Desalination 283 (2011) 237244

Contents lists available at ScienceDirect

Desalination

j ourna l homepage: www.e lsev ie r.com/ locate /desa lEnergy efficiency evaluation and economic analyses of direct contact membranedistillation system using Aspen Plus

Guangzhi Zuo a,b, Rong Wang a,b,, Robert Field c, Anthony G. Fane a,b

a School of Civil and Environmental Engineering, Nanyang Technological University, 639798 Singapore, Singaporeb Singapore Membrane Technology Centre, Nanyang Technological University, 639798 Singapore, Singaporec Department of Engineering Science, University of Oxford, UK Corresponding author at: School of Civil and EnviroTechnological University, 639798 Singapore. Tel.: +60676.

E-mail address: rwang@ntu.edu.sg (R. Wang).

0011-9164/$ see front matter 2011 Elsevier B.V. Aldoi:10.1016/j.desal.2011.04.048a b s t r a c ta r t i c l e i n f oArticle history:Received 13 December 2010Received in revised form 7 April 2011Accepted 18 April 2011Available online 19 May 2011

Keywords:DCMDAspen PlusCross-flow configurationGain output ratioWater production costA direct contact membrane distillation system (DCMD) was simulated by using Aspen Plus for the purpose ofenergy efficiency and economic analyses. A cross-flow membrane module was firstly modeled and thenincorporated into the flowsheet for system simulation. Detailed investigations have been conducted tounderstand the relationships of the water flux/production, the gain output ratio (GOR) and the waterproduction cost (WPC) with respect to various design and operation parameters of the DCMD system.Simulation results revealed that in the DCMD studied here, a critical membrane area existed, below which,significant increases of water production and the GORwere observed with increasingmembrane area, leadingto a significant drop in the WPC. Increasing feed temperature imposes positive impacts on water flux and theGOR. For the higher feed and permeate velocities, there were increases in water flux, water production andthe GOR. However for theWPC there were optimum fluid velocities beyond which the penalty of more energyinput in the form of electricity consumption for pumping was significant. It was also found that when thetemperature difference in the heat exchanger was increased to 6 C, the WPC can be reduced considerably bycutting down the heat exchanger cost.nmental Engineering, Nanyang5 6790 5327; fax: +65 6791

l rights reserved. 2011 Elsevier B.V. All rights reserved.1. Introduction

Membrane distillation (MD) is an emerging thermally drivenmembrane process, where micro-porous hydrophobic membranesserve as a barrier to separate hot feed and cold permeate, and watervapor generated in the hot feed transfers across the membrane andcondenses in the cold permeate. Among the four types of MDconfigurations, direct contact membrane distillation (DCMD) is themost widely used [13].

The concept of MD was developed for seawater desalination asearly as the 1960s [1], but it has not drawnmuch attention for decadesbecause of energy demand for heating the feed and the problem ofmembrane wetting during the operation [27]. The resurgence ofinterest inMD as a potential means for seawater desalination in recentyears is being driven by several factors, which include advancement innovel polymer materials, breakthroughs in membrane fabricationtechnology [811] and severe global fresh water scarcity. At present,the main challenge for large-scale MD desalination is energyconsumption, which was estimated to be more than 40 kWh/m3 incomparison with energy consumption of around 7 kWh/m3 for ROand 40 kWh/m3 for multiple effect distillation (MED) and multi-stageflash (MSF), leading to a relatively high cost for water production[1215]. However, MD only requires moderate temperature togenerate a thermal driving force across the membrane, whichmakes it viable to utilize the waste heat to reduce the waterproduction cost. Furthermore, it has been demonstrated that saltconcentration has relatively little effect onmass flux forMD process incomparison with RO process, indicating that MD can effectively dealwith high concentration brine [16]. It can be economically competitivewhen low-grade waste heat or renewable energy resources such assolar energy are available for use [2,16,17]. In addition, many effortshave been made to improve the hydrodynamic conditions in MDmodules [1822], as the hydrodynamics are closely associated withthe thermal condition for water transfer. It was reported thattemperature polarization in a DCMD module can be minimizedthrough proper module design and hydrodynamic improvement[18,23,24]. Moreover, system optimization, by incorporating a heatexchanger in the MD system for heat recovery, can further reduce theenergy requirement [7,14,23,25]. The concept of gain output ratio(GOR), which is defined as the ratio of heat associated with masstransfer to the energy input, has been applied to reflect the energyefficiency of the MD process decades ago. Nevertheless, little researchhas been focused on the influence of the design and operatingparameters on the GOR value [7,26]. Currently, no large scalemembrane distillation system is reported in use for water production.

http://dx.doi.org/10.1016/j.desal.2011.04.048mailto:rwang@ntu.edu.sghttp://dx.doi.org/10.1016/j.desal.2011.04.048http://www.sciencedirect.com/science/journal/00119164

Fig. 1. Schematic of DCMD module.

238 G. Zuo et al. / Desalination 283 (2011) 237244Economic analyses on pilot MD system have been performed main-ly for the situations where all the design and operation variablesare fixed [14,17,24,27]. There is a scarcity of economic analysis interms of the influences of the key design and operating parameters[26].

In order to achieve system optimization, simulation of the MDprocess has to be carried out. The membranemodule, which is the coreelement of the MD system, has been modeled extensively [21,2832].Several researchers have applied Matlab to simulate the heat and masstransfer in the MD module and the results were compared with theexperimental data [26,33,34]. In contrast, limited reports are availablefor system design and optimization [14,26]. The commercial software,Aspen Plus, which is a widely used simulation platform in chemicalengineering processes, has been applied recently to study the MDsystem [35,36]. In these studies, a customized membrane module wasprogrammed by FORTRAN, and then incorporated as a user-definedunit into the system. The mass fluxes based on different operatingconditions were predicted for a small membrane unit (membrane area:0.286 m2), but no direct linkage of operation and design parameterswith economic analysis was provided, though this is very important forpractical MD applications.

In the present study, a DCMD system was simulated using theprocess simulation platform, Aspen Plus, for the purpose of energyefficiency and economic analyses. A cross-flow membrane module,which is not available in Aspen Plus, was firstly defined and modeled,and then incorporated into the flow sheet for system simulation.Detailed investigations have been conducted to understand therelationships between the water flux/production and design andoperation parameters, including membrane area, feed temperature,feed and permeate velocities, etc. Moreover, this study examined thevariation of the GOR with the design and operation parameters. Withthe assumption of using an alternative energy such as waste heat inthe MD system, economic analysis was also performed. It is expectedthat based on this study, a guideline for optimal design and operationof DCMD system in terms of energy efficiency and water productioncost can be attained, whichwill benefit practical applications of DCMDfor desalination.

2. Theory and methodology

2.1. Heat/mass transfer in DCMD

2.1.1. Heat transfer in DCMDThe heat transfer from the feed side to the permeate side consists

of two components: (a) latent heat associating with the water vaporacross the membrane (Qv), and (b) conduction heat by membranematrix (Qc). Fig. 1 illustrates the heat transfer processes, which can beexpressed for unit area as follows:

Qf = hf TfTfm

1

Qp = hp TpTpm

2

Qm = Qv + Qc = NH +km

TfmTpm

: 3

At steady state, the heat balance guarantees that the threeconsecutive heat transfers satisfy the following equation:

Qf = Qm = Qp: 4

The heat transfer coefficients in the feed and permeate boundarylayers, hf and hp can be calculated using a number of methods [34].Here, one simple and effective form, which has been proved to bevalid in the cross-flow configuration of MD module, was taken fromthe literature [37,38]:

Nup =hpdikp

= 1:86diL

0:33RepPrp 0:33 p

pm

!0:14

Rep =diuppp

; Prp =Cppkp

5

Nuf =hf dokf

= 0:71Re0:5f Pr0:36f

PrfPrfm

!0:25Fc

Ref =douff f

; Prf =Cf fkf

; Prfm =Cfm fmkfm

:

6

where Nusselt number (Nu) is correlated with Reynolds number (Re)and Prandtl number (Pr), the former is related to hydrodynamicconditions, while the latter is only temperature dependent. Thedefinitions of the symbols can be found in the nomenclature.

2.1.2. Mass transfer in DCMDIn the DCMD process, the vapor pressure difference arising from

the temperature difference between the two surfaces of themembrane is the driving force for water vapor transfer across themembrane. From Eqs.(1) to (4), the membrane surface temperaturescan be derived as follows [39]:

Tfm = Tf TfTp 1 = hf di = do

1= hf di = do + 1= NH = TfmTpm

+ km

+ 1= hp7

Tpm = Tp + TfTp 1 = hp

1= hf di = do + 1= NH = TfmTpm

+ km

+ 1= hp8

p = exp 23:20 3816:44T46:13

: 9

Although various mathematic models such as Knudsen, Poiseuille andmolecular diffusion models have been developed for the predicationof mass flux, it is generally accepted that the water flux across themembrane can be expressed empirically as [7]

N = C pfmppm

: 10

Table 1Basic characteristics of the membrane module [38].

Length (mm) 254Inside diameter (mm) 0.33Outside diameter (mm) 0.63Packing factor 0.22Membrane material PolypropyleneThickness (m) 150Porosity 0.04MD coefficient (kgm2h1kPa1) 8.6107

239G. Zuo et al. / Desalination 283 (2011) 237244Previous studies have suggested that C is only slightly dependent ontemperature and can be considered as a relatively constant number inthe calculation [18,33,38]. Here, the C value was taken as 8.6107

kgm2h1kPa1 as recommended in the literature [38].

2.2. Modeling of cross-flow DCMD module

A cross-flow membrane module as shown in Fig. 2 (a), wasselected as a model DCMD unit for simulation. Several studies on thistype of configuration suggested that it has less temperaturepolarization effect and the mass transport coefficient can be enhancedsignificantly for Reynolds numbers that are not very high [26,40].Fig. 2 (b) shows the simulation strategy. The DCMD module wasdivided into many small regions (NN). The heat and mass transfermodels discussed in Section 2.1 were applied to each region andprogrammed in Matlab, and the differential equations for each regionwere solved using the fsolve function in Matlab.

The results, including the outlet temperature and flow rate at boththe feed and permeate sides in the MD module, served as the inputparameters in the subsequent system simulation on the Aspen Plusplatform. The experimental data reported in the literature [38] for thesame MD configuration were used to verify the modeling results. Thebasic characteristics of the membrane module used for modeling arelisted in Table 1.

2.3. Simulation of DCMD system

A DCMD system, which includes a cross-flow membrane module,one heat exchanger, two centrifugal pumps, one heater, one cooler,four valves and two tanks, was designed to simulate an entire DCMDsystem using Aspen Plus. Fig. 3 shows the flowsheet of the system. Afeed solution (model seawater: 3.0 wt.% sodium chloride) washeated to an initial temperature in the range 6090 C, and thencirculated through the shell side of the membrane module a velocitybetween 0.022 and 0.055 m/s, followed by heat exchange to recoverheat from the hot permeate side. Meanwhile, the permeate side(pure water) was maintained at 25 C and circulated through thelumen of the membrane fibers at a velocity in the range 0.250.83 m/s, and then flowed into the heat exchanger for energyrecovery.

To run the system, the self-defined cross-flow DCMD module wasconnectedwith Aspen Plus as a usermodular through four parametersof outlet temperature and flow rate at the feed and permeate sides.Other units such as the heat exchanger are built-in modules in AspenPlus. Since the feed stream is seawater (3.0 wt.%), the electrolyte(a) Feed (Tfi, qfi)

Feed (Tfo, qfo)

Permeate(Tpi, qpi)

Permeate(Tpo, qp

Fig. 2. (a) Cross-flow configuration of MD module. (b) DivisioNRTL, which is a thermodynamic model in Aspen Plus, was selectedfor the simulation.

2.4. Energy efficiency based on the GOR

Based on the system simulation in Aspen Plus platform, the GOR,expressed in terms of the water produced per steam equivalent, as afunction of various design and operating parameters can be found.The following equation was used for the GOR calculation [26]:

GOR = Taxmd

Tmd + THX11

Taxmd = TfiTfo 12

Tmd = TfiTpo 13

THX = TfoT p : 14

2.5. Economic analysis of water production

Several assumptions were made for the economic analysis [27]:(1) the system life is expected to be 20 years; (2) the membrane life5 years; (3) the system availability is 90%; (4) operation andmaintenance costs are not considered; and (5) the interest rate is5%. The omission of (4) is balanced by the relatively low value in (3).

It was also assumed that water production cost (WPC) ($/m3)involves five main components, namely membrane module, heatexchanger, electricity consumption for pumping, waste steam andinstallation of the system. Similar considerations can be found in theliterature [14,17,24,27]. The replacements of membrane and heatexchanger were considered by estimating the depreciation rates. Thepumping cost was caused by the pressure drop in the circulation loop,which can be calculated in Aspen Plus. Table 2 lists the estimated costsE-1 E-2 E-3 E-4

E-5 E-6 E-7 E-8

E-9 E-10

E 11 E 12

E-13 E-15

E 14 E 16

P-7

P-8

P-9

P-10

P-11

P-12

P-13

P-14

P-15

P-16

P-17

P-18

P-19 P-20

P-21 P-22 P-23 P-24

P-25P-26

P-27 P-28 P-29 P-30

P-31

P-32

P-33

P-34

(b)

o)

n of MD module into small regions (NN) for modeling.

Feed Permeate

Heatexchang er

Mem

b ranem

od ul e

Cooler

Feedpump

Permeatepump

Taxmd

Tmd

THX

Brine

Fig. 3. The schematic flow chart of the MD system.

10.00

20.00

30.00

40.00

50.00

60.00

Mas

s fl

ux (

kg/m

2 h)

Experimental data

Simulation data

240 G. Zuo et al. / Desalination 283 (2011) 237244of the above mentioned items in the MD system. The detailedcalculation of WPC is based on following equation:

WPCD=m3

=a AMEM PMEM + b AHX PHX + W Pelectricity + c IF AMEM PMEM + AHX PHX

MwpPwastesteam =GOR 15

where, a, b, c are depreciation rates of membrane module, heatexchanger and related installation components, respectively. AMEMand AHX are the areas of the membrane and heat exchanger,respectively. PMEM, PHX, Pelectricity and Pwastesteam refer to the costs ofunit membrane area, heat exchanger area, electricity consumptionand waste steam respectively, which are listed in Table 2. Mwp is theannual water production (kg/year). IF is the installation factor andGOR is the gain output ratio of the DCMD system.

3. Results and discussion

3.1. Verification of cross-flow DCMD model

Fig. 4 depicts the water flux as a function of feed temperatureranging from 50 to 90 C in a cross-flow DCMD module obtained bymodeling in the present work and the experiment reported in theliterature [38]. It can be seen that the simulated results were in goodagreement with the experimental data under the same conditions.This suggests that the MD model can be used to predict theperformance of MD modules in the cross-flow configuration. Thus,with simulated results under various operating conditions, thecross-flow DCMD module was incorporated into the flowsheet ofthe designed MD system for further energy efficiency and economicanalyses.Table 2Costs of the items used for WPC calculationa.

Membrane Heatexchanger

Electricity Wastesteam

Installationfactor

$ 36/m2 [27] $ 92.3/m2[26] $ 0.06/kWh [26] $ 0.73/t [26] 1.67[26]

a Amortization factor: 0.08 year 1 [27]. Plant availability: 90% [27].3.2. Effects of membrane area and feed temperature on water flux andproduction

The DCMD module performance was studied as a function ofmembrane area, feed temperature, feed velocity as well as permeatevelocity. Since permeate temperature is usually maintained at arelatively constant level of roomtemperature (25 C), it is not consideredas an operational variable in the simulation. As shown in Fig. 5 (a) and(b), an increasing feed temperature had a positive effect on the waterflux and production rate (mass fluxmultiplied bymembrane area). Thisresult is, of course, in agreementwith the general acknowledgement thata high feed temperature can enhance water flux significantly because ofthe exponential relationship between vapor pressure and temperature,as shown in Eq. (9). Whilst it is recommended that the DCMD moduleshould operate at a feed temperature as high as possible, it is recognizedthat when utilizing low-grade waste heat as the heating source, amoderate feed temperature may be expected.

From Fig. 5, it can also be observed that with an increase ofmembrane area, the water flux gradually decreased, and then0.0040 50 60 70 80 90 100

Feed Temperature (C)

Fig. 4. Comparison of modeling results with experimental data [38] in a cross-flowDCMD module (feed temperature: 5090 C, feed velocity: 0.04 m/s, feed concentra-tion: 3.0 wt.%, permeate temperature: 25 C, permeate velocity: 0.48 m/s, membranearea: 0.286 m2).

24

68

10

0

10

20

30

40

50

60

70

80

90

Mas

s fl

ux(k

g/m

2 h)

Feed

Temp

eratur

e(o C)Membrane area(m 2)

Feed

Temp

eratur

e(o C)Membrane area(m 2)

24

68

10

0

200

400

600

800

1000

1200

1400

1600

60

70

80

90

Prod

uctio

n(kg

/day

)a

b

Fig. 5. (a). Effects of membrane area and feed temperature on water mass flux. (b).Effects of membrane area and feed temperature on water production (feed velocity:0.04 m/s, feed concentration: 3.0 wt.%, permeate velocity: 0.48 m/s, permeate temper-ature: 25 C).

0.30.4

0.50.6

0.70.8

100

200

300

400

500

600

700

0.02

0.03

0.04

0.050.06

Prod

uctio

n(kg

/day

)

Feed

veloci

ty(m/

s)Permeate velocity(m/s)

Fig. 6. Effects of permeate velocity and feed velocity on water production (feedtemperature: 60 C, feed concentration: 3.0 wt.%, permeate temperature: 25 C,membrane area: 4 m2).

241G. Zuo et al. / Desalination 283 (2011) 237244approached to an asymptotic value. In contrast, the water productionpresented a different trend of a gradual increase followed by atendency towards an asymptotic value. This arises because there wasa loss in the driving force with increasing membrane area, which wasless exploited for mass and heat transfers. To understand this point,we may consider the following case. Increasing membrane area isequivalent to increasing fiber length given a fixed fiber number anddimension.With an increase in fiber length, Tfm eventually approachesTpm resulting in an extremely low temperature gradient across themembrane at the end of the fiber, which would make a negligibledifferential contribution to theflux [21]. However, thewater productionis proportional to the membrane area. For the given layout andparameters a critical point corresponding to the membrane area ofabout 4 m2 was identified in the present study. When the membranearea was beyond this critical point, the changes in water flux andproductionwerenot so significant. Obviously,morewater production ina largermembranemodule is at the penalty of more capital investment,which will be discussed in Section 3.5. Therefore, it is desirable tooperate a DCMD system with moderate membrane area for the sake ofreasonable water flux and production, which is consistent with thefinding of Gilron et al. [26].3.3. Effects of feed and permeate velocities on water flux and production

Fig. 6 presents the effects of feed and permeate velocities on waterproduction for a membrane module with fixed membrane area. In thiscase, the trend of water flux over the feed and permeate velocities issimilar to thewater production curve. Itwasobserved that a quasi-linearincrease of water production as a function of feed and permeatevelocities occurredwhen the permeate velocitywas below0.48 m/s andfeed velocity was below 0.04 m/s, but the slopes of the curve graduallyflattened when the velocities were further increased.

It has been widely reported that higher feed and permeatevelocities can help to reduce the thickness of the boundary layeradjacent to the membrane surface, thus the concentration andtemperature polarizations can be mitigated, leading to a higherdriving force between the feed and permeate sides. However, whenthe flows of the feed and permeate streams have reached a certainlevel, the heat transfer in the boundary layer is no longer thecontrolling step and so the mass flux/water production is not assensitive as before to increases in feed and permeate velocities. Onething worth noting is that, for the cross-flow MD process, the heattransfer coefficients can approach an asymptotic value before thefluids reached turbulence (ReN2000) [37,38,41]. Thus moderatepermeate velocity (0.48 m/s) and feed velocity (0.04 m/s) in thisstudy are preferred for maintaining a reasonable water productionwith less pumping energy, as further increase in fluid velocities wouldnot benefit the mass transfer significantly.

3.4. The GOR analysis

The effects of membrane area and feed temperature on the GORvalue based on the system simulation using Aspen Plus are shown inFig. 7. As discussed in Section 3.2, the feed temperature has asignificant influence on the water flux and production. To operate aDCMD system at a high feed temperature, high energy efficiency isanticipated. This is in accordance with the finding from Fane et al. [7].However, the energy source is of concern if DCMD systems are to beeconomically viable. In the present study, waste heat was assumed tobe used in the MD system, which would limit the feed temperature toa moderate level.

A higher GOR value can also be attained with an increase inmembrane area, and a similar result was found in the literature [26].However, more capital costs are required for providing moremembranes and making larger membrane modules, which will be

24

68

10

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

60

70

80

90

GO

R

Feed T

emper

ature(

o C)

Membrane area(m 2)

Fig. 7. Effects of membrane area and feed temperature on the GOR (feed velocity:0.04 m/s, feed concentration: 3.0 wt.%, permeate velocity: 0.48 m/s, permeate temper-ature: 25 C).

24

6

8

10

1.2

1.3

1.4

1.5

1.6

2

4

6

8

10

Wat

er p

rodu

ctio

n co

st($

/m3 )

Memb

rane a

rea(m

2 )Temperature difference in

heat exchanger( oC)

Fig. 9. Effects of temperature difference in heat exchanger and membrane area on WPC(Feed velocity: 0.04 m/s, feed concentration: 3.0 wt.%, permeate velocity: 0.48 m/s,permeate temperature: 25 C).

242 G. Zuo et al. / Desalination 283 (2011) 237244further discussed in Section 3.5. It was also observed that theincrease of the GOR was slowed down with increasing membranearea over 4 m2. This result implies that an optimal membrane areadoes exist.

Fig. 8 depicts the relationships of the GOR with the feed andpermeate velocities. It can be seen that the GOR increased with theincrease of both velocities, and finally reached a stable value. Thepermeate velocity of around 0.48 m/s seems to be a critical value,beyond which, no significant improvement in the GOR can beobserved, whilst the change in the GOR caused by the feed velocitywas relatively steady. These results corresponded to the impacts offluid velocities on the water production, shown in Fig. 6. It isrecommended that a DCMD system be operated at the feed velocity ashigh as possible from the concern of the GOR enhancement. However,the GOR only reflects how well the energy is utilized for the waterproduction. For economic analysis, operation and capital costsassociated with the design and operation strategy should be takeninto account.0.30.4

0.50.6

0.70.8

0.3

0.4

0.5

0.6

0.7

0.8

0.02

0.03

0.04

0.050.06

GO

R

Feed

veloci

ty(m/

s)Permeate velocity(m/s)

Fig. 8. Effects of permeate velocity and feed velocity on the GOR (feed temperature:60 C, feed concentration: 3.0 wt.%, permeate temperature: 25 C, membrane area:4 m2).3.5. Water production cost (WPC)

Fig. 9 illustrates the effects of temperature difference (THX) in theheat exchanger and membrane area on theWPC. The estimatedWPCsare within the range reported in the literature [26]. AlthoughTHXwasnot included in the Eq. (15) directly, the higher THX is, the less heatexchanger area is required for heat recovery based on the resultsprovided by Aspen Plus. Thus, the WPC decreased significantly withan increase in THX because of cost saving in reducing the heatexchanger size. However, the reduction of the heat exchanger area islimited when THX is over a certain value. In addition, according toEq. (11), a high THX will render a high energy cost. Hence, it is notsurprising to see that WPC tended to be stabilized when THX waslarger than 6 C. In addition, it is observed that the lowest WPC ofaround $1.5/m3 can be achieved for DCMD in comparison with the ROdesalination cost of $0.5/m3. Though it is not comparable, the cost forDCMD in pilot scale can be reduced for large scale production.Moreover, it is expected that WPC of DCMD can be further reducedwith cheaper waste heat available.

From Fig. 9 it was also found that there was a minimum point of$1.1/m3 for the WPC when the membrane area was around 4 m2.When the membrane area was smaller than 4 m2, the WPC droppedconsiderably with increasing membrane area because of significantincreases of water production (Fig. 5b) and the GOR value (Fig. 7),though the membrane cost was also increased. On the other hand,when the membrane area was further increased over 4 m2, theincrease in the water production (Fig. 5b) and the GOR value (Fig. 7)were slowed down. Consequently, the gain from higher waterproduction and higher energy efficiency was not sufficient to surpassthe expense of extra membrane area, leading to the increase in theWPC.

The relationship of the WPC with both feed and permeatevelocities is illustrated in Fig. 10. The WPC reduced gradually andreached a plateau with increasing fluid velocities. It seems the powerconsumption for pumping the fluids played aminor role in total WPCat low feed and permeate velocities. Referring back to Figs. (6) and(8), the water production and the GOR increased remarkably as thefeed and permeate velocities increased from a low level, and thisdominates the GOR result. At high velocities, however, the feed andpermeate velocities have limited impacts on the water productionand the GOR. The WPC would not decrease any more as the penaltyof more energy input in the form of pump electricity consumption

0.30.4

0.50.6

0.70.8

1.0

1.2

1.4

1.6

1.8

0.02

0.03

0.04

0.050.06

Wat

er p

rodu

ctio

n co

st($

/m3 )

Feed

veloci

ty(m/

s)Permeate velocity(m/s)

Fig. 10. Effects of feed velocity and permeate velocity onWPC (feed temperature: 60 C,feed concentration: 3.0 wt.%, permeate temperature: 25 C, temperature difference inheat exchanger (THX): 10 C, membrane area: 4 m2).

243G. Zuo et al. / Desalination 283 (2011) 237244will not be off-set. Thus, it is favorable to operate the DCMD systemat the feed velocity of 0.04 m/s and permeate velocity of 0.48 m/s forthe benefit of the lowest WPC, when membrane area and THX arefixed.

4. Conclusions

Energy efficiency and economic analyses of a direct contactmembrane distillation system have been performed by using AspenPlus in combination with the modeling of a cross-flow membranemodule, which was incorporated into the flowsheet as a user modulefor system simulation. Detailed investigations have been conductedto understand the relationships of the water flux/production, theGOR value and the WPC with the design and operation parameters,including the membrane area, the size and operating temperature ofthe heat exchanger, feed temperature and feed/permeate velocity,etc.

i. The size of membrane module/membrane area has beeninvestigated. There was a critical value of membrane area,below which, significant increases of water production and theGOR were observed with increasing membrane area, leading toa significant drop in theWPC. However beyond this point extramembrane area had a limited impact on water production andthe GOR value, and the increase in the membrane cost becamethe dominant factor causing the WPC to increase.

ii. Increasing feed temperature gave positive impacts on thewater flux, water production and the GOR value, which is wellin agreement with previous studies in the literature. However,when utilizing low-grade waste heat as the heating source, amoderate feed temperature may be expected.

iii. Feed and permeate velocities played an important role inwater flux and production. The higher the feed and permeatevelocities, the higher the resulting water flux, water produc-tion and GOR. However, when the flows of the feed andpermeate streams reached a certain level, the water flux/water production/the GOR is not so sensitive to furtherincreases in the feed and permeate velocities. Also the WPCdoes not decrease any more as the penalty of more energyinput in the form of electricity consumption for pumpingbecome important. It is favorable to operate the DCMD systemat moderate feed and permeate velocities for the benefit of thelowest WPC.iv. Energy recovery is very important for the operation of DCMDsystem. When the temperature difference (THX) in the heatexchanger was increased to 6 C, the WPC was reducedconsiderably by cutting down the heat exchanger cost. Therange 2 to 10 C was investigated and as the heat exchangerarea does not change much at a higherTHX 6 C of THX isrecommended for the conditions studied.

NomenclatureAMEM Membrane area, m2

AHX Heat exchanger area, m2

a, b, c Depreciation rate ofmembranemodule, heat exchanger andrelated installation components, respectively

Cf Specific heat capacity of feed side, Jkg1K1

Cp Specific heat capacity of permeate side, Jkg1K1

C Membrane distillation coefficient, kgm2h1kPa1

dh Hydraulic diameter of the flowing channels, mmd Diameter of the hollow fiber, mmFc Tube-row correction factorhf Boundaryheat transfer coefficients fromfeedside,Wm2K1

hm Heat transfer coefficient of the membrane, Wm2K1

hp Boundary heat transfer coefficients from permeate side,Wm2K1

H Latent heat of evaporation, kJkg1

k Thermal conductivity, Wm1K1

L Effective fiber length, mmN Vapor flux, kgm2h1

Nu Nusselt numberP Partial pressure of the water vapor, kPaPr Prandtl number,

cpk

PMEM Cost of unit membrane area, $/m3

PHX Cost of unit heat exchanger area, $/m3

Pelectricity Cost of electricity, $/kWhPwastesteam Cost of waste steam, $ 4/tQ Heat flux, Wm2

Qc Conductive heat flux through the membrane, Wm2Qv Latent heat associate with mass flux across the membrane,

Wm2

qf Feed circulating flow rate, L min1

qp Permeate circulating flow rate, L min1

Re Reynolds number,dh

T Temperature, KTp Temperature of the permeate side at the outlet of the heat

exchanger, KTaxmd Temperature drop in membrane module, KTmd Temperature difference in membrane module, KTHX Temperature difference in heat exchanger, KU Stream velocity, m/sW Electricity work, kWh

Greek letters Membrane porosity, % Module packing density, % Membrane tortuosity Membrane thickness, m Viscosity of the fluids, Pas1

Liquid density, g/cm3

Thermal efficiency of the DCMD module

Suffixf Feed sidefi Inlet of the membrane module in the feed side

244 G. Zuo et al. / Desalination 283 (2011) 237244fm Membrane surface at feed sidefo Outlet of the membrane module in the feed sidei Lumen sidem Membrane matrixo Shell sidep Permeate sidepm Membrane surface at permeate sidepo Outlet of the membrane module in the permeate side

Acknowledgements

We would like to thank the Environment and Water IndustryProgramme Office (EWI) of Singapore for funding support under theproject #0901-IRIS-02-03.We are also grateful to Singapore EconomicDevelopment Board for funding Singapore Membrane TechnologyCenter.

References

[1] S. Kubota, et al., Experiments on seawater desalination by membrane distillation,Desalination 69 (1) (1988) 1926.

[2] K.W. Lawson, D.R. Lloyd, Membrane distillation, Journal of Membrane Science 124(1) (1997) 125.

[3] K.W. Lawson, D.R. Lloyd, Membrane distillation. I. Module design and performanceevaluation using vacuum membrane distillation, Journal of Membrane Science 120(1) (1996) 111121.

[4] F.A. Banat, J. Simandl, Theoretical and experimental study in membranedistillation, Desalination 95 (1) (1994) 3952.

[5] A. Criscuoli, M.C. Carnevale, E. Drioli, Evaluation of energy requirements inmembranedistillation, Chemical Engineering and Processing: Process Intensification47 (7) (2008) 10981105.

[6] A. Criscuoli, M.C. Carnevale, E. Drioli, Energy requirements in membranedistillation: evaluation and optimization, Desalination 200 (13) (2006)586587.

[7] A.G. Fane, R.W. Schofield, C.J.D. Fell, The efficient use of energy in membranedistillation, Desalination 64 (C) (1987) 231243.

[8] E. Drioli, V. Calabro, Y. Wu, Microporous membrane in membrane distillation,Pure and Applied Chemistry 58 (12) (1986) 16571662.

[9] M. Khayet, T. Matsuura, Preparation and characterization of polyvinylidenefluoride membranes for membrane distillation, Industrial and EngineeringChemistry Research 40 (24) (2001) 57105718.

[10] M.M. Teoh, T.S. Chung, Membrane distillation with hydrophobic macrovoid-freePVDF-PTFE hollow fiber membranes, Separation and Purification Technology66 (2) (2009) 229236.

[11] S. Bonyadi, T.S. Chung, Flux enhancement in membrane distillation by fabricationof dual layer hydrophilichydrophobic hollow fiber membranes, Journal ofMembrane Science 306 (12) (2007) 134146.

[12] M. Khayet, M.P. Godino, J.I. Mengual, Possibility of nuclear desalination throughvarious membrane distillation configurations: A comparative study, InternationalJournal of Nuclear Desalination 1 (1) (2003) 3046.

[13] L. Martnez-Dez, F.J. Florido-Daz, M.I. Vzquez-Gonzlez, Study of evaporationefficiency in membrane distillation, Desalination 126 (13) (1999) 193198.

[14] S. Al-Obaidani, et al., Potential of membrane distillation in seawater desalination:Thermal efficiency, sensitivity study and cost estimation, Journal of MembraneScience 323 (1) (2008) 8598.

[15] S.A. Avlonitis, K. Kouroumbas, N. Vlachakis, Energy consumption and membranereplacement cost for seawater RO desalination plants, Desalination 157 (13)(2003) 151158.

[16] A.M. Alklaibi, N. Lior, Membrane-distillation desalination: Status and potential,Desalination 171 (2) (2005) 111131.[17] F. Banat, R. Jumah, M. Garaibeh, Exploitation of solar energy collected by solarstills for desalination by membrane distillation, Renewable Energy 25 (2) (2002)293305.

[18] R.W. Schofield, A.G. Fane, C.J.D. Fell, Heat and mass transfer in membranedistillation, Journal of Membrane Science 33 (3) (1987) 299313.

[19] M.S. El-Bourawi, et al., A framework for better understanding membranedistillation separation process, Journal of Membrane Science 285 (12) (2006)429.

[20] D. Zhongwei, L. Liying, M. Runyu, Study on the effect of flow maldistribution onthe performance of the hollow fiber modules used in membrane distillation,Journal of Membrane Science 215 (12) (2003) 1123.

[21] L.H. Cheng, P.C. Wu, J. Chen, Modeling and optimization of hollow fiber DCMDmodule for desalination, Journal of Membrane Science 318 (12) (2008) 154166.

[22] M.M. Teoh, S. Bonyadi, T.S. Chung, Investigation of different hollow fiber moduledesigns for flux enhancement in the membrane distillation process, Journal ofMembrane Science 311 (12) (2008) 371379.

[23] K. Schneider, et al., Membranes and modules for transmembrane distillation,Journal of Membrane Science 39 (1) (1988) 2542.

[24] P.A. Hogan, et al., Desalination by solar heated membrane distillation, Desalination81 (13) (1991) 8190.

[25] Z. Ding, et al., Analysis of a solar-powered membrane distillation system,Desalination 172 (1) (2005) 2740.

[26] J. Gilron, L. Song, K.K. Sirkar, Design for cascade of crossflow direct contactmembrane distillation, Industrial and Engineering Chemistry Research 46 (8)(2007) 23242334.

[27] F. Banat, N. Jwaied, Economic evaluation of desalination by small-scale autonomoussolar-powered membrane distillation units, Desalination 220 (13) (2008)566573.

[28] A.O. Imdakm,T.Matsuura, AMonteCarlo simulationmodel formembranedistillationprocesses: Direct contact (MD), Journal of Membrane Science 237 (12) (2004)5159.

[29] M. Khayet, A.O. Imdakm, T. Matsuura, Monte Carlo simulation and experimentalheat and mass transfer in direct contact membrane distillation, InternationalJournal of Heat and Mass Transfer 53 (78) (2010) 12491259.

[30] S. Bouguecha, R. Chouikh, M. Dhahbi, Numerical study of the coupled heat andmass transfer in membrane distillation, Desalination 152 (13) (2003) 245252.

[31] A.O. Imdakm, T. Matsuura, Simulation of heat and mass transfer in direct contactmembrane distillation (MD): The effect of membrane physical properties, Journalof Membrane Science 262 (12) (2005) 117128.

[32] Z. Ding, R. Ma, A.G. Fane, A new model for mass transfer in direct contactmembrane distillation, Desalination 151 (3) (2003) 217227.

[33] V.A. Bui, L.T.T. Vu, M.H. Nguyen, Simulation and optimisation of direct contactmembrane distillation for energy efficiency, Desalination 259 (13) (2010)2937.

[34] M. Qtaishat, et al., Heat and mass transfer analysis in direct contact membranedistillation, Desalination 219 (13) (2008) 272292.

[35] H. Chang, et al., Modeling and optimization of a solar driven membranedistillation desalination system, Renewable Energy 35 (12) (2010) 27142722.

[36] H. Chang, et al., Simulation of membrane distillation modules for desalination bydeveloping user's model on Aspen Plus platform, Desalination 249 (1) (2009)380387.

[37] B. Li, K.K. Sirkar, Novel membrane and device for direct contact membranedistillation-based desalination process, Industrial and Engineering ChemistryResearch 43 (17) (2004) 53005309.

[38] L. Song, et al., Direct contact membrane distillation-based desalination: Novelmembranes, devices, larger-scale studies, and a model, Industrial and EngineeringChemistry Research 46 (8) (2007) 23072323.

[39] V.A. Bui, L.T.T. Vu, M.H. Nguyen, Modelling the simultaneous heat and masstransfer of direct contact membrane distillation in hollow fibre modules, Journalof Membrane Science 353 (12) (2010) 8593.

[40] S.R. Wickramasinghe, M.J. Semmens, E.L. Cussler, Mass transfer in various hollowfiber geometries, Journal of Membrane Science 69 (3) (1992) 235250.

[41] B. Li, K.K. Sirkar, Novel membrane and device for vacuum membrane distillation-based desalination process, Journal of Membrane Science 257 (12) (2005)6075.

Energy efficiency evaluation and economic analyses of direct contact membrane distillation system using Aspen Plus1. Introduction2. Theory and methodology2.1. Heat/mass transfer in DCMD2.1.1. Heat transfer in DCMD2.1.2. Mass transfer in DCMD

2.2. Modeling of cross-flow DCMD module2.3. Simulation of DCMD system2.4. Energy efficiency based on the GOR2.5. Economic analysis of water production

3. Results and discussion3.1. Verification of cross-flow DCMD model3.2. Effects of membrane area and feed temperature on water flux and production3.3. Effects of feed and permeate velocities on water flux and production3.4. The GOR analysis3.5. Water production cost (WPC)

4. ConclusionsAcknowledgementsReferences