See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/238620636

Evaluation of Municipal Water Distribution System Reliability Using

Minimum Cut-set Method

Article  in  Journal of King Saud University · January 2006

DOI: 10.1016/S1018-3639(18)30822-5

CITATIONS READS

15 1,208

2 authors, including:

Muhammad A. Al-Zahrani
King Fahd University of Petroleum and Minerals
45 PUBLICATIONS   1,020 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Optimization of Residual Chlorine in Water Distribution Systems View project

All content following this page was uploaded by Muhammad A. Al-Zahrani on 02 July 2014.

The user has requested enhancement of the downloaded file.

 J. King Saud Univ., Vol. 18, Eng. Sci. (1), pp. 67-82, Riyadh (1426H./2005)

Evaluation of Municipal Water Distribution System

Reliability Using Minimum Cut-set Method
Muhammad A. Al-Zahrani and Juned Laiq Syed
Department of Civil Engineering, College of Engineering,
King Fahd University of Petroleum & Minerals (KFUPM), P.O. Box 686,
Dhahran 31261, Saudi Arabia

(Received 20 April, 2004; accepted for publication 25 April, 2005)

Abstract. Reliability analysis of water distribution systems is a complex task, as it requires both definition and
calculation of reliability measures. In this paper, a methodology for evaluating water distribution system
reliability is developed and demonstrated on a real water distribution network. The methodology comprises of
two steps: (1) nodal pressures are calculated using hydraulic simulation program (EPANET), and (2) the
minimum cut-set method is applied to calculate nodal and system reliabilities of Al-Khobar water distribution
network. The results show that the hydraulic reliability of the central part of Al-Khobar water distribution
system is 69.73%.

Introduction

Water distribution system plays a vital role in preserving and providing a desirable life
quality to the public, of which the reliability of supply is a major component. Reliability
analysis of a water distribution system is concerned with its ability to deliver water to
individual consumers in the required quantity and under a satisfactory pressure.

In general, reliability is defined as the probability that a system performs its

mission within specified limits for a given period of time in a specified environment.
Reliability of water distribution system is defined by Kaufmann et al. [1] as the
probability that the system will perform its specified tasks under specified conditions and
during a specified time. Goulter [2] and Cullinane et al. [3] defined reliability of water
distribution system as the ability of the system to meet the demands that are placed on it.
The demands are specified in terms of the flow to be supplied and the range of pressure
at which these flow rates must be provided.

67
68 Muhammad A. Al-Zahrani and Juned Laiq Syed

Literature Review

In order to analyze the reliability of water distribution system, different approaches

are presently being employed by different researchers and analysts. Germanopoulos et
al. [4] presented a methodology for assessing the security of supply from a water
distribution system associated with different network failure events such as burst mains
or source failures. They used a network simulation model to study the operation of the
network both under normal operating conditions and conditions arising from crisis
events. They identified the effect of different failure events on the supplies of the area
and they used probability analysis of the occurrence of such events to provide an
assessment of the security of water supply. They also identified the operational
responses that should be triggered by crisis events. They concluded that the assessment
of supply reliability obtained using the adopted methodology is considerably different
from that suggested by the conventional approach, which simply relates supply
reliability to the amount of emergency storage available in the network.

Lansey et al. [5] presented a constrained model for the minimum cost design of
water distribution networks. Their methodology attempted to account for the
uncertainties in required demands, required pressure heads, and pipe roughness
coefficients. They formulated an optimization problem as a non-linear programming
model which is solved using a generalized reduced gradient method. Their results show
that uncertainties in future demands, pressure head requirements, and pipe roughness can
have significant effects on the optimal design and cost. They observed that cost versus
reliability relationship is convex, which means an incremental amount at a higher
reliability level will result in a greater increase in the system cost than for an incremental
change at a lower level.

Quimpo and Shamsi [6] developed a strategy for prioritizing decisions for the
maintenance of a water distribution system. Using component and network reliability
based on the time-varying connectivity concepts, the probabilities that the water will be
available at demand points in the system are calculated to determine a reliability surface.
At any time, this surface is used to locate low reliability areas, which identify parts of the
system that need maintenance priority. The specific components that must be repaired or
replaced are determined using a component importance criterion that measures the
overall effect of component maintenance on the system reliability.

Mays [7] computed the reliability of water distribution system by treating the
demand, pressure head, and pipe roughness as random variables. He assumed that water
demand and pipe roughness coefficient follow a probability distribution, and then used a
random number generator to generate the values of random variables for each node and
pipe. Then, he performed hydraulic simulation and computed the pressure heads at the
demand nodes, provided the demands are satisfied. Finally, he computed the nodal and
system hydraulic reliabilities.
 Evaluation of Municipal Water Distribution System Reliability … 69

Calvin et al. [8] investigated the capacity reliability which is defined as the
probability that the carrying capacity meets the flow demand. They described the use of
capacity reliability for networks with more than one demand node through finding the
probability of a feasible flow, given the probability distributions of flow capacities in
pipes and fixed nodal demand. The solution procedure generates a set of inequalities that
represents a necessary and sufficient condition for feasible flow. They proposed a
solution procedure for evaluating the probability that all the inequalities are satisfied by
eliminating redundant inequalities and by determining bounds for the probability of
feasible flow. They developed a decision-making framework that applies both the
capacity reliability measure and the solution approach for maintenance and rehabilitation
decision making.

Ostfeld [9] developed a tailor-made reliability methodology for the reliability

assessment of regional water distribution and applied it to the regional water distribution
system. The methodology comprised of two interconnected stages: the analysis of
storage/conveyance properties of the system and implementation of stochastic simulation
through the use of the software “US Air Force Rapid Availability Prototyping for
Testing Operational Readiness” (RAPTOR).

Shinstine et al. [10] applied reliability models to large-scale municipal water

distribution systems based on minimum cut-set method and examined the reliability
levels that engineers implicitly design into their systems.

Reliability of water distribution system can be analyzed based on different

criteria’s. For example, it can be correlated to (1) inability of the system to supply
proper water quality and quantity, (2) maintain reasonable heads throughout the
distribution network, or (3) pumps failures. In this study, the focus is mainly towards the
hydraulic failure of the water distribution system, which considers the system’s failure
due to demands and the inability of the pressure heads to satisfy or meet the
requirements. Here, the hydraulic reliability associated with the central part of Al-
Khobar city water distribution system is analyzed using the minimum cut-set approach.
A method for calculating the complete pipe failure probability using Poisson method is
used, which takes into account the number of breaks in the pipe. Finally, nodal and
system reliability is calculated.

Development of Methodology

The proposed methodology for calculating nodal and system reliability is based on
“minimum cut-sets” [10] in which nodal demand, pipe roughness, tank and reservoir
water levels are considered as deterministic values. In order to explain this methodology,
the following discussion is necessary.

“Hydraulic Availability” is defined as the ability of the water distribution system to

provide service with an acceptable level of interruption in spite of abnormal conditions [3].
70 Muhammad A. Al-Zahrani and Juned Laiq Syed

Availability is evaluated in terms of developing the required minimum pressure.

Pressures between 20 psi and 80 psi [10] are considered to be desirable pressures under
normal daily demands.

Goulter and Coals [11] proposed the use of discrete relationship between
availability and pressure as shown in Fig. 1. The availability during a time period t can
be expressed by the following mathematical relationship:

HA jt = 1............. for NP jt ≥ PR ⎫
⎬ at time t (1)
HA jt = 0 ............. for NP jt < PR ⎭

where

HAj = hydraulic availability of node j,

NPj = pressure at node j,
PR = required minimum pressure,
t = time during which hydraulic availability was evaluated.

Cullinane et al. [3] formulated an approach that describes availability index as a

continuous “fuzzy” function. Using this concept, a significant index value may be
assigned to pressure values slightly less than the arbitrary assigned required minimum
pressure value, PR. Accordingly, a curve similar to Fig. 2 can be developed which
resembles the curve of a normal distribution. Thus, the hydraulic availability function
can be described mathematically as:

HA j = ( PR ≤ NP j )

( H −µ H )
σH t2
1 − ⎡ (H − µ H ) ⎤
HA j =
2π ∫
−∞
e 2 dt = P ⎢
⎣ σH
⎥
⎦
(2)

where

H = NPj = value of nodal pressure,

µH = mean nodal pressure,
σH = standard deviation of pressure,
P[●] = probability function.
 Evaluation of Municipal Water Distribution System Reliability … 71

1
Hydraulic Availability

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100 120
Pressure (psi)

Fig. 1. Discrete hydraulic availability step function during time (t).

0.8
Hydraulic Availability

0.6

0.4

0.2

0
20 30 40 50 60 70

Pressure (psi)

Fig. 2. Continuous hydraulic availability function.

72 Muhammad A. Al-Zahrani and Juned Laiq Syed

Calculation of nodal and system reliability

The minimum cut-set approach is adopted to calculate the nodal and system
reliability, Rnode and Rs, respectively. According to Su et al. [12], the minimum cut-set
can be defined as “a set of system components (e.g., pipes) which, when failed, causes
failure of the system”. However, system failure will not occur if any component of the
set does not fail [13].

Assuming that a pipe break can be isolated from the rest of the system, the
minimum cut-sets are determined by closing a pipe or combination of pipes in the water
distribution system and using a hydraulic simulation model to determine the values of
pressure head at each demand node of the system. In this study, EPANET was used [14].
By comparing these pressure heads with the minimum pressure head requirements, the
reliability model can determine whether or not this pipe or combination of pipes is a
minimum cut-set of the system or an individual demand node. A minimum cut-set for a
node is the one that causes reduced hydraulic availability at that node, while a minimum
cut-set for the system is a cut-set that reduces the hydraulic availability for any node in
the system. To calculate the number of combinations for pipe closure for the cut-set
determination, it is observed that the failure of two or three pipes is purely a “random”
phenomenon. Therefore, in order to determine the pipe combinations for the cut-set
determination, subsets of pipe combinations should be determined by applying a random
approach. For instance, if there are K numbers of pipes in the water distribution system,
then out of those K pipes, T subsets should be randomly generated and each sub-set
could have only one pipe or a combination of two or three pipes. A flow chart of the
procedure is shown in Fig. 3.

According to Shinstine et al. [10], for n components (pipes) in the ith minimum cut-
set of a water distribution system, the failure probability of the ith minimum cut-set (MCi)
is:

n
P ( MC i ) = ∏P = P • P
i =1
i 1 2 • P3 • ....... • Pn (3)

Using the step function for hydraulic availability and assuming that the occurrence of the
failure of the components within a minimum cut-set is statistically independent, for a
water distribution network with four minimum cut-sets (MCi) with the system reliability,
Rs, the failure probability of the system Ps is then defined [13] as:

Ps = P( MC1 ∪ MC 2 ∪ MC3 ∪ MC 4 ) (4)

Ps = P ( MC1 ) + P ( MC 2 ) + P ( MC3 ) + P( MC 4 ) (5)

Evaluation of Municipal Water Distribution System Reliability … 73

Fig. 3. Minimum cut-set flow chart.

74 Muhammad A. Al-Zahrani and Juned Laiq Syed

4
Ps = ∑ P(MC )
i =1
i (6)

In general form,

M
Ps = ∑ P(MC )
i =1
i (7)

The system reliability, Rs, is expressed as:

M
Rs = 1 − Ps = 1 − ∑ P(MC )
i =1
i (8)

where M = number of minimum cut-sets in the system.

It is possible to weigh the nodal terms as the function of the nodal demand. Nodal
reliabilities can be computed with the same relationship including only failures that
affect the individual node.

Using the continuous hydraulic availability concept, a true minimum cut-set does
not exist. The probability of a cut-set occurring is consistent; however, reliability is
defined as the nodal hydraulic unavailability (1–HA). The system reliability is then
expressed as [15]:

M
Rs = 1 − Ps = 1 − ∑ (1 − HA
i =1
i
net ) P ( MC i ) (9)

i
where HAnet is the network hydraulic availability and expressed as:

J
HAnet = ∏ HA j =1
j (10)

where HAj = hydraulic availability of node j.

If HAnet equals one, the failure probability of the cut-set is not included in Eq. (9);
thus, it is identical to Eq. (8) for the step function hydraulic availability case. To
compute the system reliability with continuous hydraulic availability, all cut-sets are
included.
 Evaluation of Municipal Water Distribution System Reliability … 75

Application of the Methodology

The developed methodology was applied on Al-Khobar water distribution system to

evaluate its hydraulic reliability. The water distribution system, shown in Fig. 4, consists
of 191 pipes, 131 junctions, 12 reservoirs, 2 tanks and 12 pumps. In the analysis, nodal
demand, Chezy’s roughness coefficient for pipe, tank and reservoir water level are
assumed and considered as deterministic values. A sample data are summarizes as
shown in Table 1.

Table 1. Sample characteristics data for Al-Khobar water distribution network

a. Pipe Characteristics
Pipe No.
Parameter
2 8 15 20

Length (m) 498 714 948 1390

Diameter (mm) 300 200 250 1000

Roughness Coefficient 100 100 100 100

b. Junction Characteristics

Junction No.
Parameter
1 3 40 90

Elevation (m) 30.1 25.71 14.57 6.98

Demand (lps) 0.926 2.130 3.016 12.035

c. Tanks Characteristics

Parameter
Base Elevation Initial Max. Tank Diameter
Tank
(m) Elevation Elevation (m)
(m) (m)
Makkah Tank 30.0 4.0 20.0 30.0

Al-Yarmouk Tank 27.5 10.0 24.5 40.0

76 Muhammad A. Al-Zahrani and Juned Laiq Syed

The pipe failure combinations required for the cut-set calculations are determined
by assuming randomness in the simultaneous failure of two or three pipes. Then, a
steady-state hydraulic analysis is performed using the hydraulic simulation software
EPANET [14], and nodal pressures are calculated for the pipe closure combinations
shown in Table 2.

Table 2. Pipe closure combinations for Al-Khobar sub-network

Closure Comb Pipe ID Pipe ID Pipe ID

No.
1 160 111 20
2 83 - -
3 146 - -
4 150 - -
5 112 119 -
6 164 21 -
7 37 - -
8 175 44 -
9 12 - -
10 150 - -
11 114 27 -
12 158 142 -
13 17 96 -
14 141 73 -
15 7 - -
*
16 7 YRMK-900 -
17 166 158 -
18 50 132 -
19 10 142 -
20 21 - -
21 41 26 -
22 39 116 163
23 170 - -
24 24 25 -
25 37 127 -
*
Al-Yarmouk reservoir tank.
 Evaluation of Municipal Water Distribution System Reliability … 77

Table 2. Contd.

Closure Comb Pipe ID Pipe ID Pipe ID

No.
26 170 26 55
27 7 - -
28 127 - -
29 158 150 128
30 10 26 -
31 10 - -
32 178 115 -
33 27 163 -
34 50 - -
35 170 20 -
36 162 - -
37 110 128 -
38 96 132 -
39 164 172 -
40 180 146 -
41 83 150 -
42 119 - -
43 114 86 -
44 6 43 -
45 41 172 -
46 83 - -
47 178 50 -
48 83 - -
49 160 - -
50 19 21 -

The results of this study show that the system reliability of Al-Khobar network is
69.73% and the nodal reliabilities of the system are ranging from a minimum value of
74.32% to a maximum value of 99.99%. This means that the probability of Al-Khobar
water distribution system to have a required minimum pressure of 33 psi at all
junctions is 69.73%, and the probability that each junction will have a required
minimum pressure of 33 psi varies from 74.32% to 99.99% depending upon the
individual junction.
78 Muhammad A. Al-Zahrani and Juned Laiq Syed

Table 3 summarizes the calculated reliability at selected nodes. It reveals that

among all the 131 junctions of Al-Khobar network, junction J-57 has the lowest
reliability value equal to 74.32% while junction J-520 has the highest reliability value of
99.99%. The reason for lower reliability value at junction J-57 is due to the larger
number of cut-sets resulting in higher cut-set failure probability. On the other hand, the
high reliability value observed at junction J-520 is due to its location, which is near to
YRMK (Al-Yarmouk water reservoir tank) where high pressure is expected during the
day.

Table 3. Nodal and system reliability values of Al-Khobar water distribution network
Junction ID Reliability (%)
J-14 84.47
J-16 84.45
J-31 84.47
J-116 86.85
J-57 (Min) 74.32
J-520 (Max) 99.99
System 69.73

In the evaluation process of water distribution system reliability, the mean and
standard deviations at the nodes are required to estimate hydraulic availability needed to
calculate nodal and system reliabilities. The results indicate that higher values of mean
and standard deviations of nodal pressures will increase the variation of the generated
values of pressure at the nodes, which will consequently result in low values of nodal
and system reliabilities.

The number of pipe closure combinations also affects the nodal and system
reliabilities. A high number of pipe closure combinations will result in a large number
of cut-sets that will result in low values of nodal and system reliabilities. Therefore,
appropriate number of pipe closure combinations should be selected based on field
experience to get realistic values of nodal and system reliabilities.

Finally, it is recommended to use the developed methodology for large water

distribution networks, such as the entire Al-Khobar water distribution system. This is
because the possibilities of pipe closure combinations are high compared to the case
dealing with small water distribution network.
 Evaluation of Municipal Water Distribution System Reliability … 79

Conclusion
Reliability analysis of a water distribution system is concerned with measuring its
ability to meet consumers’ demands in terms of quantity and quality, under normal and
emergency conditions. This paper developed a methodology based on minimum cut-set
method, which can be applied to evaluate the hydraulic reliability of water distribution
systems. The methodology applied on Al-Khobar water distribution system shows that
the system reliability is 69.73%, which means that the probability of Al-Khobar water
distribution system to have a required minimum pressure of 33 psi at all the junctions is
69.73%. The reason for low reliability of Al-Khobar network, compared to similar
distribution networks, is due to the high annual failure rates of the pipes resulting in
higher cut-set failure probabilities. If this reliability value is compared with the
reliability value of Tucson, Arizona, U.S.A. which is 96.0 %, it is observed that the Al-
Khobar network is less reliable. The reason of high reliability value of Tucson water
distribution network is the proper maintenance of the network by the private water
business companies, as in U.S.A. most of the municipal water distribution networks are
managed and maintained by private water business companies. The system reliability
value of Al-Khobar network could be improved by reducing the annual failures of the
pipes through proper maintenance and replacement of older pipes in the network. The
improvement in system reliability could be easily achieved by simply improving the
reliability of individual junctions. Therefore, as a preliminary step in improving system
reliability values, critical junctions of the network should be identified and the pipes
affecting the reliability values of these junctions should be properly maintained and
replaced, if necessary.

Depending upon the residential and commercial consumers’ requirements of Al-

Khobar network, “levels of reliability” should be established and water agencies should
be made responsible for its proper implementation and monitoring.

Acknowledgment. This study was supported by the Department of Civil Engineering and
the College of Graduate Studies at King Fahd University of Petroleum and Minerals
(KFUPM), under SABIC grant No. SABIC 2003/01. This generous support is gratefully
acknowledged.

References
[1] Kaufmann, A.D., Grouchko, D. and Croun, R. Mathematical Models for the Study of the Reliability of
Systems. New York, N.Y.: Academic Press, 1977.
[2] Goulter, I.C. “Analytical and Simulation Models for Reliability Analysis in Water Distribution
Systems.” In: E. Cabrera and A. Vela (Eds.), Improving Efficiency and Reliability in Water Distribution
Systems. Dordrecht: Kluwer Academic Publishers, 1995.
[3] Cullinane, M.J., Lansey, K.E. and Mays, L.W. “Optimization Availability-based Design of Water
Distribution Networks.” ASCE J. Hydraulic Eng., 118, No. 3 (1992), 420-441.
[4] Germanopoulos, G., Jowitt, P.W. and Lumbers, J.P. “Assessing the Reliability of Supply and Level of
Service for Water Distribution Systems.” Proc. Instn. Civ. Engrs., Part 1, Water Engineering Group,
(1986), 413-428.
80 Muhammad A. Al-Zahrani and Juned Laiq Syed

[5] Lansey, K.E., Duan, N. and Mays, L.W. “Water Distribution System Design under Uncertainties.”
ASCE J. Water Resources Planning & Management, 115, No. 5 (1989), 630-644.
[6] Quimpo, R.G. and Shamsi, U.M. “Reliability-based Distribution System Maintenanace.” ASCE J. Water
Resources Planning & Management, 117, No. 3 (1991), 321-339.
[7] Mays, L.W. “Methods for Risk and Reliability Analysis.” Conf. Proc., Part of Risk-based Decision
Making in Water Resources VI, (1993), 26-44.
[8] Calvin, R.S., Yacov, Y.H., Duan, L. and James, H.L. “Capacity Reliability of Water Distribution
Networks and Optimum Rehabilitation Decision Making.” Water Resources Res., 32, No. 7 (1996),
2271-2278.
[9] Ostfeld, A. “Reliability Analysis of Regional Water Distribution Systems.” Urban Water, 3 (2001), 253-
260.
[10] Shinstine, D.S., Ahmed, I. and Lansey, K.E. “Reliability/Availability Analysis of Municipal Water
Distribution Networks: Case Studies.” ASCE J. Water Resources Planning & Management, 128, No. 2
(2002), 140-151.
[11] Goulter, I.C. and Coals. “Quantitative Approaches to Reliability Assessment in Pipe Networks.” ASCE
J. Transportation Eng., 112, No. 3 (1986), 287-301.
[12] Su, Y.C., Mays, L.W. and Lansey, K.E. “Reliability-based Optimization Model for Water Distribution
Systems.” ASCE J. Hydraulic Eng., 113, No. 12 (1987), 1539-1556.
[13] Billinton, R. and Allan, R.N. Reliability Evaluation of Engineering Systems Concepts and Techniques.
London: Pitman, 1983.
[14] Rossman, L.A. EPANET – User’s Manual. Cincinnati, Ohio: United States Environmental Protection
Agency (USEPA), 2000.
[15] Fujiwara, O. and De Silva, A.U. “Algorithm for Reliability Based Optimum Design of Water
Networks.” ASCE J. Environmental Eng., 116, No. 3 (1990), 75-87.
 Evaluation of Municipal Water Distribution System Reliability … 81

‫ﺗﻘﻴﻴﻢ ﻣﻮﺛﻮﻗﻴﺔ ﺃﺩﺍﺀ ﻧﻈﺎﻡ ﺷﺒﻜﺔ ﺗﻮﺯﻳﻊ ﺍﳌﻴﺎﻩ ﺑﺎﺳﺘﺨﺪﺍﻡ‬

‫ﻤﻮﻋﺎﺕ ﺍﻟﺼﻐﺮﻯ ﺍﳌﺘﻘﻄﻌﺔ‬‫ﻃﺮﻳﻘﺔ ﺍ‬

‫ﳏﻤﺪ ﻋﺒﺪﺍﷲ ﺍﻟﺰﻫﺮﺍﱐ ﻭ ﺟﻨﻴﺪ ﻟﻴﻚ ﺳﻴﺪ‬

 ‫א‬،‫א‬‫א‬
 ‫א‬‫א‬‫א‬،‫א‬‫א‬،‫א‬‫א‬

E٢٠٠٥L٠٤L٢٥‫؛‬٢٠٠٤L٠٤L٢٠F

‫א‬‫א‬‫א‬‫א‬‫א‬K‫א‬
‫א‬‫א‬K‫א‬‫א‬
 ‫א‬‫א‬E١FW‫א‬‫א‬‫א‬،‫א‬ ‫א‬‫א‬
 ،‫א‬‫א‬‫א‬‫א‬‫א‬‫א‬ (EPANET)
K‫א‬‫א‬‫א‬‫א‬‫א‬‫א‬‫א‬‫ א‬E٢F
‫א‬‫א‬‫א‬‫א‬‫א‬‫א‬
‫א‬‫א‬‫א‬‫א‬‫א‬ K‫א‬‫א‬
 K‫א‬
 82 Muhammad A. Al-Zahrani and Juned Laiq Syed

View publication stats