Contents lists available at ScienceDirect
Desalination
journal homepage: www.elsevier.com/locate/desal
Global optimization of MSF seawater desalination processes
Chandra Sekhar Bandi, R. Uppaluri ⁎, Amit Kumar
Department of Chemical Engineering, Indian Institute of Technology Guwahati, North Guwahati, Assam 781039, India
HIGHLIGHTS
• For MSFBR and MSFM processes, DE provided global optimal solutions.
• Cost based MSF process ranking is MSFBR > MSFM > MSFOT* (* refers solution with penalty).
• For important MSF process parameters, obtained solutions improved by 2.31%, 3.9%, 2.92%, 20.24%, 3.53% and 5.2%.
article info
Article history:
Received 2 February 2016 Received in revised form 5 March 2016 Accepted 6 April 2016 Available online 12 May 2016
Keywords:
Differential evolution algorithm Global optimization Seawater desalination Multi stage ﬂash (MSF) MSFBR MSFOT MSFM SQP Modeling Optimization
abstract
This article addresses the global optimal design of multistage ﬂash desalination processes. The mathematical for mulation accounts for nonlinear programming (NLP) based process models that are supplemented with the nondeterministic optimization algorithm. MSFonce through, simple mixture (MSFM) and brine recycle (MSFBR) process conﬁgurations have been evaluated for their optimality. While freshwater production cost has been set as the objective function for minimization, mass, energy and enthalpy balances with relevant sup plementary equations constitute the equality constraints. Differential evolution algorithm (DE/rand/bin) was adopted to evaluate the global optimal solutions. Further, obtained solutions have been compared with those obtained with MATLAB optimization toolbox solvers such as SQP and MSSQP. The global optimal solution corre sponds to a variable value set of [2794.4 m ^{3} /h, 1.0499, 7.62 m, 3.359 kW/m ^{2} ∙ K, 3.297 kW/m ^{2} ∙ K, 3.042 kW/m ^{2} ∙ K and 22] for decision variables [W _{M} , R _{H} , L _{T} , U _{B} , U _{R} , U _{j} , N _{R} ] in the MSFBR process to yield an optimal freshwater production cost of 1.0785 $/m ^{3} . Compared to the literature, the obtained global solution from DE is 2.31% better. Further, inequality constraint resolution has been excellent for DE but not other methods such as MSSQP, SQP and DESQP.
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
Among several technologies viable for potable water production, the desalination of sea and brackish water is an established technol ogy in several countries including the USA, Persian Gulf and European countries [15,35] . Based on the working energy principle, desalination processes are further classi ﬁ ed primarily into two clas ses namely thermal processes that involve phase change due to addi tion of heat and membrane processes that involve pressure energy. While thermal processes are primarily classi ﬁ ed into multieffect evaporation (MEE), MSF and vaporcompression (VC) processes, membrane processes are primarily classi ﬁ ed into RO and electrodial ysis (ED) processes. Among various alternate technologies for sea water desalination, MSF processes have the promising features of
⁎ Corresponding author. Email address: ramgopalu@iitg.ernet.in (R. Uppaluri).
00119164/© 2016 Elsevier B.V. All rights reserved.
large scale operation and ability to deliver good quality potable water (5 – 50 ppm total dissolved solids). A typical MSF process involves brine heating followed with ﬂ ash distillation in multiple stages and subsequent heat recovery. Thereby, a MSF process plant has three important sections namely brine heater, heat rejection and heat recovery sections. Design variations in the MSF process systems refer to either once through (OT) or simple mixer (M) or brine recirculation (BR) process con ﬁ gurations to yield MSFOT or MSFM or MSFBR processes respectively. Among these, while MSFOT is the simplest in design, it is not as ef ﬁ cient as the MSFBR system. The design of efﬁcient MSF processes invariably requires simulation and optimization studies. Several researchers have conducted simula tion studies to obtain insights upon the process performance of MSF processes. These have been contributed by Mandil and Abdel Ghafour [19], Helal et al. [2], AlMutaz and Soliman [14], Rossol et al. [26], Thom as et al. [29] , AbdelJabbar et al. [28] , Hawaidi and Mujtaba [6] , and Tayyebi and Alishiri [34]. Many of these literatures emphasized upon
C.S. Bandi et al. / Desalination 394 (2016) 30–43
31
stagetostage calculations and deployed Newton–Raphson method or tridiagonal matrix (TDM) formulations solved with Thomas algorithm (TA) for evaluation of MSF process performance. Further, optimization studies have also been conducted by several
researchers. These include MSFOT processes [3–5,21–23]; MSFM pro cesses [20] and MSFBR processes [1,3–5,6,7,8,9,16,18,24,25,31,32,33]. Considering the minimization of water production cost as objective function, the literature refer to the deployment of either one of the fol lowing methods: genetic algorithm (GA) [24,32]; sequential quadratic programming (SQP) method [22,23] , deterministic optimization methods built in gPROMS [6,25], generalized reduced gradient (GRG) [1,3–5,20] and in DICOPT++ [22,23]. Further, MATLAB programming environment has also been used in several engineering applications as a competent modeling tool for simulation and optimization studies
[36,37,38,39].
A critical analysis of the available literatures in optimization studies
refers to the following. Firstly, earlier research works mostly addressed
either MSFBR or MSFM or MSFOT processes for process optimization based insights. Only [3–5] addressed MSFOT and MSFBR process opti
mization but not the MSFM process. The authors adopted GRG optimi zation method which is a local optimization tool. It is well known that GRG might provide local solutions whose quality could not be judged in conjunction with the global optimality. Further, GRG is well known to be nonrigorous and fails to solve problems with larger number of in equality constraints, as the method needs the satisfaction of all inequal ity constraints in each iteration. While SQP method foregoes such limitation, the SQP also could not provide insights upon the quality of generated optimal solutions. On the other hand, nondeterministic models such as GA were only investigated for the MSFBR but not MSFOT and MSFM processes. Thus, it is apparent that global optimiza tion methods have not been applied till date for the comparative assess ment of MSFBR, MSFM and MSFOT processes. Secondly, a critical issue with respect to alternate optimization methods such as GRG, SQP, and GA, is with respect to the satisfaction of inequality constraints. The traditional approach to couple a penalty function with cost function may or may not yield feasible solutions using GRG and SQP methods, given the fact that these algorithms may require additional ﬁ ne tuning of optimization algorithm parameters such as maximum number of iterations, maximum function evalua tions, and penalty parameters, to obtain feasible solutions. Thus, it might be the case that an engineer may have to spend a signiﬁ cant amount of time in ﬁne tuning these parameters for the deterministic op timization methods. On the other hand, such insights may not be appli cable for the nondeterministic optimization methods due to random nature of solution search. Therefore, an important issue that also needs to be addressed is the ability to fetch feasible solutions with sim ilar penalty function parameters for both deterministic and non deterministic optimization methods.
A third and essential insight is to visualize upon the sensitivity of
process and operating parameters using global optimization ap proaches. While such sensitivity analysis might be possible with local optimization methods, they may not provide the most stringent sensi
tivity analysis. Therefore, the sensitivity analysis conducted with non deterministic methods needs to be judged with that conducted with de terministic methods.
In summary, this work addresses three major objectives. The ﬁrst ob
jective refers to comparative assessment of MSFM, MSFOT and MSFBR processes using nondeterministic optimization. The second objective re fers to the evaluation of inequality constraint resolution ability for both deterministic and nondeterministic methods. The ﬁnal objective is to conduct sensitivity analysis of all MSF processes in the light of global and local optimization. Differential evolution (DE) has been chosen as
the global optimization tool as it has not been studied for MSF process op timization despite being proven effective for other engineering optimiza tion problems. Thereby, suitable benchmarks are expected to be set for the engineering optimization of MSF processes.
2. Process conﬁgurations
A schematic representation of the MSFOT, MSFM and MSFBR pro cess conﬁgurations is presented in Fig. 1(a)–(c). Among these processes, while MSFOT limits the temperature of the last stage to 30–40 °C for winter and summer operations, the ﬂashing operation on several ﬂash stages requires vacuum pressure conditions to achieve operating tem peratures below 100 °C. As indicated in the ﬁgure, the common features of these process conﬁgurations are brieﬂy summarized as follows:
– The feed seawater ( W _{M}_{F} ) at temperature T _{S}_{e}_{a} , is deaerated and chemically treated before being introduced into the condenser/pre heater tubes of the last ﬂashing stage in the heat recovery section.
– The preheated feed seawater at temperature T _{2} enters the brine heater tubes, where the heating steam ( W _{S} ) is condensed on the outside surface of the tubes. Eventually, the seawater reaches the maximum design temperature value also known as the top brine temperature (T _{3} ).
– The feed seawater ﬁnally enters the ﬂashing stages, where a small amount of fresh water vapor is generated by brine ﬂashing in each stage. In each stage, the ﬂashed off vapor condenses on the outside surface of the condenser tubes, where the feed seawater ( W _{M}_{F} ) ﬂ ows inside the tubes from the cold to the hot side of the plant. Thereby, the heat recovery process enables an increase in the feed seawater temperature. The condensed fresh water vapor outside the condenser tubes accumulates across the stages and forms the distillate product stream (W _{M}_{D} ).
Fig. 1(b) illustrates that the MSFM process essentially consists of a brine heater, heat recovery section and brine recycle mixing tank. Hence, the MSFM process con ﬁ guration facilitates a brine recycle stream to reduce fresh seawater requirements and associated chemical pretreatment costs. This is achieved by mixing part of the blowdown brine stream (W _{M}_{R} ) with the feed stream ( W _{M}_{S}_{C} ), thereby generating a mixed stream (W _{M}_{F} ) with higher salinity than that of the fresh seawa ter (set as 70,000 ppm for the upper bound according to ElDessouky et al. [11]. It can be further observed in Fig. 1(c) that the MSFBR desalination plant has heat rejection, recovery section and brine heater section. The ﬁ nal reject stream from the heat recovery section is being split into two streams which serve as cooling seawater stream ( W _{M}_{C}_{W} ) and makeup stream (W _{M} ). The makeup stream is further chemically treated and mixed in the brine pool of the last ﬂashing stage in the heat rejec tion section. The mixed stream is sent to blowdown splitter S _{2} from which the brine recycle stream (W _{M}_{R} ) is introduced into the condenser tubes of the last stage in the heat recovery section. The stream after ab sorbing the latent heat of condensation from ﬂashing vapor in several stages leaves the last stage and enters the brine heater, where its tem perature is enhanced to saturation temperature (i.e., top brine temper ature) at the prevalent system pressure.
3. Methodology
Process optimization of alternate MSF con ﬁ gurations has been targeted by coding a competent simulation model that is supplemented with a nondeterministic optimization algorithm. For comparison pur poses, deterministic optimization algorithms have also been considered to evaluate upon the efﬁcacy of the nondeterministic optimization al gorithm. The following subsections summarize the simulation and op timization models.
3.1. Simulation model
The simulation models for MSFOT and MSFBR processes were adopted from Helal et al. [3] . For the MSFM process, the simulation
32
C.S. Bandi et al. / Desalination 394 (2016) 30–43
model was taken from Abduljawad and Ezzeghni [20]. Other than these literatures, a comprehensive account of the models for MSF processes is not apparent elsewhere in other literature. The process simulation model accounts for the mass and energy bal ance equations applied for each stage including the brine heater, recov ery and rejection sections. The process models are nonlinear in nature and consist only algebraic expressions. Further details of the process models can be obtained from the cited literature and this section brieﬂy outlines various important features of the same. As outlined in the liter atures, the following assumptions are usually applicable for the process models on a theoretical basis [3,10]:
a) Temperature proﬁles of all streams ﬂowing within the plant are linear.
b) Each section has a constant value for heat transfer coefﬁcient, heat transfer area, boiling point rise and speciﬁc heat capacity (C _{P} ) of brine solution.
c) The latent heat of vaporization of water (λ) is constant, and indepen dent of temperature.
d) The speciﬁc heat capacity of brine solution is a weak function of salt concentration.
e) Distillate product is salt free; noncondensable gases have negligible effect on the heat transfer process.
f) Effect of the boiling point rise and nonequilibrium losses on the stage energy balances is negligible.
g) The average speciﬁc heat capacity of brine solution is equal to that of the distillate.
h) The boiling point rise at the exit from the last recovery for MSF is neg ligible.
i) Heat loss is negligible.
The nonlinear system of algebraic equations involving mass and en ergy balances has been deduced by carrying out overall mass balances across blowdown splitter, rejected sea water splitter (for the MSFBR process), salt balances across mixer, energy balances across brine heater and condenser, enthalpy balances on the heat recovery section, heat
Fig. 1. Schematic of (a) oncethrough MSF (MSFOT), (b) brinemixing MSF (MSFM), and (c) brine recycle MSF (MSFBR) seawater desalination processes.
C.S. Bandi et al. / Desalination 394 (2016) 30–43
33
Fig. 1 (continued).
rejection section (MSFBR process), brine heater, mixers and ﬂashing brine. Supplementary expressions for the process model include ex pressions for distillate product temperature (equilibrium correlation), overall heat transfer coef ﬁ cients in various process sections, inside ﬁlm resistance, average tube side brine temperature and thermal resis tance. The thermal resistance is accounted as a function of steamside condensing ﬁlm, steamside fouling, tube metal and brine side fouling resistance components. Table 1(a) summarizes various parametric and design speciﬁcations required for the process simulation model for cross ﬂow type MSF pro cesses. Table 1(b) presents the process cost parameter data, which is adopted from Helal et al. [3]. For various process models, the indepen dent simulation model variables that need to be speciﬁed are (W _{M}_{F} , L _{T} , U _{B} , U _{R} and N _{R} ) for MSFOT; (L _{T} , U _{B} , U _{R} , W _{M}_{S}_{C} , C _{M}_{S}_{M}_{F} and N _{R} ) for MSF M and ( W _{M} , L _{T} , U _{B} , U _{R} , N _{R} , R _{H} and U _{j} ) for the MSFBR process. Hence forth, these variables are treated as independent (decision) variables during process optimization.
3.2. Optimization model
The MSF process simulation model consisting of nonlinear set of al gebraic equations is formulated as a nonlinear programming (NLP) op timization model deﬁned as:
MinOF ¼ ψ Subjected to f ðxÞ¼ 0 g ðx Þ≤ 0
where ψ refers to the objective function deﬁned as the sum of annual ized freshwater production cost and associated penalties. The process simulation model is speciﬁed as f(x) = 0 for a speciﬁc MSF processes and g(x) refers to the set of inequality constraints. The total annualized freshwater production cost function is evaluat ed using relevant expressions for direct capital less intake investment (C _{D}_{C}_{L}_{I}_{C} ), intakeoutfall cost (C _{I}_{C} ) (evaluated as a sum of costs of electro chemical equipment, civil work, electro chlorination, brine disposal cost and annual plant intakeoutfall), direct capital investment cost (C _{D}_{C}_{C} ), indirect capital investment (C _{I}_{C}_{C} ) and operating and maintenance cost ( C _{O}_{M}_{C} ) (evaluated as a sum of costs of steam, chemical treatment, power, labor and spares costs). Relevant expressions have been adopted from Helal et al. [3]. Thereby, the objective function for MSF is modiﬁed and expressed as:
^{ψ} MSF ^{¼}

C 
IC 
þ C 
DCLIC þ C ICC 
þC 
OMC 


$ y 


m y 
3 


W Y 
þ ðpenalty of gð xÞÞ:
ð1Þ
Inequality constraints refer to lower bound and upper bound speci ﬁcations for makeup ﬂow rate for MSFBR (W _{M} ), feed ﬂow rate for MSF OT ( W _{M}_{F} ) and MSFM ( W _{M}_{S}_{C} ), tube length ( L _{T} ), number of recovery stages (N _{R} ), heat transfer coefﬁcients in various sections, brine loading, brine velocities in various sections, rejected brine concentration values ( C _{M}_{B}_{D} , C _{l}_{a}_{s}_{t} ) and absolute values of various heat transfer coef ﬁ cients. The inequality constraint parameters and values have been presented in Table 2(a –b). The penalty function has been evaluated using large
34
C.S. Bandi et al. / Desalination 394 (2016) 30–43
Table 1 
Table 2 

(a) Parametric and design speci ﬁcations data for cross ﬂow type MSF processes. (b) A 
(a) 
Summary of lower and upper bound values for the optimal design of MSF processes. 
summary of MSF process cost parameter data. 
(b) 
A summary of inequality constraint parameters for MSF optimization model. 
(a) 

Variable 
Unit MSFOT 
MSFM 
MSFBR 

a 
– 0.88 
0.88 
0.88 

C _{M}_{F} 
ppm 42,000 
42,000 
42,000 

C _{P}_{R} 
kJ/kg ∙ K kJ/kg ∙ K kJ/kg ∙ K 4.18513 4.17658 4.18513 
4.18513 
4.18513 

C _{p}_{j} 
4.17658 
4.17658 

C _{p} 
4.18513 
4.18513 

ID 
m 0.02199 
0.02199 
0.02199 

OD 
m 0.024069 
0.024069 
0.024069 

N _{j} 
–– °C °C °C °C °C °C m ^{3} /h °C °C J/kg kg/m kg/m – – 110 – 30 114 25 1122 1.4 2346.3 1060 1000 ^{3} ^{3} 
– 
3 

T _{1} T _{3} 
– 
34 

110 
110 

T _{6} T _{D} 
– 
28 

30 
30 

T _{S} 
114 
114 

T _{S}_{e}_{a} 
25 
25 

W _{M}_{D} 
1122 
1122 

α _{R} 
1.4 
1.4 

α _{j} λ 
– 
1.78 

2346.3 
2346.3 

ρ ρ _{B} _{W} (b) 
1060 1000 
1060 1000 

Unit 
Value 

Current work capacities MSF plant section 
m ^{3} /d 
26,928 

Reference capacities and ﬂ ow rates Ref. feed rate for intake calculation Ref. blowdown rate for outfall calculations 
m ^{3} /d m ^{3} /d 
1000 

750 

MSF plant cost data Capital recovery factor Parameter Φ in ﬁxed capital cost term Chemicals Energy (pumping) Spares Labor 
0.0963 

5500 

$/m ^{3} prod $/m ^{3} prod $/m ^{3} prod $/m ^{3} prod 
0.024 

0.03 

0.082 

0.1 
positive penalty parameters that are set to realize the satisfaction of all inequality constraints.
3.3. Optimization algorithm
Differential evolution (DE) algorithm has been applied for the non deterministic optimization of NLP process models. Introduced by Storn and Price [30], the DE optimization algorithm is a stochastic population based direct search optimization method that essentially involves the generation of new candidate solutions by combining the parent individ ual and several other individuals of the same population. This is facilitat ed by adding the weighted difference between any two population vectors to a third population vector [13]. Further, the parent vector is re placed with the mutant vector only when the mutant vector provides a better ﬁtness value [17]. Thus, DE is an effective, fast, simple, robust, and inherently parallel technique that has few control parameters and needs less effort to tune and adopt optimization parameters. The DE has been applied as a sequence of mutation, crossover and selection operations for all populations. The maximum number of permitted gen erations has been set as the termination criteria. The efﬁcacy of the DE algorithm has been evaluated by comparing the results obtained from sequential quadratic programming (SQP) method with and without multistart (MS) approach. Further, hybrid optimization approach involving the combination of both DE and SQP was also considered. For this case, the solution generated from DE is being provided as an initial guess value for the SQP method. Thus, the optimization methodology involves the application of either one of DE, SQP, MSSQP and DESQP methods. Table 3 (a) summarizes the
(a) 

Variable 
Unit 
MSFBR 
MSFOT 
MSFM 

Lower 
Upper 
Lower 
Upper 
Lower 
Upper 

W _{M} 
m ^{3} /h m ^{3} /h m ^{3} /h – m – kW/m ^{2} ∙ K kW/m ^{2} ∙ K kW/m ^{2} ∙ K 
2000 
3000 
–– 
–– 

W _{M}_{F} 
– 
– 
5500 
10,000 
– 
– 

W _{M}_{S}_{C} 
–––– 
5500 
10,500 

R _{H} 
0.8 
2 
–– 
–– 

L _{T} 
7 
10 
7 
15 
7 
15 

N _{R} 
18 
30 
15 
50 
15 
50 

U _{B} 
2.7 
3.7 
2.7 
3.7 
2.7 
3.7 

U _{j} 
2.7 
3.7 
–– 
–– 

U _{R} 
2.7 
3.7 
2.7 
3.7 
2.7 
3.7 

(b) 

Constraints 
Unit 
Lower 
Upper 

Brine loading 
m ^{3} /h m width m/s m/s m/s ppm – – – 
900 
1200 

V _{B} 
0.9144 
1.8288 

V _{j} 
0.9144 
1.8288 

V _{R} 
0.9144 
1.8288 

C _{M}_{B}_{D} or C _{l}_{a}_{s}_{t} 
– 
80,000 

ABS (1 − U _{B}_{c}_{a}_{l} / U _{B} ) ABS (1 − U _{j}_{c}_{a}_{l} / U _{j} ) ABS (1 − U _{R}_{c}_{a}_{l} / U _{R} ) 
– 
0.001 

– 
0.001 

– 
0.001 
parametric settings that were deﬁned for the initialization, propagation and termination strategies of various optimization methods for the MSF optimization problem. Further, it shall be noted that for several methods including SQP, MSSQP and DESQP, initial vector values (independent variables) had to be deﬁned for MSFOT, MSFBR and MSFM processes. Table 3(b) summarizes the initial vector values set for the optimization studies. All simulations were conducted in MATLAB programming envi ronment. SQP has been implemented from MATLAB optimization tool box [27,36].
4. Model validation
Model validation precisely refers to the validation of developed code for simulation and optimization models. Since appropriate data has been available only for the MSFBR system, the simulation model code validation has been carried out using the data provided by Helal et al. [3] . Based on the input simulation variable data set of [2790 m ^{3} /h (2,790,000 kg/h), 1.0129, 18, 3.26 kW/m ^{2} ∙ K, 3.443 kW/m ^{2} ∙ K, 2.864 kW/m ^{2} ∙ K, 7.62 m] for [W _{M} , R _{H} , N _{R} , U _{B} , U _{R} , U _{j} , and L _{T} ], the obtained results of dependent variables are [101.7 °C, 32.41 °C, 70,178 ppm, 135.5 m ^{3} /h (135,446 kg/h), 3662.8 m ^{2} , 2515.5 m ^{2} , 3355.4 m ^{2} , 9040 m ^{3} /h (9,040,000 kg/h)] for variables [T _{2} , T _{B}_{D} , C _{M}_{B}_{D} , W _{S} , A _{B} , A _{R} , A _{j} , and W _{M}_{F} ] respectively. Precisely, the same values have been reported in the literature [3] and hence the model code validation is inferred to be successful. The DE algorithm code has been tested for standard optimization model test function such as Rosenbrock Banana Function. For the algo rithm parameters [F, CR, NG, and NP] speci ﬁ ed as [0.8, 0.8, 100 and 100], the DE provided optimal solution of [1.000, 1.000] with a standard deviation of 10 ^{−} ^{6} . The obtained optimal solution is in complete agree ment with that available as the global optimal solution for Rosenbrook (Banana) function [12]. Hence, the DE algorithm is inferred to be effec tive for the optimization of alternate MSF process conﬁgurations.
5. Results and discussions
5.1. Efﬁcacy of the DE algorithm
Fig. 2 depicts the comparative performance of various optimization methods to obtain optimal solutions for the MSFM, MSFOT and MSF
C.S. Bandi et al. / Desalination 394 (2016) 30–43
35
Table 3 (a) Parameters for various optimization techniques. (b) Initial independent variable set data for MSF processes and alternative optimization methods.
(a)
Parameter 
DE 
SQP 
MSSQP 
DESQP 

NP/no. of starting points NG max/function evaluations DEstepsize, F CR Termination criteria 
350 
400 
500 
500 

1000 
700 
1000 
1000 

0.9 
– 
– 
0.9 

– NG max (DE) or 10 ^{−} ^{6} (SQP) 0.9 – 
0.9 

(b) 

Optimization 
MSFBR [ W _{M} , R _{H} , N _{R} , U _{B} , U _{R} , U _{j} , 
MSFM [ W _{M}_{S}_{C} , N _{R} , U _{B} , U _{R} , L _{T} ] 
MSFOT [ W _{M}_{F} , N _{R} , U _{B} , U _{R} , 

method 

L 
T ] 
L 
T ] 

SQP 
[2449, 1.02, 19, 3.5124, 3.1937, 2.9223, 8.51] [2767, 1.17, 18, 3.1673, 3.4619, 3.4454, 8.21] 
[6168, 21, 3.4599, 3.5266, 7.4, 99.21, 
[9543, 26, 

3.4517, 3.6275, 

69,944] 
8.12] 

[6014, 23, 3.1432, 3.5221, 8.12, 103.23, 
[9561, 28, 

3.5454, 

69,021] 
3.5514.7.65] 

[6546, 22, 3.3565, 3.4526, 8.85, 101.45, 
[9756, 27, 

[2439, 1.16,21, 3.3351, 3.2680, 3.5171, 7.12] 
3.4624, 

69,584] 
3.6454.7.69] 

[2566, 1.03, 23, 3. 3513, 3.4259, 
[6059, 25, 3.2452, 3.4411, 7.85, 102.56, 
[9628, 26, 

3.3515, 3.5554, 

3.6637.7.51] 
69,885] 
7.18] 

MSSQP 
[2787, 1.07, 21, 3.4673, 3.4119, 3.4454, 7.61] [2790, 1.05, 22, 3.3624, 3.3003, 3.0448, 7.62] 
[6356, 25, 3.5232, 3.6221, 8.25, 101.23, 
[9643, 28, 

3.3217, 3.3275, 

69,981] 
7.98] 

DESQP 
[5857, 25, 3.7, 3.6719, 9.74, 101.48,70,772] 
[9500, 23, 

3.4108, 3.7398, 

8.62] 
BR plant con ﬁ gurations. It has been evaluated that the optimal fresh water production cost using DE, SQP, MSSQP and DESQP methods are 1.2251, 1.2856, 1.2785 and 1.2534 $/m ^{3} for MSFOT, 1.198, 1.2445, 1.22 and 1.2135 $/m ^{3} for MSFM and 1.0785, 1.1, 1.0852 and 1.0843 for MSFBR respectively. Hence, it can be concluded that the DE provides the lowest fresh water production cost for MSFBR. Based on the obtain ed results, the optimal cost based ranking is MSFBR b MSFM b MSF OT*. For the MSFBR plant, the efﬁcacy of various methods for the opti mization is in the order of DE N SQP N MSSQP* N DESQP*. The solution with asterisk (*) corresponds to solution with penalty (and hence infeasible solution). An infeasible solution for the DESQP is due to rounding of solutions generated from DE, which were then supplied to SQP as initial guess. A summary of the important independent and dependent variable values obtained after optimization of MSFBR, MSFM and MSFOT
Fig. 2. Bar chart depicting the performance of optimization methods for MSF desalination processes.
using various methods is presented in Tables 4, 5 and 6 respectively. In these methods, optimization methods with asterisk indicate the exis tence of penalties in the ﬁnal solutions obtained and hence these solu tions cannot be recommended as appropriate design solutions for the respective MSF process conﬁgurations. For several cases, U _{B} inequality constraint could not be resolved without penalty. Further, the literature data for MSFBR has also been presented in
Table 4 to reﬂect upon the comparative optimality of obtained variable values with those presented in the literature. As outlined in Table 4, in comparison with the available literature data for the MSFBR process, the optimal solution obtained with DE refers to marginal combinations of higher feed ﬂow rate (2794.4 but not 2790.0 m ^{3} /h), higher speciﬁc heat ratio (1.0499 but not 1.0001), lower tube length (7.62 m but not
7.92 m), higher overall heat transfer coefﬁcient in brine heater section
(3.359 kW/m ^{2} ∙ K but not 3.260 kW/m ^{2} ∙ K), lower overall heat transfer coef ﬁ cient in recovery section (3.297 kW/m ^{2} ∙ K but not 3.443 kW/ m ^{2} ∙ K), higher overall heat transfer coef ﬁ cient in brine heater (3.042 kW/m ^{2} ∙ K but not 2.864 kW/m ^{2} ∙ K), higher number of recovery stages (22 but not 18), lower steam ﬂ ow rate (130.365 m ^{3} /h but not 135.446 m ^{3} /h), lower feed ﬂow rate, cooling water and cost. It is further interesting to note that the optimal concentration of rejected stream
leaving the heat rejection section is precisely the same for this case and literature data. Compared to the literature reported optimal cost of 1.104 $/m ^{3} , the DE generated solutions for the MSFBR process to be 2.31% lower (1.0785 $/m ^{3} ), MSFM process to be 7.85% higher (1.198 $/m ^{3} ) and MSFOT process to be 9.88% higher (1.2251 $/m ^{3} ). No relevant data was available in the same literature to compare the solutions obtained for the MSFM and MSFOT process conﬁgurations. It can be observed that marginal improvement in solutions can be obtain ed for the MSFBR plant conﬁguration in comparison with the literature using the DE algorithm. For the MSFBR process con ﬁgurations, solu tions with penalty were obtained using the DESQP method. For the MSFM process model, only DE provided solutions without penalty. Fur ther, all optimization methods can be observed to provide solutions with penalty for the MSFOT conﬁgurations. This conveys the efﬁcacy of the DE algorithm to obtain solutions without penalty (feasible solu tions) for both MSFBR and MSFM conﬁgurations. Hence, it is impor tant to envisage that DE has potential to obtain high quality solutions even for the MSFM process. The best set of optimal decision variable
values are [ W _{M} , R _{H} , L _{T} , U _{B} , U _{R} , U _{j} and N _{R} ] for [2794.4 m ^{3} /h, 1.0499,
7.62 m, 3.359 kW/m ^{2} ∙ K, 3.297 kW/m ^{2} ∙ K, 3.042 kW/m ^{2} ∙ K and 22].
For the MSFBR, the optimal variable values corresponded to pro duction cost, thermal performance, speci ﬁ c heat transfer area and plant recovery of 1.0785 $/m ^{3} , 8.61, 0.009 m ^{2} /h/kg and 0.1287 respec tively. These corresponded to an improvement of about 2.31, 3.90,
2.92 and 20.24% with respect to those reported in the literature [3]
(1.104 $/m ^{3} , 8.283, 0.008487 m ^{2} /h/kg and 0.107 respectively). For the
same case, the optimal feed ﬂ ow rate and cooling water ﬂ ow rate were evaluated as 8720 m ^{3} /h and 5930 m ^{3} /h respectively, which corre spond to a reduction of about 3.53% and 5.2% with respect to the best known optimal values in the literature (9040 m ^{3} /h and 6250 m ^{3} /h respectively). These results indicate marginal improvement in the optimal solutions obtained and thereby convey the competence and ef ﬁ cacy of the DE algorithm for desalination process design and analysis.
5.2. Optimality of other dependent variables
Based on thermodynamics, heat transfer and process economics based insights, several dependent variables have been deﬁned to indi cate upon the optimality of various MSF processes. These refer to specif ic heat transfer area, speciﬁc feed ﬂow rate, speciﬁc cooling water rate and overall plant recovery. De ﬁ nitions and formulae to evaluate the same are presented as follows:
36
C.S. Bandi et al. / Desalination 394 (2016) 30–43
Table 4 Optimal independent and dependent variable values for the MSFBR process and alternate optimization methods.
Variable Unit 
DE 
SQP 
MSSQP* 
DESQP* 
Literature [3] 

Independent variables 
W _{M} 
C _{M}_{B}_{D} m ^{3} /h – m kW/m ^{2} ∙ K kW/m ^{2} ∙ K kW/m ^{2} ∙ K – °C °C ppm m ^{3} /h m m m ^{2} ^{2} ^{2} 
2794.4 
2809.8 
2617.8 
2784.4 
2790.0 

R _{H} 
1.0499 
1.0001 
1.0339 
1.0149 
1.0129 

L _{T} 
7.62 
7.92 
7.98 
7.61 
7.62 

U _{B} 
3.359 
3.362 
3.7 
3.254 
3.260 

U _{R} 
3.297 
3.161 
3.294 
3.436 
3.443 

U _{j} 
3.042 
3.114 
3.687 
2.858 
2.864 

N _{R} 
22 
19 
20 
18 
18 

Dependent variables 
T _{2} 
102.00 
102.03 
102.03 
101.50 
101.70 

T _{B}_{D} 
32.35 
32.38 
32.34 
32.34 
32.41 

70,178 
67,092 
67,093 
70,038 
70,178 

W _{S} 
130.365 
139.860 
129.900 
135.175 
135.446 

A 
3470.5 
3165.7 
3304.4 
3655.47 
3662.8 

_{B} 

A 
3297.1 
3137.9 
2588.7 
2510.47 
2515.5 

_{R} 

A _{j} 
3044.7 
2992.8 
2868.4 
3348.69 
3355.4 

W _{M}_{F} m ^{3} /h 
8721.0 
8732.6 
8856.5 
9021.9 
9040.0 

W _{M}_{C}_{W} 
5926.6 
6257.5 
6238.8 
6237.5 
6250.0 

W _{M}_{B}_{D} m ^{3} /h m ^{3} /h m ^{2} $/m ^{3} 
1672.4 
1353.1 
1495.8 
1662.4 
1668.0 

A _{T} 
85140.8 
74902.1 
63683.6 
58799.61 
59008.8 

Cost 
1.0785 
1.1000 
1.0852 
1.0843 
1.104 
The speci ﬁ c heat transfer area is de ﬁ ned as the ratio of total heat transfer area to the total amount of fresh water produced i.e.
Specific heat transfer area
¼
total heat transfer area
total amount of fresh water produced ^{¼} A
_{W} MD :
ð2Þ
The speciﬁc ﬂow rate is expressed as the ratio of total seawater fed to plant to the total amount of fresh water produced i.e.
Specific feed flow rate
¼
total seawater intake fed to plant
total amount of fresh water produced ^{¼} W MF
W MD
:
ð3Þ
The speci ﬁc cooling water ﬂow rate is de ﬁned as the ratio of total cooling water ﬂow rate of the plant to the total amount of fresh water produced i.e.
Specific cooling water flow rate
¼
total cooling water flow rate of plant
total amount of fresh water produced ^{¼} W MCW
W MD
:
ð4Þ
The overall plant recovery is deﬁned as the percentage of total sea water intake that gets converted to the fresh water i.e.
Overall plant recovery
_{¼} total amount of fresh water produced
total seawater intake fed to plant
100 ¼ ^{W} ^{M}^{D} 100 :
W MF
ð5Þ
Along with thermal performance, Table 7 summarizes the optimal dependent variable values for various MSF processes. The thermal per formance values of MSFOT, MSFM, MSFBR and MSFBR literature data are evaluated to be 10.3457, 7.5918, 8.6066 and 8.2837 respective ly. Fresh water production cost is lowest for MSFBR plant (1.0785 $/ m ^{3} ) and the corresponding water recovery is highest (12.28). For these dependent variables, the DE approach used in this work provided better results than those reported in literature [3] using the Newton– Raphson method (1.10 $/m ^{3} and 10.7 respectively). The speci ﬁc feed ﬂow rate value is also the lowest for MSFBR case (7.77) and this indi cates lower processing cost involved with the feed pretreatment plant.
5.3. Contributions of various cost functions
Fig. 3(a), (b), (c) and (d) presents the pie charts that depict the per centage cost contributions of ICM, DCLIC, ICC, FCC and OMC to the total cost for MSF processes. For the literature data, the pie chart depicting
Table 5 Optimal independent and dependent variable values for the MSFM process and optimization methods.
Variable Unit 
DE 
SQP* 
MSSQP* 
DESQP* 

Independent variables 
W _{M}_{S}_{C} C _{M}_{B}_{D} W _{M}_{F} W _{M}_{B}_{D} m ^{3} /h – m kW/m ^{2} ∙ K kW/m ^{2} ∙ K °C ppm °C ppm m ^{3} /h m m m ^{3} /h m ^{3} /h m ^{2} $/m ^{3} ^{2} ^{2} 
5852.7 
5475.1 
5617.8 
5784.4 

N _{R} 
26 
22 
22 
22 

L _{T} 
8.32 
9.99 
8.15 
9.98 

U _{B} 
3.699 
3.626 
3.7 
3.625 

U _{R} 
3.671 
3.665 
3.499 
3.664 

T _{2} 
101.48 
100.28 
101.42 
100.23 

C _{M}_{F} 
70,772 
56,342 
75,829 
56,314 

Dependent variables 
T _{B}_{D} 
36.52 
37.18 
36.57 
37.16 

70,803 
56,399 
75,905 
56,370 

W _{S} 
147.790 
171.429 
148.434 
171.343 

A 
3661.3 
3907.2 
3496.5 
3905.2 

_{B} 

A 
1964.7 
1992.9 
2700.0 
1991.9 

_{R} 

9727.1 
9745.3 
9755.3 
9740.4 

8605.1 
8623.3 
8633.3 
8618.5 

A _{T} 
54743.5 
47751.0 
62896.5 
47705.2 

Cost 
1.198 
1.2445 
1.22 
1.2135 
C.S. Bandi et al. / Desalination 394 (2016) 30–43
37
Table 6 Optimal independent and dependent variable values for the MSFOT process and optimization methods.
Variable Unit 
DE* 
SQP* 
MSSQP* 
DESQP* 

Independent variables 
W _{M}_{F} C _{M}_{B}_{D} W _{M}_{B}_{D} m ^{3} /h m kW/m ^{2} ∙ K kW/m ^{2} ∙ K – °C °C ppm m ^{3} /h m m m ^{3} /h m ^{2} $/m ^{3} ^{2} ^{2} 
10043.2 
10061.9 
10072.1 
10056.9 

L _{T} 
10.18 
8.0 
11.8 
11.79 

U _{B} 
3.7 
3.410 
3.70 
3.691 

U _{R} 
3.456 
3.468 
3.498 
3.495 

N _{R} 
23 
21 
21 
21 

Dependent variables 
T _{2} 
101.48 
103.6 
103.6 
103.51 

T _{B}_{D} 
31.4 
31.39 
31.4 
31.37 

48,035 
47,624 
47,623 
47584.9 

W _{S} 
153.450 
153.490 
158.450 
157.363 

A 
3053.3 
3094.4 
3285.1 
3249.8 

_{B} 

A 
3530 
4397.0 
3596.9 
3594.0 

_{R} 

1672.4 
1353.1 
1495.8 
1662.4 

A _{T} 
84243.3 
95431.4 
78820.0 
78663.9 

Cost 
1.2251 
1.2856 
1.2785 
1.2534 
the cost contributions of these components is presented in Fig. 3 (d). From Fig. 3(c) and (d), for the MSFBR process conﬁguration, it can be observed that DCLIC is 2% lower for the DE based optimal solution in comparison with that reported in the literature. For the same case, the DE based optimal solution indicated the DCLIC, ICC, OMC and optimal water production cost to be 5.44, 0.70, 22.40 M$/y and 1.0785 $/m ^{3} re spectively. Compared to the literature data [3], these values correspond to a reduction of 12.71, 7.29, 3.22 and 2.31% respectively. Further, with respect to the literature data, it can be also analyzed that the areas of heat rejection and brine heater sections are respectively marginally lower and marginally higher for the heat recovery section for the MSFBR process conﬁguration.
5.4. Optimality of DE algorithm parameters
The effect of DE algorithm parameters (F and CR) on the solution quality was investigated. According to Storn and Price [30] , among F and CR, DE is much more sensitive to F. For optimal DE algorithm perfor mance, they further suggested that F, CR and NP can be set as F ∈ [0,2], CR ∈ [0, 1] and NP = 10 ∗ D , where D is dimensionally of the problem. Typically, F and CR are speciﬁed as 0.9 and 0.8 respectively for engineer ing optimization problems. Thereby, algorithm optimality is critically investigated for the optimality of NG and NP. Signi ﬁ cant tradeoffs exist for the optimality of NG and NP. A very low value of NG and NP may terminate the algorithm before it could reach the global optimal and a very high value of NG and NP may take signiﬁcantly a long time to achieve all solutions very close to the global optimal solution. Solu tion clustering phenomena are typically addressed as the criteria to set optimal values of NP and NG. In this work, NP is varied from 25 to 340 and NG is varied from 50 to 2000 for the optimization of MSF processes. Fig. 4 panels (a) and (b) respectively summarize the variation of total optimal objective function value (including penalty) and total penalty function value with population size (NP = 35 to 340) and gen eration size. As shown, signiﬁcant penalties existed for lower combina tions of NG and NP for the MSFBR process. Solutions without penalty have been achieved using a critical speci ﬁ cation of NG and NP, over
Table 7 Optimality of thermodynamic, heat transfer and cost function variable values for MSF de salination processes.
Performance model 
TPR 
sA, m ^{2} /(kg/h) 
sF 
sW _{C}_{W} 
OPR 
MSFOT* MSFM MSFBR MSFBR literature [3] 
10.34 
0.005867 
8.95 
– 
11.17 
7.59 
0.005014 
8.66 
– 
11.78 

8.61 
0.008745 
7.77 
5.28 
12.87 

8.28 
0.008497 
8.06 
5.57 
10.70 
and above which solutions were found to exclusively cluster around the global optimal domain. For all cases of NP, a critical value of NG =
200 has been evaluated to be relevant to yield optimal solutions without
penalty. Further, above an NG value of 800, almost all solutions remained fairly constant, thus indicating that the solutions generated are very close to the global optimal domain. Further, for few cases of NP and generation size where few best solutions were obtained, addi tional investigations were carried out to evaluate upon the solution
quality in the context of the global optimal domain. For this purpose, standard deviations were evaluated for the obtained solutions. The standard deviation of the solutions obtained for a population and gener ation size of 270 and 800 respectively is 10 ^{−} ^{5} for the best 100 solutions. For this case, the lowest optimal solution of 1.0785 $/m ^{3} has been ob tained. However, for the case of the population and generation size of
100 and 500 respectively, a standard deviation of 10 ^{−} ^{3} was obtained
for the best 100 solutions. The achievement of lower standard deviation for lower combinations of NG and NP in comparison with the higher combinations of NG and NP is expected, given the fact that higher NG and NP combinations facilitate better search of the solution space. A similar explanation could be provided for the results indicated in Fig. 4(c)–(f) for MSFM and MSFOT processes. Thus, the optimal DE al gorithm parameter combinations for MSFBR refer to F, CR, NG and NP values of 0.9, 0.8, 800 and 270 respectively. For the MSFOT process con ﬁ guration, Fig. 4 panels (c) and (d) respectively depict the variation of optimal total objection function value and optimal total penalty value with NP and generation size. The obtained trends are similar to those obtained for MSFBR case and the solutions converged to optimal value for all populations (NPs) for a gen eration size of 800. The results reported for the MSFOT system refer to the existence of penalties for few constraints and the reported values indicate solutions with lowest penalties. A careful analysis of the opti mization results for MSFOT indicated that brine velocity constraint was the violated beyond the speciﬁed upper bound value (6 ft/s). Even tually, with higher brine velocity, the inequality constraint presented as
abs ð1− ^{U} ^{B}^{c}^{a}^{l} Þ−0 : 001 did not get satisﬁed, as U _{B}_{c}_{a}_{l} is a function of brine
U
B
density, which is in turn a function of brine velocity. The penalties for the MSFOT have been successfully eliminated by targeting the follow ing alterations in the parametric and design speciﬁcations:
a) The brine velocity is ﬁxed as 6 ft/s for heat recovery section and the inequality constraint has been speciﬁed to have a constraint param eter value of 10 ^{−} ^{2} but not 10 ^{−} ^{3} . For such a scenario, the minimal fresh production cost has been evaluated using the DE algorithm as 1.1249 $/m ^{3} .
b) The brine velocity is ﬁxed as 6 ft/s and the brine heater width is re duced to 30 ft from 35 ft. No additional changes have been carried out for the abovementioned inequality constraint speciﬁcation. For such a scenario, solutions without penalty were obtained for the
38
C.S. Bandi et al. / Desalination 394 (2016) 30–43
Fig. 3. Piecharts representing the cost contributions of various cost functions to optimal freshwater production cost; (a) MSFOT, (b) MSFM, (c) MSFBR and (d) MSFBR literature data [3].
MSFOT system to indicate the optimal fresh water production cost of 1.1154 $/m ^{3} .
Fig. 4 panels (e) and (f) respectively illustrate the variation of total optimal objective function value (inclusive of penalties) and total opti mal penalty function values for various cases of NP and generation values. The obtained trends are similar to those presented for MSFBR case. Also, it can be observed that for MSFM, solutions without penalty (feasible solutions) have been obtained using the DE algorithm. The MSFOT con ﬁ guration provided infeasible solutions among the three different process conﬁgurations. Using DE, the optimal water produc tion cost for the MSFM process is 1.1980 $/m ^{3} respectively.
5.5. Sensitivity analysis
5.5.1. Effect of feed concentration Sensitivity analysis involves the evaluation the optimal freshwater production cost critical dependence on with various operating parame ters of the MSF process. Typically, feed concentration is varied from 20,000–50,000 ppm for MSF processes [3]. Varying feed concentration in this range, the optimal freshwater production cost was evaluated using DE for MSF processes. Fig. 5(a) summarizes the results obtained for the feed concentration effect on optimal fresh water production cost. It can be observed that the optimal water production cost in creased from 1.2125 to 1.2455, from 1.1876 to 1.2132, from 1.0656 to 1.1051 and from 1.09 to 1.13 $/m ^{3} respectively for MSFOT, MSFM, MSFBR and MSFBR literature processes. From a variation in feed con centration 20,000 ppm to 40,000 ppm, the optimal water production cost varied from 1.0656 to 1.0708 for the MSFBR process which afﬁrms that the cost remained fairly constant. From 40,000–50,000 ppm varia tion in feed concentration, water product cost increased linearly from 1.0708 to 1.1051 for the MSFBR process. The slope of the graph is about 3.433 × 10 ^{−} ^{6} . The insensitivity of the water production cost
with feed concentration up to 40,000 ppm is due to the insigniﬁcant ef fect of feed concentration in inﬂuencing the product ﬂow rates, concen trations and temperature. This might not be the case for reverse osmosis process where feed concentration will have a signiﬁcant effect on the water production cost. Above 40,000 ppm, the feed concentration can be observed to have a signiﬁcant effect on the water production cost. The DE based optimal cost pro ﬁ les lowered by 2.23 – 2.65% than those reported in the literature for the variation in feed seawater con centration. In comparison with the literature data, it can be observed that the lowest cost trends have been obtained for the MSFBR process. This once again conﬁrms the efﬁcacy of DE to obtain high quality solu tions for MSF process optimization.
5.5.2. Thermal performance The thermal performance is deﬁned as the ratio of total amount of fresh water produced to total steam intake of the MSF process:
Thermal performance
_{¼} total amount of fresh water produced
total steam intake fed to plant
W MD
_{¼}
W MS
:
ð6Þ
In general, the thermal performance ratio varies from 6 to 12 for the MSF desalination system and below 1 for the single stage ﬂash desalina tion system [10]. Similarly, the top brine temperature (TBT) is varied from 90 to 100 °C for a variation in total number and thermodynamic loss range of 18–29 and 0.5–2 °C [10]. The variation of thermal performance with TBT for MSF processes is presented in Fig. 5(b). For a variation in TBT from 90 to 110 °C, the ther mal performance ratio varied from 6.66 to 7.31, from 6.75 to 7.59, from 7.26 to 8.61 and from 7.02 to 8.28 for MSFOT, MSFM, MSFBR and MSFBR literature data. It can be observed that the thermal performance is marginally sensitive with TBT for MSFOT and MSFM processes but not MSFBR. Further, it can be observed that the thermal performance
C.S. Bandi et al. / Desalination 394 (2016) 30–43
39
Fig. 4. Effect of DE algorithm parameters (NG and NP) on the solution optimality for various MSF processes — (a) total optimal objective function value for MSFBR, (b) total penalty function value for MSFBR, (c) total optimal objective function value for MSFM, (d) total penalty function value for MSFM, (e) total optimal objective function value for MSFOT and (f) total penalty function value for MSFOT.
increased linearly with increasing TBT and the corresponding slope values are 3.2 × 10 ^{−} ^{2} , 4.0 × 10 ^{−} ^{2} , 6.7 × 10 ^{−} ^{2} , and 6.4 × 10 ^{−} ^{2} respec tively for the said process sequence. Also, in comparison with the liter ature data, it is apparent from the ﬁ gure that the highest thermal performance values were obtained for the MSFBR process using the DE algorithm. The thermal performance of the MSF system has been
varied from 6 to 9, where the DE based optimal costs have been evalu ated to be 3.42–3.99% lower than those reported in the literature.
5.5.3. Chemical cost multiplier Typically, cost multipliers are varied from 0.25 to 2 for MSF desalina tion processes [3] . Fig. 5 (c) presents the variation of optimal water
40
C.S. Bandi et al. / Desalination 394 (2016) 30–43
Fig. 5. Sensitivity of MSF processes with respect to various process and operating parameters; (a) feed concentration, (b) TBT on TPR, (c) chemical cost multiplier, (d) steam cost multiplier. (e) Labor cost multiplier, (f) power cost multiplier, (g) spares cost multiplier, (h) TBT on cost (+[3]).
C.S. Bandi et al. / Desalination 394 (2016) 30–43
41
production cost as a function of chemical cost multiplier for MSF pro cesses. It can be observed that the water production cost increased nonlinearly from 1.1345 to 1.2958, from 1.0866 to 1.12798, from
1.0015 to 1.1365 and from 1.022 to 1.161 $/m ^{3} with increasing chemical
cost multiplier (0.25–2) for MSFOT, MSFM, MSFBR and MSFBR liter ature data. As it is apparent in other sensitivity analyses, the lowest
water production cost trends have been obtained in this work in com parison with the existing data trends of the MSFBR process in the liter ature. The freshwater production cost varied linearly with chemical cost multiplier in the range of 0.25 – 2. For such a case, the corresponding slopes of the trend lines are 0.09, 0.11, 0.071, and 0.072. These slope values are indicative towards water production cost sensitivity with re spect to chemical cost multiplier. Thus the water cost ($/m ^{3} ) will in crease by 9 ¢/m ^{3} , 11 ¢/m ^{3} , 7.1 ¢/m ^{3} , and 7.2 ¢/m ^{3} respectively for a unit increase in chemical cost multiplier for MSFOT, MSFM, MSFBR and MSFBR literature data [3] . For a variation in the cost multiplier from 0.25 to 0.5, the corresponding optimal water cost varied non linearly from 1.1348 to 1.1752, from 1.0866 to 1.1271, from 1.0015 to
1.0495 and from 1.022 to 1.076 $/m ^{3} . A further increase in the chemical
cost multiplier from 0.5 to 2.0 enabled a linear enhancement for the op timal water cost from 1.1752 to 1.2958, from 1.1271 to 1.2798, from
1.0495 to 1.1365 and from 1.076 to 1.161 $/m ^{3} for the said sequence
of processes. Compared to the literature, the obtained optimal cost is
2–2.46% better for variation in the chemical cost multiplier.
5.5.4. Steam cost multiplier
For the steam cost multiplier sensitivity analysis, the steam cost was varied from 0.00104 to 0.00832 $/kg of steam (corresponding to a vari ation of steam cost multiplier from 0.25 to 2.0) at a constant TBT value of 110 °C [3]. The obtained sensitivity analysis based cost trends with re spect to steam cost multiplier are presented in Fig. 5(d). As shown, for
a variation in steam cost multiplier from 0.25 to 2.0, the minimal
water cost varied from 0.8656 to 1.7123, from 0.8202 to 1.687, from
0.7284 to 1.4876 and from 0.739 to 1.5212 $/m ^{3} for MSFOT, MSFM,
MSFBR and MSFBR literature data [3]. The obtained cost trends with respect to steam multiplier indicate a linear variation in water cost with steam multiplier. Such a trend is expected, given the fact that the MSF is a thermal process and its performance is a strong function of steam as a heat source for ﬂash operation. The sensitivity parameters can be obtained from the slope of the obtained data trends. These values have been evaluated correspondingly as 0.48, 0.49, 0.43 and 0.44. Thus, the water cost can be evaluated to increase by 48 ¢/m ^{3} , 49 ¢/m ^{3} , 43 ¢/m ^{3} and 44 ¢/m ^{3} respectively for a unit increase in steam cost for the said se quence of processes. The obtained simulation based trends indicate that MSFBR provides the lowest water cost trends which are placed margin ally below the cost trends reported in the literature [3]. For a variation in steam cost multiplier from 0.2 to 2.0, the DE based optimal cost is 1.43– 2.21% lower than that reported in the literature.
5.5.5. Labor cost multiplier
Fig. 5(e) shows that the optimal water production cost evaluated for MSF processes is sensitive with respect to labor cost multiplier for all processes. As shown, for a variation in labor cost multiplier from 0.25 to 2.0, the minimal water cost varied from 1.1395 to 1.3396, from 1.209 to 1.3008, from 1.0014 to 1.1813 and from 1.0239 to 1.2001 $/
m ^{3} for MSFOT, MSFM, MSFBR and MSFBR literature data [3] .
Hence, signi ﬁ cantly higher cost trends can be observed for MSFOT and MSFM processes but not for MSFBR and MSFBR literature data. The obtained cost trends with respect to labor cost multiplier indicate that the variations are linear with slopes of 0.11, 0.10, 0.10 and 0.10
for MSFOT, MSFM, MSFBR and MSFBR literature [3] respectively. Thus the water cost ($/m ^{3} ) will increase by 11 ¢/m ^{3} , 10.3 ¢/m ^{3} , 10 ¢/
m ^{3} , and 10 ¢/m ^{3} respectively for a unit increase in labor cost multiplier
for the said sequence of processes. The lowest data trends refer to the
data obtained with DE for the MSFBR process. For a variation in labor
cost multiplier from 0.2 to 2.0, the DE based optimal cost is 1.57–2.2% better than that reported in the literature.
5.5.6. Power cost multiplier
For a variation in power cost multiplier from 0.25 to 2.0, Fig. 5(f) presents the variation of optimal water production cost trends for various MSF desalination processes. The observed trends are similar to those obtained for labor cost multiplier. As shown, for a variation in power cost multiplier from 0.25 to 2.0, the minimal water cost varied from 1.1286 to 1.3589, from 1.1139 to 1.3101, from 0.9915 to 1.1921 and from 1.02 to 1.213 $/m ^{3} respectively for MSFOT, MSFM, MSFBR and MSFBR literature data [3]. The corresponding slopes of the obtain ed cost trends are 0.13, 0.11, 0.11 and 0.11. Thus the water cost ($/m ^{3} ) will increase by 12.9 ¢/m ^{3} , 11.2 ¢/m ^{3} , 10.9 ¢/m ^{3} , and 10.9 ¢/m ^{3} respec tively for a unit increase in labor cost multiplier for the said sequence of processes. The lowest cost trend corresponds to that obtained with MSFBR and the DE algorithm in this work. For a variation in power cost multiplier from 0.2 to 2.0, the DE based optimal cost is 1.72 – 2.79% better than that reported in the literature.
5.5.7. Spares cost multiplier
Fig. 5(g) illustrates the variation of minimal water production cost as
a function of spares cost multiplier for alternate MSF processes. As
shown, for the variation in spares cost multiplier from 0.25 to 2.0, the
minimal water cost varied from 1.1456 to 1.3102, from 1.1348 to 1.2823, from 1.0203 to 1.1628 and from 1.0375 to 1.1825 $/m ^{3} for MSFOT, MSFM, MSFBR and MSFBR literature data [3]. The cost trends are similar to those obtained for labor and power cost multipliers. The slopes of the linearized trends are 0.09, 0.08, 0.08 and 0.08 respectively for the said sequence of processes. Thereby, the sensitivity of the spares cost multiplier has been evaluated in terms of an increase by 9.2 ¢/m ^{3} ,
8.4 ¢/m ^{3} , 8.2 ¢/m ^{3} , and 8.2 ¢/m ^{3} respectively for a unit increase in
labor cost multiplier for the said sequence of processes. For a variation in spares cost multiplier, the DE based optimal cost is 1.66% lower
than that reported in the literature.
5.5.8. Effect of top brine temperature
Fig. 5 (h) presents the variation of optimal water production cost with variation in top brine temperature for various cases. As shown, for a variation in TBT from 90 to 110 °C, the costs varied linearly from 1.1051 to 1.2251 and from 1.0845 to 1.1980 $/m ^{3} for MSFOT and MSFM processes. However, for the MSFBR processes (reported in this work and in literature), up to a temperature of 100 °C, the optimal water production cost was not at all affected with variation in TBT. Above 100 °C, the optimal water production cost increased with in creasing TBT. Based on these increasing trends, the slopes of the various
plots have been evaluated (5.9 × 10 ^{−} ^{3} , 5.6 × 10 ^{−} ^{3} , 2.5 × 10 ^{−} ^{3} , and
2.8 × 10 ^{−} ^{3} respectively for the said sequence of processes). Thus, for a
unit increase in top brine temperature, the water cost will increase by 0.59 ¢/m ^{3} , 0.56 ¢/m ^{3} , 0.25 ¢/m ^{3} , and 0.28 ¢/m ^{3} for the said sequence
of processes. Overall, the lowest data trends have been obtained for the MSFBR and DE algorithm case. For a variation in TBT from 95 to 110 °C, the DE based optimal cost is 1.66% lower than that reported in the literature. In summary, the DE based sensitivity analysis of MSFBR enables one to infer that the slope based ranking of various process and
operating parameters is as per the following order: steam cost multipli er (43 ¢/m ^{3} ) N labor cost multiplier (10 ¢/m ^{3} ) N power cost multiplier (8.2 ¢/m ^{3} ) N spares cost multiplier (8.2 ¢/m ^{3} ) N chemical cost multiplier (7.1 ¢/m ^{3} ) N TBT (0.25 ¢/m ^{3} ) N feed concentration (fairly constant). In other words, the optimal freshwater production cost for the MSF process
es is highly sensitive to steam cost multiplier, marginally sensitive with all
other process parameters but not feed concentration and TBT. Since the MSF process is highly energy intensive, the highest sensitivity of the MSFBR process optimal cost with steam multiplier is expected.
42
C.S. Bandi et al. / Desalination 394 (2016) 30–43
6. Conclusions 
MEE 
multieffect evaporator 

MSSQP 
multistartsequential quadratic programming 

Based on the global optimization approach involving differential 
MSF 
multistage ﬂash 

evolution 
algorithm, this work provided signiﬁcant insights and infer 
MSFBR 
brine recycle (BR) multistage ﬂash system (MSF) 
ences with respect to the comparative assessment of alternate MSF pro 
MSFM 
brinemixing (M) multistage ﬂash system (MSF) 

cesses, nondeterministic/deterministic optimization methods and 
MSFOT 
once through (OT) multistage ﬂash system (MSF) 

pertinent 
sensitivity analysis. The modeling approach adopted in this 
NG 
maximum number of generations 
work might refer to design solutions under stringent uncertainty, 
NLP 
nonlinear programming 

given the inability to generate feasible solutions with methods other 
NP 
population size 

than DE. The following conclusions are applicable from the insights de 
OF 
objective function 

duced in this work. 
OPR 
overall plant recovery 

Firstly, DE has been proven to be effective to generate feasible 
ppm 
parts per million 

optimal design variable values for MSFBR, MSFM but not MSFOT pro 
RO 
reverse osmosis 

cesses. Compared to the literature optimal value, DE provided a reduc 
sA 
speciﬁc heat transfer area 

tion of about 2.31% in the optimal freshwater production cost. This is 
sF 
speciﬁc feed ﬂow rate 

due to the identiﬁ cation of better optimal decision variable value set 
sW _{C}_{W} 
speciﬁc cooling feed ﬂow rate 

of [2794.4 m ^{3} /h, 1.0499, 7.62 m, 3.359 kW/m ^{2} ∙ K, 3.297 kW/m ^{2} ∙ K, 
SQP 
sequential quadratic programming 

3.042 kW/m ^{2} ∙ K and 22] for variable set [W _{M} , R _{H} , L _{T} , U _{B} , U _{R} , U _{j} and N _{R} ] re 
TA 
Thomas algorithm 

spectively 
where the optimal freshwater production cost corresponds 
TBT 
top brine temperature 
to 1.0785 $/m ^{3} . Secondly, the deterministic optimization algorithms 
TDM 
tridiagonal matrix 

such as SQP, MSSQP and DESQP could not provide better solution 
TDS 
total dissolved solids 

than the DE. This is primarily due to the dependence of the optimal var 
TPR 
thermal performance 

iable value set and objective function on the initial guess values. Thus, 
USA 
United States of America 

compared 
to other optimization methods, DE would provide better ini 
VC 
vaporcompression 
tialization 
strategies and is expected to serve better for problems with 

greater complexity in terms of decision variables. Another important insight that has been deduced in this work is that 
Symbols 
the MATLAB based optimization toolbox uses default optimization algo rithm parameters and they cannot be as such used for MSF optimization
problems studied in this work. Thirdly, the sensitivity analysis afﬁrmed
that DE based analysis provided 1.41–3.99% better proﬁles than those available in the literature. While such improvement could be regarded to be optimal, it is important to note that the freshwater production cost related improvement is signiﬁcant, given the fact that optimization studies that allow even 1% reduction in water production cost could turn out in terms of a huge amount of savings. Fourthly, inequality con straint resolution appears to be better tackled by DE than any other op timization method. A further resolution of the generated solutions has also been demonstrated in this work i.e., to alter certain design param eter value for chamber width. Fifthly, the chosen literature might be rel atively old in the existing stateoftheart, but the trends obtained in this work appear to be generic to afﬁrm upon the efﬁcacy of DE as the most versatile optimization method to yield feasible solutions under strong conditions of uncertainty. In summary, it is inferred that DE
based optimization is highly effective to obtain feasible global optimiza tion solutions in conjunction with SQP, MSSQP and DESQP. It is antic ipated that DE would be able to provide con ﬁ dence in the solutions generated with complex and hybrid process conﬁgurations involving MSF process con ﬁgurations. This will be addressed in subsequent re search articles.
Nomenclature
Abbreviations
CR 
cross over ratio 
DE 
differential evolution 
ED 
electrodialysis 
F 
mutation factor 
DICOPT++ DIscrete and Continuous OPTimizer
GA 
Genetic algorithm 
GAMS 
general algebraic modeling system 
GOR 
gained output ratio 
GRG 
generalized reduced gradient 
ID 
inside diameter of condenser tubes, m 
IDA 
International Desalination Association 
a coef ﬁ cient to account for using average latent heat of vaporization A _{B} heat transfer area of brine heater, m ^{2}
A _{j}
A
A _{T} total heat transfer area of the MSF process, m ^{2} C _{D}_{C}_{C} direct capital investment, $ C _{D}_{C}_{L}_{I}_{C} direct capital less intake investment, $ C _{I}_{C} annual plant intakeoutfall cost, $ C _{I}_{C}_{C} indirect capital investment, $/y C _{l}_{a}_{s}_{t} concentration of brine stream from last stage of the heat re
jection section, ppm C _{M}_{B}_{D} concentration of reject stream leaving the heat rejection sec tion, ppm C _{M}_{F} feed (seawater) concentration, ppm C _{M}_{R} concentration of recycle stream (from splitter to heat recov ery section), ppm C _{M}_{S}_{M}_{F} concentration of feed stream to the MSFM process, ppm C _{O}_{M}_{C} operating and maintenance cost, $/y C _{p} heat capacity, kJ/kg ∙ K C _{P}_{R} avg. speciﬁc heat capacity, recovery section, kJ/kg ∙ K C _{p}_{j} avg. speciﬁc heat capacity, rejection section, kJ/kg ∙ K ID inside diameter of condenser tubes, m L _{T} tube length, m M _{1} mixer 1 N _{j} number of rejection stages N _{R} number of stages in heat recovery stages OD outside diameter of condenser tubes, m R _{H} speciﬁc heat ratio (W _{M}_{R} C _{p}_{R} /W _{M}_{F} C _{p}_{j} ) S _{1} _{–} _{2} splitter 1–2 respectively sW _{C}_{W} speciﬁc cooling water ﬂow rate T _{1} temperature of reject coolant stream in the MSFBR process, °C T _{2} temperature of brine stream entering the brine heater in the MSF process, °C T _{3} top brine temperature, ^{o} C T _{4} temperature of brine stream leaving last stage in the heat recovery section of the MSFBR process, °C
heat transfer area of the rejection section, m ^{2} heat transfer area of recovery section, m ^{2}
_{R}
C.S. Bandi et al. / Desalination 394 (2016) 30–43
43
T _{5} temperature of brine stream leaving last stage in the heat rejection section of the MSFBR process, °C T _{6} temperature of recycle stream in the MSFBR process, °C T _{B}_{D} temperature of reject stream in the MSFBR process, °C T _{D} temperature of distillate product stream in the MSF process, °C T _{S} steam temperature, °C T _{S}_{e}_{a} seawater temperature/feed temperature, °C U _{B} overall heat transfer coefﬁcient in brine heater, kW/m ^{2} ∙ K U _{j} overall heat transfer coef ﬁ cient in the rejection section, kW/m ^{2} ∙ K U _{R} overall heat transfer coefﬁcient in recovery section, kW/m ^{2} ∙ K U _{B}_{c}_{a}_{l} calculated overall heat transfer coef ﬁ cient in brine heater, kW/m ^{2} ∙ K U _{j}_{c}_{a}_{l} calculated overall heat transfer coef ﬁ cient in the rejection
section, kW/m ^{2} ∙ K U _{R}_{c}_{a}_{l} calculated overall heat transfer coefﬁcient in recovery section, kW/m ^{2} ∙ K V _{B} brine velocity in brine heater, m/s V _{j} brine velocity in the rejection section, m/s V _{R} brine velocity in recovery section, m/s
ﬂow rate of ﬂashing brine stream leaving the last stage of the heat rejection section in the MSFM and MSF processes, m ^{3} /h
W _{l}_{a}_{s}_{t}
W _{M} ﬂow rate of makeup stream in the MSF process, m ^{3} /h W _{M}_{B}_{D} ﬂow rate of rejected stream in the MSF process, m ^{3} /h
Molto più che documenti.
Scopri tutto ciò che Scribd ha da offrire, inclusi libri e audiolibri dei maggiori editori.
Annulla in qualsiasi momento.