Integrated membrane systems for seawater desalination: energetic and exergetic analysis, economic...

Desalination 203 (2007) 260–276

0011-9164/07/$– See front matter © 2007 Elsevier B.V. All rights reserved

Presented at EuroMed 2006 conference on Desalination Strategies in South Mediterranean Countries: Cooperationbetween Mediterranean Countries of Europe and the Southern Rim of the Mediterranean. Sponsored by theEuropean Desalination Society and the University of Montpellier II, Montpellier, France, 21–25 May 2006.

*Corresponding author.

Integrated membrane systems for seawater desalination:energetic and exergetic analysis, economic evaluation,

experimental study

Francesca Macedonioa, Efrem Curcioa,b, Enrico Driolia,b*

aDepartment of Chemical Engineering and Materials, University of Calabria,Via P. Bucci CUBO 45/A, 87030 Arcavacata di Rende (CS), Italy

Tel. +39 (0984) 492039; Fax +39 (0984) 402103; email: [email protected], [email protected] on Membrane Technology ITM-CNR c/o University of Calabria,

Via P. Bucci CUBO 17/C, 87030 Arcavacata di Rende (CS), Italy

Received 2 February 2006; accepted 15 February 2006

Abstract

Currently over 12,500 industrial desalination plants, with an average production rate of 22.8 million m3/d, areoperating worldwide to alleviate water shortage problem. In this contest, pressure driven membrane operations aretoday largely utilized. They compete successfully with the more traditional thermal systems due to their lowerenergy consumption. However, critical issues still remain open: improvement of water quality and plant globalrecovery factor, reduction of desalted water cost and brine disposal problem. Recent developments make realisticto affirm that integrated membrane systems might offer one of the most reliable solution to solve these problemsand also a sustainable development of desalination processes. Integrated membrane processes represent an attractiveopportunity because of the synergic effects that can be reached, the simplicity of these units, the possibility ofadvanced levels of automation and remote control. In this logic well fit in also temperature driven membraneoperations, such as membrane distillation and membrane crystallization: they allow increasing fresh water recoveryfactor, recovery of the crystals naturally present in the concentrated streams of the desalination plants, reduction ofbrine disposal cost and its negative environmental impact. In this work seven flow sheets for seawater desalinationhave been proposed, analysed and compared. In each of them different membrane units have been integrated inpre-treatment and post-treatment stages for reaching the process intensification goals, overcoming the limits of thesingle units and, thus, improving the performance of the overall desalination process.

Keywords: Process intensification; Seawater desalination; Membrane distillation; Membrane crystallization;Exergetic analysis; Economical evaluation

F. Macedonio et al. / Desalination 203 (2007) 260–276 261

1. Introduction

Together with the supply of energy and the en-vironmental protection, fresh water is one of thethree key elements for the sustainable develop-ment of every society. Fresh water is needed inagriculture, as drinking water, and as process waterin a large variety of industries. Where the avail-ability of water cannot be carried out by usingconventional sources, unavoidable appears theresort to the major water source: the sea. As amatter of fact, desalination processes represent avalid solution to the water shortage problem andtheir application has completely changed the situ-ation in many arid zones in the world. The prin-cipal desalination processes are divided into twomain categories: thermal and membrane separa-tion methods. At present, reverse osmosis (RO)is the most widely used process, whose contractsaccount for 54% of the total market share of seaand brackish water desalination [1]. The successof membrane operations is due to their intrinsiccharacteristics of efficiency and operational sim-plicity, high selectivity and permeability for thetransport of specific components, compatibilitybetween different membrane operations in integ-rated systems, low energetic requirement, goodstability under operative conditions, environment-compatibility, easy control and scale-up, and largeflexibility. Further it is possible to reach advancedlevels of automation and remote control. Forreaching the goals of still decreasing costs andincreasing water quality in the field of membraneoperations, these could be obtained in two dif-ferent ways:• discovering new materials for making mem-

branes more stable and resistant to chemicalsin order to increase their performance and life-time and, then to reduce the replacement andmaintenance costs;

• coupling several membrane processes in orderto overcome the limits of the single units andto use their synergic effects in terms of betterperformance of the overall system.

In fact, in the last years the reliability of ROhas greatly increased also as a consequence of thedevelopment of other various membrane opera-tions (such as microfiltration (MF), ultrafiltration(UF) and nanofiltration (NF)), that can be com-bined with RO in the pre-treatment steps: MF isan obvious technique for the removal of suspendedsolids and for lowering the silt density index(SDI). It generally provides an RO feedwater ofgood quality, with a lower COD/BOD. Furtherimprovement of the RO feedwater can be obtainedby replacing MF by UF. In ultrafiltration, not onlysuspended solids and large bacteria are retained,but also dissolved macromolecules, colloids andsmaller bacteria. The introduction of the NF pre-treatment step leads to significant improvementin the reliability of RO because it has implicationson the desalination process itself: turbidity, micro-organisms and hardness are strongly reduced, aswell as most of multivalent ions; monovalent spe-cies are also retained by 10–50% depending onthe membrane properties. As a consequence, theosmotic pressure of the RO feed stream is de-creased, thus allowing the unit to operate at higherrecovery factors without scaling problems.

At present, the most interesting developmentsfor these industrial membrane technologies arevery probably related to the possibility of integ-rating various membrane operations with all theimportant benefits in the logic of Process Inten-sification. It consists in the development of novelapparatuses and avant-garde techniques that, com-pared to those commonly used today, are expectedto bring drastic improvements in manufacturingand processing, substantially decreasing equip-ment size/production capacity ratio, energy con-sumption, or waste production [2]. In this logicwell fit in also membrane distillation (MD) andmembrane crystallization (MCr) techniques. MDis a temperature-driven membrane operationwhich allows obtaining fresh water also from high-ly concentrated aqueous solutions. Since it ope-rates on the principles of vapour–liquid equilib-rium, 100% (theoretical) of ions, macromolecules,

262 F. Macedonio et al. / Desalination 203 (2007) 260–276

colloids and other non-volatile components arerejected. MD is not limited by concentration pola-rization phenomena as it is the case in pressure-driven processes. Therefore, when MD operateson RO brine, more fresh water can be produced.MCr has been recently proposed as one of the mostinteresting and promising extensions of the MDconcept [3], an innovative process for the quasitotal recovery of the desalted water combined tosolid salts production. As a matter of fact, this in-novative technology uses evaporative mass trans-fer of volatile solvents through microporous hyd-rophobic membranes in order to concentrate feedsolutions above their saturation limit, thus attain-ing a supersaturated environment where crystalsmay nucleate and grow. When an MCr followsan NF or an RO stage, the highly concentratedbrine does not represent waste but the motherliquor in which crystals could nucleate and grow.Membrane crystallizer has another advantage: itreduces until to eliminate brine disposal problem.In effect, seawater desalination processes are posi-tively contributing to solving the problem of watershortage but, at the same time, they cause locallysome negative impacts on the environment thatneed to be minimized. At present, the majority ofdesalination facilities discharge their concentratewaste streams into surface waters or the oceans.This disposal method currently represents the mosteffective and least expensive option for both smalland larger systems located near coastal regions.However, the promulgation of more and morestringent environmental protection regulationswill progressively reduce this opportunity. It istherefore necessary to develop alternative methodsto ensure a more sustainable growth of desalina-tion processes.

In the case of sea and brackish water desalina-tion process, the optimization and the integrationof various membrane units could contribute toreaching the goals of better water quality, an in-crease of the recovery factor and marginal or zerobrine production, with consequent reduction ofthe environmental problems related to its disposal.

2. Integrated membrane systems for seawaterdesalination: case study

Seven different flow sheets for seawater desali-nation process have been analyzed in the presentwork (see Figs. 1–7). In each of them, except forthe first (FS1) constituted by only the RO unit,different membrane units have been integrated:RO operates on NF permeate in FS2; both MFand NF have been introduced for the feed waterpre-treatment and load reduction to the followingreverse osmosis unit in FS3. In the remaining fourflow sheets, at the basic diagram represented byFS3, a membrane crystallizer module has beenadded. It has the role to increase the quantity ofdesalted fresh water produced in combination withsolid salts production: MCr operates on NF brinein FS4, on RO brine in FS5, both on RO and NFbrines in FS6. In the last flow sheet (FS7), MCrhas been introduced on NF brine while MD ope-rates on RO brine.

In all the flow sheets the feed water composi-tion considered has been found in the box ofGEOL 206.3 Seawater Chemistry which reportedthe standard seawater composition [4]. Seawateris a solution of more than 70 elements but only thefollowing reported in Table 1 make up more than99% of all the dissolved salts.

The RO performance has been taken from [5].In all the flow sheets the RO permeate water com-position has been calculated considering the samerejection values (see Table 2). The feed water flowrate and the operation conditions for the RO unithave been also taken from [5].

For the NF unit the rejection values are takenfrom [5] and are also reported in Table 2. The NFacts on the feed seawater in FS2, on the MF per-meate in all the other flow sheets. In all cases itspermeate is fed to the RO unit and has a water re-covery of 75.3%.

By introducing the NF unit for the feed waterpre-treatment, the RO permeate increases due tothe lower osmotic pressure of the water fed to theunit (related to the removal of bivalent ions by


Fig. 1. Flow sheet 1 (FS2): RO unit alone.

High Pressure Pump

Feed RO

Valve

Brine

Permeate

Fig. 2. Flow sheet 2 (FS2): RO operating on NF perme-ate.

PumpPumpPump

RO

Valve Valve

NFMF

Fig. 3. Flow sheet 3 (FS3): MF/NF/RO.

Fig. 4. Flow sheet 4 (FS4): MF/NF/RO and MCr on NF retentate.

Spurge

Fresh Water

PumpPumpPump

RO

Valve Valve

NFMF

Salts

Distillate recycle Precipitator

MCr

PumpPump

NF

Valve

RO

Valve


Pump

Valve

NaCl

PumpPump

RO

Valve

NFMF

Spurge

Fresh water

Distillate recycle

MCr

Precipitator

Fig. 5. Flow sheet 5 (FS5): MF/NF/RO and MCr on RO retentate.

Precipitator

Precipitatorrecycle

NaCl

spurge

water

recycle salts

spurge

Valvewater

Pump PumpPump

RONFMF

MCr

MCr

Fig. 6. Flow sheet 6 (FS6): MF/NF/RO and MCr both on NF and RO retentate.


brine

fresh water

recycle

recycle

salts

spurge

water

Precipitator

Valve

Pump

MCr

Pump Pump

RONFMF

MD

Fig. 7. Flow sheet 7 (FS7): MF/NF/RO, MCr on NF retentate and MD on RO retentate.

Table 1Standard seawater composition

Table 2Rejection values

NF). In fact coupling the two units (FS2), a globalrecovery factor of 52% [6] has been obtained (seeTable 3).

From the third flow sheet on, the MF unit actson the feed seawater and has a recovery factor of94.7% [7]. In FS3 and in FS2 the quantity of freshwater globally produced is higher and more purewith respect to FS1. The introduction of MF in

TDS 35,000 mg/l pH 8.1 Chloride, mg/l 19,345 55.03 wt % Sodium, mg/l 10,752 30.59 wt % Sulphate, mg/l 2,701 7.68 wt % Magnesium, mg/l 1,295 3.68 wt % Calcium, mg/l 416 1.18 wt % Potassium, mg/l 390 1.11 wt % Bicarbonate, mg/l 145 0.41 wt %

Ion NF (%) RO (%) HCO3

– 62.0 98.4 Na+ 22.0 98.9 Cl– 12.8 99.0 SO4

2– 90.0 99.6 Ca2+ 88.4 99.7 Mg2+ 89.0 99.6

FS3 weakly decreases global recovery factor withrespect to FS2 because the RO unit works at thesame degree of efficiency of FS2 but on a lowerNF permeate flow rate. The MF introduction leadsto benefits in terms of reduction of membranefouling with consequent extension of the life timeof NF/RO membranes and reduction of their main-tenance and replacement costs.

In all the three flow sheets until now analyzed


Table 3Product characteristics for the flow sheets analyzed

FS1 FS2 FS3 FS4 FS5 FS6 FS7 Brine flow rate, m3/h 629.9 504.5 531.9 296.6 309.9 74.60 118.5 Brine concentration, g/l 57.60 71.91 68.02 95.94 76.53 214.4 240.0 Fresh water flow rate, m3/h 421.2 547.0 517.6 753.0 739.6 974.9 931.5 Fresh water concentration, g/l 0.3385 0.2699 0.2699 0.1856 0.1889 0.1433 0.1500 Fresh water recovery, % 40.10 52.00 49.20 71.60 70.40 92.80 88.60 CaCO3 flow rate, kg/h — — — 893.9 85.37 979.3 893.9 NaCl flow rate, kg/h — — — 7,371 12,470 19,840 7,371 MgSO4·7H2O flow rate, kg/h — — — 938.6 — 938.6 938.6

the quantity of brine globally produced is higherand very much concentrated. It would be betterto concentrate brine in order to obtain somethingfrom the waste water, with advantages in termsof both production of further products and reduc-tion of disposal costs. Therefore, in the followingfour integrated systems a membrane crystalliza-tion unit has been introduced on NF and/or onRO retentate streams with the triple purpose toincrease the plant global recovery factor, to re-cover crystals naturally present in the concentratedstreams and to reduce the brine disposal problem.

By analysing NF and RO brine composition,it is possible to recover calcium sulphate, magne-sium sulphate and sodium chloride. The preli-minary objective is to limit calcium sulphate pre-cipitation that causes the reduction of SO4

2– contentin the solution and drastically limits the recoveryof magnesium sulphate. In this work, Ca2+ ionshave been precipitated as carbonates by reactionwith Na2CO3 [8].

In each MD/MCr process, the fresh water com-ing out from the RO has been used as cold waterstream. Tables 3 also reports the results of theanalysis for the last four flow sheets in terms ofproduct characteristics and quantity of salts pro-duced in the case when the crystallization tankoperates at 25°C.

By introducing an MCr unit on NF brine (FS4),the quantity of fresh water produced is higher andmore pure with respect to when it is on RO brine

(FS5) because MCr works at the same degree ofefficiency but on a lower brine flow rate. The in-troduction of the membrane crystallizer unit onboth retentate streams increases plant recoveryfactor so much as to reach 92.8% in FS6, higherthan that of an RO unit alone (about 40%) and muchhigher than that of a typical MSF (about 10%).With respect to FS6, in FS7 the quantity of freshwater produced decreases because MD has a re-covery factor of 77% [5,6] while that of MCr isabout 100%.

As far as salts are concerned, the highest pro-duction there is when the membrane crystallizerunit operates on both retentate streams.

3. Energy and exergy analysis — basic criteria

In order to estimate the energy saving in theprocesses, a methodology based on energy andexergy analysis can be used for establishing, res-pectively, the energy requirements of the mem-brane integrated systems and their exergetic effici-ency evaluated in terms of entropic losses [5,9].

The word “exergy” has been introduced to ex-press the quality of energy. As a matter of fact,the total energy is divided in two parts: exergyand anergy [9]. The anergy is the part of energythat is forced to be given to the environment asheat in conditions of complete degradation. Theexergy is the part of energy that, by reversibletransformations, can be completely converted


from one form to another. Therefore, it can be de-fined as the maximum amount of work obtainedby the evolution of a system with reversible trans-formations from the initial state to the equilibriumstate with the environment. In Table 4 the basicequations used to carry out exergetic analysis arereported. Table 5 shows the results of the analysisfor the seven proposed flow sheets.

From an energetic point of view (see Table 5),the FS1, FS2 and FS3 processes use only electricalenergy. The introduction of MCr and/or MD (inFS4, FS5, FS6 and FS7) introduces a thermal ener-gy requirement (PE), due to the retentate flow ratewhich has to be heated and which increases theglobal energy demand.

FS1, FS2 and FS3 exergetic efficiencies areinteresting because the entropic losses are mode-

Table 4Basic equations for exergy analysis

The mathematical definition of exergy for a fluid stream is given by the following equation [5,9]:

( ) ( )00 0 0 1

0

Ex ln ln Ex Ex ExT p cp p s

P PTG c T T c T N RT x

T−

= − − + − = + +ρ

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

where Ex is the exergy; G = mass flow rate and cp = specific heat of the solution; the subscript 0 indicates the reference state that, for the systems considered, is pure water at T0 and P0;

Ns = moles of solvent per mass unit of the solution =

1000

MW

ci

s

− ∑ρ

⎛ ⎞⎜ ⎟⎜ ⎟⎝ ⎠ ; x1 =

MW

s

i is

i

Nc

Nβ

+ ∑ρ

⎡ ⎤⎛ ⎞⎜ ⎟⎢ ⎥⎝ ⎠⎣ ⎦

;

βi = number of particles generated from dissociation of species i; ρ = density of the liquid solution; ci = mass concentration of the ith chemical component per liter of solution; MWs, MWi = molecular weight of the solvent and of the ith chemical component, respectively.

Exergetic balance: ∆Ex = –T0 · RS + WU + W′U, where T0 · RS = total exergy destroyed and transformed in the production of entropy, , ,x x i x k

i k

E E E∆ = −∑ ∑ = exergy variation between outlet and inlet streams, 3600UW E= ⋅ electrical

exergy and ( ) ( )[ ]0 U V v c v cW G h h T s s′ = − − − thermal exergy supplied to the system. GV is the required steam mass flow rate which can be calculated from the equation:GV = Q/λv where Q = G · cp · (T2 – T1) is the heat required to warm up the fluid G from temperature T1 to temperature T2.

Primary energy (PE) is the energy supplied by fuel combustion to produce thermal energy: PE = GV · 0.8 where 0.8 is the primary energy (Mcal) needed in the boiler for producing 1 kg of steam.

rate with respect to the other flow sheets. In thelast four cases, instead, MCr and or MD increasesexergy destruction: in FS5, where the RO brinehas to be further concentrated in a membrane crys-tallization unit, the thermal energy necessary isreduced with respect to the other systems becausea lower flow rate has to be heated; as a conse-quence, the entropic losses are also lower withrespect to FS4, FS6 and FS7. In FS4, NF brinehas been sent to a MCr: the thermal energy nece-ssary increases (due to the higher flow rate whichhas to be heated) and the entropic losses alsoincrease with respect to the FS5. In FS6, whereboth retentate streams have to be heated, the totalexergy destroyed and transformed in irreversibleproduction of entropy is the highest. The samething occurs in FS7. However, if the water streams


Table 5Energy and exergy for the flow sheets analyzed

(a) If thermal energy is available in the plant or the stream is already at the operating temperature of the MCr unit.

FS1 FS2 FS3 FS4 FS5 FS6 FS7 Elect. energy, kWh/h

2206 1986 1913 1913 1913 1913 1913

Wu, kJ/h 7.942×106 7.148×106 6.886×106 6.886×106 6.886×106 6.886×106 6.886×106 Wu

’, kJ/h 0 0 0 7.129×106 6.640×106 13.77×106 13.77×106 ∆Ex, kJ/h –1.972×104 –3.454×104 –3.270×104 –7.876×105 –11.75×105 –17.78×105 –5.637×105 RST0, kJ/h 7.962×106 7.183×106 6.919×106 14.80×106 14.70×106 22.43×106 21.22×106 RST0, kJ/h (a) 7.962×106 7.183×106 6.919×106 7.674×106 8.061×106 8.664×106 7.450×106 GV, kg/h 0 0 0 13,430 12,510 25,950 25,950 PE, Mcal/h 0 0 0 10,750 10,010 20,760 20,760

are already available at the temperature neededfor carrying out the MCr/MD operation or thethermal energy is available in the plant, than theentropic losses and the energy requirements of theintegrated systems with MC (membrane con-tactor) decrease, reaching competitive values withthose of the other desalination processes.

Since membrane operations utilize in pre-valence electrical energy, the estimate benefit canbe done using the method of the “substitution coef-ficient” (CS) introduced by Electricité de France;this coefficient compares the primary energy savedto the electrical energy consumed in cycles thatutilize electricity-consuming operations in substi-tution of conventional thermal operations [9]:

( )( )

1 2

2 1

CSC CE E

−=

−(1)

where C is the consumption of thermal primaryenergy (MJ or Mcal), E the consumption of elec-trical energy (kWh), and 1,2 the relative index ofthe conventional and innovating process, respec-tively. An innovating process results are conve-nient when the CS is greater than 10.5 MJ/kWh(2.5 Mcal/kWh).

The CSs calculated for FS1, FS2 and FS3 withrespect to MSF and TVC indicate that all the three

integrated membrane systems are more ener-getically convenient than thermal processes(Table 6).

Among flow sheets with MCr unit, FS5 is themost energetically convenient (CS like 99.3, 50.4and 71.9 MJ/kWh with respect FS4, FS6 and FS7,respectively).

4. Exergy destruction distribution

The use of exergy analysis in actual desalina-tion processes from a thermodynamic point ofview is of growing importance to identify the sitesof greatest losses and to improve the performanceof the processes [10]. For reaching this aim, theseven processes have been described in detail andthe exergies across the major components of eachplant have been calculated in an attempt to assessthe exergy destruction distribution. It can be seenthat in each flow sheet the primary locations of

Table 6Substitution coefficient (MJ/kWh) for FS1, FS2, FS3 withrespect to MSF and TVC

FS1 FS2 FS3 MSF 14.9 52.6 47.9 TVC 11.6 23.0 22.1


exergy destruction are the membrane modules inwhich the saline water is separated into the brineand the permeate, and the throttling valves wherethe pressure of liquid is reduced. Meanwhile thereis nothing that can be done to eliminate or decreasethe lost of exergy in the membrane module; themost reasonable and practical way to increase theefficiency or reduce the power input of the plantsignificantly has been shown to be replacing thethrottling valves on the brine stream by an energyrecovery system (like a Pelton turbine or a pres-sure exchanger system). In fact, the pressure ofthe RO brine could be recovered by introducing apressure exchanger system or a Pelton turbinewhich lead to a reduction of the energy consump-tion for all the systems analyzed transferring thebrine pressure to low-pressure feed water whiledischarging the brine at low pressure [11]. Theconvenience of a process with an energy recoverydevice with respect another without ERD can beevaluated also in terms of their exergetic efficiencyε, defined as follows:

output

input

Ex(%) 100

Exε = ⋅ (2)

The analysis of the alternative design was al-ways based on the exergy analysis.

Diagrams in Figs. 8 and 9 allow a comparison

28,0 30

,0

30,8

48,5

36,5

37, 2

48,7

36,6

37,3

0

10

20

30

40

50

60

FS1 FS2 FS3

ε [%

] without ERDwith Peltonwith PES

of exergetic efficiencies of the seven flow sheetsin three different situations: 1) when the seawaterdesalination processes do not use any energy re-covery devices (ERDs), 2) when a Pelton turbineis introduced in order to minimize the specific en-ergy consumption (SEC), or 3) in the case in whicha pressure exchanger system (PES) is used.

In each case, the introduction of an energy re-covery system allows increase of exergetic effici-ency. The process that gains more profit from theuse of an ERD is FS1 because is in this systemthe turbine or the pressure exchanger, workingwith the highest RO retentate flow rate, permitsthe major reduction of energetic requirements.Exergetic efficiency of the integrated systems withthe MCr unit is lower than the other; this is due tothe increase of exergy inlet because of the thermalexergy necessary to heat the retentate streams.

In any case, exergetic efficiency of membranedesalination systems is greater than that of thermalsystems (such as MSF) which are highly irrever-sible and with exergetic efficiencies in the range1.12–10.4% due to the high energy consumption[12].

The estimation of energy saving can be an im-portant but not a determining factor in the choiceof the flow sheet to be used in the desalinationoperations. As already said, the fresh water pro-duction and the brine disposal are also two im-portant factors to be taken into account. Referring

Fig. 8. Exergetic efficiencies of FS1, FS2 and FS3.


Fig. 9. ε for the flow sheets with the MCr unit.

Table 7Quantity of energy required per m3 of fresh water produced (kWh/m3)

(a) If thermal energy is available in the plant or the stream is already at the operating temperature of the MCr unit.

MSF FS1 FS2 FS3 FS4 FS5 FS6 FS7 Without ERD 26.4 [5]

25.74 [13] 5.24 3.63 3.70 2.54a/19.1 2.59a/18.3 1.96a/26.7 2.05a/28.00

With Pelton — 2.69 2.87 2.93 2.02a/18.61 2.05a/17.79 1.56a/26.32 1.63a/27.54 With PES — 2.58 2.84 2.90 1.99a/18.59 2.03a/17.77 1.54a/26.30 1.61a/27.52

to this, in Table 7 the quantity of energy requiredper m3 of fresh water produced for the seven flowsheets analyzed and for a typical MSF plant isreported. The results in the table show that all theproposed flow sheets present better performancewith respect to the MSF process. A comparisonamong the different membrane system leads tothe conclusion that if the thermal energy is alreadyavailable in the plant, the FS4, FS5, FS6 and FS7become very attractive for carrying out the processalso from an energetic point of view.

5. Economic evaluation

For each flow sheet proposed, an economicevaluation has been made to determine the unitcost of fresh water produced and the gain for thesalts sale. All calculations are based on recent eco-nomic data extracted from actual field data and

from design studies in literature (see Table 8). Theachieved results for the seven proposed flowsheets are reported in Figs. 10 and 11.

In all the flow sheets examined, the fresh watercost is lower than that of thermal desalination pro-cesses (about 1.5 $/m3) [15] and ranges from0.39 $/m3 for FS3 with Pelton turbine like energyrecovery system, to 0.74 $/m3 for FS7. The higherwater cost in the integrated system with the MCrmodule is due to the thermal demand of the mem-brane crystallizer unit needed for heating the brine.If the water stream is already available at the temp-erature needed for carrying out the MCr operationor the thermal energy is available in the plant, thewater cost is reduced to about 0.47–0.55 $/m3.However, it should be pointed out that in the integ-rated system with an MCr unit there are two im-portant advantages:• The quantity of produced salts is high enough

19,4

15,8

8,8

19,1

22,7

18,6

10,1

21,922

,7

18,6

10,1

21,9

0

5

10

15

20

25

FS4 FS5 FS6 FS7

ε [%

] without ERDwith Peltonwith PES


Table 8Equations and assumptions for economic evaluation

Equations to be used for economic calculations:

• Amortization factor: (1 )

(1 ) 1

n

n

i ia

i+

=+ −

;

• Annual fixed charges: Afixed = a · DC where DC=direct capital cost;

• Indirect capital cost Ain equal to 10% of total direct capital cost [14];

• Annual electric power cost: Aelectric = c · w · f · m · 365 where w = specific consumption of electric power (kWh/m3), m = plant capacity (m3/d);

• Annual steam cost: Asteam = s· G · f · 365; • Annual cost for chemicals: Achemicals = k · f · m · 365; • Annual labour cost: Alabor = γ · f · m · 365; • Annual brine disposal cost: Abrine = b · B · f365; • Annual maintenance and spare parts cost: Amaintenance =

p · m · f · 365; • Annual membrane replacement cost: Amembrane; • Annual Na2CO3 cost: ANa2CO3

= salt price flowrate · f · 365; • Total annual cost: Atotal = Afixed + Ain + Aelectric + Achemicals +

Alabor + Asteam + Amaintenance + Amembrane + ANa2CO3; • Unit product cost: Aunit, p = Atotal / (f · m · 365). For the systems with the MCr unit, it is necessary to add the equation to calculate the profit for the sale of the salts: • Annual profit for the sale of the salts: Annual profit = Salt

price · Salt flow rate · Plant availability · 365 ($/y).

The calculations are based on the following assumptions: • Interest rate i = 5% [15], Plant life n = 30 y [15] and

amortization factor a = 0.0651 y–1 [15]; • Plant availability f = 0.9 [15]; • Electric cost c = 0.09 $/kWh [15] and heating steam cost

s = 0.0032 $/lb [16]; • Specific chemicals cost k = 0.025 $/m3 [11,15]. By

introducing MF as pre-treatment k = 0.018 $/m3 [17]; • Specific cost of operating labour γ = 0.05 $/m3 [15]. By

introducing MF as pre-treatment γ = 0.03 $/m3 [11]; • Specific cost of brine disposal b = 0,0015 $/m3 [15]

while specific maintenance and spare parts cost p = 0.033 $/m3 [11,14];

• Na2CO3 price = 0,068 $/kg [18]; NaCl price = 30$/t and MgSO4·7H2O price = 570 $/t [19]; CaCO3 price = 62 $/t [20];

• Membrane replacement rate varies between 20% per year for membranes treating high-salinity seawater to 10% per year for membranes treating low-salinity water supported by pre-treatment systems [15];

• Pumping and energy recovery system cost [14]: Costpump

or turbine = 0.0151 · Ge · P where = pressure (atm), Ge = flow rate (kg/h);

• Heat exchanger cost [15, 17]: Cost ($) = [(M&S) / 280] · 101.3 · (2.29 + Fc) · A0.65;

• NF/RO membrane cost = 30$/m2, MF/MD membrane cost = 90 $/m2 [15].

Fig. 10. Desalted water unit cost comparison for FS1, FS2 and FS3.

0,61

0,47 0,460,40 0,40 0,39

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

FS1 FS2 FS3

cost

[$/m

3 ]

without ERDwith Pelton turbine


0,68

0,59

0,73 0,7 4

0,55

0,4 7

0,54 0,55

0,63

0,5 4

0,69 0,

71

0,51

0,43

0,51

0,51

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

FS4 FS5 FS6 FS7

cost

[$/m

3 ]

without ERD and thermal energy available without ERD, with thermal energy availablewith ERD, without thermal energy available with ERD and thermal energy available

Fig. 11. Comparison of the desalted water unit cost for FS4, FS5, FS6 and FS7 without considering the gain from the saltssale.

so that the gain for the salts sale covers morethan entirely the cost of the desalination pro-cess (see Table 9). Therefore, the overall integ-rated desalination process becomes very attrac-tive also from an economical point of view.

• The environmental problems related to thebrine disposal are totally eliminated. Theseproblems are more evident when a desalinationplant has to be constructed away from a saltwater body. Any existing ground or surfacewater has to be prevented against the pollutionwith the salts of the concentrate streams. Thetransport by pipeline to a suitable disposalpoint could add up to the total costs.

Flow rate and temperature of the MCr feed areparameters that influence thermal energy con-sumption, heater and membrane crystallizer sizeand, as a consequence, membranes purchase and

replacement cost. Without considering the gainfor the salts sale, it can be seen that:• The unit water cost increases when MCr feed

flow rate rises (due to the increase of thermalenergy requirement and membrane replace-ment cost);

• When the temperature of the MCr feed grows,the fresh water cost has a minimum (Fig. 12)due to the increase in annual steam cost andtransmembrane flux and to the decrease inmembrane replacement cost.

6. Experimental

Considering the rejection values of Table 2 andthe scheme of FS5, synthetic aqueous solutionwith the composition of the RO retentate havebeen prepared by dissolving the appropriateamount of salts in bi-distillate water. It has been


Table 9Profit for the salts sale and desalted water unit cost

FS4 FS5 FS6 FS7 CaCO3, $/y 437,000 41,730 478,700 437,000 NaCl, $/y 1,743,000 2,949,000 4,693,000 1,743,000 MgSO4⋅7H2O, $/y 4,218,000 — 4,218,000 4,218,000 Total annual profit, $/y 6,398,000 2,991,000 9,389,000 6,398,000 Total annual cost, $/y 4,024,000 3,440,000 5,593,000 5,445,000 Total annual cost per m3 of fresh water produced,

$/m3 –0.40/–0.52a 0.077/–0.041a –0.49/–0.68a –0.13/–0.32a

Total annual cost per m3 of fresh water producedb, $/m3

–0.44/–0.57a 0.032/–0.086a –0.53/–0.71a –0.16/–0.36a

(a) If thermal energy is available in the plant or the stream is already at the operating temperature of the MCr unit.(b) If Pelton turbine is used as energy recovery device.

Fig. 12. Desalted water cost vs. temperature of MCr feed.

Fig. 13. Schematic representation of the experimentalplant: (A) pumps; (B) flowmeters; (C) balance; (D) per-meate tank; (E) cooling bath; (F) crystallization tank; (G)heater; (H) membrane module.

further concentrated, up to crystals formation, ina membrane crystallization unit. The design of thelaboratory apparatus has been made according tothe scheme shown in Fig. 13.

The membrane module MD020CP-2N, suppli-ed by Microdyn, contains 40 hydrophobic poly-propylene hollow fibers of 0.1 m2 total interfacialarea. The crystallization plant was fed with theRO retentate; the permeate stream, consisting in

T

T

T

T

P

P

AA

B B

C

D

E

F

G

H

demineralized water, was recirculated in a counter-current mode. After a transitory state, the tempera-tures observed at the module entrance for retentateand permeate side were 35±1°C and 16±2°C res-pectively. The crystallization tank worked at 25°Cand atmospheric pressure.

The experimental tests have allowed testingthe fluid-dynamic effect on membrane crystalliza-tion operation and, in particular, on trend of sol-

0,53

0,63

0,73

50 60 70 80 90

T [°C]

Cos

t [$/

m3 ]

FS4 without ERD FS4 with turbineFS5 without ERD FS5 with turbine


vent transmembrane flux, on crystal size distribu-tion, on their shape, dimension and growth rate.

Fig. 14 shows the transmembrane flux and theconcentration obtained during laboratory tests atthe operation condition reported above. Whilesalinity exhibits a linear trend (Fig. 14a), trans-membrane flux, apart an initial transitory stage,shows an almost constant trend (Fig. 14b). Thisis because the temperature of feed and the thermal

Fig. 14. Trend of concentration (a) and water transmembrane flux (b) vs. time.

30

35

40

45

50

0 50 100 150 200 250time [min]

conc

entr

atio

n [g

NaC

l /10

0g H

2O] ]

100 l/h120 l/h200 l/h250 l/h

(a)

(b)

difference between the two streams are highenough to contain the decrease in the driving forcedue to the concentration rise.

Further, both transmembrane flux and concen-tration are higher when feed flow rate is high, asa consequence, the time for reaching supersatura-tion and crystals formation decreases when thefeed flow rate increases. Crystals distribution hasbeen characterized by the middle diameter (which

0

0,002

0,004

0,006

0,008

0,01

0 50 100 150 200 250time [min]

tran

smem

bran

e flu

x [k

g /m

2 m

in

100 l/h120 l/h200 l/h250 l/h

(b)


ranges from 25.5 to 78.9 µm) and by the coeffici-ent of variation (see Table 10). The value of CVof about 50% obtained when the product is remov-ed from a conventional mixed-suspension crystal-lizers provides a first comparison between thetraditional equipment and the membrane crystal-lizer.

Crystals obtained have shown the characteris-tic cubic form in accordance with the expectedgeometry of the NaCl crystallized (Fig. 15). Thelab tests have proved that the growth rate decreaseswhen feed flow rate increases, therefore the crystalgrowth is limited by the integration of new mate-rial on crystal surface (Fig. 16).

The results obtained showed that membranecrystallization is a valid method in order to pro-duce salts of good quality and fresh water from

Table 10Values of coefficients of variation (CV) at different crys-tallization conditions

100 l/h 120 l/h 200 l/h 250 l/h Time (min)

CV (%)

Time (min)

CV (%)

Time (min)

CV (%)

Time (min)

CV (%)

170 45.97 160 58.4 154 55.2 140 49.2 200 45.98 190 57.1 184 58.3 170 50.0 230 52.6 220 42.6 214 49.1 200 42.3

Fig. 15. Solid salts produced by MCr.

Fig. 16. Growth rate vs. feed flow rate.

0

0,2

0,4

0,6

0,8

0 50 100 150 200 250 300

feed flow rate [l/h]

Gro

wth

rate

[ m

m/m

in ]

]

the RO brine.

6. Conclusion

In this work the potentialities of different inte-grated membrane desalination systems have beenanalyzed. The results achieved confirm the interestin operating by combining different membraneunits for seawater desalination. In particular, thepresence of NF as pre-treatment allows the in-crease of the water recovery of the RO unit up to52%; the introduction of a MCr unit, on one or onboth retentate streams, increases the plant recovery


factor so much as to reach 92.8% in FS6, higherthan that of a RO unit (about 40%) and muchhigher than that of a typical MSF (about 10%).The presence of a membrane crystallization and/or membrane distillation unit introduces an addi-tional thermal energy requirement which increasesthe global energy demand. However, if the waterstreams are already available at the temperatureneeded for carrying out the MCr/MD operationor the thermal energy is available in the plant, theenergy requirement of the integrated system withMC decreases, reaching competitive values withthose of the other desalination processes.

The recourse to the MD and/or MCr allowsutilizing the added value of the retentate, increas-ing the plant recovery factor, reducing the brinedisposal problem and its environmental impact,and producing valuable crystals for medical, do-mestic or agricultural use. In the integrated mem-brane systems analysed in this work, the qualityand the quantity of produced crystals are highenough that the gain for the salts sale covers morethan entirely the cost of the desalination process,particularly in the case of FS6 (see Table 9). There-fore, the overall desalination process becomesvery attractive also from an economical point ofview.

From this analysis, adoption of the integratedmembrane systems appears an interesting possibi-lity for improving desalination operations andmeeting the increasing pure water demand.

Acknowledgement

The authors acknowledge the financial supportof the Italian Ministry of University and Research,MIUR, for the grant to the FIRB research projectMicrosensori e Nanofiltri per il Controllo di Ag-gregati Molecolari e di Micro-Organismi in FaseLiquida: Applicazione alla Potabilizzazionedell’Acqua Firb-RBNE01MBTC.

References[1] H. El-Dessouky, Sustainability Assessment of Water

Desalination Technologies. Vilamoura, Portugal,2000, pp. 1–13.

[2] A.I. Stankiewicz and A.J. Moulijn, Chem. Eng.Prog., 96 (2000) 1–22.

[3] E. Curcio, A. Criscuoli and E. Drioli, Ind. Eng.Chem. Res., 40 (2001) 2679–2684.

[4] www.usask.ca/geology/classes/geol 206.[5] A. Criscuoli and E. Drioli, Desalination, 124 (1999)

243–249.[6] M. Turek, Desalination, 153 (2002) 173–177.[7] S. Ebrahim, S. Bou-Hamed, M. Abdel-Jawad and

N. Burney, Desalination, 109 (1997) 165–175.[8] E. Drioli, E. Curcio, A. Criscuoli and G. Di Profio,

J. Membr. Sci., 239 (2004) 27–38.[9] R. Molinari, R. Gagliardi and E. Drioli, Desalination,

100 (1995) 125–137.[10] Y. Cerci, Desalination, 142 (2002) 257–266.[11] M. Wilf, Proc. International Conference on Desalina-

tion Costing, Cyprus, Dec. 2004, pp. 18–31.[12] F.A. Al-Sulaiman and B. Ismail, Desalination, 103

(1995) 265–270.[13] M.A. Darwish, F. Al Asfour and N. Al-Najem, Desa-

lination, 152 (2002) 83–92.[14] A.M. Helal, A.M. El-Nashar, E.S. Al-Katheeri and

S.A. Al-Malek, Desalination, 154 (2003) 43–66.[15] H.M. Ettouney, H.T. El-Dessouky, R.S. Faibish and

P.J. Gowin, CEP Magazine, December (2002) 32–39.

[16] J.M. Douglas, Conceptual Design of Chemical Pro-cess. McGraw-Hill, 1998.

[17] A.M. Helal, A.M. El-Nashar, E.S. Al-Katheeri andS.A. Al-Malek, Desalination, 160 (2004) 13–27.

[18] Data from Sole 24 Ore, 2005.[19] T.E. Borgate, Value Adding to Salts Recovered from

Saline Waters in Disposal Basins in the Murray-Darling Basin. Murray-Darling Basin Commission,Canberra City, Australian Capital Territory (web sitehttp://www.mdbc.gov.au).

[20] Data from Retail Market Price Trends (IndustrialChemicals).

Integrated membrane systems for seawater desalination: energetic and exergetic analysis, economic...

Documents

Transcript of Integrated membrane systems for seawater desalination: energetic and exergetic analysis, economic...