Desalination 478 (2020) 114286

Contents lists available at ScienceDirect

Desalination
journal homepage: www.elsevier.com/locate/desal

Risk evaluation of large-scale seawater desalination projects based on an T

integrated fuzzy comprehensive evaluation and analytic hierarchy process
method
Yin Zhang, Ruihao Wang, Pengfei Huang , Xiaoli Wang, Shenghui Wang
⁎

The Institute of Seawater Desalination and Multipurpose Utilization, MNR (Tianjin), Tianjin 300192, PR China

ARTICLE INFO ABSTRACT

Keywords: Desalination projects play a vital role in the water supply of coastal regions with scarce water resources. The
Desalination project risks associated with desalination projects are worth investigating, especially for large-scale projects. This paper
Risk identification presents the risk identification and evaluation processes of large-scale desalination projects. Two levels of risk
Risk evaluation indicators are identified and the first-level risks include water intake and outfall risk, processing risk, financial
Fuzzy comprehensive evaluation
risk and circumstance risk. With the identified risk indicators, an integrated fuzzy comprehensive evaluation
Analytic hierarchy process
(FCE) and analytic hierarchy process (AHP) method is introduced to conduct quantitative risk evaluations for
large-scale desalination projects. Twenty experts in desalination-related fields are invited to vote to determine
the weighting vectors for the FCE through the AHP. They also participate in deciding the membership matrixes
in the FCE for three practical desalination projects. The evaluation results indicate that the overall risks of all the
considered projects are at the “Very low” level. Finally, to diminish the potential risks, several instructions and
recommendations are suggested that depend on the evaluation outcomes. It is expected that the current risk
evaluation research will make remarkable contributions to the risk management and control of large-scale de-
salination projects and further promote the development of the desalination industry.

1. Introduction membrane cleaning of membranes used for filtration were highlighted

by Nghiem and Schäfer [7]. Endocrine-disrupting chemicals (EDCs)
As global urbanization, industrialization and population growth were the particular concern, and the results revealed that the mem-
continue to accelerate, the demand for a reliable water supply has be- brane could be considered an enormous reservoir for EDCs and that a
come a serious global concern. Desalination is broadly acknowledged as high concentration of EDCs might exist in the spent cleaning solution.
a promising approach to deal with the problem of water resource Stewart [8] presented a supply risk decision support tool for water al-
shortage [1–3]. Large-scale desalination plants contribute to the ma- location strategies, which combined a sensitivity analysis with a water
jority of global desalination production [4]. Due to the complex tech- balance model. The proposed tool was proven suitable for estimating
nology and possible uncertainties, these plant projects generally require the risks of future over-supply or under-supply scenarios in an opera-
an enormous investment and long construction time, possibly resulting tional environment. Baggett et al. [9] investigated the discrepancy in
in various risks to desalination projects throughout the life cycle. risk perceptions from different stakeholders when facing water reuse
Several scholars have studied the possible risks of water treatment planning through a questionnaire survey. An obvious variation in the
or supply in different applications, including seawater desalination, expected levels of agreement regarding risk was observed in the survey,
water reuse and stormwater reuse. Habib and Fakhral-Deen [5] con- and some pronounced variability was also noticed. Salgot et al. [10]
ducted a risk assessment on a widely applied desalination plant mate- proposed the specific microbial and chemical limits for various re-
rial, namely copper alloys, under a polluted marine environment with claimed water quality categories and suggested additional microbial
oil spills and oil fires. Furthermore, Habib et al. [6] carried out a similar and chemical risk assessments. Zafra-Cabeza et al. [11] introduced a
assessment on the susceptibility of copper alloy to stress corrosion risk management algorithm for reverse osmosis plants through a mixed
cracking in sulfide-polluted seawater. The potential risks of the treat- integer quadratic programming formulation. Several common risks
ment and disposal processes of the concentrate and wastewater from were identified and corresponding mitigation actions were suggested.

⁎
Corresponding author.
E-mail address: [email protected] (P. Huang).

https://doi.org/10.1016/j.desal.2019.114286
Received 10 October 2019; Received in revised form 12 December 2019; Accepted 16 December 2019
Available online 23 December 2019
0011-9164/ © 2019 Elsevier B.V. All rights reserved.
Y. Zhang, et al. Desalination 478 (2020) 114286

Eriksson et al. [12] performed a risk assessment of stormwater in terms depicted in Fig. 1. The possible risks of large-scale desalination projects
of xenobiotics. The main purpose of the assessment was to identify the are classified first. With the risk identification results, several FCE steps
most relevant xenobiotics among the numerous xenobiotics for the are processed to attain the ultimate risk evaluation consequence. In
determination of appropriate source control options for stormwater. particular, the AHP procedure with a few steps is utilized for de-
Chang [13] summarized previous literature on the ecological risk and/ termining the weighting factor of the FCE procedure. Detailed de-
or impact assessment of desalination plants and proposed several ne- scriptions of the entire FCE-AHP procedure are introduced in this sec-
cessary characteristics for a potential ecological indicator. Troldborg tion.
et al. [14] performed a quantitative risk assessment to evaluate the
potential damage to human health from reclaimed irrigation water. The 2.1. Risk identification of large-scale desalination projects
assessment was demonstrated to be effective establishing the quality
standards of reclaimed water considering the possible harm caused by According to the literatures and the extensive experience of the
both microbial and chemical hazards. Goodwin et al. [15] explored the Institute of Seawater Desalination and Multipurpose Utilization, MNR
challenges of risk management for a non-potable water reuse project (Tianjin) (ISDMU), the potential risks that might be encountered during
through a case study in London. The results showed that collaboration the large-scale desalination project are classified into four categories,
and learning opportunities were feasible approaches to enhance risk namely, the water intake and outfall risk, processing risk, financial risk
management performance. and circumstance risk.
Summarizing the reported researches, previous studies on the de-
salination project risk only focus on several certain aspects, yet very few
studies have provided a comprehensive and thorough risk assessment 2.1.1. Water intake and outfall risk
method for a desalination project, especially for large-scale ones. In this Water intake and outfall are important components of the entire
paper, the characteristics of large-scale desalination projects are ana- desalination project. Water intake is the first step of the desalination
lyzed, and the risk factors are identified. A risk evaluation system of process, with the aim of guaranteeing sufficient and continuous raw
large-scale desalination project based on the integrated fuzzy compre- water supply for the whole desalination process, while the brine outfall
hensive evaluation-analytic hierarchy process (FCE-AHP) method is is the terminal segment of the desalination process. The detailed factors
established and adopted to assess three different kinds of practical that should be contemplated when evaluating the intake risk or outfall
large-scale desalination projects. Several suggestions and re- risk include the following:
commendations are presented depending on the evaluation results. The
proposed evaluation system can provide useful references for the de- (1) The risk of changes in raw water quality
velopment of desalination, especially in the industrial practice aspect.
The raw water quality refers to the physical, chemical and biological
characteristics of raw water for the desalination process, which are
2. Evaluation method reflected by indexes such as total dissolved solids (TDS), pH, tem-
perature and chemical oxygen demand (COD). The pre-treatment and
To create meticulous risk assessments of large-scale desalination desalination processes are closely associated with raw water quality,
projects, an integrated FCE-AHP method is applied in the proposed and the change in raw water quality leads to an unpredictable influence
study. A flowchart of the integrated FCE-AHP method procedure is on fresh water production.

Fig. 1. Flowchart of the integrated FCE-AHP method procedure.

2
Y. Zhang, et al. Desalination 478 (2020) 114286

(2) The risk of the intake scheme is categorized into financing risk, cost risk and profitability risk in the
present study.
The determination of the intake approach and the construction of
the intake system have a critical impact on the economic performance (1) Financing risk
of fresh water production, the operational stability of desalination
plants and the marine environment near desalination projects [16]. An Large-scale desalination projects require vast capital and a long-
appropriate construction of an intake system requires deep hydro- term need for funds. Hence, an irrational financing scheme or an im-
logical, geological and climatic surveys before deciding the specific practical loan plan may bring insufficient capital or have excessive
intake technology. The capital investment, plant scale, water quality capital costs, thus constituting the financing risk [19].
requirement, material corrosion and biofouling should be considered
thoroughly during the design period of the intake system. (2) Cost risk

(3) The risk of the outfall scheme Numerous factors account for the cost fluctuation of desalination
projects, which primarily consist of material costs, fuel and power costs,
The negative effects of concentrated brine on the adjacent en- salary and welfare costs, maintenance costs, management costs and so
vironment is a prime concern for desalination projects and they can be on. In addition, changes in project standards, design or progress in-
predominantly mitigated with pertinent planning of outfalls [17]. crease the potential cost risk.
Outfall planning usually encompasses the decision of outfall location
and configuration [18] on the basis of design regulations, environ- (3) Profitability risk
mental protection policies and geological conditions.
A profitability risk exists in the construction and operating periods
2.1.2. Processing risk of desalination projects, which may lead to the deterioration or even
The processing risk, which refers to the risk caused by the desali- failure of the project economic performance. The causes relating to the
nation process, is one of the major risks of the desalination project. The profitability risk involve the delay of project construction, suspension of
processing risk is assumed to be comprised of the design risk, the project operation and the overproduction of fresh water.
construction risk, the risk of an imperfect municipal water network and
the risk of equipment compatibility in the present study. 2.1.4. Circumstance risk
The circumstance risk of large-scale desalination plants proposed in
(1) Design risk the current study principally contains natural risk, policy risks, social
risks and energy supply risks.
An inaccurate design may cause adverse consequences for the de-
salination project. Therefore, a meticulous design of the desalination (1) Natural risk
process is crucial for the desalination project, which should compre-
hensively cover the project site, geological and hydrological survey The natural risk represents the risk of force majeure such as flood,
results and stakeholder requests. typhoon (hurricane), tornado, earthquake, tsunami and debris flow.
These catastrophes will likely cause severe damage to desalination
(2) Construction risk plants (either constructed or under construction) and affect the clean
raw water supply [20–22].
The construction risk is the risk occurring during the construction
period, which can be caused by improper construction technology, in- (2) Policy risk
correct operation by the construction workers, delay of the construction
schedule and so on. The policy risk is indicated by administrative and legislative vari-
abilities. In many countries and regions, desalination projects are en-
(3) The risk of an imperfect municipal water network couraged and supported by incentive polices including direct financial
aid, subsidized loans and tax concessions [23,24]. However, since de-
The majority of global desalination plants produce fresh water for salination projects usually have long construction and operation dura-
municipal purposes [4]. For metropolitan municipal sectors with few tions, adjustment or even reversal of supporting policies may happen.
experiences dealing with desalinated water, the municipal water net- With this change, the desalination project will face possible losses in
work might face a corrosion problem with desalinated water entering terms of construction and management.
the existing water supply pipes in view of the chemical discrepancy
between the desalinated water and regular municipal water. (3) Social risk

(4) The risk of equipment compatibility To a great extent, social risk reflects the social demand of desali-
nated water and the attitude of the public toward desalinated water.
The desalination equipment is the decisive factor of the whole The social demand of desalinated water depends on the condition of the
project, and it directly determines the plant efficiency, project eco- local water resource shortage and the accessibility of a desalinated
nomic feasibility and energy consumption performance. It is worth water supply. Moreover, the public consciousness and attitude toward
mentioning that for many countries and regions, desalination equip- desalination are significant for desalination projects [25]. Currently,
ment is imported from other countries, and thus, the incompatibility there is still public concern or doubt regarding desalination production,
risk between the various setups should also be considered for equip- especially the safety of drinkable desalinated water. This can be as-
ment risk. cribed to the inadequate amount of desalination propaganda.

2.1.3. Financial risk (4) Energy supply risk

The financial aspect of the desalination project should be considered
carefully [19]. The financial risk refers to the risk of a declining ex- For the three most dominant desalination technologies, a heat
pected yield of the desalination project due to financial reasons, which supply is necessary for MED and MSF, while an electricity supply is

3
Y. Zhang, et al. Desalination 478 (2020) 114286

Fig. 2. The risk identification and classification of large-scale desalination projects.

indispensable for RO [26,27]. Therefore, the instability, interruption or or quantified value, which is utilized to describe the grading of in-
even cessation of an energy supply will disturb the operation of desa- dividual evaluation indicators. The judgment set of the present study is
lination plants. composed of five criteria, namely Very low, Low, Moderate, High and
The relationships among all the mentioned risks with their in- Very high, which are denoted by v1, v2, v3, v4 and v5 respectively.
dividual symbols are depicted in Fig. 2.
(3) Determination of the membership matrix R for first-level indicators
2.2. The integrated FCE-AHP method
The membership matrix R of the first-level indicators can be as-
The integrated FCE-AHP method is widely adopted in numerous certained as:
research fields to develop assessment criteria, such as mineral pro- r11 … r1n
spectivity mapping [28], acoustical environment [29], teaching per- R=
formance evaluation [30,31], coastal reclamation suitability assessment rm1 rmn (1)
[32], coal stockpiles [33], human resource management [34], mine
n
safety training [35], and sustainable energy planning [36]. In the cur-
rij = 1, i = 1, 2, …, m
rent study, the integrated FCE-AHP method is utilized to conduct a risk j=1 (2)
assessment for large-scale desalination projects.
where m and n denote the amounts of corresponding second-level in-
2.2.1. Establishment of the FCE model dicators and judgment criteria, respectively. The element of R, rij, re-
The FCE theory was firstly presented by Zadeh [37,38], the essence presents the possibility of appraising the ith second-level indicator with
of which is to decide the value of an evaluation objective element the jth criterion. In the current study, the membership matrixes of the
through the membership matrix and factor weightings. The model es- four first-level indicators (R1, R2, R3 and R4) are calculated and nor-
tablishment of FCE complies with the following steps in the current malized through expert consultation for the considered desalination
study: projects.

(1) Determination of the evaluation indicator set U (4) Determination of the weighting factor W

The two specific levels of identified risks in section 2.1 are the exact The weighting vector refers to the relative prominence degree of the
evaluation indicators. The first-level and second-level indicators have first-level evaluation indicator. In the current study, the weighting
been labelled in Fig. 2, which illustrates the structure of the indicators vector W = {w1, w2, w3, w4} is calculated through the AHP, which is
as well. introduced later in this paper. Each single element of W contains several
second-level evaluation indicators, and the secondary indicators are
(2) Foundation of the judgment set V also decided by the AHP.

The element of judgment set V = {v1, v2, …, vn} can be a qualitative (5) Determination of the comprehensive grading vector B

4
Y. Zhang, et al. Desalination 478 (2020) 114286

The comprehensive grading vector represents the eventual com- Table 2

prehensive result of the risk assessment based on fuzzy arithmetic. The Random index values for different scales.
fuzzy arithmetic principle for the comprehensive grading vector can be Scale 1 2 3 4 5 6 7 8 9
expressed as
RI 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.45
B=W R (3)
where ° denotes all kinds of fuzzy operators. The fuzzy operators have
various calculation modes and the frequently used modes include M The consistency index (CI) is used to measure the consistency
(∧,∨), M (∨, ●), and M (●,⊕). In the current study, the M (●,⊕) mode among pairwise comparisons and can be calculated by the following
is adopted because each evaluation indicator affects the desalination equation:
project [30]:
max n
CI =
m m n 1 (8)
bj = wi·rij = min 1, wi·rij , j = 1, 2, …, n
i=1 i=1 (4) where λmax is the eigenvalue and n is the number of indicators in the
pairwise comparison matrix problem.
Substituting the membership matrix R in Eq. (3) by R1, R2, R3 and
The consistency ratio CR is utilized to judge the consistency of the
R4, and replacing W in Eq. (3) by the practical weighting factors w1, w2,
matrix and can be expressed by:
w3 and w4 successively, the comprehensive grading vectors of the four
first-level indicators (B1, B2, B3 and B4, respectively) can be obtained. CR =
CI
The ultimate comprehensive grading vector of the entire evaluation can RI (9)
be acquired by where RI is the random index and is given in Table 2 according to Saaty.
B = W [B1 , B2 , B3 , B4 ]T (5) Generally, the upper limit of the satisfactory range of CR is 0.1.
In line with the AHP procedure, twenty experts were invited to
support the rigorous consultation process. As shown in Table 3, the
2.2.2. Utilization of the AHP to determine weighting vectors
experts include senior engineers from ISDMU, researchers from well-
The weighting vectors for the first-level and second-level evaluation
known universities, senior executives of construction companies that
indicators are obtained through the AHP algorithm, which was initially
have desalination projects and the managing executives representing
put forward by Saaty [39–41]. The analytic hierarchy process (AHP) is
desalination plant owners. The comprehensive coverage of various as-
an analytical method that can characterize the significance of applic-
pects of desalination projects ensures the reliability of expert con-
able factors with natural numbers between 1 and 9. The detailed AHP
sultants.
procedures are described as follows:
All the questionnaires from the experts are valid, and the calculated
weightings of all the first-level and second-level indicators are verified
(1) Construction of the hierarchical structure
since they passed the consistency check. The arithmetic means of all the
indicator weightings given by the experts are assumed to be the ade-
In this step, the influential factors of the problem are determined
quate weightings, as shown in Table 4.
and logically decomposed into several levels. In the current study, the
evaluation indicators are the very influential factors, which have been
identified and classified into two levels in section 2.1. The orderly 3. Case study
hierarchical structure of all the evaluation indicators is described in
Fig. 2. With the risk identification and integrated FCE-AHP method in-
troduced in section 2, the risk assessments are performed for three
(2) Establishment of the pairwise comparison matrix large-scale desalination projects. The assessment results are compared
among the three projects, and some suggestions as well as re-
With the construction of the hierarchical structure accomplished, a commendations are proposed in terms of the assessment results.
pairwise comparison matrix A of each evaluation indicator, is generated
by 3.1. Brief information of the desalination projects

A = (aij )k × k , i = 1, 2, …, k, j = 1, 2, …, k (6) Three large-scale desalination projects in China are selected as the
The term aij refers to a quantified judgment for a pair of indicators case study objectives, which represent three typical uses of desalination
of the same level (either first-level or second-level for the current study) plants.
and the value of aij is expressed by Saaty's 1–9 point scale method de- The desalination project of the Beijiang power station (abbreviated
scribed in Table 1 and adheres to the following parameters: as Project A) is representative of desalination projects for municipal
purposes, and it is located in Hangu Subdistrict, Binhai New District,
aij > 0, aij = 1 aji (i j), aij = 1(i = j ) Tianjin, China (39.2°N, 117.9°E). The desalination plant of Project A
(7)
belongs to Tianjin SDIC Jinneng Electric Power Co., Ltd. and has pro-
(3) Consistency check duced fresh water since 2009. The present capacity of Project A is ap-
proximately 20,000 t/d, and the desalinated water is distributed into
the municipal network of the Binhai New District for the residential
Table 1 water supply. The desalination technology of Project A is multi-effect
Saaty's 1–9 point scale. distillation and all the steam is generated by the nearby Beijiang power
Point scale Reciprocal Description
station.
The Xinquan desalination project (abbreviated as Project B) located
1 1 Equally preferred in Dagang Subdistrict, Binhai New District, Tianjin, China (38.8°N,
3 1/3 Moderately preferred 117.5°E) produces desalinated water for industrial purposes. The an-
5 1/5 Strongly preferred
ticipated capacity of Project B is 10,000 t/d, and the desalinated water
7 1/7 Extremely preferred
9 1/9 Completely preferred is transported to a chemical factory that produces ethylene. Reverse
osmosis desalination technology is adopted in Project B. The electricity

5
Y. Zhang, et al. Desalination 478 (2020) 114286

Table 3
Background of the consulted experts.
Affiliation Number of experts Professional level

ISDMU 8 Six senior engineers and two engineers

University 4 Three professors and one vice professor
Construction company 4 Three senior engineers and one engineer
Project owner 4 Two senior economists and two senior engineers

Table 4
AHP results of the present study.
Target level First-level indicator Weighting Second-level indicator Weighting Comprehensive weighting

Risk evaluation of large-scale desalination project u1 0.1002 u11 0.5619 0.0563

u12 0.2506 0.0251
u13 0.1875 0.0188
u2 0.1617 u21 0.3939 0.0637
u22 0.3381 0.0547
u23 0.1473 0.0238
u24 0.1207 0.0195
u3 0.2599 u31 0.1977 0.0514
u32 0.3696 0.0961
u33 0.4327 0.1125
u4 0.4782 u41 0.4843 0.2316
u42 0.1997 0.0955
u43 0.1704 0.0815
u44 0.1456 0.0696

from the power grid and warm seawater ejected from the cooling indicators are expressed in Eqs. (10)–(13):
towers of nearby power plants are the driving energy and raw water
origin of Project B, respectively. It should be noted that the chemical 0.45 0.40 0.10 0.05 0.00
R1A = 0.30 0.45 0.20 0.05 0.00
factory and the desalination plant belong to the same company.
0.25 0.30 0.35 0.10 0.00 (10)
The desalination project of Yongxing Island (abbreviated as Project
C), Sansha, Hainan Province, China (16.8°N, 117.5°E) is an off-grid
0.75 0.20 0.05 0.00 0.00
project on a Pacific island. The reverse osmosis desalination plants of 0.35 0.45 0.20 0.00 0.00
Project C provide domestic water for residential and military facilities R2A =
0.10 0.15 0.40 0.25 0.10
with a capacity of 1000 t/d. The operation of the desalination plants is 0.40 0.35 0.25 0.00 0.00 (11)
supported by a set of diesel generators and fresh water flows inside
newly built distributing pipes. 0.60 0.35 0.05 0.00 0.00
R3A = 0.30 0.40 0.25 0.05 0.00
3.2. Comparison of the evaluation results of the desalination projects 0.65 0.35 0.00 0.00 0.00 (12)

All the experts voted on the judgment of the second-level indicators 0.15 0.40 0.35 0.10 0.00
after the project information was presented to them. The experts judged 0.35 0.45 0.20 0.00 0.00
R 4A =
0.20 0.30 0.25 0.15 0.10
the indicators by the judgment set introduced in section 2.2, and the
0.60 0.35 0.05 0.00 0.00 (13)
voting outcome of project A is assembled in Table 5.
The normalized membership matrixes of the four first-level Using Eqs. (3)–(4), Eqs. (10)–(13) and the weightings of first-level
indicators, the comprehensive grading vectors of the four first-level
Table 5 indicators are calculated as follows:
Voting outcome for the judgment level of Project A.
B1A 0.3749 0.3938 0.1719 0.0594 0.0000
Indicator Number of votes for the judgment levels
B2A 0.4768 0.2953 0.1764 0.0368 0.0147
=
v1 v2 v3 v4 v5 B3A 0.5108 0.3684 0.1023 0.0185 0.0000
B4A 0.2640 0.3857 0.2593 0.0740 0.0170 (14)
u11 9 8 2 1 0
u12 6 7 4 2 1
Combining Eq. (5), Eq. (14) and the weightings of the second-level
u13 5 6 4 3 2
u21 15 4 1 0 0
indicators, the ultimate normalized comprehensive grading vector of
u22 7 9 4 0 0 Project A is obtained as follows:
u23 2 3 8 5 2
u24 8 7 5 0 0 BA = [0.3737 0.3674 0.1964 0.0521 0.0105] (15)
u31 12 7 1 0 0
u32 6 8 5 1 0 Similarly, the judgment voting results, normalized membership
u33 13 7 0 0 0 matrixes of the four first-level indicators, comprehensive grading vec-
u41 3 8 7 2 0 tors of the four first-level indicators as well as the ultimate normalized
u42 7 9 4 0 0
comprehensive grading vector of Project B and Project C can be ac-
u43 4 6 5 3 2
u44 12 7 1 0 0 quired, which are listed in Table 6, Table 7, Eqs. (16)–(21) and Eqs.
(22)–(27), respectively.

6
Y. Zhang, et al. Desalination 478 (2020) 114286

Table 6 0.80 0.15 0.05 0.00 0.00

Voting outcome for the judgment level of Project B. R2C =
0.35 0.40 0.25 0.00 0.00
0.85 0.15 0.00 0.00 0.00
Indicator Number of votes for the judgment levels 0.60 0.25 0.15 0.00 0.00 (23)
v1 v2 v3 v4 v5
0.65 0.35 0.00 0.00 0.00
u11 10 7 2 1 0 R3C = 0.50 0.40 0.05 0.05 0.00
u12 6 10 3 1 0 0.70 0.30 0.00 0.00 0.00 (24)
u13 5 6 7 2 0
u21 14 5 1 0 0 0.20 0.45 0.30 0.05 0.00
u22 8 8 4 0 0 0.60 0.35 0.05 0.00 0.00
u23 17 3 0 0 0 R 4C =
0.50 0.40 0.10 0.00 0.00
u24 10 5 5 0 0 0.15 0.20 0.35 0.15 0.15 (25)
u31 9 7 2 1 1
u32 6 8 4 1 1
u33 7 6 5 1 1
B1C 0.5717 0.3532 0.0627 0.0125 0.0000
u41 3 7 8 2 0 B2C 0.6311 0.2466 0.1223 0.0000 0.0000
=
u42 7 7 4 1 1 B3C 0.6162 0.3468 0.0185 0.0185 0.0000
u43 8 6 5 1 0 0.3237 0.3851 0.2233 0.0461 0.0218
B4C (26)
u44 11 7 2 0 0

BC = [0.4743 0.3496 0.1376 0.0281 0.0104] (27)

Table 7 With the principle of maximum membership, the overall risks of all
Voting outcome for the judgment level of Project C. three projects are at the “Very low” level. The degrees of “Very low” for
Indicator Number of votes for the judgment levels Project A, Project B and Project C are 0.3737, 0.3672 and 0.4743 in-
dividually. In terms of first-level risks, Project A has two “Very low”
v1 v2 v3 v4 v5 degrees and two “Low” degrees while Project B and Project C both have
three “Very low” degrees and one “Low” degree. The various char-
u11 13 7 0 0 0
u12 6 8 5 1 0 acteristics of the projects can be attributed to the differences in their
u13 14 6 0 0 0 locations, scales and ownerships.
u21 16 3 1 0 0 For the water intake and outfall risk, it can be revealed from
u22 7 8 5 0 0 membership matrix B1A, B1B and B1C that the risks of Project A and
u23 17 3 0 0 0
u24 12 5 3 0 0
Project B are worse than that of Project C. The reason for this situation
u31 13 7 0 0 0 is that the locations of Project A and Project B are at the edge of the
u32 10 8 1 1 0 Bohai Sea, which suffers from severe pollutant discharge of adjacent
u33 14 6 0 0 0 industrial enterprises and is considered as one of the most contaminated
u41 4 9 6 1 0
coastal areas in China [42–44]. In contrast, Project C is located in the
u42 12 7 1 0 0
u43 10 8 2 0 0 Xisha Islands, which is far from the mainland of China and there are no
u44 3 4 7 3 3 large industrial activities. Therefore, the inlet seawater is of high
quality for the desalination process of Project C and the water intake
risk is naturally lower. On the other hand, the desalination plant ca-
0.50 0.35 0.10 0.05 0.00 pacity of Project C is only approximately 1% of the capacity of Project B
R1B = 0.30 0.50 0.15 0.05 0.00 and Project A. Thus, the influence of desalination outfall on the sur-
0.25 0.30 0.35 0.10 0.00 (16)
rounding environment of Project C is much smaller than that of the
other two projects.
0.70 0.25 0.05 0.00 0.00 In terms of the processing risk, the only noticeable factor is the risk
0.40 0.40 0.20 0.00 0.00 of an imperfect municipal water network for Project A, since Project A
R2B =
0.85 0.15 0.00 0.00 0.00
0.50 0.25 0.25 0.00 0.00 (17) provides water for the municipal network. Part of the municipal net-
work has been built for years and concern from some experts about this
risk is reasonable. Otherwise, there is no other serious concern in terms
0.45 0.35 0.10 0.05 0.05
of the processing risk for all the projects because the design and con-
R3B = 0.30 0.40 0.20 0.05 0.05
0.35 0.30 0.25 0.05 0.05 (18) struction groups responsible for the projects are trustworthy.
Regarding the financial risk, the situations for Project A and Project
C are more satisfactory than that of Project B, which can be deduced
0.15 0.35 0.40 0.10 0.00
0.35 0.35 0.20 0.05 0.05 from membership matrices B3A, B3B and B3C. This can be explained by
R 4B = the different ownership properties of the projects. Project A and Project
0.40 0.30 0.25 0.05 0.00
0.55 0.35 0.10 0.00 0.00 (19) C are owned by the government, and Project B is owned by a private
company. The constructions of Project A and Project C are government-
B1B funded to guarantee the residential water supply and the profit of the
0.4030 0.3782 0.1594 0.0594 0.0000
B2B projects is not the priority of the government. In contrast, the financial
0.5965 0.2860 0.1175 0.0000 0.0000
B3B
=
0.3513 0.3468 0.2019 0.0500 0.0500 reliability of a private company cannot be as stable as the government
0.2908 0.3415 0.2908 0.0669 0.0100 in China. Moreover, a robust and continuous profit is significant for the
B4B (20)
owner of Project B. Hence, the financial risk and profitability risk are
lower for Project A and Project C than for Project B.
BB = [0.3672 0.3376 0.2265 0.0510 0.0178] (21)
The circumstance risk is non-negligible for all three projects, espe-
cially the natural risk. The earthquake risk is considered by the experts
0.65 0.35 0.00 0.00 0.00 in the natural risk evaluation because both Project A and Project B are
R1C = 0.30 0.40 0.25 0.05 0.00
located at the Zhangjiakou–Bohai seismic belt [45], where catastrophic
0.70 0.30 0.00 0.00 0.00 (22)
earthquakes have occurred including the well-known Tangshan

7
Y. Zhang, et al. Desalination 478 (2020) 114286

earthquake in 1976 [46]. The typhoon risk is a natural risk considered 4. Conclusions
by the experts for Project C, as this project lies in the South China Sea,
facing the typhoon threat several times every year [47]. Policy support This paper focuses on the risk evaluation of large-scale desalination
is important for desalination projects especially for Project B, which is projects. With two levels of risks identified, an integrated FCE-AHP
controlled by a private company. In China, private companies are more model is built to create a quantitative risk assessment for large-scale
susceptible than government-run companies to policy changes. Thus, desalination projects. Three typical projects are chosen as the case
the policy risk is more obvious for Project B than for the other two study targets and the evaluation results suggest that the risks of all the
projects. More attention is paid to the social risk for Project A because studied projects are at the “Very low” level. The performance variations
of the public concern over the safety of desalinated water. The energy of the projects in specific first-level risk factors are also analyzed.
supply risk is relatively high for Project C because the electricity is Additionally, several corresponding suggestions and recommendations
provided by a set of diesel generators. Hence, the electricity generation aiming to alleviate the possible risks of large-scale desalination projects
is largely restrained by the climate, and the electricity supply is possibly are proposed. The present study provides a feasible and comprehensive
intermittent, which is hazardous for desalination plants. approach to evaluate the risks of desalination projects for robust
planning, construction and management of future desalination projects.
3.3. Recommended strategies to mitigate potential risks Overall, the proposed study can promote the development of desali-
nation, especially the industrial desalination activities.
With the analysis of the evaluation results, several strategies are
recommended to mitigate the potential risks of desalination projects. CRediT authorship contribution statement

(1) It is necessary to conduct a risk evaluation for desalination projects. Yin Zhang:Writing - original draft, Writing - review & editing,
In this paper, the risk identification for large-scale desalination Visualization, Investigation.Ruihao Wang:Methodology,
projects is presented. However, the identification result might not Software.Pengfei Huang:Conceptualization, Supervision.Xiaoli
be directly extended to other types of desalination projects, such as Wang:Project administration.Shenghui Wang:Funding acquisition.
small-scale desalination projects for marine purpose or offshore
petroleum platforms. For these purposes, the convenience of in- Declaration of competing interest
stallation and maintenance must be considered during the risk
identification process, which is different from large-scale desalina- All the authors of the present paper declare that there is no conflict
tion projects. of interest.
(2) The government, academic institutions and desalination enterprise
are encouraged to disseminate information about the current water Acknowledgement
situations, including desalinated water, to the public. The attitude
of citizens toward an irregular water supply is still questionable due The financial support from the National Key R&D Program of China
to the lack of education on water resources. For instance, according (no. 2017YFC0403502), the Natural Science Foundation of Tianjin (no.
to a survey conducted in Tianjin [48], most people in this city have 17JCQNJC04900), the Special Fund for Basic Scientific Research
no knowledge about the specific information concerning the city Business of Central Public Research Institutes (nos. K-JBYWF-2017-T15,
water supply and consumption. Although the utilization of re- K-JBYWF-2017-T16, K-JBYWF-2018-ZT01, K-JBYWF-2018-T08 and K-
claimed water is highly accepted by the public, reclaimed water for JBYWF-2019-ZD01), the Key Technologies R&D Program of Tianjin (no.
domestic and potable use is still not widely accepted by citizens. 18YFZCSF00310) and the Tianjin Science and Technology Program (no.
The prevalence of water resource knowledge will enhance the 18ZXSZSF00190) is deeply appreciated.
consumption of desalinated water and further promote the devel-
opment of the desalination industry. References
(3) For future municipal network construction in coastal areas and
oceanic islands, the possibility of distributing desalinated water [1] P.S. Goh, T. Matsuura, A.F. Ismail, N. Hilal, Recent trends in membranes and
should be taken into consideration, regardless of whether there is membrane processes for desalination, Desalination 391 (2016) 43–60, https://doi.
org/10.1016/j.desal.2015.12.016.
currently a constructed desalination project. The utilization of [2] F. Ahmed, B.S. Lalia, V. Kochkodan, N. Hilal, R. Hashaikeh, Electrically conductive
proper materials for the network will ameliorate the risk of pipeline polymeric membranes for fouling prevention and detection: a review, Desalination
erosion caused by desalinated water. 391 (2016) 1–15, https://doi.org/10.1016/j.desal.2016.01.030.
[3] M.C. Duke, B. Zhu, C.M. Doherty, M.R. Hill, A.J. Hill, M.A. Carreon, Structural
(4) Facing the crisis of water resource shortages, desalination policies effects on SAPO-34 and ZIF-8 materials exposed to seawater solutions, and their
including legislative, administrative and financial policies should potential as desalination membranes, Desalination 377 (2016) 128–137, https://
be consistent to help the desalination industry overcome economic doi.org/10.1016/j.desal.2015.09.004.
[4] E. Jones, M. Qadir, M.T.H. van Vliet, V. Smakhtin, S. mu Kang, The state of desa-
and management difficulties.
lination and brine production: a global outlook, Sci. Total Environ. 657 (2019)
(5) More attention should be paid to desalination plants with an off- 1343–1356, https://doi.org/10.1016/j.scitotenv.2018.12.076.
grid energy supply. This kind of desalination plant generally has no [5] K. Habib, A. Fakhral-Deen, Risk assessment and evaluation of materials commonly
used in desalination plants subjected to pollution impact of the oil spill and oil fires
heat source such as waste steam and uses osmosis reverse tech-
in marine environment, Desalination 139 (2001) 249–253, https://doi.org/10.
nology. To decrease the pressure of diesel consumption, the adop- 1016/S0011-9164(01)00316-2.
tion of renewable energy including solar energy and wind energy is [6] K. Habib, O. Almatar, B. Alfeeli, Risk assessment and evaluation of materials
accessible and promising. However, the incorporation of a reverse commonly used in desalination plants subjected to stress corrosion cracking in
polluted marine environment, Desalination 153 (2003) 217–221, https://doi.org/
osmosis setup and renewable energy has mostly been investigated 10.1016/S0011-9164(02)01138-4.
theoretically, and the capacity of experimental setups has seldomly [7] L.D. Nghiem, A.I. Schäfer, Critical risk points of nanofiltration and reverse osmosis
exceeded 1000 t/d [49]. Future research on the practical integra- processes in water recycling applications, Desalination 187 (2006) 303–312,
https://doi.org/10.1016/j.desal.2005.04.089.
tion of a reverse osmosis setup and renewable energy will greatly [8] J. Stewart, Assessing supply risks of recycled water allocation strategies,
improve the stability and reduce the energy supply risk of off-grid Desalination 188 (2006) 61–67, https://doi.org/10.1016/j.desal.2005.04.102.
desalination projects. [9] S. Baggett, P. Jeffrey, B. Jefferson, Risk perception in participatory planning for

8
Y. Zhang, et al. Desalination 478 (2020) 114286

water reuse, Desalination 187 (2006) 149–158, https://doi.org/10.1016/j.desal. 292–296, https://doi.org/10.1016/j.apacoust.2017.06.022.

2005.04.075. [30] J. Yang, L. Shen, X. Jin, L. Hou, S. Shang, Y. Zhang, Evaluating the quality of si-
[10] M. Salgot, E. Huertas, S. Weber, W. Dott, J. Hollender, Wastewater reuse and risk: mulation teaching in fundamental nursing curriculum: AHP-fuzzy comprehensive
definition of key objectives, Desalination 187 (2006) 29–40, https://doi.org/10. evaluation, Nurse Educ. Today 77 (2019) 77–82, https://doi.org/10.1016/j.nedt.
1016/j.desal.2005.04.065. 2019.03.012.
[11] A. Zafra-Cabeza, M.A. Ridao, E.F. Camacho, A mixed integer quadratic program- [31] J.F. Chen, H.N. Hsieh, Q.H. Do, Evaluating teaching performance based on fuzzy
ming formulation of risk management for reverse osmosis plants, Desalination 268 AHP and comprehensive evaluation approach, Appl. Soft Comput. J. 28 (2015)
(2011) 46–54, https://doi.org/10.1016/j.desal.2010.09.048. 100–108, https://doi.org/10.1016/j.asoc.2014.11.050.
[12] E. Eriksson, A. Baun, P.S. Mikkelsen, A. Ledin, Risk assessment of xenobiotics in [32] L. Feng, X. Zhu, X. Sun, Assessing coastal reclamation suitability based on a fuzzy-
stormwater discharged to Harrestrup Å, Denmark, Desalination 215 (2007) AHP comprehensive evaluation framework: a case study of Lianyungang, China,
187–197, https://doi.org/10.1016/j.desal.2006.12.008. Mar. Pollut. Bull. 89 (2014) 102–111, https://doi.org/10.1016/j.marpolbul.2014.
[13] J.S. Chang, Understanding the role of ecological indicator use in assessing the ef- 10.029.
fects of desalination plants, Desalination 365 (2015) 416–433, https://doi.org/10. [33] Z. Song, H. Zhu, G. Jia, C. He, Comprehensive evaluation on self-ignition risks of
1016/j.desal.2015.03.013. coal stockpiles using fuzzy AHP approaches, J. Loss Prev. Process Ind. 32 (2014)
[14] M. Troldborg, D. Duckett, R. Allan, E. Hastings, R.L. Hough, A risk-based approach 78–94, https://doi.org/10.1016/j.jlp.2014.08.002.
for developing standards for irrigation with reclaimed water, Water Res. 126 (2017) [34] L. Abdullah, N. Zulkifli, Integration of fuzzy AHP and interval type-2 fuzzy
372–384, https://doi.org/10.1016/j.watres.2017.09.041. DEMATEL: an application to human resource management, Expert Syst. Appl. 42
[15] D. Goodwin, M. Raffin, P. Jeffrey, H.M. Smith, Collaboration on risk management: (2015) 4397–4409, https://doi.org/10.1016/j.eswa.2015.01.021.
the governance of a non-potable water reuse scheme in London, J. Hydrol. 573 [35] H. Zhang, X. He, H. Mitri, Fuzzy comprehensive evaluation of virtual reality mine
(2019) 1087–1095, https://doi.org/10.1016/j.jhydrol.2017.07.020. safety training system, Saf. Sci. 120 (2019) 341–351, https://doi.org/10.1016/j.
[16] M.P. Shahabi, A. McHugh, G. Ho, Environmental and economic assessment of beach ssci.2019.07.009.
well intake versus open intake for seawater reverse osmosis desalination, [36] Y.A. Solangi, Q. Tan, N.H. Mirjat, S. Ali, Evaluating the strategies for sustainable
Desalination 357 (2015) 259–266, https://doi.org/10.1016/j.desal.2014.12.003. energy planning in Pakistan: an integrated SWOT-AHP and fuzzy-TOPSIS approach,
[17] D. Malcangio, A.F. Petrillo, Modeling of brine outfall at the planning stage of de- J. Clean. Prod. 236 (2019) 117655, , https://doi.org/10.1016/j.jclepro.2019.
salination plants, Desalination 254 (2010) 114–125, https://doi.org/10.1016/j. 117655.
desal.2009.12.005. [37] L.A. Zadeh, Fuzzy sets, Inf. Control. 8 (1965) 338–353, https://doi.org/10.1016/
[18] T.M. Missimer, R.G. Maliva, Environmental issues in seawater reverse osmosis de- S0019-9958(65)90241-X.
salination: intakes and outfalls, Desalination 434 (2018) 198–215, https://doi.org/ [38] L.A. Zadeh, Fuzzy sets as a basis for a theory of possibility, Fuzzy Sets Syst. 1 (1978)
10.1016/j.desal.2017.07.012. 3–28, https://doi.org/10.1016/0165-0114(78)90029-5.
[19] M.H.A. El-Saie, Y.M.H.A. El-Saie, A.D. Moneer, Financial, economical and technical [39] T.L. Saaty, A scaling method for priorities in hierarchical structures, J. Math.
aspects of establishing remote desalination plants, Desalination 135 (2001) 25–42, Psychol. 15 (1977) 234–281, https://doi.org/10.1016/0022-2496(77)90033-5.
https://doi.org/10.1016/S0011-9164(01)00136-9. [40] T.L. Saaty, G. Hu, Ranking by eigenvector versus other methods in the analytic
[20] M.T. Amin, M.Y. Han, Water environmental and sanitation status in disaster relief of hierarchy process, Appl. Math. Lett. 11 (1998) 121–125, https://doi.org/10.1016/
Pakistan’s 2005 earthquake, Desalination 248 (2009) 436–445, https://doi.org/10. S0893-9659(98)00068-8.
1016/j.desal.2008.05.085. [41] T.L. Saaty, Decision-making with the AHP: why is the principal eigenvector ne-
[21] W.B.G. Fernando, A.H. Gunapala, W.A. Jayantha, Water supply and sanitation cessary, Eur. J. Oper. Res. 145 (2003) 85–91, https://doi.org/10.1016/S0377-
needs in a disaster - lessons learned through the tsunami disaster in Sri Lanka, 2217(02)00227-8.
Desalination 248 (2009) 14–21, https://doi.org/10.1016/j.desal.2008.05.032. [42] X. Xu, Z. Cao, Z. Zhang, R. Li, B. Hu, Spatial distribution and pollution assessment of
[22] Koutseris, A. Filintas, P. Dioudis, Antiflooding prevention, protection, strategic heavy metals in the surface sediments of the Bohai and Yellow Seas, Mar. Pollut.
environmental planning of aquatic resources and water purification: the case of Bull. 110 (2016) 596–602, https://doi.org/10.1016/j.marpolbul.2016.05.079.
Thessalian basin, in Greece, Desalination 250 (2010) 318–322, https://doi.org/10. [43] X. Liang, C. Tian, Z. Zong, X. Wang, W. Jiang, Y. Chen, J. Ma, Y. Luo, J. Li, G. Zhang,
1016/j.desal.2009.09.049. Flux and source-sink relationship of heavy metals and arsenic in the Bohai Sea,
[23] H. Mahmoudi, O. Abdellah, N. Ghaffour, Capacity building strategies and policy for China, Environ. Pollut. 242 (2018) 1353–1361, https://doi.org/10.1016/j.envpol.
desalination using renewable energies in Algeria, Renew. Sust. Energ. Rev. 13 2018.08.011.
(2009) 921–926, https://doi.org/10.1016/j.rser.2008.02.001. [44] H. Li, X. Gao, Y. Gu, R. Wang, P. Xie, M. Liang, H. Ming, J. Su, Comprehensive large-
[24] P.D. Welle, J. Medellín-Azuara, J.H. Viers, M.S. Mauter, Economic and policy dri- scale investigation and assessment of trace metal in the coastal sediments of Bohai
vers of agricultural water desalination in California’s central valley, Agric. Water Sea, Mar. Pollut. Bull. 129 (2018) 126–134, https://doi.org/10.1016/j.marpolbul.
Manag. 194 (2017) 192–203, https://doi.org/10.1016/j.agwat.2017.07.024. 2018.02.022.
[25] B. Haddad, N. Heck, A. Paytan, D. Potts, Chapter 14 - social issues and public ac- [45] P. Li, F. Ren, M. Cai, Q. Guo, S. Miao, Journal of Asian Earth Sciences present-day
ceptance of seawater desalination plants, in: V.G. Gude (Ed.), Sustain. Desalin. stress state and fault stability analysis in the capital area of China constrained by in
Handb, Butterworth-Heinemann, 2018, pp. 505–525, , https://doi.org/10.1016/ situ stress measurements and focal mechanism solutions, J. Asian Earth Sci. 185
B978-0-12-809240-8.00014-9. (2019) 104007, , https://doi.org/10.1016/j.jseaes.2019.104007.
[26] I. Khoshrou, M.R. Jafari Nasr, K. Bakhtari, New opportunities in mass and energy [46] H. Guo, J. Zhao, The surface rupture zone and paleoseismic evidence on the seis-
consumption of the multi-stage flash distillation type of brackish water desalination mogenic fault of the 1976 Ms 7.8 Tangshan earthquake, China, Geomorphology 327
process, Sol. Energy 153 (2017) 115–125, https://doi.org/10.1016/j.solener.2017. (2019) 297–306, https://doi.org/10.1016/j.geomorph.2018.11.006.
05.021. [47] B. Huang, Z. Ouyang, H. Zheng, H. Zhang, X. Wang, Construction of an eco-island: a
[27] D. Saldivia, C. Rosales, R. Barraza, L. Cornejo, Computational analysis for a multi- case study of Chongming Island, China, Ocean Coast. Manag. 51 (2008) 575–588,
effect distillation (MED) plant driven by solar energy in Chile, Renew. Energy 132 https://doi.org/10.1016/j.ocecoaman.2008.06.007.
(2019) 206–220, https://doi.org/10.1016/j.renene.2018.07.139. [48] Q. Gu, Y. Chen, R. Pody, R. Cheng, X. Zheng, Z. Zhang, Public perception and ac-
[28] N. Zhang, K. Zhou, X. Du, Application of fuzzy logic and fuzzy AHP to mineral ceptability toward reclaimed water in Tianjin, Resour. Conserv. Recycl. 104 (2015)
prospectivity mapping of porphyry and hydrothermal vein copper deposits in the 291–299, https://doi.org/10.1016/j.resconrec.2015.07.013.
Dananhu-Tousuquan island arc, Xinjiang, NW China, J. African Earth Sci. 128 [49] M.T. Mito, X. Ma, H. Albuflasa, P.A. Davies, Reverse osmosis (RO) membrane de-
(2017) 84–96, https://doi.org/10.1016/j.jafrearsci.2016.12.011. salination driven by wind and solar photovoltaic (PV) energy: state of the art and
[29] D. Yang, C.M. Mak, An assessment model of classroom acoustical environment challenges for large-scale implementation, Renew. Sust. Energ. Rev. 112 (2019)
based on fuzzy comprehensive evaluation method, Appl. Acoust. 127 (2017) 669–685, https://doi.org/10.1016/j.rser.2019.06.008.