DESALINATION

ELSEVIER Desalination 169 (2004) 43-60

www.elsevier.corrdlocate/desal

Optimal design of hybrid RO/MSF desalination plants

Part III: Sensitivity analysis
A.M. Helal, A.M. E1-Nashar, E.S. A1-Katheeri, S.A. A1-Malek
Power and Desalination Research Center, Abu Dhabi Water and Electricity Authority (ADWEA),
Abu Dhabi, UAE, P.O. Box 54111
Tel. +971 (2) 508-1504; Fax +971 (2) 508-1506; email: alya123@emirates.net.ae

Received 21 March 2002; accepted 18 November 2003

Abstract
This is the last paper in a series of three parts entitled"Optimal design o fhybrid RO/MSF desalination plants". This
research is concerned with exploring the feasibility of hybridization of multi-stage flash (MSF) and reverse osmosis
(RO) technologies in order to improve the performance characteristics and process economics of the conventional MSF
process. The research project involved an optimization study where the water cost perunit product is minimized subject
to a number of constraints. In the first part, the design and cost models were presented, the optimization problem
formulated and solutions for a number of cases were outlined. In the second part, results were presented and discussed.
In this paper we discuss the sensitivity of water cost from the alternative plant designs to variations in some cost
elements and operating conditions. In general, it is concluded that, for the same desalting capacity, hybrid RO/MSF
plants can produce desalted water at a lower cost than brine recycle MSF plants, while hybrid plants are characterized,
by lower specific capital costs and higher water recovery fractions. Reduction in steam cost allows MSF to compete
more with hybrid RO/MSF plants. This result explains the advantage of coupling MSF plants and steam power plants
where the exhaust steam from the back pressure turbine represents a relatively cheaper source of heat for the MSF
process. Results showed that the RO technology exceeds all other designs over the whole range of energy, chemicals
and membrane costs studied here. However, water cost of the RO process was the most sensitive to variations in
membrane and electricity costs compared to other hybrid configurations.

Keywords: Hybridization; Optimization; Hybrid RO/MSF; Desalination economics; Sensitivity analysis

1. Introduction • Steam cost

• Membrane cost
In this sensitivity analysis, the following cost
• Top brine temperature (TBT)
elements and operating variables were considered
• Cost o f chemicals
to study their impact on th m i n i m u m water cost:
• Cost o f electrical energy
*Corresponding author.

0011-9 t 64/04/$- See front matter © 2004 Elsevier B.V. All rights reserved
doi: 10.1016/j.desa1.2004.08.006
44 A.M. Helal et al. / Desalination 169 (2004) 43-60
The objective function, W ($/m 3) in this opti-
mization study is given by:
W = [(CDR+C,R+CoR)+(CDM+CIM+CoM)] $/Y
Wy m3/y
where Cm = RO direct capital cost including
membrane cost, civil work and intake cost (S/y);
CIR = RO indirect capital cost (S/y); COR = RO
operation and maintenance cost including labor,
membrane replacement, parts, chemicals and
energy costs (S/y); CoM -- MSF direct capital
investment cost (S/y); CIM= MSF indirect capital
investment cost (S/y); Cora = MSF operation and
maintenance cost including labor, parts, steam,
energy and chemicals costs (S/y); and Wy = total
production capacity per year (m3/y).
In all the cases presented here, the plant
capacity of desalted water is the same (3366
metric t/h; 21.37 mgd), and the final product
quality is less than 500 ppm.
The plants presented in the following text are
single purpose (water only). They receive elec-
trical power and steam from an external source at
specified costs. In the case of hybrid plants, the
MSF to RO production ratio is 1:2. In all the
cases including brine recycle MSF plants, the
number of rejection stages was taken to be three.
Calculations are based on a feed concentration of
42,000 ppm and a seawater temperature of 25°C.
The objective of the computations was to
design the desalination plant for a given
configuration so that the water cost is minimized
while the following constraints are satisfied:
• The final product concentration should be less
than 500 ppm.
• The brine velocities inside tubes in the heat
recovery, heat rejection and the brine heater
lie between 3 and 6 ft/s.
• The brine loading in the MSF plant section
should be maintained between 1000 and
1200 m3/h per meter of stage width.
• The maximum concentration of the flashing
brine at the exit from the last rejection stage is
limited to 80,000 ppm maximum.
• The reject brine concentration from the first
stage in the two-stage RO plant should be less
than 67,000 ppm. The maximum reject brine
concentration off the second RO stage is
30,000 ppm.
• In all hybrid plants, the reject concentration
from the RO (single-stage) plant section is
limited to a maximum of 67,000 ppm.
• Operating pressure in the RO plant section in
the hybrid designs should not exceed 80 atm.
The same condition is imposed on the operat-
ing pressure in the first stage in the case of a
two-stage RO plant. Pressure in the second
stage is limited to 35 arm maximum.
In the cases including MSF plants, an amount
of seawater which is equal to 145 metric t/h is
considered to be used at the venting condenser.
The overall plant recovery is defined as the
percent of the total seawater intake, including the
water to the vent condenser, converted to fresh
water when an MSF plant is included.
2. Results and discussions
Figs. 1-9 represent the optimal designs ob-
tained through computations based on the
engineering and cost data given in Tables 1 and 2.
The output plant characteristics corresponding to
those flow sheets are given in Table 3.The output
values reported in Table 4 are based on the cost
data outlined in Table 2. These cost data are taken
as reference where a cost multiplier equal to one
is assigned to each cost item. In Tables 5-9,
representing the results of the sensitivity study,
the output values corresponding to a cost mul-
tiplier of 1 are emphasized in bold.
It was explained earlier in Part II of this work
that the high capital investment cost of the MSF
process represents a major economic disadvan-
tage of that technology. According to the cost
 A.M. Helal et al. / Desalination 169 (2004) 43-60 45

1st Stage RO Membrane

( )
Wm .8.97E+O6 kg/ht ---~'( ~ - - - ~ ' - - - ~ ' - - ~ --
Cartridge Crol= 419tSppm -- ~ --d
Filters ---¢~-~"'~'~...~.....~..~.~..-
( ) I ~
wp=~.gsE+o6
J-'--lC~-42000 ppm
I .5 C
....
~lgn 2~ressure
I
. , .~'--.-~-'---'].__

Intake and
Feed Pump
( ) u
$
~ 1st Stage
Product

Multi Media W,o~=1,03E+06 kg/hr I Energy Recovery Turbine

Filters Cr, ffi525 p~m I ~ Wm~3.38E+06 kg/hr
Co, =525ppm

"~ Cv~f7ppm Cz,~-525 ppm

W,~=l.78E+04 k e r r ~tqf=5~SgE+06kg/hr
W ~g- 3.3 7E+OSkg/kr Cr~=3OOOOppm )q~-67000 ppm
2ridStageRO "1
[ G#g'370 ppm
Reject
Membrane ¢ ___~....._.~ Brine

Fig. 1. Schematic diagram of a seawater reverse osmosis desalination plant.

Reject Cooling watea"

Wa~24E+O6 kg/hr
OF---42OOOppm
ToEjectors r,=34 c
4.99E+03kg/hr w~--a.e4E+o6ka~r W~--a.O4E+O6 kg~ar
MSFPlant Or--42000ppra CF---42000ppm
4/s=l'35E+OS kgdzr T, =lOl.7 CI ]T6=32.4 t "(. T=34C r ]Ts~--25 C
r,=114c 2~ ....................... 1- I L Seawater
v | I reed
~ ......... | | ~ Distillate
'vu=z'/Yw"~~g/~] ~ / o =l.12E+OdkO'hr
:r~2ooo p ~ l [r/, =3o c
r,=34c I_____] "_
:~ Blowdown
1.40E+OSP-g/hr t~4ne Henter Heat Recovery Seca'on ~vp ~ IW~l.am~o6
J [Cn--70178pt~2
Steam ]Te--32,4C
Wn,,~a=l.OgE'c'O7kg/l~ Woa--1.67E+O6kg/hr Recycle WR= 9.14E+O6kg/hr
Wt~t=8.02E.t.Od~ CR=70178 pPra CR=7017817pra OR-- 70178ppm
C~t_.8OOOOpl~ , ~ T~=32.4 C ~ C Ta--32.4 C
Heat Rejection S ~
T,=Sl.SC T T
W ~ 2 . ZPE40d kg/hr ] I WR-- 9.14E+Odkg/kr
c~2ooo m,~ l lc ~ mm ~ m
r,=34 Cl ~r~32. 4 c

Fig. 2. Brine recycle multi-stage flash desalination plant (typical one of three identical modules).
46 A.M. Helal et al. / Desalination 169 (2004) 43-60

Wcw=7.O1E+05kg/hr
CF=42000ppm
T1=34[C
MSF Feed
WF= 9. 80E+OSkg/hr

~,, Tso,=25C
Brine Recycle
M S F Plant

'1
Distillate
[ Wo=I. 12E+05 kg/hr
Final Product ~ To=30C
Wr=3.37E+OS
kg~ I TwoStage
CT=264ppm I Permeate I s w R o Plant.~ / RO Feed
TT=26.7 C W =2.24E+05 kg/hd i . . _ Wro=5.99E+06 kg/hr ( ~ _ - -
P C =395ppm CF-- 42000 ppm
Tsea=25 C

¢.=3. 74E+06kg/hr W.~=I.57E+05 kg/hr

d'~57000ppm C~d~= CR=70178 ppm
~e~=25C ( ~ Tbd= T6=32.4 C

"~'~ W _.=5 41E+06 kg/hr

Out fall I

Fig. 3. Case 1: Independent two-stage SWRO and brine recycle MSF plants with common intake-outfall facilities.
 A.M. Helal et al. / Desalination 169 (2004) 43-60 47

Wm=5.95E+O6ksChr Single Stage RO Plant with

Ce=4eO00/~ra ~ Reoovery Sk~tem

Tt=34C ~~--~4E+05

ro,~tors
Ctl=dTOOOtlnn;
w~2z~o~ ~e,~ T,=~4c ~_p=750ppm
72=lO/.g c= ] ~ F Plant T~=J?.4C C~=42000ppra IT1=34 C w~.92E+Od kg/hr
r ~ ......... _. ~ - - ---J ~ J ~ C~'=42000ppm
v.=~..~,~.wsk~--~ ...................... ~-' (
"~ T,=114c/'~ I.......... I, ~__ . T~=25Cscawatcr
~) ~

~ rg~oc Wr=3.3Z~+O~ke/~r
C~ 500R~M
Car~'n~
Product

, ~ 3= "[~11 I i I]'~1 ......' ....

$tnam /~t ~Seahm
W~=2.YBE+05~ Cba=57137ppm
Ma=e, C~.42000m,~
Brme//~ta- T1=34C T~33.5 C
Wnv~=I. I OE+C7k8~ l Ib~3cle
Wt=t=~.O2-~+O6~vlir C,==57411ppra Wn2=1.g~E+O51c~W#=3.ZOE+Odkg/ltr Heat Rq~/on ~ a n
Ct.=76846ppm - TS=31.SC CR=d7411ppm ~ C~=57000pptn
r,=,l., C .(~ )~ r , - - 3 4 c
Fig. 4. Case 2: SWRO-brine recycle MSF
W~--2.98~+061~1~l WR=~.14E+Od~g/hr[
hybrid plant. Heated MSF cooling water
c~.4a~0om=t c.=ez4Hm=t reject is fed to the RO plant.
T,=34'CI T,=32~4Cl

Single Stage RO Plant

with Energy Recovery
System
W~o=6.25E+O6kg/hr
Cro=43167ppm
Wt~f, C~ T1 ~ Tro=33.9 C

To Ejectors
4.99~+03 kg/hr
ICr=42000 ppm [ "~1 f ~ - ....
MSF.Plant Ir - I I I w~=J.oaz+o~kg/hr
r, =~oo.~_~ ,~ _Lr"34 ~ -~....... - - ; I ) ) o,=,:ooo~m
--- ~ (s?~----1. J 7" =2~ c . S~,,,ate~
w,=l.ssz+os,g/~ I I . . . . . . . . . . . . . . . . . . . . . . . . " I I .L - - P~ed

C, II
I I I
I
I
I IW,=l.t2E÷Oeke~, Wvvr='J-a/z~'u°Kg/nr
I m =~n("
............ P,odo,,

" = z H c ~ II I~ = , , , I I ' "~-I .L Cr=362pPm

%['-)'--'4 ..... . ~ ~ .... ~ , $2 Tr =32,5C ) Combined
Brine ~eater ......... *
1.63E+05 kg/hr
Steam W ~ = 5,26E+05 kg/hr Recycle
Cto=43157ppra
Tm=33.9 C )
Wml=5.53E+OS kg/br )
Cf=42000 ppm W m2=6.25E+05 kg/hr
T~=34 C CR=53723 ppm
Wr.,~ = 1.32E+07 kg/h/ Te=32.6C
I Wd =4.0 IE+05 kg/hr
W~,t=8.02E÷O6 kg/hr CR =53723 ppm Wb2=3.4OE+OSkg~ r ICtl=67000ppm
Ci, t =61247 ppm T~ =32.5 C Ca =53723ppm ~T~=33.9 C
T,=31.8C )(~W k;i) '
Fig. 5. Case 3: SWRO-brine recycle MSF
u = 5.14E+06 Wt~=9.14E+05k g / h r ~ lE+O5 kg/hr hybrid plant with hot parts of the MSF
I CF=42000ppm Ct~ =53723ppm Cba=dO914ppm blowdown and cooling reject forming the
I Tf=34 C r.=32.6c T~=3~.3 C
feed to the RO plant.
48 A.M. Helal et al. / Desalination 169 (2004) 43-60

Single Stage RO Plant

with EnergyRecovery
System
We= 2,24E+O6kg/hr
C/,= 511ppm
T~=34 C

To~ ~ootor~

T 4.99e~.O3*~b" M S F Plant W,~,Y..Y#.~+06k,f/~r

CF" 42000ppm
) r,.-2~ c SaewatcrFeed
I W,-t.41E+O~ t,f/hr I I J
.~. ~ Combined Product
~/ W r- 3.37E+06"kg/'~
Or= 341ppm
re.3zTc

,-3--6,,3~" W~¢- 6.20~'+06kg/M . •

V ~" CM" 64630p/am Rejoct-Blowdown
steam~ Brine Heater t l ~ t Reoove.ry ~ t l o n
I T~=33.1C
] W/~-9,14E+Oakg/lu"
WC-3.72e÷06kg,9~r I C ~ 61839ppm
Cd.aTOOOpp= t Wm"5.97H'tOdkg/hr R6wycJe ~, T~-32.~"C
| Ct,'42000ppm
Te'34 C W~z'3.33~+06k ~ r
( W~,t'8.02B+O5
G~t" 70494ppm
Wrwbd-l.2OE+07kg/'m"
C~-~lS39ppm
Wn~-2.84~+06 k ~
Ce.#1839ppm
]C~. 67000ppm
J T~34 C
Wq,-3. 726+0Jkl/In [W~-3.JSg+o6k~Vhr
I c~-6zooo9p= Tg"31.8C )(~r~) T~,.32,5C ) , Q Te.,32.5C ~,(~ }-
Cq'67000ppt=
To=34 IT;-3,,~
W~-3.97B+O5kg/hr T l w~, 6.20~+0~k ~
| I WR" 9.14E+06kg/hr C~¢-64630ppm
w~-3.gze.oe~A "= ~#.E÷06
',..,-3.59B÷06k,f~t _ :~..^ kg//u" Cu- 44347ppm
T1- 34 C
[ I C~- 61839ppm Te¢-33.3 C
~ = 44347ppm t,--..='420""
~ . m~ppm ~t~ ~uuv ppm
Tt- 34 C Tt.34 C T~'34 C

Fig. 6. Case 4: Hybrid SWRO--brine recycle MSF plant fully integrated with part of the RO reject forming a portion of
the MSF make-up.

agP/aa
gcrg.t az rys) =m

/1 c~twmn
Ty---"3/.4 C

~=101¢c Ts=31.4C
• .. T"=Zrcsan~erFa ~
w,~.OC-~tr ~ ' -
r,=n4c ~~
T,p3oc ;@

, rr--~gc
Ta=llOC
,(,
za~uostgtr
amra r~,-~z4c w ~ . ~ z ~ r ~qz4c
c,,,,~mzem-,, On/ir,M
Bineti:nar r~z 4c 11~~

Fig. 7. Case 5:Once-through MSF-SWRO hybrid plant with the MSF unit upstream of the SWRO unit.
 A.M. Helal et al. / Desalination 169 (2004) 43-60 49

Single Stage RO Plant

with Energy Recovery
System w~=~.41c-~
~ wn=t.~sz+oz~r
A C~=42000pten CF-4200OPPm
_ ~ I ~ , =2SC = rm=25C
Wp=224E+06kg/~
Cp--471ppm Seawater
T~ =25 C Feed

ToEjectors
4.~E+O3kg/br
Once ThroughMSF Plant W~--~.O~E+06lce/lr [ C0-'64362 ppm
T==I03., ......................
1 l
Tm=2Y C < (

w,=!.o4E+os,~
(.~.=,,c WT= 3 . 3 7 ~ 0 6 k g / ~
Cambincd
~ Product
w~ =m 2E~oa~a/~ CT=SHm ~
ra=30 C Tr=ed.7C

¢. Blouffown
r,=no c l W== ZPTg+Odkg/hr
1.09g~5 k~ar IIe~R~overv~.'tt~ CtL~t=59501ppm
Steam BrineHeater T5 =31.4C

Fig. 8. Case 6:Once-through MSF-SWRO hybrid plant fully integrated with the SWRO unit upstream of the MSF unit.

Wp=214E+06kg/~
Cf 75oppta
Single Stage RO Plant with T~=J4c
Energy Recovery System
ToEjectors
4,P)E+OJ
W~=Wf=8.32E+06kg/hr cf42o~pt~
MSF P&m C~=42000ppm T~-2Y C
Te=70C 2.7C T~=34C
u ----~ ?
¢
(
W¢-~.OSE+06k~
C~37226ppm
- r,=S4C
T4=,4OC
6.Y#E+oske/hr
Steam
T3-HO C
Heat Reco~ry S~ct~'on H e s t Rejwtion
~+
T
0I6c~=7o178~ kg,~
Brine Heater ~,~h't~ "J t Tbd=32"7C
c,,.To178
m'n
Wn,~C"I.41E+07~
T~,3Z7 C ~ , Q
Wt~=4,95~+05
c..rTolzg m~
Tt~,32.;C
WR--e.14E+Odkg/hr
Cao=70178ppm
r~-3e.z c
Tg.31.8C W
IW.l~. / 4e+o6eg4~
w~.os~+o6t ~ l
c1-~7e26m~l T~=32. 7 C
T~-34C1

Fig. 9. Case 7: SWRO-brine recycle MSF hybrid plant with the SWRO unit upstream of the MSF heat recovery section.
All the reject from the SWRO is used as make-up to the MSF plant.
50 A.M. Helal et al. / Desalination 169 (2004) 43-60

Table 1 Table 2
Input data for design calculations Basic input cost data

Variable Symbol Value Capacity

BR-MSF plant section Current work capacities
Coefficient to account a 0.88 RO plant section, m3/d 53,856
for using avg. latent MSF plant section, m3/d 26,928
heat of vaporization
Hybrid plant total capacity, m3/d 80,784
Feed seawater cone., ppm CF 42,000
Avg. specific heat capacity Cpg 4,185
Capacity or
rec. see., kJ/kg.K
Avg. specific heat capacity Cpj 4,176 flow rate
rej. see., kJ/kg.K
Reference capacities and
Inside dia. of condenser ID 0.022
tubes, m
flowrates
Outside dia. of condenser OD 0.024 Feed rate for intake calc., m3/h 1,000
tubes, m Blowdown rate for outfall talc., 750
No. of rejection stages N/ 3 m3/h
MSF reject coolant temp., TI 34 RO plant capacity, m3/d 29,484
°C MSF plant capacity, m3/d 27,216
Top bring temp., °C T3 110
Distillate temp., °C Td 30 Value
Steam temp., °C Ts 114
Feed seawater temp., °C Ts~ 25 MSF plant cost data
Distillate flowrate, kg/h Wa 1.122E6 Capital recovery factor 0.0963
Avg. boiling point rise 0cR 1.4 Chemicals, $/m3 prod 0.024
ht. rec. see., °C Power (pumping), $/m3 prod 0.109
Avg. boiling point rise ~j 1.78 Steam cost (95-125°C), 0.00291-
ht. rej. see., °C $/kg steam 0.0045
Avg. latent heat of ~ 2346.3
Spares, $/m3 prod 0.082
vaporization, kJ/kg
Avg. brine density, kg/m 3 PB 1060 Labor, $/m3 prod 0.1
Avg. pure water density, Pw 1000
kg/m3 SWRO plant cost data
Capital recovery factor 0.0963
SWRO plant section Membrane cost, S/module 4,000 (stage 1)
Pure water permeability, m/s.atm Aw 8.33E-8 3,000 (stage 2)
Molar density of pure water, C 55.56 Chemical treatment, $/m3 prod 0.018
kg mol/m3 Power, S/kWh 0.03
Solute transport permeability, K 3.51E-8 Spares, $/m 3 prod 0.033
m/s Labor, $/m 3 prod 0.1
Mass transfer coefficient for k 2.7E-5 Intake/outfall, $/m 3 prod 0.808
salt, m/s
Module area (stage 1 or hybrid), 100
No. of ions on salt dissociation n 2
m2
Salt molecular weight Ms 58.5
Pure water molecular weight Mw 18 Module area (stage 2-two-stage 100
Universal gas constant, R 8.21E-5 RO plant only), m 2
m3.atm/mol.K
Permeate mass flowrate, kg/h Wp 2.244E6
 A.M. Helal et al. / Desalination 169 (2004) 43-60
data given in Table 2 and the cost equations
applied in this work [1 ], it was found that the cost
of water produced by the MSF process is the
highest whereas the RO technology has the most
favorable economics or least water cost. The
other hybrid processes produce water at costs
more or less between those of the RO and MSF,
depending on the capacity ratio of the RO to the
MSF plant sections (see Table 4).
It is important to notice that, in the following
discussions, prices of steam, energy, membranes
and chemicals were varied by multiplying the
corresponding values given in Table 2 (where the
cost multiplier equals one) by multipliers varying
between 0.25 and 2. As an example, for an MSF
plant operated at a TBT of 110°C and at a cost
multiplier of 2 (as shown in Fig. 10), the cost of
1 kg of steam will be: 2x0.00415 or 0.0083 $/kg
or 8.3 S/ton.
2.1. M i n i m u m water cost vs. steam cost
Steam cost was varied between 0.00104 and
0.00832 $/kg of steam. This corresponds to a
variation of steam cost multiplier from 0.25 to 2
when the TBT is kept at 110°C. Correspondingly,
the minimum water cost was calculated for each
design alternative shown in Figs. I-9. Results are
represented graphically in Fig. 10, and numerical
values are given in Table 5. For better readability,
the data of the MSF or the RO plant were plotted
together with those corresponding to only one or
two of the other hybrid configurations.
Since the MSF is a thermal process that
depends mainly on steam as a source of heat for
distillation, the cost of water produced by this
process is most affected by variations in steam
cost. Linear regression of the data given in
Table 5 corresponding to the MSF process,
results in a straight line with a slope of0.11. This
slope represents the sensitivity of the specific cost
of water produced by this process to variation in
steam cost. This value means that the water cost
51
($/m 3) will be increased by 11 ¢ for an increase of
1 $/t of steam. In other words, the cost per
thousand gallons (Imperial gallons) will be
increased by about 50¢ for an increase in steam
cost equal to 1 $/t.
An investigation of Fig. 10 reveals that for the
whole range of steam cost, water cost from the
MSF process remains to be higher than that
obtained from all hybrid plants. However, as the
specific cost of steam goes down, water costs
from different plants get closer. Table 5 shows
that the MSF process may be favorable to the
hybrid plants of cases 5-7 only at values of steam
cost equal or below 0.001 $/kg (cost multiplier =
0.25). This explains the reason why the water cost
is less when produced from a cogeneration plant
rather than a water only plant where steam is
obtained by burning fuel in an auxiliary boiler, tn
cogeneration plants, the steam cost is allocated
partly to the desalination plant and partly to the
power plant rather than being loaded solely on the
desalination plant. From Fig. 10 it is seen that the
behavior of the water cost from the two hybrid
plants of cases 5 and 6 is more or less the same.
Hybrid plants assume lower sensitivity of
water cost to variation in steam cost. Fitting the
corresponding data of each one of these plants to
a straight line gives slopes of 0.04 for cases 1--4
0.033 for cases 5 and 7. Case 6 is the least
sensitive; linear correlation of its data gives a
slope of 0.02 or 2 (¢/m3)/(S/t).
2.2. M i n i m u m water cost vs. m e m b r a n e cost
Computational results of this part are shown in
Table 6 and plotted on Fig. 11. Except for the RO
plant, the other cases assume approximate linear
relationships. Linearization of the RO line results
in a straight line with a slope of 0.19 (S/m3)/
(100% increase in membrane cost). This result
tells that the water cost from the RO process is
most sensitive to variations in membrane cost.
This is expected since the RO process is purely
52 A.M. Helal et al. / Desalination 169 (2004) 43-60

0
0

t"q
¢q
~ 0 ~ ~ ~ ~ ° . °

0
0
¢q
¢q

°
0
¢q
¢',1

0
0
0
¢4
¢q

0
0
0
r4
t"q

0
0
0
e4
("4

0
0
0 0
,.4
t",l

f~

r13

q.,

t~
t~
 A.M.. H e l a l et al. / Desalination 169 (2004) 4 3 - 6 0 53

oh
~'o6 ~0

0 ¢'q ¢~I e4 ~o

tN oh
~0
0 • ~ qD
r,.) 0 ,.--, ¢".1
oO

0
tt~
¢N 0
¢xl • ¢'e5
tll oh
0 ¢xl e4

0 oh

0 0
0
0
0
ooo
0
o oO ~,0 c3
oh ~"
¢xl e4

¢.Q
0

,7_,
0
0
¢,.q t",l

• oh 0,1
0 e~ 0 ~ oh
0

oh kO 0
oo oh
o ~.. oO
0
txl e4 ~5 e~

r,.)
0
o
0

0 ~oh

etO
0
e~ ¢-q

0
0 o

[-..,
o~ -- oo
~ ~ ~ oo o
0
¢..)

e..)
0 e.d,

0
t~

0 o
0
,.,±± •
~,<<~ ~ ,~ o -~ ~ ~ ~ o'~
54 A,M. Helal et al. /Desalination 169 (2004) 43-60

• . • , •

. 0 • . •

:.9

. ° • 0~. .
0 0 o 0 ~

ddod~

¢.e)
t.~ 0 ¢q00 I~

0 0 ~ 0 , ' "

¢q

0
°

o
•

0
°

o o
• .
r,d O 0 0 0 0

o
0

~A ,A
• • ° . . l>

o 0
0
0

Otr;~
t,atr~ w). , 0 ,.t'~ 0 C , .,m.,4
e5c5~,-~
 A.M.. Helal et al. / Desalination 169 (2004) 43-60 55

t"...

~3

~ 0 ~

0 0 ~ 0 0 ~ o ~ 0 0

• . ° ° ° • . • . °

0
;:,..,.

rA

0 0
e~

0
¢.)

b-,
56 A.M. Helal et al. / Desalination 169 (2004) 43-60

1.6
1.6

,-, 1.4
1.4

J.2 ' ~ 1.2

L)
~ 1.0

.~ 0.8

0.6
"i
~ 0.8

0.6 .~-............................................
,................................................
,..........................................~.........................................,
0.5 1 1.5 2 0 0.5 1 1.5 2
Steam Cost Multiplier Steam Cost Multiplier
1.6 1.6 i ~Case 7 - o - - MSF

~ 1.2 ~ 1.2
o
r..) , rj

.~ 0.8 i .~ 0.8-i

0.6 ~ , , , 0.6 '

0 0.5 i 1.5 2 0 0.5 l 1.5 2
St~un Cost Multiplier Steam Cost Multiplier
Fig. 10. M i n i m u m w a t e r c o s t vs. s t e a m cost.

I.!
1.2
* RO -" Casel--O--Case2 ~RO x C.ase3 A Case4
g-, 1.0 ~ ,.-, 1.1

0.9 ¸ ~ 1.0
~ 0.9
8 o.s
0.8
0.7
0.7
.E 0.6 - .fl 0.6
0.5 0.5 . . . . . . . . . . . . . . . -r ......................... l . . . . . . . . . . . . . . . . ~ ......................... ---q

0.5 1 1.5 2 0 0.5 1 1.5 2

Module Cost Multiplier Module Cost Mukiplier
1.2
* RO ¢ C a s e 5 -+-..-Case6 1"27 , RO~Case7
1.1.,
1.0~ g,.o
"~0.9 ~ 0.9
0

~0.8 i ~ 0.8
0.7 ~ 0.7
•~ 0.6 - .~ 0.6
0.5 , , , 0.5 ~ ............... ~. . . . . . . . . . . . . . . . . . . . . ~ ........................ r ,
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Module Cost Multiplier Module Cost Multiplier

Fig. 11. M i n i m u m w a t e r c o s t vs. m o d u l e cost.

 A.M. Helal et al. / Desalination 169 (2004) 43-60 57

a membrane-based technology. Case 7 comes
next to the RO process with a slope of 0.15. Case
6 is the least sensitive regarding water cost as a
function of membrane cost. The line representing
this relationship has a slope of 0.08.
With a onefold increase in membrane cost
(changing from a cost multiplier of 1 to 2), water
cost at the RO plant (0.90 $/m 3) approaches the
cost of water produced at the hybrid plants repre-
sented by cases 3 and 4 (0.939 and 0.921 $/m3).
2.3. M i n i m u m water cost vs. T B T
the MSF plant continued to be the highest over
the whole temperature range. From Table 7 it is
seen that water cost for the MSF process shows a
continuous, yet slight, increase (from 1.05 to 1. I
$/m3) with increasing TBT from 90 to 110°C. The
same behavior can be noticed for cases 1-4 and 6,
with case 2 showing the minimum water cost
(0.786 $/m 3) amongst all plants. Hybrid plants 5
and 6 will produce water at the same cost (0.881
$/m3) when the TBT is kept at 110°C. In all cases
the steam temperature used at the brine heater
was 4°C higher than the selected TBT.

In this study the TBT at the MSF plant section

was varied between 90 and 110°C. For each
temperature, the minimum water cost corres-
ponding to each plant configuration was com-
puted and the output data are plotted in Fig. 12.
The calculated values are given in Table 7.
Obviously, the two-stage RO plant was excluded
from this part of the study. The water cost from
2.4. M i n i m u m water cost vs. cost o f chemicals
Table 8 shows numerical values for the cal-
culated water costs from the different plants.
Fig. 13 represents plots of these values against the
chemicals cost multiplier. From these figures it is
seen that over a range of 0.25 to 2 for the cost
multiplier of chemicals, the water cost from the

t.2
---e--MSF • Case 1 ..o..... Case 2 1.2
--¢---MSF + C a s e 3 A Case4
~ - - " - ' - ~ ~ " 1.1
, ~ I ~
1.0 ~ 1.0

~, 0.9 ~ 0.9

0.8 ................. 0.8

0.7 0.7
90 95 100 105 110 90 95 100 105 110
Top Brine Ternperattre ( C ) Top Brine Temperature ( C )
1.2 7
, --¢'--MSF o Case5 --+-Case6
1.2 ~ .-¢---MSF ~ C a s e 7
-.'-" ], I i
~" 1.0
e !
p~ 0.9 ~0.9

.~ 0.8 1
' o.s J
0.7 0.7 -~ .................................. ~......................................... , ...................................... ~............................................... .
90 95 100 105 110 90 95 100 105 110
TopBrine Temperattre ( C ) Top Brine Temperature ( C )
F i g . 12. M i n i m u m water cost vs. top brine temperature.
58 A.M. Helal et al. / Desalination 169 (2004) 43-60

+RO.--4t--Casel o Case2 • MSF 1.2 • RO x Case3 • Case4 t3 MSF ..---.o

1.2 1.1 ------.o--.-----'--
1.1
.t-'-- ~ 1.0
~ 1.0
0,9
'a 0.9 ill
O - ~ ~0.8
0.8

i 0.7
•~ 0.6
- t " ~ 0,7

0.6

0.5 0.5
0 0.5 I 1.5 2 I 1.5 2
Chemical Cost Multiplier 0.5 Chemical Cost Multiplier
1.2 1.2 • RO~Case7 [] MST
~ R O O Case5 +-Case6--a--MSF _-.--a------'---
---a-- -----'~'---
1.1
1.0
0.9 0.9
O

0.8 t. 0.8 ~,~

~: 0.7 0.7 *---"*---- ~ *
•-~ 0.6 0.6
0.5 0.5 ...............................................•............................................... ~-.........................................................,........................................................
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Chemical Cost Multiplier Chemical Cost Multiplier

F i g . 13. M i n i m u m water cost vs. chemical cost.

1.3 • RO - Casel -O-Case2 MSF 1.3 ~RO, X Case3 • Case4 [] MSF

"-, 1.2
1.1
"-" 1.0 1.0
0.9 0.9
I.
0.8 o
o~ 0.8
0.7 0.7
.GI 0.6 .~ 0.6
0.5 0.5
0 0.5 I 1.5 2 0.5 1 1.5 2
Energy Cost Multiplier Energy Cost Multiplier

~
1,3 j t RO e ~ e 5 + ~ e 6 o MSF ~ R O ~Case7 O MSF
1.3
"-" 1.2 -~
~"" 1.2
I.I 4
1.1
i.oi 1.0
0.9-~ 0.9
0.8
0.7 i 0.7
0.6 i .~ 0.6
0.5 i 0.5
0.5 I 1.5 2 0.5 1 1.5 2
EaerSy Cost M u l t i p l i e r Ener~ Cost Multiplier

F i g . 14. M i n i m u m water cost vs. energy cost.

 A.M. Helal et aL / Desalination 169 (2004) 43-60
MSF plant remains to be the highest while that of
the two-stage RO continues to be the least.
Linearization of the individual plant data given in
Table 8 shows that the specific water cost from
the MSF process is the least sensitive to chemical
cost variation. The corresponding line shows a
slope of 4.4 (¢/m 3) per unit cost multiplier or
100% change in chemicals cost. For the same
change in chemicals cost, the water cost from the
RO plant will be increased by 7.2 ¢/m 3. Cases 5
and 6 assume the highest sensitivity to chemical
costs. Increasing the chemical cost by onefold
will result in an increase in water cost equal to 10
and 9,3 ¢/m 3 for cases 5 and 6, respectively. For
the same change in chemical cost, case 7 assumes
an 8.3 ¢/m 3 increase in water cost. The higher
sensitivity of cases 5 and 6 to changes in
chemical costs is attributed to their configuration
where the MSF part of the plant is of the
once-through type which by virtue calls for a
higher rate of chemical consumption.
2.5. M i n i m u m w a t e r cost vs. cost o f energy
As shown in Table 2, the reference case cor-
responds to an energy cost of 3 C/kWh. Table 9
shows the values of minimum water cost cal-
culated for the different cases vs. the energy cost
multiplier. The same data are plotted in Fig. 14.
The following discussions are based on lineari-
zation of the water cost data and electricity
energy cost multiplier given in Table 9. Slopes of
these lines show that water cost from the RO
plant is the most sensitive to changes in energy
cost. The corresponding line assumes a slope of
17.7 (¢/m 3) for a one-fold increase in energy cost.
Over the whole range of energy cost variation,
the water cost from the MSF plant remains to be
the highest while that of the RO plant continues
to be the least. Water cost of the different hybrid
plants will be increased by values varying
between 13.4 and 17.2 (¢/m 3) for a one-fold
increase in energy cost. However, water cost from
59
the MSF plant is the least sensitive to energy cost
variation where the corresponding line shows a
slope of only 7.9 (¢/m 3) for a one-fold increase in
energy cost.
3. Conclusions
1. In the first two parts of this study [1,2], the
feasibility of the hybridization of RO and MSF
technologies to improve the MSF economics
were investigated through an optimization study
based on minimum water cost. In this study, nine
different water only plants, having the same
desalting capacity, were considered for com-
parison. Besides a brine-recycle MSF plant, a
two-stage RO plant and seven other hybrid
RO/MSF plant configurations were studied.
2. Comparison revealed that two-stage RO
has the most favorable economics and perfor-
mance characteristics from the standpoint of
water cost, specific capital cost and recovery
fraction. The study concluded that in designing
new desalination plants, hybridization of the RO
and MSF processes would result in better eco-
nomics and operation characteristics than those
corresponding to conventional MSF plants.
3. The MSF process proved to have high
specific capital costs, and this is the main reason
for the high cost of water produced by this
process. For the MSF process to compete, it is
suggested that the process design may be revised
for possible reduction in its capital investment.
4. We investigated the sensitivity of the mini-
mum water cost to: steam cost, membrane cost,
top brine temperature, cost of chemicals and cost
of electrical energy. Results showed that RO
technology has the most favorable economics
regardless of the cost element varied and the
range of variation. A sensitivity study showed
that RO technology is the most sensitive to
variations in the cost of electricity, while the
MSF process is the least sensitive. Also,
60 A.M. Helal et al. / Desalination 169 (2004) 43-60
membrane cost variation will affect RO water
cost more significantly than is the case with
hybrid RO/MSF plants con-sidered in this study.
The MSF process is the most sensitive
technology to variation in steam cost. Chemical
cost variation has the least effect on water cost
from the studied plants.
References
[1] A.M. Helal, A.M. EI-Nashar, E.S. A1-Katheeriand
S.A. A1-Malek,Optimal design of hybrid RO/MSF
desalination plants, Part I: Modeling and algorithms,
Desalination, 154 (2003) 43-66.
[2] A.M. Helal, A.M. EI-Nashar, E.S. A1-Katheeriand
S.A. A1-Malek,Optimal design of hybrid RO/MSF
desalination plants, Part II: Results and discussion,
Desalination, 160 (2004) 13-27.