshown in Fig. 1. It consists of two sub-systems namely (i) the Multi-Effect Distillation (MED) system and (ii) the Adsorption Desalina-tion (AD) cycle. The vapor uptake from the AD cycle is maintained

Fig. 1. The schematic diagram o

by the sorption properties of the adsorbent where the direct vaporcommunication is controlled by a set of vapor valves located on thevapor ducts connecting the adsorber beds with the evaporator andthe condenser. Water vapor in a MED stage emanates from thefilm-boiling of seawater which is being sprayed onto the externalsurfaces of heat exchanger tubes. The heat source for film-boilingis derived from the condensation of the vapor that is generatedin the preceding stage or effect where the saturation pressureand temperature are higher. Although there are several operationmodes of a MED cycle, we consider only a backward-feed typeMED cycle for the integration exercise.

In the proposed cycle, the de-aerated seawater is supplied tothe last stage of the ADMED system. The saturation condition ofthe stage is maintained two factors: (i) the condensation energyof water vapor emanating from the preceding effect of the MEDsystem and (ii) the rate of vapor uptake by the adsorbent in theAD cycle. Being backward-feed, the saline feed flows to the nextstage of higher temperature and pressure via a pressure balancingU-tube, and the process for the solution feed is repeated until theconcentrated brine is rejected from the TBT stage. The vapor, onthe other hand, flows in opposite direction to the saline feed. Vaporis generated in the TBT or brine heater at temperatures of 65–90 �Cdepending on the heat source temperature. The produced vapormigrates to the inner tube surfaces in the next MED stage, and

f the hybrid ADMED cycle.

K. Thu et al. / Applied Energy 104 (2013) 810–821 813

the latent heat of condensation is deposited onto the tubes. The la-tent heat is re-utilized in multiple times for vapor generationacross every MED stage of lower saturation pressures and temper-atures. Thus, the feed flows downwards whereas the vapor travelsupwards and condenses at every MED stage in a backward-feedconfiguration. Such a configuration has the simplicity of operationwhere the internal heat recovery from the vapor stream to the feedis guaranteed. Thus, the numerical simulation code is simpler inoperation.

The Adsorption Desalination (AD) cycle comprises three majorcomponents: namely (i) the pseudo ‘‘evaporator’’, (ii) the adsorberor reactor beds and (iii) the condenser. The pseudo evaporator ofAD cycle is deemed as an additional MED effect resulted fromthe hybridization. In a conventional MED cycle, its condensationtemperature at the last stage is subjected to the ambient or seawa-ter temperature, typically around 35–40 �C, and consequently,there exists limitation in the number of MED stages when theTBT and the temperature difference across adjacent effects are lim-ited. With a hybrid design, there are more flexibility in design andoperation. The combined ADMED cycle lowers the last stage satu-ration temperatures to below the ambient, allowing one to incor-porate more MED stages at the same TBT or lower the brineheater temperature to reduce fouling and scaling problemsimproving water production rate and the performance.

The adsorber/desorber beds in AD cycle are simply heatexchangers where the adsorbent material (silica gel) is packedaround the finned-tubes inside which heating or cooling fluidflows, alternatively depending on desorption or adsorption cycle.AD cycles are inherently batched-type cycles since the adsorbentcannot perform both adsorption and desorption (regeneration)simultaneously. Thus, AD cycles utilize at least two or multi-bedconfiguration for continuous production of useful effects. Duringadsorption process, the evaporator is in communication with theadsorber bed and external cooling water circuit is used to maintainthe adsorption process by rejecting the heat of adsorption. Duringdesorption process, the saturated adsorber is regenerated using hotwater typically between 85 �C and 60 �C. Thus, in AD cycles, one ora pair of adsorber beds performs adsorption process whilst anotherpair undergoes desorption in the first half-cycle time. The beds ex-change their roles in the next half cycle.

The transition from adsorption to desorption or vice versa canbe readily achieved with valve opening and closing arrangements;Prior to interchanging their roles, the adsorber bed needs to be pre-heated whereas the desorber bed requires precooling to attain thecondensation and evaporation pressures, respectively. In the pres-ent system, four physical adsorber beds are employed where a pairof two adsorber beds perform sorption process operating as a 2-bed AD cycle. This configuration is adopted as the mathematicalmodel and simulation results have been rigorously validated withexperimental data at various operating conditions [36–39]. The re-jected water vapor from the desorber beds is condensed on thetube surfaces of the water-cooled condenser where the condensa-tion heat is rejected to the ambient through a cooling tower. Prac-tically, seawater can be used for the condensation where theenergy is recovered to the preheating of the feed. However, asthe objective of this work is to demonstrate the hybridization ofthe MED system with AD cycle, the simplified operational configu-ration is implemented.

The present hybrid desalination cycle is configured in such away that the driving heat source for the MED and AD cycle canbe independent or the outlet hot water from the MED cycle canbe channeled to operate the AD cycle with some valve opening/closing arrangements. It should be noted that the AD cycle at-tached to the MED system can be configured in such a way thatthe operation of these two systems can be fully separated. In suchscenario, the AD cycle can perform as dual-purpose (cooling and

desalination) plant [40,41] whilst the MED system continues asdesalinating unit. This operating condition can easily be achievedwith slight alteration in the valve sequence.

3. Mathematical modeling and simulation

The hybridized ADMED cycle consists of two sub-systemsnamely (i) the Multi-Effect Distillation (MED) system and (ii)adsorption desalination (AD) cycle. The mathematical models fora backward-feed counter flow MED cycle and adsorption desalina-tion cycle are developed based on the mass, energy and concentra-tion balances along with the sorption model for silica gel–waterpair. The general assumptions in the present models are: (1) thehomogenous temperature in the heat exchangers and adsorbentmaterial, (2) complete condensation of water vapor inside the ef-fects and the condenser, (3) heat loss or gain from the ambient isnegligible and (4) ideal de-aeration with no non-condensablegases. It is noted that only cyclic-steady state conditions can beachieved in the AD cycle and thus the transient modeling of theMED system is developed to investigate the performance of theADMED cycle.

3.1. Multi-Effect Distillation (MED)

A numerical model is developed using mass, concentration andenergy balances of the major components in the MED system witha backward-feed configuration. Seawater is supplied to the last ef-fect of the MED system and it gets preheated successively by thecondensation of the water vapor while flowing through the inter-mediate effects.

3.1.1. Brine heater or vapor generatorThe vapor generator or brine heater produces water vapor using

hot water as heat source. It is a simple liquid to vapor shell andtube heat exchanger with the liquid in the tube side. The energybalance of the hot water that circulates inside the brine heater isgiven as,

qcpVdThw

dt¼ _mðhf ðThw;InÞ � hf ðThw;OutÞÞ � hi Ai ðThw � TtÞ ð1Þ

Thermal energy for the evaporation of the seawater is trans-ferred through metal heat exchanger tubes and the energy balancefor the metal tubes is written as,

qcp VdTt

dt¼ hi Ai ðThw � TtÞ � hoAoðTt � ToÞ ð2Þ

The heat transfer coefficient for hot water inside the tubes iscalculated using Dittus–Boelter’s correlation as [42],

Nu ¼ 0:023Re0:8 Pr0:4 ð3Þ

The mass balance for the seawater inventory in the evaporatorside of the brine heater is given as,

dme

dt¼ _mf � _mb � _mv ð4Þ

Here me is the mass of seawater in the evaporator side, _mf is thefeed water flow rate, _mb and _mv are the brine and vapor flow ratesleaving the heat exchanger. The falling film type evaporation isadopted in the brine heater while the seawater concentration is cal-culated as,

dðmeXeÞdt

¼ _mf Xf � _mbXe � _mvXv ð5Þ

where X is the total dissolved solids (TDSs) in parts per million(ppm) and the TDS of the distillate is taken as 10 ppm. The energybalance for the evaporator side can be written as,

814 K. Thu et al. / Applied Energy 104 (2013) 810–821

dðmeueÞdt

¼ _mf hf ðTf Þzfflfflfflfflffl}|fflfflfflfflffl{Feed

� _mbhf ðTbÞzfflfflfflfflfflffl}|fflfflfflfflfflffl{Brine discharge

� _mvhfgðTvÞzfflfflfflfflfflfflffl}|fflfflfflfflfflfflffl{Vapor

þhoAoðTt � ToÞzfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflffl{Heat transfer

ð6Þ

where Tv is the temperature of the vapor inside the evaporatorwhich can be calculated by subtracting boiling point elevation(BPE) at specific concentration of the saline water from the evapo-rator temperature. The present design adopts the falling film evap-oration of seawater over a bundle of horizontal tubes. The Han andFletcher’s correlation [43] is applied for the boiling of seawater overcircumferentially groove tubes and it is given as,

hom2

k3 g

!1=3

¼ 0:0007Re0:2 Pr0:65q000:4 ð7Þ

3.1.2. Intermediate effectsThe vapor generated from the brine heater travels upward to

the tube side of the first intermediate effect where the latent heatof condensation is recovered to evaporate the seawater. The gener-ated vapor is reutilized for the evaporation of the seawater in thesuccessive effects. The intermediate effects are shell and tube heatexchangers for vapor to vapor heat transfer with horizontal tubearrangement. The energy balance for the condenser side of each ef-fect assuming complete vapor condensation is given by,

qcp VdTcond

dt

� �j

¼ ½ _mv hfgðTvÞ�j�1 � ½hiAiðTcond � TtÞ�j ð8Þ

The evaporation of the water vapor in the intermediate effect ismaintained by the energy released from the condensation of watervapor from the previous effect. Classical Nusselt film condensationcorrelation is applied to calculate the heat transfer coefficient forthe condensation of the water vapor inside the condenser tubesand it is given as [44],

hi ¼ 0:728ql g k3ðql � qvÞ ½hfg þ 3

8 cplðTs � TtÞ�Dl ðTh � TtÞ

" #0:25

ð9Þ

The energy balance for the metal tubes is similarly expressed asEq. (2) using the condensation and evaporation energy. The massand salt balance equations for the seawater inventory in the evap-orator side of each effect are given as,

dme

dt

� �j

¼ _mf ;j � _mb;j � _mv;j ð10Þ

Table 1Thermophysical properties and adsorption parametersof Type A++ silica gel.

Parameter Value

Pore size (nm) 0.8–7.5Porous volume (cm3/g) 0.476Surface area (m2/g) 863.6Average pore diameter (nm) 2.2Apparent density (kg/m3) 700pH 4.0Specific heat capacity (kJ/kg K) 0.921Thermal conductivity (W/m K) 0.198q0 (kg/kg) 0.5E (kJ/mole) 3.105n (–) 1.1Ds0 (m2/s) 2.54 � 10�4

Ea (kJ/kg) 4.2 � 10�4

Rp (mm) 0.4

dðmeXeÞjdt

¼ ð _mf Xf Þj � ð _mbXeÞj � ð _mvXvÞj ð11Þ

Similarly, the energy balance for the evaporator side the effect iswritten as,

dðmeueÞjdt

¼ ½ _mf hf ðTf Þ�zfflfflfflfflfflfflffl}|fflfflfflfflfflfflffl{Feed

j � ½ _mbhf ðTbÞ�zfflfflfflfflfflfflffl}|fflfflfflfflfflfflffl{Brine discharge

j�½ _mvhfgðTvÞ�zfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflffl{Vapor

j

þ ½hoAoðTt � ToÞ�zfflfflfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflfflffl{Heat transfer

j ð12Þ

Finally, the overall heat transfer coefficient Ui is calculated as,

Ui ¼1hiþ

ri l n rori

kþ Ai

Ao

1ho

" #�1

ð13Þ

Thermodynamic properties of seawater such as specific heatcapacity, enthalpy and boiling point elevation are accounted for[45].

3.2. Adsorption desalination cycle

The adsorption cycle is coupled to the last effect of the MED sys-tem replacing the down condenser. The water vapor generated inthe last effect is adsorbed by adsorbent materials inside the reactorbeds of the AD cycle. The water vapor from the saturated adsorberbed is then regenerated using low-temperature hot water and it iscondensed using a water-cooled condenser. The performance orsorption characteristic of the adsorbent is governed by the equilib-rium uptake and kinetic behavior of the silica gel–water pair.Water vapor uptake by the silica gel is experimentally evaluatedat varied adsorption conditions and these data are fitted using D-A isotherm model i.e. [46],

q� ¼ q0 � exp � RTE� ln

P0

P

� �� �n� �ð14Þ

where q⁄ and q0 are the equilibrium and limiting uptake by theadsorbent, E is the activation energy and R is gas constant while p

p0

is the relative pressure and n is a D-A constant. The transient uptakeat specific temperature and pressure condition is given by linerdriving force equation as,

dqdt¼ 15Ds0e�

ERT

R2p

ðq� � qÞ ð15Þ

where Ds0 denotes a pre-exponential factor of the efficient waterdiffusivity in the adsorbent, and Rp is the average radius of theadsorbent grains. Kinetic data were taken from [47,48]. As migra-tion of water vapor from the last effect to the pore surfaces of theadsorbent, kinetic energy is released in terms of heat of adsorption.Thus, adsorption process is an exothermic process and externalcooling is required to maintain the sorption process. The heat rejec-tion process involved in the adsorption is denoted as isosteric heatof adsorption and is calculated as,

Qst ¼ hfg � E� � lnqq0

� �� �1=n

þ T � vg �@P@T

� �g

ð16Þ

where hfg is latent heat and vg is the specific volume of the gaseousphase [49]. The thermophysical properties and adsorption parame-ters of silica gel associated with water vapor sorption in the AD cy-cle are listed in Table 1. Adsorption (AD) cycles are inherentlybatched-operated since adsorption and regeneration processes can-not be performed simultaneously by a single adsorber bed. Thus, atypical AD cycle consists of one or a pair of adsorber beds that per-forms adsorption process and another pair executing desorption

Table 2Parameters used in the simulation program.

Heating, cooling and feed input dataHot water inlet temperature (�C) 90–65Seawater inlet temperature (�C) 30Number of intermediate effects 6Brine flow rate from the last effect (kg/s) 0.1Concentration of the feed seawater (mg/L) 35,000Brine heater input dataHot water flow rate (kg/s) 7.14Area of brine heater (m2) 65.79Tube length (m) 2Number of tubes 128Diameter of tubes (mm) 16Intermediate effect input dataArea of intermediate effects (m2) 33.4Length (m) 2Number of tubes 100

K. Thu et al. / Applied Energy 104 (2013) 810–821 815

process. This multi-bed arrangement provides continuous produc-tion of useful effect such as potable water or cooling energy. The ad-sorber bed performing adsorption of water vapor from the lasteffect of the MED cycle is mathematically described as,

½msg cp;sg þmHX cp;HX þmabe cp;a ðTadsÞ�dTads

dt

¼ msgdqads

dtQ stðTads; PadsÞ � _mcw;bed cpðTcwÞðTcw;bed;Out

� Tcw;bed;InÞ ð17Þ

The first term in the left hand side of Eq. (17) represents thethermal mass by the adsorbent materials, the second term standsfor that by the heat exchanger assembly whilst the last term de-notes the thermal mass contributed by the adsorbed phase. Thespecific heat capacity of the adsorbed phase is calculated as,

cp;a ¼ cp;g þ1T� 1

vg

@vg

@T

� �P

� �Q st�

@Q st

@T

P

ð18Þ

here cp,g and vg are the specific heat and specific volume of the gas-eous phase [50]. Similarly, the transient temperature of the desorb-er bed that is in communication with the condenser is given as,

½msg cp;sg þmHX cp;HX þmabe cp;a ðTdesÞ �dTdes

dt

¼ �msgdqdes

dtQ stðTdes; PdesÞ þ _mhw;bed cpðThwÞ ðThw;In � Thw;OutÞ

ð19Þ

The temperature of the outlet water from the adsorber beds isestimated using log mean temperature difference method for heatexchanger as,

TOut ¼ T0 þ ðTin � T0Þ exp�UA

_mcpðT0Þ

� �ð20Þ

where T0ð¼ Tads=TdesÞ is the temperature of the heat exchanger. Therejected water vapor from the desorber bed is condensed on thesurfaces of the water-cooled condenser and the condensate is col-lected as potable water. The temperature of the condenser is thusgiven as,

½mw;cond cpðTcondÞ þmHX cp;HX �dTcond

dt

¼ �hf ðTcondÞdmd

dtþ msg hgðTcondÞ

dqdes

dt� _mcw;cond cpðTcondÞ ðTcw;cond;Out � Tcw;cond;InÞ ð21Þ

The evaporator of the AD cycle or pseudo ‘‘evaporator’’ of theMED system is connected to the adsorber beds via vapor ductsand the energy balance for the evaporator side of this device is gi-ven as,

dð _meueÞdt

¼ _mf hf ðTf Þzfflfflfflfflffl}|fflfflfflfflffl{Feed

� _mbhf ðTbÞzfflfflfflfflfflffl}|fflfflfflfflfflffl{Brine discharge

� _msghgðTvÞdqads

dt

zfflfflfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflfflffl{Vapor to adsorber

þhoAoðTt � ToÞzfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflffl{Heat transfer

ð22Þ

3.3. Performance modeling

In ADMED cycle, thermal energy is supplied to the brine heateror vapor generator and the desorber beds whilst hot water is usedas heating medium. Thermal energy is rejected at the adsorber bedduring vapor uptake process and the condenser to the ambient. Inthe proposed hybrid desalination system, the input thermal energyto the brine heater is rejected at the adsorber bed via the adsorp-tion process in contrast to a conventional MED cycle. The energyinput to the brine heater can be calculated using the inlet and out-let temperatures of the hot water as,

Qbh ¼ _mhw;bh cpðTbhÞðThw;bh;In � Thw;bh;OutÞ ð23Þ

In adsorption cycle, pellet-type silica gels are packed around theheat exchanger tubes of the adsorber bed and isosteric heat ofadsorption is supplied/rejected by the hot water/the circulatingcooling water. Thus, the desorption and adsorption heats are calcu-lated using inlet and outlet temperatures of the hot and coolingwater streams as,

Qdes ¼ _mhw;bed � cpðTdesÞ � ðThw;bed;In � Thw;bed;OutÞ ð24Þ

Qads ¼ _mcw;bed � cpðTadsÞ � ðTcw;bed;Out � Tcw;bed;InÞ ð25Þ

Similarly, heat rejection at the condenser is given by,

Qcond ¼ _mcw;cond � cpðTcondÞ � ðTcw;cond;Out � Tcw;cond;InÞ ð26Þ

The amount of water vapor produced in each effect ð _mw;iÞ andthat in the condenser are obtained as,

_mw;i ¼Ui AiLMTDi

hfgðTiÞð27Þ

_mw;cond ¼U � A� LMTD

hfgðTcondÞð28Þ

Finally, the overall water production rate (WPR) and the perfor-mance ratio (PR) by the plant are given as,

WPR ¼XN

i¼1

_mw;i þ _mw;cond ð29Þ

PR ¼PN

i¼1 _mw;i hfgðTiÞ þ _mw;cond hfgðTcondÞ_mhw;bh cpðThw;bhÞ ðThw;br;In � Thw;bh;OutÞ þ Q des

ð30Þ

The values of the parameters used in the simulations are sum-marized in Table 2. These sets of equations are simultaneouslysolved using IVPAG routines of the IMSL library functions calledfrom the simulation code written in FORTRAN.

Fig. 2. Temporal temperature profiles of the MED system: (heat source at 90 �C and seawater inlet temperature at 30 �C).

Fig. 3. Temporal water production rate of the MED system: (heat source at 90 �Cand seawater inlet temperature at 30 �C).

816 K. Thu et al. / Applied Energy 104 (2013) 810–821

4. Results and discussion

Three configurations of the desalination systems namely: (i) aconventional MED system with 8 effects (Config.#1), (ii) a hybridADMED cycle where the heat source temperature to the AD cycleis maintained at 85 �C (Config.#2) and a hybrid ADMED cyclewhere the AD cycle is operated using the outlet hot water fromthe brine heater of the MED system (Config.#3). Config.#2 is appli-cable to hybridize steam-driven MED systems with AD cycle thatcan be operational using separate low-temperature waste heatsources. Thus, the two heat sources in this configuration are inde-pendent. However, Config.#3 optimizes the energy extraction fromthe heat source by further utilizing the thermal energy of the outlethot water from the brine heater of the MED plant to drive the ADcycle. In this setup, the heat source to the AD cycle is dependentto the MED system and hence the performance of the attachedAD cycle varies with heat source temperature to the MED system.In the present work, a backward-feed type MED system is adoptedsince such configuration offers implicit heat recovery between thevapor and the feed seawater as it flows from the low temperatureeffects to the brine heater, successively and yet the efficacy of theproposed hybridization with adsorption (AD) cycle can be demon-strated. The performances of such desalination systems are ana-lyzed for various heat source temperatures ranging from 65 �C to90 �C in terms of key performance indicators such as Water Pro-duction Rate (WPR), Gain Output Ratio (GOR) and Performance Ra-tio (PR). The temporal temperature of a conventional MED system(Config.#1) with eight effects is shown in Fig. 2 where the heatsource (hot water) and seawater inlet temperatures are maintainedat 90 �C and 30 �C, respectively. The system approaches the steadystate conditions about 45 min after commencing the operation.From Fig. 2, substantial drop in the temperature (25 �C) betweenthe brine heater and the Effect#1 is observed. This is due to thenature of a typical backward-feed MED system where the largeamount of seawater inventory is supplied at the lower temperatureeffects resulting huge quenching of the vapor. Similar scenario isreported by Darwish et al. [51,52].

Transient water production rate by each effect of Config.#1 isgiven in Fig. 3 where it is observed that the water production rate

decreases asymptotically as effect temperature dropped. This phe-nomenon occurs in a backward feed MED system with no flushingtype evaporation of seawater since the feed travels from low pres-sure to the higher pressure effects [51,52].

Fig. 4 depicts the concentration of the sea water in each effect.The total water production rate (WPR) from Config.#1 is found tobe 0.221 kg/s at steady state conditions whilst the Gain Output Ra-tio (GOR) and the Performance Ratio (PR) are 3.38 and 4.31,respectively.

The temperature profiles of the condenser side, metal tube andevaporator side of each effect in the hybrid ADMED cycle for Con-fig.#2 are given in Fig. 5. It is noted that the heat source tempera-ture for the AD cycle in this configuration is fixed at 85 �C whilethat for the brine heater remains the same as Config.#1 i.e.,90 �C. The AD cycle is attached to the MED cycle with eight effectswhere the vapor from the last effect is condensed inside the tube-side of the AD evaporator. Thus, the evaporator of the AD cycle

Fig. 4. Temporal concentration, GOR, PR and WPR of the MED system: (heat source at 90 �C and seawater inlet temperature at 30 �C).

Fig. 5. Temporal temperature profile of the ADMED plant.

K. Thu et al. / Applied Energy 104 (2013) 810–821 817

performed as an additional effect (9th effect). The results show thatmost of the effects operate at temperature levels lower than ambi-ent (30 �C) due to the contribution by the vapor uptake of the ADcycle. The adsorbent materials in the AD cycle adsorb water vaporfrom the last effect resulting in the temperature drop due to themass transfer effect. Hence the condensation temperature of thevapor from the last effect decreases and this cold front is propa-gated to the higher-temperature effects. The AD cycles are bat-ched-type that switches vapor adsorption and desorptionprocesses at specific cycle time interval with a switching periodin between. This kind of operation nature results in the tempera-ture fluctuation i.e., a sharp drop in the evaporator temperatureat the commencement of the half-cycle time (adsorption) as dryor dehydrated adsorbent draws up water vapor rapidly. However,the adsorption amount becomes diminished as the adsorbent getsaturated and the evaporator temperature rises slightly at laterperiod of the half-cycle time. It is observed that the temperaturefluctuation is detected significantly in the last three effects whilstthis effect propagates mildly until the fifth effect. The delay inthe temperature wave contributed by the respective thermal massis detected as well.

Water production rate by the ADMED cycle with Config.#2 ispresented in Fig. 6. It should be noted that the water productionrate by the condenser reflects the batched operation nature ofthe AD cycle. The cycle average water production rate by the con-denser is higher than those by most of the effects of the MED sys-tem. This is because the water production rate in AD cycle iscontrolled by the uptake properties and the amount (inventory)of the adsorbent. The water production rate reflects the degree ofsaturation of the adsorbent. Again, the effect of the AD cycle is de-tected in the water production rate that propagates until last threeeffects of the MED system.

The concentration of brine in each effect and the performance(GOR, PR, WPR) of the ADMED cycle is given in Fig. 7 where the var-iation in the brine concentration due to the fluctuation of waterproduction by the AD cycle is observed. It is noted that the perfor-mance of the combined cycle is improved as compared to the con-ventional MED system where the WPR increases from 0.222 to0.385 kg/s whilst the PR is about 6.06. The improvement in waterproduction rate by the Hybrid cycle with Config.#2 is found to be74%.

The temperature and pressure profiles of the AD cycle with theConfig.#2 are given in Figs. 8 and 9. It is noted that the AD cycle inthis configuration employs four physical adsorber beds operatingas a 2-bed machine where a pair of two adsorber beds performsadsorption and another pair as desorber beds. This type of AD cycleconfiguration is selected due to the adaptation of numerical mod-eling and simulation for the AD cycle that has been validated withexperimental data [36–39].

The comparison on the effect temperatures of three types ofconfigurations (Config.#1, Config.#2 and Config.#3) at steady stateconditions are shown in Fig. 10 where the heat source temperaturefor the MED system is fixed at 90 �C and that for the AD cycle ismaintained at 85 �C in the Config.#2. However, the AD cycle inConfig.#3 is driven by the outlet hot water from the MED system.

The results highlight that the evaporation and condensation offeed and distillate in both Config.#2 and Config.#3 occur at signif-icantly lower temperature as compared to the conventional MEDsystem. Due to the constant heat source i.e., 85 �C for the AD cyclein the Config.#2, the operating temperatures in the effects areslightly lower than that in Config.#3 since sorption processes aremuch efficient at higher desorption temperatures. However,ADMED cycle with Config.#3 is applicable in practical aspects

Fig. 6. Water production rates by each effect of the ADMED cycle.

Fig. 7. Brine concentration and the performance of the ADMED plant.

818 K. Thu et al. / Applied Energy 104 (2013) 810–821

where the further utilization of the outlet hot water or condensedsteam from the brine heater of the MED system can be realized. Itis observed that the condensation temperature in both hybrid cy-cles is relatively higher than that in the conventional MED system.This is because the amount of energy that is adsorbed from theevaporator of the AD is required to reject at the condenser in addi-tion to that is supplied at the brine heater.

The water production rates by each effect in the aforemen-tioned cycles are given in Fig. 11. As a result of hybridization withthe AD cycle, the cycle provides higher water production rate ineach effect. However, the increment in water production issignificantly attributed to the condenser of the AD cycle wherewater vapor uptake is responsible to the sorption phenomena by

the adsorbent and yet it essentially provides cooler mediafor the improved condensation of the water vapor from the MEDsystem.

The performance of the desalination cycle with aforesaidconfigurations is analyzed at assorted heat source temperatures(from 90 �C to 65 �C). The water production rates by each effectare summarized in Table 3. As a general trend, larger waterproduction rate is obtained at higher heat source temperatures.This situation is applicable to all the effects and the condensersexcept the condenser of the Config.#2 where the water produc-tion rate does not vary significantly. This is because the heatsource temperature for the AD cycle in this configuration isfixed at 85 �C translating that the performance of the AD cycle

Fig. 8. The temporal temperature profiles of the major components of the AD cyclehybridized with a conventional MED system.

Fig. 9. The temporal pressure profiles of the major components of the AD cyclehybridized with a conventional MED system.

Fig. 10. The comparison of the temperatures of the convent

Fig. 11. The comparison on the water production rates by each effect of MED andADMED cycle.

K. Thu et al. / Applied Energy 104 (2013) 810–821 819

is largely dependent to the regeneration temperature that re-stores the chemical potential of the adsorbent to perform vaporuptake in next adsorption cycle.

The performance comparison (GOR, PR, WPR) on the MED andADMED cycles are listed in Table 4 for assorted hot water inlettemperatures. It is observed that the water production rate bythe Config.#2 is from 74% to twofolds as compared to that by theconventional MED system whilst the improvement by Config.#3is about 60–66%. As the heat source temperature to the MED sys-tem is increased, the percentage improvement in terms of waterproduction rate by Config.#2 decreases while that by Config.#3 lin-early increases. At lower heat source temperatures (at 65 �C), theeffect of the AD cycle with constant heat source temperature(Config.#2) is significant and the water production rate by AD cycleis comparable to that of its counterpart i.e., MED system. As theheat source temperature for the MED system increases, itswater production rate increases linearly as well while that by ADcycle remains more or less the same due to the unchangedregeneration temperature. However, the temperature of the heat

ional MED system and the hybrid cycle for each effect.

Table 3Water production rate by each effect from desalination system with different configurations.

90 �C 85 �C 80 �C

Config.#1 Config.#2 Config.#3 Config.#1 Config.#2 Config.#3 Config.#1 Config.#2 Config.#3

(A)Effect#1 0.0532 0.0672 0.0654 0.0494 0.0639 0.0611 0.0456 0.0605 0.0566Effect#2 0.0393 0.0536 0.0520 0.0367 0.0510 0.0486 0.0340 0.0484 0.0451Effect#3 0.0303 0.0431 0.0419 0.0285 0.0412 0.0393 0.0267 0.0393 0.0366Effect#4 0.0241 0.0354 0.0344 0.0229 0.0340 0.0324 0.0216 0.0325 0.0303Effect#5 0.0198 0.0296 0.0288 0.0189 0.0285 0.0272 0.0179 0.0274 0.0255Effect#6 0.0167 0.0252 0.0245 0.0160 0.0243 0.0232 0.0153 0.0235 0.0219Effect#7 0.0144 0.0218 0.0212 0.0139 0.0211 0.0202 0.0133 0.0204 0.0191Effect#8 0.0127 0.0189 0.0184 0.0123 0.0184 0.0176 0.0118 0.0178 0.0168Effect_AD 0.0155 0.0152 0.0150 0.0146 0.0146 0.0140Condenser 0.0114 0.0731 0.0662 0.0110 0.0726 0.0620 0.0107 0.0721 0.0575

75 �C 70 �C 65 �C

(B)Effect#1 0.0416 0.0571 0.0520 0.0377 0.0537 0.0472 0.0336 0.0501 0.0422Effect#2 0.0314 0.0458 0.0415 0.0286 0.0432 0.0378 0.0258 0.0404 0.0340Effect#3 0.0248 0.0373 0.0339 0.0229 0.0353 0.0310 0.0208 0.0333 0.0281Effect#4 0.0202 0.0310 0.0281 0.0188 0.0295 0.0259 0.0173 0.0279 0.0237Effect#5 0.0169 0.0262 0.0239 0.0159 0.0250 0.0221 0.0148 0.0238 0.0203Effect#6 0.0145 0.0226 0.0206 0.0137 0.0216 0.0192 0.0129 0.0207 0.0178Effect#7 0.0127 0.0197 0.0181 0.0121 0.0189 0.0170 0.0115 0.0182 0.0158Effect#8 0.0114 0.0173 0.0160 0.0109 0.0167 0.0151 0.0104 0.0161 0.0142Effect_AD 0.0142 0.0134 0.0137 0.0128 0.0132 0.0122Condenser 0.0103 0.0716 0.0527 0.0100 0.0711 0.0477 0.0096 0.0704 0.0424

Table 4The performance comparison of the MED and ADMED plants at different heat source temperatures.

T (�C) GOR PR WPR Improvement in WPR (%)

Config.#1 Config.#2 Config.#3 Config.#1 Config.#2 Config.#3 Config.#1 Config.#2 Config.#3 Config.#2 Config.#3

65 3.95 5.30 4.94 4.75 6.56 6.14 0.157 0.316 0.251 101.8 60.370 3.79 5.16 4.85 4.63 6.43 6.08 0.171 0.330 0.276 93.8 61.975 3.67 5.03 4.78 4.53 6.31 6.03 0.184 0.344 0.300 87.3 63.380 3.56 4.92 4.72 4.44 6.22 5.99 0.197 0.358 0.324 82.0 64.585 3.46 4.82 4.67 4.37 6.13 5.96 0.209 0.372 0.346 77.4 65.490 3.38 4.73 4.63 4.31 6.06 5.94 0.222 0.385 0.368 73.6 66.0

Table 5The energy balances of the MED and ADMED cycle for assorted heat source temperatures.

T (�C) Config.#1 Config.#2 Config.#3

QIn (kW) QOut (kW) Energy balance (%) QIn (kW) QOut (kW) Energy balance (%) QIn (kW) QOut (kW) Energy balance (%)

65 387.9 385.8 0.53 868.7 859.1 1.11 656.1 648.5 1.1770 414.8 412.3 0.61 896.8 886.7 1.13 717.9 709.4 1.1975 441.8 438.8 0.68 925.0 914.2 1.16 778.7 769.3 1.2180 468.9 465.4 0.74 953.1 941.9 1.18 838.6 828.3 1.2385 496.0 492.0 0.81 981.3 969.5 1.21 897.5 886.4 1.2490 523.1 518.6 0.86 1009.6 997.2 1.23 955.6 943.6 1.25

Config.#1: Conventional MED cycle.Config.#2: ADMED cycle with Thot = 85 �C to AD cycle.Config.#3: ADMED with outlet of MED cycle is supplied to AD cycle.

820 K. Thu et al. / Applied Energy 104 (2013) 810–821

input to the AD cycle in Config.#3 is dependent to that of the MEDsystem and the system water production rate reflects the heatsource temperature and hence the percentage increment in WPRin this configuration improves with the increase in the heat sourcetemperature.

Finally, the energy balance statistics of the systems are eval-uated to counter check the numerical simulation. The analysesfor all the heat source temperatures are conducted and the val-ues are listed in Table 5. It is observed that the system energybalance for all the configurations is well below the acceptablepercentage with 1.25% to be the highest whilst the lowest fig-ure is 0.53%.

5. Conclusion

A simple hybrid ADMED desalination system has been success-fully modeled and analyzed for three configurations, i.e., a base-line 8-stage MED, an ADMED with constant temperature heatsource and a temperature-cascaded ADMED. Owing to the directvapor communication between the systems and water vapor up-take properties of the adsorbent, the operating temperature ofthe combined cycle is brought down to lower temperatures witheffects operating at below ambient temperature. The evaporationof saline water at low temperature reduces scaling and fouling.The results show that the water production rate of the ADMED

K. Thu et al. / Applied Energy 104 (2013) 810–821 821

cycle is significantly increased up to twofolds as compared to aconventional MED system whilst the PR and GOR improve by40%. The quantum increase in water production rate and loweringthe evaporation temperature of the seawater in the effects couldcontribute to lowering the desalting cost with reduced operatingcost due to less fouling and corrosion. However, the additional cap-ital investment is required to implement the hybridization to thecurrent MED systems. It is expected the proposed system rejuve-nates the existing MED plants with higher water production ratewith simple modification. The bonus by the hybridization is theproduction of cooling power (chilled water temperature between5 and 20 �C) from the AD system with slight operation alterationof the valve control scheme.

Acknowledgements

The authors gratefully acknowledge the financial supportgiven by Grant (No. R33-2009-000-101660) from the World ClassUniversity (WCU) Project of the National Research Foundationand the generous research grant from Office of CompetitiveResearch Funding (OCRF) of KAUST under the account No.7000000411, 2012.

References

[1] Rogers P, Silva Rd, Bhatia R. Water is an economic good: how to use prices topromote equity, efficiency, and sustainability. Water Policy 2002;4:1–17.

[2] Spiegler KS, El-Sayed YM. The energetics of desalination processes.Desalination 2001;134:109–28.

[3] Elimelech M, Phillip WA. The future of seawater desalination: energy,technology, and the environment. Science 2011;333:712–7.

[4] Plappally AK, Lienhard V JH. Energy requirements for water production,treatment, end use, reclamation, and disposal. Renew Sustain Energy Rev2012;16:4818–48.

[5] Peñate B, García-Rodríguez L. Current trends and future prospects in the designof seawater reverse osmosis desalination technology. Desalination2012;284:1–8.

[6] Dundorf S, MacHarg J, Sessions B, Seacord TF. optimizing lower energyseawater desalination: the affordable desalination collaboration, IDA WorldCongress, Dubai UAE; 2009.

[7] Tonner J. Barriers to thermal desalination in the United States, desalination andwater purification research and development Report No. 144; 2008<www.usbr.gov/pmts/water/publications/reports.html>.

[8] Ophir A, Lokiec F. Advanced MED process for most economical sea waterdesalination. Desalination 2005;182:187–98.

[9] Al-Shammiri M, Safar M. Multi-effect distillation plants: state of the art.Desalination 1999;126:45–59.

[10] Alasfour FN, Darwish MA, Bin Amer AO. Thermal analysis of ME-TVC + MEEdesalination systems. Desalination 2005;174:39–61.

[11] Kamali RK, Abbassi A, Sadough Vanini SA. A simulation model and parametricstudy of MED–TVC process. Desalination 2009;235:340–51.

[12] Amer AOB. Development and optimization of ME-TVC desalination system.Desalination 2009;249:1315–31.

[13] Ameri M, Mohammadi SS, Hosseini M, Seifi M. Effect of design parameters onmulti-effect desalinationsystem specifications. Desalination2009;245:266–83.

[14] Ophir A, Lokiec F. Review of MED fundamentals and costing. IDE TechnologiesLtd.; 2006. <http://www.ide-tech.com/media-center/articles/Review-med-fundamentals-and-costing>.

[15] Wang Y, Lior N. Performance analysis of combined humidified gas turbinepower generation and multi-effect thermal vapor compression desalinationsystems Part 1: the desalination unit and its combination with a steam-injected gas turbine power system. Desalination 2006;196:84–104.

[16] Shakib SE, Amidpour M, Aghanajafi C. A new approach for processoptimization of a METVC desalination system. Desalin Water Treat2012;37:84–96.

[17] Khawaji AD, Kutubkhanah IK, Wie JM. Advances in seawater desalinationtechnologies. Desalination 2008;221:47–69.

[18] Sayyaadi H, Saffari A. Thermoeconomic optimization of multi effect distillationdesalination systems. Appl Energy 2010;87:1122–33.

[19] Jana AK. Heat integrated distillation operation. Appl Energy 2010;87:1477–94.[20] Shu L, Chen L, Sun F. Performance optimization of a diabatic distillation-

column by allocating a sequential heat-exchanger inventory. Appl Energy2007;84:893–903.

[21] Alarcón-Padilla D-C, García-Rodríguez L. Application of absorption heat pumpsto multi-effect distillation: a case study of solar desalination. Desalination2007;212:294–302.

[22] Alarcón-Padilla DC, García-Rodríguez L, Blanco-Gálvez J. Experimentalassessment of connection of an absorption heat pump to a multi-effectdistillation unit. Desalination 2010;250:500–5.

[23] Alarcón-Padilla DC, García-Rodríguez L, Blanco-Gálvez J. Designrecommendations for a multi-effect distillation plant connected to a double-effect absorption heat pump: a solar desalination case study. Desalination2010;262:11–4.

[24] Wang Y, Lior N. Proposal and analysis of a high-efficiency combineddesalination and refrigeration system based on the LiBr–H2O absorptioncycle—Part 1: system configuration and mathematical model. Energy ConversManage 2011;52:220–7.

[25] Wang Y, Lior N. Proposal and analysis of a high-efficiency combineddesalination and refrigeration system based on the LiBr–H2O absorptioncycle—Part 2: thermal performance analysis and discussions. Energy ConversManage 2011;52:228–35.

[26] Wang Y, Lior N. Thermoeconomic analysis of a low-temperature multi-effectthermal desalination system coupled with an absorption heat pump. Energy2011;36:3878–87.

[27] Al-Ansari A, Ettouney H, El-Dessouky H. Water–zeolite adsorption heat pumpcombined with single effect evaporation desalination process. Renew Energy2001;24:91–111.

[28] El-Dessouky HT, Ettouney HM. Multiple effect evaporation vapor compression.In: Fundamentals of salt water desalination. Amsterdam: Elsevier Science B.V;2002 [chapter 5].

[29] Ng KC, Chua HT, Chung CY, Loke CH, Kashiwagi T, Akisawa A. Experimentalinvestigation of the silica gel–water adsorption isotherm characteristics. ApplTherm Eng 2001;21:1631–42.

[30] Thu K, Chakraborty A, Saha BB, Ng KC. Thermo-physical properties of silica gelfor adsorption desalination cycle. Appl Therm Eng 2011. http://dx.doi.org/10.1016/j.applthermaleng.2011.09.038.

[31] Chua HT, Ng KC, Chakraborty A, Oo NM, Othman MA. Adsorptioncharacteristics of silica gel + water system. J Chem Eng Data 2002;47:1177–81.

[32] Ng KC. Recent developments in heat-driven silica gel–water adsorptionchillers. Heat Transfer Eng 2003;24:1–3.

[33] Ng KC, Saha BB, Chakraborty A, Koyama S. Adsorption desalination quenchesglobal thirst. Heat Transfer Eng 2008;29:845–8.

[34] Wu JW, Biggs MJ, Pendleton P, Badalyan A, Hu EJ. Experimentalimplementation and validation of thermodynamic cycles of adsorption-based desalination. Appl Energy 2012;98:190–7.

[35] Wu JW, Hu EJ, Biggs MJ. Thermodynamic cycles of adsorption desalinationsystem. Appl Energy 2012;90:316–22.

[36] Thu K, Myat A, Kim YD, Chakraborty A, Ng KC. Performance investigation ofadvanced adsorption desalination cycle with condenser–evaporator heatrecovery scheme. Desalin Water Treat 2012. http://dx.doi.org/10.1080/19443994.2012.69365.

[37] Thu K, Chakraborty A, Kim YD, Myat A, Saha BB, Ng KC. Numerical simulationand performance investigation of an advanced adsorption desalination cycle.Desalination 2012. http://dx.doi.org/10.1016/j.desal.2012.04.021.

[38] Thu K, Saha BB, Chakraborty A, Chun WG, Ng KC. Study on an advancedadsorption desalination cycle with evaporator–condenser heat recoverycircuit. Int J Heat and Mass Transfer 2011;54:43–51.

[39] Ng KC, Thu K, Kim Y, Chakraborty A, Amy G. Adsorption desalination: anemerging low-cost thermal desalination method. Desalination 2012. http://dx.doi.org/10.1016/j.desal.2012.07.030.

[40] Ng KC, Thu K, Chakraborty A, Saha BB, Chun WG. Solar-assisted dual-effectadsorption cycle for the production of cooling effect and potable water. Int JLow Carbon Technol 2009;4:61–7.

[41] Ng KC, Thu K, Saha BB, Chakraborty A. Study on a waste heat-driven adsorptioncooling cum desalination cycle. Int J Refrig 2012;35:685–93.

[42] Dittus FW, Boelter LMK. Heat transfer in automobile radiators of the tubulartype. Int Commun Heat Mass Transfer 1985;12:3–22.

[43] Han J, Fletcher L. Falling film evaporation and boiling in circumferential andaxial grooves on horizontal tubes. Ind Eng Chem Process Des Dev1985;24:570–97.

[44] Collier JG, Thome JGCJR. Convective boiling and condensation. ClarendonPress; 1996.

[45] El-Dessouky HT, Ettouney HM, Mandani F. Performance of parallel feedmultiple effect evaporation system for seawater desalination. Appl Therm Eng2000;20:1679–706.

[46] Dubinin MM. The potential theory of adsorption of gases and vapors foradsorbents with energetically nonuniform surfaces. Chem Rev1960;60:235–41.

[47] Tien Chi. Adsorption calculations and modeling. Boston: Elsevier Science &Technology Books; 1994.

[48] Saha BB, Kashiwagi T. Experimental investigation of an advanced adsorptionrefrigeration cycle. ASHRAE Trans 1997;103:50–8.

[49] Chakraborty A, Saha BB, El-Sharkawy II, Koyama S, Srinivasan K, Ng KC. Theoryand experimental validation on isosteric heat of adsorption for an adsorbent +adsorbate system. High Temp – High Press 2008;37:109–17.

[50] Chakraborty A, Saha BB, Koyama S, Ng KC. Specific heat capacity of a singlecomponent adsorbent–adsorbate system. Appl Phys Lett 2007;90:171902.

[51] Darwish MA, Abdulrahim HK. Feed water arrangements in a multi-effectdesalting system. Desalination 2008;228:30–54.

[52] Kim YD, Thu K, Myat A, Ng KC. Numerical simulation of solar assisted multieffect distillation (SMED) desalination systems. Desalin Water Treat 2012.