10th International Conference on Hydroinformatics

HIC 2012, Hamburg, GERMANY

Risk Assessment and Management of Wastewater Treatment

Plants (WTPs) Using a Fuzzy Multi-Attribute Decision-Making
(FMADM) Approach
EBRAHIM BADALI BAVANI (1), MASSOUD TABESH (2)
(1): School of Civil Engineering, College of Engineering, University of Tehran, Tehran, Iran.
(2): Associate Professor, Center of Excellence for Engineering and Management of
Infrastructures, School of Civil Engineering, College of Engineering, University of Tehran,
Tehran, Iran.

Abstract
Wastewater Treatment Plants (WWTPs) are infrastructures of vital importance for ensuring the
human and environmental health. Furthermore, they have a key role in environmental
sustainability by recycling the useful materials such as water and nutrients. Failures in WTPs
may lead to damaging consequences for urban dwellers such as disease outbreak and flooding;
hence, assessing the potential risks in WWTPs and designing the prevention and mitigation
plans is necessary. In this paper, an algorithm has been proposed for risk assessment and
management of WWTPs. In the proposed algorithm, operational risk of a specific element of a
WWTP is calculated by multiplying three parameters’ values comprising probability of failure,
intensity of probable damages and vulnerability of element. The values of aforementioned
parameters is quantified by means of experts and policy-makers’ viewpoints which are elicited
from questionnaires. Because of significant ambiguities and imprecision’s in human judgments,
a fuzzy approach has been employed in the proposed algorithm to considering the impacts of
uncertainties on final solutions. To convert the qualitative values into quantitative values,
applying the Triangular Fuzzy Numbers (TFNs) have proposed. Furthermore, employing a
Fuzzy Multi-Criteria Decision-Making (FMADM) method has been proposed to calculate the
values of operational risks. Based on the values of operational risk elements (probability,
intensity and vulnerability), different mitigation or prevention plans are recommendable to
reduce the total value of risks. In this paper, the proposed algorithm has been implemented for a
WTP as a test example. The results of the study show that the proposed algorithm can be
employed by managers and policy-makers of WWTPs as a robust and practical decision-
making tool.
Keywords: Risk assessment, Risk management, Wastewater Treatment Plants (WWTPs),
Fuzzy Multi-Criteria Decision-Making (FMADM) method, Triangular Fuzzy Numbers (TFNs)

1. Introduction

Wastewater Treatment Plants (WWTPs) are infrastructures of vital importance for ensuring the
human and environmental health. Furthermore, they have a key role in environmental
sustainability by recycling the useful materials such as water and nutrients. Environmental
protection, prevent contamination of underground water resources and promoting public health,
are the reasons that have become the WWTPs systems to one of the most important
infrastructure of modern civilized societies. Failures in WTPs may lead to adverse outcomes
such as untreated wastes exclusion containing various chemical and biological pollutants from
WWTPs and their entry into the urban environments and agricultural lands. There is a wide
range of threats that may affect wastewater systems. These threats can be divided into three
categories [1]: Malevolent threats, those caused by human activities for the express purpose of
damaging or destroying assets and terrorizing people. Natural threats, those caused by natural
phenomenon such as extreme weather, wildfires, and earthquakes. Unintentional threats, those
caused by human activities, but whose consequences are accidental or inadvertent. This paper
focused on the analysis of risks associated with unintentional threats. Risk assessment process
includes the following essential parts: scope definition, hazard/risk identification, accident
scenario development, probability and consequences estimation, risk estimation and risk
assessment [3]. Several methods have previously been employed and developed for
vulnerability assessment. These include: the comprehensive evaluation method [4]; the artificial
neural-network evaluation method [5]; the landscape evaluation method [6]; and the principal
component analysis (PCA) method [7]. Among them, the analytic hierarchy process (AHP) is
one of the most commonly used method of assessment [8], which works on a premise that
decision-making of complex problems can be handled by structuring the complex problem into
a simple and comprehensible hierarchical structure. Despite of its wide range of applications,
the conventional AHP approach may not fully reflect a style of human thinking, in which
human’s judgments are represented as exact numbers. However, in many practical situations,
decision makers usually feel more confident to give interval judgments rather than expressing
their judgments in the form of exact numeric values. Therefore, AHP technique involves
subjectivity in pair-wise comparisons and vagueness and uncertainty dominate in this process.
In view of that, [9] have evolved AHP into the FAHP, bringing the triangular fuzzy-number of
the fuzzy set theory directly into the pair-wise comparison matrix of the AHP. In this paper, an
algorithm has been proposed for risk assessment and management of WWTPs. In the proposed
algorithm, the risk of a specific element of a WWTP is calculated by multiplying three
parameters’ values comprising probability of failure, intensity of probable damages and
vulnerability of element. The values of aforementioned parameters is quantified by means of
experts and policy-makers’ viewpoints which are elicited from questionnaires. Based on the
values of operational risk elements (probability, intensity and vulnerability), different
mitigation or prevention plans are recommendable to reduce the total value of risks. The
proposed algorithm is outlined in Figure 2.

2. Methodology
The analytic hierarchy process (AHP) method introduced by [10] shows the process of
determining the priority of a set of alternatives and the relative importance of attributes in a
multi-criteria decision-making (MCDM) problem [11]. The primary advantage of the AHP
approach is the relative ease with which it handles multiple criteria and performs qualitative and
quantitative data [12]. To improve the AHP method and to determine the relative weight of
criteria for risk assessment, this study applies the fuzzy analytic hierarchy process (FAHP) and
uses triangular fuzzy numbers to express the comparative judgments of decision-makers. Some
calculation steps are essential and explained as follows: Step 1. Establishing the hierarchical
structure: Construct the hierarchical structure with decision elements, decision-makers are
requested to make pair-wise comparisons between decision alternatives and criteria using a nine
point scale. All matrices are developed, and all pair-wises comparisons are obtained from each
n decision-maker. Step 2. Calculating the consistency: To ensure that the priority of elements is
consistent, the maximum eigenvector or relative weights and k max are calculated. Then the
consistency index (CI) for each matrix order n using Eq. (1) is computed. Based on CI and
random index (RI), the consistency ratio (CR) is calculated using Eq. (2). The CI and CR are
defined as follows [13]:

 max  n
CI  (1)
n 1
I .I
CR  (2)
I .I .R

where, n is the number of items being compared in the matrix, and RI is a random index, the
average consistency index of randomly generated pair-wise comparison matrix of similar size,
as shown in Table 1. If the consistency test is not passed, the original values in the pair-wise
comparison matrix must be revised by the decision maker. Step 4. Calculate fuzzy weights
value: According to the Lambda- Max method proposed by [14] the fuzzy weights of the
hierarchy can be calculated. This process is described as follows:

 wk 
Slk  min  imk 1  i  n,
 wil 
(4)
 wk 
Suk  min  imk 1  i  n,
 wiu 

where Slk is upper and Suk is lower bound of the weight matrix. Step 5. Integration of the
opinions of decision makers. A geometric average was applied to combine the fuzzy weights of
different decision makers:
 1/ K
~  K ~ 
W    Wi k  , k  1,2,..., k (5)
 k 1 
where WÞ‾: combined fuzzy weight of decision element i of decision makers, WÞ i k: fuzzy
weight of decision element i of decision maker k, K: number of decision makers.
Step 6. defuzzification. Based on the equation proposed by [26], a proximity coefficient was
defined to obtain the ranking order of the decision elements. The proximity coefficient is
defined as follows:

1
~
Y ( Ai ) 
 g ( x)   dx
0
Ai ( x )
(6)
1
  dx
0
Ai ( x )

where Y(Ãi): Longitudinal coordinates of the center of membership function  A ( y) , g(x): A

weighting factor to indicate the degree of importance of the x.

3. Case study

Tehran South Wastewater Treatment Plant was chosen as a case study. This large
environmental unit occupied about 70,000 square meters (m2) of Sharerey area located in south
of Tehran. The final population of 3150000 people covered and the input flow is 670,000 cubic
meters per day (m3/d). This WWTP uses conventional active sludge as treatment mechanism.

3. Proposed Algorithm

The general approach of this research will be presented, in the form of an algorithm as
"algorithm assessment and risk management systems, wastewater treatment plant" Figure 2
shows the algorithm. The algorithm consists of seven main steps in more detail in the form of a
numerical example will be described.
Step 1. System Identification, Wastewater treatment plant for the main part of the
evaluation process in this research are: 1. Site, 2. Office buildings, 3. Control systems, 4. Inlet
pumping station, 5. Pipes, 6. Screening, 7. Grit removal, 8. Aeration ponds, 9. Primary and
secondary settling ponds, 10. Electrical utilities, 11. Treatment and sludge processing unit 12.
Disinfection unit 13. Storage tanks of gas (biogas), 14. Operational personnel, 15. The return
sludge pumping station.
Step 2. Threat Identification, The second step of the algorithm is identifying the type of
threats that are affecting the system. This article have considered only unintentional threats: 1)
Pollution of water resources, 2) power failure, 3) cut off communications and communications
systems, 4) fire, 5) design problems, 6) Explosion, 7) leakage, 8) physical damage to
equipment, 9 ) corrosion of equipment, 10) fluctuations in the incoming sewage flow.
 Step 3. Threat Rating, To evaluate and rank the threats identified, specialized questionnaire
was used as the basis for the work. Then AHP method and a geometric mean method (Eq. 5)
were implemented. Eq. (7) is used to calculate "the threats".

~ ~ ~ )7(
Ti  Pi  Wi

where T̃i: the value of the threat i, P̃i: the potential of the threat i, and W̃i: the relative
weight of the overall threat i. In the present study, for each identified threat, the
probability values (Eq. 5) and relative weights are calculated. To calculate the relative
weight and analyze the threats hierarchical matrix of paired comparisons is used. Table
(2) shows the final results of the calculation of the parameter "threat" to the threats
identified by the system. These results are obtained based on opinions of experts and
decision makers and also Eqs. (5 and 7). Here a sample of the threats to the
"contamination" is given.
T̃= (1,1,2) × (0.16,0.22,0.32) = (1.8,2.67,3.23)

System Identification

Unintentional Threats
Threats Identification
Criteria: Death toll, environmental
damage, replaces ability, Threat Rating
functional capability, Financial

Consequences Assessment
Criteria: potential change in
the structure, weakness of the
Vulnerability Assessment
structure, reconstruction
capabilities
Risk Analysis
No

Risk Control
Consider mitigation options

Options technical and economic assessment Is risk

Aacceptabl
?e?
Yes
Decision making

Accept Risk
Monitoring

Figure 2. The proposed algorithm for risk assessment and management of the WWTPS
Table 2. Threat value of unintentional threats

NO Threat Likelihood (pt) Relative Wight (Wi) Threat value (T)

1 Pollution of water resources (1,1,2) (0.16,0.22,0.32) (1.8,2.67,3.23)
2 power failure (1.2,0.23,2.1) (0.002,0.003,0.006) (0.02,0.12,0.03)
3 cut off communications systems (1.36,0.78,1.98) (0.02,0.31,0.16) (0.032,0.12,0.04)
4 fire (1.2,2,1.1) (0.12,0.13,0.002) (1.2,0.23,0.67)
5 design problems (1.3,2.1,1.23) (0.21,0.13,0.1) (2.1,1.2,0.2)
6 Explosion (0.6,0.32,1.1) (0.78,1.2,2.34) (3.43,2.1,3.2)
7 leakage (0.76,0.21,3.5) (0.12,0.001,0.02) (2.8,1.3,2.6)
8 physical damage to equipment (0.23,1.2,0.56) (0.003,0.001,0.023) (0.012,0.13,0.12)
9 corrosion of equipment (2.3,1.4,3.78) (0.23,0.002,0.007) (0.12,0.11,0.0023)
10 fluctuations in the incoming sewage (3.7,2.4,1.87) (0.001,0.21,0.32) (0.12,0.2,0.231)
flow

Step 4. Consequences assessment, Damage assessment criteria and thresholds for different
parts of the system, which are: 1) Death toll, 2) environmental, 3) finance, 4) ability to replace,
5) functional capabilities. Obtain the relative weights of criteria is the first step to complete the
damage assessment process. To do this using matrices of paired comparisons and expert
opinions in the method of AHP relative weights of these criteria is obtained. Finally, using Eq.
(8) in each segment of the system the amount of damage caused by any of the selected threats,
is obtained.

~ m ~
C  Wi  S i (8)
i 1

~
where, C̃: Fuzzy amount of consequences, Wi : Fuzzy weight of the ith criterion, S i : Fuzzy
rating the severity of the threat of damage according to the ith criterion and m: the number of
criteria. To calculate the system losses for each part of a specific threat, the threshold for a
series of criteria was developed. For example, to measure point 10 to the tragic life - death
threats for being a part of the system is considered. For all components of the system losses are
calculated and the results for a sample site are given in Table 3.
Table 3. Calculation of fuzzy consequences

criteria Dead toll environmental finance replacement functional Total score

Weight (0.58,0.2,0.11) (0.2,0.12,0.23) (0.11,0.12,0.14) (0.03,0.2,0.01) (0.06,0.02,0.3) Eq. 8
S T
(4.85,2.3,1.34) (4.32,5.2,2.1) (9.25,5.2,2.7) (3.32,1.3,3.2) (4.06,3.2,1.5) (4.67,3.2,1.56)
site fire

Step 5. Vulnerability assessment, Similar to the previous step, the first step is to provide
criteria for assessing the vulnerability of system components. Then using the matrix of paired
comparisons and expert opinions relative weight of each criterion is obtained. Using Eq. (4), for
different parts of the system vulnerability is calculated. Criteria developed for this threat are:
potential change in the structure, weakness of the structure, reconstruction capabilities. Eq. (9)
is used to calculate the "number of vulnerability".

n
~
V  Wi  vi (9)
i 1

where: Wi : the Fuzzy relative weight of the criteria i, n: number of criteria, vi : Rate of the
threat vulnerability of each criterion with respect to the criteria i, and Ṽ: the Fuzzy number of
vulnerabilities. For all components of the system amount of vulnerability are calculated and the
results for the sample site are given in Table 4.

Table 4. Calculation of fuzzy vulnerability

criteria Potentialin theWeakness of ereconstruction capabilities Total score

structure the structure
Weight (0.2,0.2,0.41) (0.4,0.22,0.23) (0.41,0.01,0.54) Eq. 8
S T
(3.81,4.3,1.34) (6.32,5.2,4.1) (5.81,2.3,4.67) (5.57,2,1.23)
site fire
Step 6. Risk assessment, At this stage of the algorithm using Eq. (10), the main risk systems
is calculated.

~ ~ ~ ~
Ri  Vi  Ci  Ti (10)

~ ~ ~
Where: Ri : Fuzzy risk of segment i, Vi : the fuzzy vulnerability of the segment i, C i : a
~
compensation intensity of the ith component and Ti : threat. Table (5) shows the risks related to
the sewage treatment plant site. The risk values are the result of expert opinions and analysis of
their comments is the 6th step of the process of risk analysis and management algorithm for
wastewater treatment plants.

Table 5. Calculated risk

 segment Threat name Threat value consequence vulnerability Risk
Site Fire (0.07,0.02,0.011) (4.67,3.2,1.2) (5.5,3.2,6.3) (1.82,4.3,2.1)

Step 7. Defuzzification,

YA = 0.014x – 0.14 , YB = -0.003x + 1.25

Y(r) = Y(r)=23.6

4. Conclusion

In this paper an algorithm for risk assessment and management of wastewater treatment was
introduced. The proposed algorithm consists of seven main steps, all steps described in detail
with a numerical example. The output of this algorithm presented a risk value to the all system
segments. Given the number of risk managers and policy makers can increase the safety level of
each part of the system.

5. References

[1] ASCE., “Interim Voluntary Security Guidance For Wastewater Utilities”, Water
Environment Federation, . (2004), pp 10-70.
[2] Mays, L. W., “Water Supply Systems Security”, McGraw-Hill, New York. (2004).
[3] Dreazen, Y. J., “Backflow Water –line Attack Feared”, Wall Street Journal, Vol. 50, No.
3, (2003), pp. 4.
[4] Bailey, K. C., “The Biological and Toxin Weapons Threat to the United States”, National
Institute for Public Policy, Farifax, VA, (2001).
[5] Grigg, N. S., “Water Security: Multiple Hazards and Multiple Barriers”, Journal of
Infrastructure Systems, ASCE, Vol. 9, No. 2, (2003), pp 81- 88.
[6] Chung, R., “January 17, 1995 Hydrogoken-Nabu (Kobe) Earthquake: Performance of
Structures, Lifelines, and Fire Protection Systems”, National Institute of Standard and
Technology, (1996).
[7] Horsley, M. B., Carlson, D. J., McCarthy, D. W., Guta, D. P., “Design lessons from the
flood of ’93”, Proceedings of Annual Conference of American Water Works Association,
Denver, USA. (1994).
[8] Wessberg, N., Molarius, R., Seppälä, J., Koskela, S., Pennanen, J., “Environmental Risk
Analysis for Accidental Emissions”, Journal of Chemical Health and Safety, Vol. 15, No. 1,
(2008),pp 24-31.
[9] Sandia Corporation.,“A Risk Assessment Methodology (RAM) for Physical Security”,
Lockheed Martin Company, US, (2004).
[10] Eldin, N. N., and Eldrandaly, K. A., “A Computer-Aided System for Site Selection of Major
Capital Investment”, Proceedings of 1st ASCAAD International Conference: e-Design in
Architecture, Dhahran, Saudi Arabia, (2004).
[11] Mathews, M., D.M., Karydasand, M.A., Delichatsios., “A Performance-Based Approach
for fire Safety Engineering”, International Association for Fire Safety Science, (1997), pp
595-606.
[12] Satty, T.L.,. “A scaling method for priorities in hierarchical structures”, J. Math. Psychol.
Vol. 15, (1977), pp. 234–281.
[13] Csutora, R., & Buckley, J. J.. “Fuzzy hierarchical analysis: The Lambda-Max method”,
Fuzzy Sets and Systems, Vol. 120, No. 2, (2001), pp. 181–195.
[14] Saaty, T. L., “The analytic hierarchy process”, New York: McGraw-Hill, (1980).