*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.

E-mail address: raido@staff.ttu.ee (R. Puust).
1
A Review of Methods for Leakage Management in Pipe Networks

R. Puust
a,*
, Z. Kapelan
b
, D.A. Savic
b
, T. Koppel
a

a
Tallinn University of Technology, Department of Mechanics, Ehitajate tee 5, Tallinn,
19086, Estonia;
b
University of Exeter, School of Engineering, Computing and Computer Science, Centre
for Water Systems, Harrison Building, North Park Road, Exeter EX4 4QF, UK

Abstract
Leakage in water distribution systems is an important issue which is affecting water companies and their
customers worldwide. It is therefore no surprise that it has attracted a lot of attention by both practitioners
and researchers over the past years. Most of the leakage management related methods developed so far
can be broadly classified as follows: (1) leakage assessment methods which are focusing on quantifying the
amount of water lost; (2) leakage detection methods which are primarily concerned with the detection of
leakage hotspots and (3) leakage control models which are focused on the effective control of current and
future leakage levels. This paper provides a comprehensive review of the above methods with the objective
to identify the current state-of-the-art in the field and to then make recommendations for future work. The
review ends with the main conclusion that despite all the advancements made in the past, there is still a lot
of scope and need for further work, especially in area of real-time models for pipe networks which should
enable fusion of leakage detection, assessment and control methods.

Keywords: Distribution system; leakage assessment; leakage control; leakage detection; pipe network;
water distribution systems; leakage model, pressure-dependent leakage.



*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
2
1. Introduction

Leakage occurs in all water distribution systems nowadays. As noted by William Hope
long time ago (1892), “there is no water-supply in which some unnecessary waste does
not exist and there are few supplies, if any, in which the saving of a substantial
proportion of that waste would not bring pecuniary advantage to the Water Authority”.
The amount of water leaked in water distribution systems varies widely between
different countries, regions and systems, from as low as 3–7% of distribution input in the
well maintained systems in Netherlands (Beuken et al., 2006) to 50+ % in some
undeveloped countries and less well maintained systems (Mamlook and Al-Jayyousi, 2003;
Lambert, 2002).
Leakage is not just an economical issue as it is often perceived and presented by
water companies but it is also an environmental, sustainability and potentially a health
and safety issue. As noted by Colombo and Karney (2002), leakages cause inefficient
energy distribution through the network (thus wasting energy used for pumping the
water) and, also, may affect water quality by introducing infection into water distribution
networks in low pressure conditions.
A number of past, review type papers exist in the field of leakage modelling and
management. One of the earliest review papers is the paper by Morris Jr. (1967) which
provided an overview of potential causal factors leading to water pipe breaks. A report
summarising different leakage control policies can be found in Goodwin (1980).
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
3
Comparisons of the key attributes of different leak detection methods are given by Cist
and Schutz (2001). Another review and classification of leakage detection methods is
reported by Liou et al. (2003). A review of calibration methods in water pipelines
(including leaks) can be also found in Kapelan (2002) and Savic et al. (2009).
Unlike the existing approaches mentioned above, which are focusing on a
particular leakage issue (usually leakage detection), this paper looks wider by considering
the overall leakage management process. The objective of this paper is to review the
methods and models developed in the past used in different phases of this process, from
becoming aware of the leak existence to controlling the level of leakage in the system. It
is hoped that this way the new promising research areas will be found as they often exist
along the boundaries of current research areas. More specifically, this review looks into
past methods and models developed that can be used to either assess, detect or control
leakage in distribution (and other) pipe systems. The main objective is to identify the
advantages and disadvantages of all existing approaches and to then use the observations
made to suggest possible future research work in the field.
The paper is laid out as follows. After this introduction, the relevant background
information is presented in section 2. This is followed by the review of leakage
assessment methods in section 3, detection methods in section 4 and leakage control
methods in section 5. Finally, the main conclusions are drawn in section 6 of the paper.

2. Background
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
4

Different definitions of leakage in distribution systems exist. The most frequently used
one defines the leakage as (amount of) water which escapes from the pipe network by
means other than through a controlled action (Ofwat, 2008). Water leakage in distribution
systems is typically classified into background and burst related leakage (O’Day, 1982).
Bursts (i.e. main breaks) represent structural pipe failures and background leaks represent
the water escaping through inadequate joints, cracks, etc. Leaks can also exist in service
reservoirs and tanks.
Leakage in distribution systems can be caused by a number of different factors.
Some examples include bad pipe connections, internal or external pipe corrosion or
mechanical damage caused by excessive pipe load (e.g. by traffic on the road above or by
a third party working in the system). Other common factors that influence leakages are
ground movement, high system pressure, damage due to excavation, pipe age, winter
temperature, defects in pipes, ground conditions and poor quality of workmanship.
Therefore, the presence of leakage may damage the infrastructure and cause third party
damage, water and financial losses, energy losses and health risks.
Leakage is dependent on system pressure. Basically, the higher the pressure, the
larger the leak flow and vice versa. Initially, an orifice type equation (Wiggert, 1968) was
used to describe this relationship. Although the orifice equation is still widely used in
many research studies, the user must be aware that the equation can lead to misleading
results when the pipe in question is not made of a rigid material (Greyvenstein and van Zyl,
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
5
2005) or when the pressure is negative (Todini, 2003). Lately, a more generalised leak
flow – pressure equation has been adopted which allows specifying leakage exponent
different from 0.5 (Germanopoulos, 1985). It has been shown that the value of this
exponent depends on the type of leak, pipe material behaviour, soil hydraulics and water
demand (Cassa et al., 2005; Greyvenstein and Van Zyl, 2005; Walski et al., 2006; Noack and
Ulanicki, 2006). For example Van Zyl and Clayton (2005) note that when leakage is
analysed as pressure dependent, demand should follow the same procedure. More on
leakage as a hole in a pipe and its characteristics can be found in Beck et al., (2005a, b) and
Coetzer et al. (2006). Various studies about the pressure dependent leakage modelling can
be found from the literature. Modelling based on leak discharge coefficient and leak area
can be found from the articles by May (1994); Vela et al. (1995); Simpson and Vitkovský
(1997); Vitkovský and Simpson (1997); Dunlop (1999); Hernandez et al. (1999); Stathis
and Loganathan (1999); Alonso et al. (2000); Rossman (2000); Ulanicki et al. (2000);
Ulanicka et al. (2001); Vitkovský et al. (2003a) and Verde (2005). Modelling that also
included pipe characteristics can be found from Germanopoulos (1985; 1995);
Vairavamoorthy and Lumbers (1998); Martinez et al. (1999); Reis and Chaudry (1999);
Tucciarelli et al. (1999); Ainola et al. (2000) and Dias et al. (2005).

3. Leakage Assessment Methods

*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
6
The objective of leakage assessment (i.e. water audit) is to estimate the quantity of water
lost in the system analysed without worrying where the leaks are actually located. The
assessment methods developed so far can be broadly classified into the following two
main groups: (a) top-down leakage assessment methods and (b) bottom-up leakage
assessment methods.


3.1 Top-down approaches
The objective of top-down leakage assessment approaches is to estimate the leakage in a
particular system by evaluating different components of the overall water balance,
primarily the water consumed for different purposes. The two main approaches used are
the IWA approach (Lambert and Hirner, 2000) and the approach used by the OFWAT in
the UK. Although quite similar, there are some differences between the two approaches
due to slightly different terminology and definitions used for some water balance
components.
More information about the general leakage assessment can be found from
Stenberg (1982), Thornton (2002), Farley and Trow (2003), and Scott and Barrufet (2003). The
latest reports about average losses in the UK (based on areas operated by different water
service companies) can be found in (AHL, 2006). The latest guidance notes on leak
location and repair are published in Pilcher (2003) and Pilcher et al. (2007).
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
7
Despite the simplicity of a top-down type leak assessment, the leakage estimate
obtained via this method is referred to as a crude estimate. Gathering such information
helps to decide what the next step in leakage studies should be for a particular network
but it does not help to bound potential leak areas, let alone locate leaks.

3.2. Bottom-up approaches
Bottom-up type leakage assessment can be considered the second part of the audit
process. This procedure is implemented when the company has confirmed the data used
in the top-down portion. It includes every area of the company’s operation: billing
records, distribution system, accounting principles etc. The audit’s main purpose is to
find out the efficiency of the water distribution system and the measures needed to
achieve these. Bottom-up audits require the most accurate and up-to-date data possible.
Bottom-up real loss assessment can be carried out in two different ways: (a) 24
Hour Zone Measurement (HZM) or (b) Minimum Night Flow (MNF) analysis. HZM
needs a temporary isolated area of the distribution network that is supplied from one or
two inflow points only. In these areas, 24 hour inflow measurements shall always be
logged along with pressure measurements. MNF in urban situations normally occurs
during the early morning period, usually between 02:00 and 04:00 hours (Liemberger and
Farley, 2004). The estimation of the real loss component is carried out by subtracting
legitimate night uses from the MNF. To get a satisfactory estimate of the daily leakage,
Stenberg (1982) has found that night leakage flow rates should then be multiplied by 20
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
8
hours. This assumption does not take into account that pressure is not constant over a
period of time. Therefore Lambert (2001) suggests using a method called fixed and
variable area discharges (FAVAD). This method uses the following equation:

β
H k q ⋅ =
(1)

Where
q
= volume rate per unit length;
( )
β
g b C k
d
2 ⋅ =
;
β
= leakage exponent; d
C
=
discharge coefficient;
b
= width of the slot;
g
= gravitational acceleration;
H
= pressure
head. Leak exponents vary, being close to 0.5 with fixed area leakage path (hole in pipe)
and 1.5 with variable leakage path (crack in pipe). The increase or decrease of real losses
due to a change in pressure can then be computed by FAVAD concept as:

( )
β
2 1 2 1
/ / H H L L =
(2)

where L
1
and L
2
are leakage rates and H
1
, H
2
are pressure heads at respective times.
Thereafter leakage can be simulated over the full 24h period (see Figure 1).
At the end of the real loss assessment process, the advantage of the combined top-
down, bottom-up and component analysis (Table 1) that were introduced in the early
1990s (Farley and Trow, 2003) becomes obvious. Several countries have had their own
measures or indicators. For example the sample measures could be: percentage of average
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
9
daily flow (USA, France); m
3
/km of mains/hours (German, Japan); litres/property/hour
(UK) and litres/service connection/hour. The problem is that those indicators do not take
account of component analysis techniques therefore additional performance measures
have to be used. Comparison of some additional performance measures can be found in
Tuhovčák et al. (2005). Performance indicators should count the possibility of consumption
decreases seasonally or annually (non-revenue water does not) and also take into account
pressure relations in the pressure zone. Therefore recommended indicators should always
indicate its robustness. Robustness can be defined with a level and a function (Table 2).
The most commonly used leak index nowadays is Infrastructure Leakage Index - ILI
(Lambert, 2003; Farley and Trow, 2003). The advantages of using ILI are that it can be
consistently applied across a range of utilities and that it is a measure of what can be
achieved given the condition of the infrastructure. Its key disadvantage is that it is not
easily understood by non-technical readers. Additionally it does not take into account the
relative costs of leakage management (and other marginal costs, like environmental costs)
and it is not able to define what level of reduction is economically feasible. An additional
advantage of calculating an ILI index is that it can be used to calculate the leakage
exponent (Thornton and Lambert, 2005):

( )
100
/ 65 . 0 1 5 . 1
p
ILI ⋅ − − = β
(3)

*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
10
where
p
= the percentage of rigid pipes in network.
Using Eq.(3) the calculation for different leakage exponents for different
networks/countries/regions can be found (see Table 3).
Although classical minimum night flow analysis (Araujo et al., 2003a; Covas et al.,
2006a; Garcia et al., 2006) can reduce real losses (leakage) considerably, there are many
other methods of leak assessments that could possibly be used depending on network
architecture (see Puust, 2007). In addition to water audits the assessment can be done also
using some statistical analysis for detecting the magnitude of leaks (Buchberger and
Nadimpalli, 2004). This is expected to be more accurate but with a cost of a need of
continuous, high resolution measurements of discharge at one or more locations within
the district metering area (DMA). This can be problematic in some cases because the
high resolution data measurements are not used very often within a DMA and the
location of data acquisition systems must be carefully planned in such case studies
(Vitkovský, et al. 2003c; Kapelan et al. 2003c, 2005; Behzadian et al. 2009).

4. Leakage Detection Methods

Historically, leakage assessment studies have been carried out to quantify total losses
including, if possible, real and apparent losses. This was followed by the development of
leakage detection methods with the aim to detect and locate leaks. Although some
leakage detection methods have been around for years, because of constant development,
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
11
they are getting increasingly high tech and sophisticated than ever before. Still, regardless
of whether the methods are equipment or non-equipment based, it is common practice to
use some leak detection method in conjunction with other methods.

4.1. Leakage awareness methods
The term ‘leak awareness’ is used to explain the discovery of a leak in a particular area
within the network. It does not give any information about its precise location. Usually a
hydraulic model is needed for the leakage awareness test. Various hydraulic models have
been proposed to detect leaks in water distribution systems. Those methods usually
involve calibration/optimisation techniques to analyse the different areas of the network.
The problem is formulated as a constrained optimisation problem of weighted least-
square type to minimise the objective function E:

( ) ( ) ( )
∑ ∑ ∑
= = =
− + − + − =
N
k
i
m
i p
Q
i
i
m
i q
P
i
i
m
i h
p p w q q w h h w E
1
2
1
2
1
2
(4)

where P and Q are the number of pressure, flow measurement respectively,
m
i
h
is the
measured head at node i, i
h
is the computed head at node
i
,
m
i
q
is the measured flow at
pipe i, i
q
is the computed flow at pipe i,
m
i
p
is the prior estimate (pseudo measurement),
i
p
is the prior estimate and N is the number of prior estimate,
w
is a weight factor for
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
12
pressure/flow and prior estimate part. Prior estimates were introduced into the
minimisation problem by Kapelan (Kapelan, 2002; Kapelan et al., 2000, 2003a, 2003b, 2004)
to avoid the ill-posed problem (that is there is no solution, no unique solution or the
solution is unstable), to improve the accuracy of the estimated calibration parameters and
to increase the speed of the convergence process. It should be noted that prior estimates
work better with pipe friction factors as these are less sensitive than leak effective areas.
The minimisation of Eq. (4) gives the solution to an inverse problem (Pudar and
Liggett, 1992; Stathis and Loganathan, 1999). Various minimisation algorithms have been
used to minimise the objective function, Eq. (4). When steady state regime is used, both
pressure and flow measurements can be used. In a transient flow regime flow
measurements are difficult to use because most flow meters do not react instantaneously
to a change in flow (Chen, 1995). Early adoptions of fluid transients for leak detection can
be found from Wiggert (1968), Nicholas (1990), Liggett (1993), Liggett and Chen (1994).
The use of fluid transients for leak detection has gained popularity over the last
decade as a massive amount of data can be gathered in a very short period of time
therefore ensuring that the inverse problem will always be overdetermined. Another good
advantage over steady state calculation is that pressure waves are less affected by friction
than the general flow and thus the precise friction values become less important to the
calculation. Therefore, using transients, the leak detection and calibration (friction
factors) can be done simultaneously, thus providing a solution to the problem of unknown
or poorly known friction. Fluid transients are used to probe the pipeline in much the same
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
13
way as radar and sonar are applied to locate and identify objects. The reason why
methods based on transients are mainly used on single, grounded pipelines is that an
uncertainty of the system does affect the results considerably (pressure wave reflections
from each feature of the pipe). For undergrounded pipes the system’s architecture can be
hardly followed thus its applicability in such situations is still questionable.
A number of hydraulic transient-based techniques for detecting and locating
existing leaks are described in the literature: leak reflection method (LRM) (Jönsson, 1995,
2003; Brunone, 1999, Brunone and Ferrante, 1999, 2001, 2004; Covas and Ramos, 1999),
inverse transient analysis (ITA) (Liggett and Chen, 1994; Liou, 1994; Vitkovský et al., 2000;
Kapelan et al., 2003a, 2003b; Covas et al., 2001a, 2003, 2005b; Covas and Ramos, 2001b;
Stephens, et al., 2004; Wang et al., 2006; Soares et al., 2007), impulse response analysis
(IRA) (Liou, 1998; Vitkovský et al., 2003b; Kim, 2005), transient damping method (TDM)
(Wang et al., 2002, 2003), frequency domain response analysis (FRM) (Mpesha et al., 2001,
2002; Stoianov et al., 2001; Ferrante and Brunone, 2001a, 2001b, 2003a, 2003b; Covas et al.,
2005a; Ferrante et al., 2005; Lee et al., 2003; 2005a; 2005b, 2006; Zecchin et al., 2005, 2006).
The main objective of all transient leak detection methods is the same – to extract
information about the presence of a leak from the measured transient trace. A transient
event is generated either by system elements (i.e. inline valves and pumps) or special
devices (for example solenoid side discharge valves).
In the leak reflection method (LRM), a transient wave is travelling along a
pipeline and it is partially reflected at the leak. The location of the leak can be then
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
14
identified from the measured pressure trace (Figure 2). The magnitude from the leak
depends on the ratio between the size of the generated transient wave and the size of the
leak orifice. LRM methods are so far used only in single pipe case studies in laboratory
conditions.
The inverse transient analysis method (ITA) was first introduced by Liggett and
Chen (1994). The ITA uses least-squares regression between modelled and measured
transient pressure traces. The leak is usually modelled at network nodes and the
minimisation of the deviation between the measured and calculated pressures produces a
solution of leak location and size (Figure 3). The ITA method is a well-researched topic
but since its introduction, the main effort has been focused on the development of the
mathematical part of the technique and not on experimental validation or field testing.
Some limited experiences from laboratory and fields tests can be found from Vitkovský et
al., (2001), Stephens et al., (2004), Covas et al. (2005b) and Saldarriaga et al. (2006). As with
LRM, the tests are made on single pipeline rather than on a network. Application
difficulties lie in the fact that ITA needs an accurate modelling of the transients and
boundary conditions of the pipe system. To address the latter, a greater emphasis should
be directed toward analysis of errors and strategies to deal with the uncertainties in
general (Vitkovský et al., 2007). Model error is the most likely limiting factor in successful
field application of ITA and its results should never be presented without quantification
of their uncertainty.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
15
The impulse response analysis (IRA) is based on the fact that the transient
propagation along the pipeline is affected by the friction of the pipe wall and other loss
elements such as leaks. This effect results in damping of the transient wave. A leak can
be detected when a transient damping in the same pipeline is compared with and without
a leak (Figure 4). The lack of information about tests in real pipeline systems where noisy
data would be used makes it a less important method when compared with LRM and
ITA. It has one advantage when the comparison should be made with TDM or FRM.
Namely, in IRA no discretization of the pipeline is needed and the shape of the generated
transient is not important.
In the transient damping method (TDM) it is analytically derived that friction
related transient damping in a pipeline without a leak is exactly exponential and the
corresponding damping in a pipeline containing a leak is approximately exponential
(Wang et al., 2002). The rate of the leak-induced damping depends on leak characteristics,
the pressure in the pipe, the location of the transient generation point and the shape of the
generated transient. Tests on a laboratory pipeline showed successful leak detection
(Figure 5) but in a real situation, friction is not the only cause of transient damping.
Transient damping can be caused by other physical elements like joints, connections, fire
hydrants and pipe wall deterioration products. The modelling of these elements can be
complicated and in some cases even impossible. Therefore it may be difficult to estimate
the leak-free damping for a real pipeline.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
16
The frequency response method (FRM) uses the analysis of transient response in
the frequency domain. Fourier transforms are used to transform time-domain data into the
frequency domain. Leak location can be obtained when the dominant frequencies of no-
leak and leaking pipelines are compared (Figures 6, 7). Performance of the method is
strongly influenced by the shape of the transient and the measurement location. As with
other methods based on transients, only pipeline applications of frequency response
analysis are presented in the literature. Some of the case studies that are based on
pressure transients are summarised in Table 4.
There are many other leak awareness methods but only three of them have been
applied to pipe networks (Saldarriaga et al., 2006; Deagle et al., 2007; Wu and Sage, 2007). It
should be noted that most of them are very rarely used and/or do not have any practical
tests made to support the idea. One of the reasons why these model based techniques are
not so widely used could be because of the low flow rates in pipelines that eliminate the
possible use of commonly used pressure measurement devices that are cheap and easily
manageable, but not effective when used in low flow conditions.
When leak awareness methods are under discussion it should be also mentioned
that very few of them are probabilistic ones. In that respect, the Bayesian system
identification methodology has been used by Poulakis et al. (2003), Rougier (2005) and
Puust et al. (2006). The main reason to use a Bayesian interface in leakage studies is that
normally we are dealing with different kinds of errors that cannot always be included in
calculations. Therefore to make more sense, the final discrete value is bounded with a
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
17
certain probability that gives us more information about the result reliability. The
drawback is that usually such procedures need a great deal of computer power for
calculations. In general, probabilistic models are used also to conduct criticality analysis,
where small cracks in pipes might cause the overall failure of much larger systems (like
cooling systems in nuclear power plants, Rahman et al., 1997).
Great effort has been made so far in the development of model based leak
detection methodologies. Whilst this development will continue, it is obvious that some
methods will be suitable for application to simple systems only (e.g. single pipelines). An
example of this is transient based methodologies. Because of their limitations when
applied to network systems it is clear that development of transient based methodologies
for leakage control will be limited to single pipelines. For a general reference about
leakage control please see section 5.

4.2 Leakage localisation methods
Leak localising is an activity that identifies and prioritises the areas of leakage to make
pinpointing of leaks easier. Some methods/techniques that belong to this group are:
acoustic logging (Moyer, 1983; Hough, 1988; Rajtar and Muthiah, 1997; Hessel et al., 1999;
Hunaidi and Chu, 1999; Miller et al., 1999; Lockwood, 2003; Shimanskiy, 2003; Bracken and
Hunaidi, 2005; Muggleton et al., 2006), step-testing (Farley and Trow, 2003, Pilcher et al.,
2007), ground motion sensors and ground penetrating radars (Hunaidi, 1998; Lockwood et
al., 2003; O'Brien et al., 2003).
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
18
The most well-known and effective leak localising method is step-testing and it
has been used by several water utility companies for quite some time. Step-testing is an
activity whereby the area is subdivided by the systematic closing of valves during the
period of minimum night flow. Depending on the methodology used, step-testing may
cause backsiphonage, the risk of infiltration of ground water and some parts of the
network can be without water for a period of time. Not all networks are planned with the
possibility of future step-testing in mind and therefore it may be difficult to apply.
Because of a need of careful planning, night work involving the step-testing has been
replaced by acoustic logging during the 1990s (Pilcher et al., 2007).
Acoustic logging (AL) is performed using vibration sensors or hydrophones,
which are temporarily or permanently attached to the pipe fittings. The distance between
each other typically varies between 200 to 500 m. As with step-testing the data is
collected at night times, usually between 2 and 4 am. Downloaded data will then be
analysed statistically for detection of leak signals (Figure 8). Although a wide area may
be covered quickly, for a successful leak detection good skill is required. The fact that
quiet leaks may not be heard and the background noise can’t be ignored makes it difficult
to apply in certain situations.
The application of ground penetrating radar (GPR) for leak location has been
given a lot of attention during the last few years (Farley, 2008). Ground penetrating radar
inspection is a non-destructive geophysical method that produces a continuous cross-
sectional profile or record of subsurface features (Figure 9). Methods like this could be
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
19
used to locate leaks in water pipes by detecting either underground voids created by
leaking water as it circulates near the pipe or by detecting anomalies in the pipe depth as
measured by radar. GPR has evolved for some years now. It has been previously
described as a time consuming methodology but recent studies show that along
transmission main routes it can be carried out at 15 – 30 km per hour, depending on
location and traffic. As GPR technology is similar in principle to seismic and ultrasound
techniques, the main disadvantage comes from the fact that anomalies like metal objects
in the ground can lead to false conclusions and it might not be applicable in cold climates.
Some developed GPR technologies have a penetration capability of up to 2 meters into
the ground. For example in northern European countries the water pipe bottom should be
laid down in some occasions at least 1.8m deep to avoid water freezing. Therefore GPR
technology can not give trustworthy results on those extreme occasions and in situations
where main pipes are excavated even deeper into the ground. It should be still noted that
this methodology is a good alternative in situations when large diameter or non-metallic
pipes need monitoring.
In summary, leakage localisation methods can be used on their own or
before/following the application of some other method. For example, if a hydraulic model
of the analysed system is available then some numerical (i.e., inexpensive) leak detection
method may be used before the leak localisation method, to narrow the area searched for
the leak. However, if the hydraulic model is not available (or not updated regularly) then
a leak localisation method could be used on its own.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
20
In general district audits are labour-intensive and costly, since they are performed
at night. A more recent trend is that permanent flow meters are installed that are
connected telemetrically to a supervisory control and data acquisition ( SCADA) system.
The transmitted flow rate data are automatically analysed to detect unusual increases in
flow patterns (Mounce et al., 2009). Based on experience with water system, the increase
in flow rate can be explained by leakage or not. District audits and step testing help
identify areas of the distribution system that have excessive leakage. No information
about the exact location of leaks is given. When step-testing or SCADA system is not
available, some other technology is needed that can be used for leak localisation with a
reasonable time. The reasonable time to detect leakage varies depending on the leak flow
rate. Small leaks are more difficult to locate, especially when using acoustic logging for
plastic pipes. As a consequence, the GPR technology was developed (any pipe material
can be surveyed) and various studies published demonstrate promising performance. GPR
is probably one of the key technologies studied in Europe currently (WATERPIPE,
2009).

4.3. Leakage pinpointing methods
Leakage pinpointing methods include methodologies that are the most accurate in today's
leak detection surveys. Three main groups described here are based on (a) leak noise
correlators (Grunwell and Ratcliffe, 1981; Cascetta and Vigo, 1992; Gao et al., 2004, 2005,
2006; Hunaidi et al., 2004; Muggleton et al., 2004, Muggleton and Brennan, 2004, 2005); (b)
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
21
gas injection (Field and Ratcliffe, 1978, Hunaidi et al., 2000; Farley and Trow, 2003); and (c)
pig-mounted acoustic sensing (EPSRC, 2002; McNulty, 2001; Mergelas and Henrich, 2005; ).
The historical appearance of leakage pinpointing methodologies is given in Figure 10.
Leak noise correlators (LNC) are the most common technique for leak location
that was first introduced commercially into the marketplace in the late 1970s (Thornton,
2002). Their technology has been improved over the last few years quite considerably.
The Water Research Centre (WRC) in England was one of the leading research
institutions to apply the methodology onto real pipelines. To correlate the sound from a
leak, two microphones are located in contact with the pipe or valve stems at the same
time, with one microphone on each side of the leak (Grunwell and Ratcliffe, 1981; Stenberg,
1982). The sound is compared in the correlator, which is capable of determining the
difference in time for sound to reach the correlator. Knowing the speed of sound in the
pipe, it is then easy to calculate the distance to the leak, which will be independent of the
geophone, traffic noise, etc. For accurate leak localisation the pipe system should be
known precisely as a leak correlating in a branched section tend to show a leak on a tee
and not at its exact location. Such misleading information affects mainly excavation costs
and man-hours needed for a repair. The latest versions of leak noise correlators can
accurately locate a leak to within 1 metre in most pipe sizes. The distance between the
sensors can be as high as 3000 m but it depends highly on pipe material. For plastic pipes
this methodology is quite questionable as distance between the sensors should be quite
small – 15 to 100 m, making this method very slow. The method works best with clean,
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
22
small diameter metallic pipes in high water pressure areas where hard pipe backfill is
used.
In a tracer gas technique (TGT), a non-toxic, water-insoluble and lighter-than-air
gas, such as helium or hydrogen, is injected into an isolated section of a water pipe. This
is followed by ground scanning with a highly sensitive gas detector which should identify
any traces of escaped gas from the leak point(s) (Figure 11). Although this method is
widely used for machinery testing, it is normally prohibitive for leak detection because of
the high cost. Its effectiveness comes from aspects that through TGT multiple leak
locations can be found in a single pipe section or at a branched pipe systems where noise
correlation techniques usually fails or gives misleading results. The main disadvantage in
addition to high costs are that the gas could be trapped near the ceiling of water-filled
pipes and thus could not escape if leaks were not near the top of the pipe.
The pipe pig-mounted acoustic (PMA) technique has also been used for leak
detection (EPSRC, 2002; McNulty, 2001). This technique requires the insertion of a
microphone (or a pair of microphones) under pressure into the main. The velocity of
water carries the microphone to the leak position whereas the noise and its position are
continuously recorded. Some latest technology examples can be found from Chastain-
Howley, (2005) and Fletcher, (2008). Inline pigs are used to carry different kinds of
sophisticated measuring devices such as magnetic flux leakage (Mukhopadahyay and
Srivastava, 2000), hydroscopes (Makar and Chagon, 1999) or ultrasonic tools (Willems and
Barbian, 1998) along the pipeline. In general these tools need clean pipes and therefore it
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
23
is difficult to apply this methodology to old pipes where there may be heavy corrosion.
Access to the inside of the pipeline is also needed. Attention must be paid to the fact that
as pigs are in contact with the pipe inner wall their effect on the water quality must be
considered before a survey is performed.
Leak pinpointing techniques are the most precise technologies currently available
for leak detection. It should be remembered that such a precision comes with very high
costs in terms of equipment owning or renting and man-hours needed for surveys to be
carried out. Considering this and the length of time needed for implementation, it is
recommended to use leak pinpointing techniques in conjunction with some leakage
awareness or localisation method. Table 5 brings out some general guidance when leak
localisation or pin-pointing technique should be chosen.

5. Leakage Control Models

Leakage control models can be generally classified into the following two main groups:
(a) passive (reactive) leakage control and (b) active leakage control. A passive leakage
control is a policy of responding only to leaks and bursts reported by the public (in some
cases also by a company's own staff). Active leakage control concerns management
policies and processes used to locate and repair unreported leaks from the water company
supply system and customer supply pipes.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
24
Many water utilities still take a passive attitude of waiting for a problem to arise
and repairing it only when leaks become self-evident. For example, the appearance of
water on the ground surface following pipe failure is visually detected by the staff or
reported by customers. Manual location techniques are then used to identify the actual
location of the failure. This presents inevitable problems for customers (Ramos et al.,
2001). Passive policy is very straightforward and simple to use but it does not involve any
systematic action. Therefore this kind of acting is reasonable only in such water systems
where there are very low leakage levels, the average loss is constantly below 10 – 15%.
Even in low loss cases it is advisable to use some more advanced technology at the same
time (like SCADA system) as when using passive policy the overall loss can easily raise
to 40%.
Active leakage policy involves the techniques like: active leakage control and
active pressure management. There are also sectorisation and economic intervention but
those are not discussed here. The most appropriate leakage control policy will mainly be
dictated by the characteristics of the network and local conditions, which may include
financial constraints on equipment and other resources (Farley and Trow, 2003). The final
choice of the method is also based on economic considerations. Term ‘economically
viable’ can be defined with an economic curve of leakage (ELL) analysis that is described
in Figure 12.
The most widely used active leakage control methodology on single pipelines is
based on pressure transients (Misiunas et al., 2003, 2005a, 2005b, 2006). The lack of their
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
25
commercial availability gives them more attention from the research side and any system-
wide conclusions are hard to be made. The main drawbacks using transients were already
discussed in section 3.2.1. Additional comment made here is that in on-line leak detection
situations there are always pressure signal from normal operations (for example transients
caused by pump start-up, valve closures etc.) and those should be carefully eliminated
from automatic reports. A sample of an on-line leak analysis through pressure traces is
given in Figure 13. Active policy applications on network studies are much more difficult
to apply. From recent research papers the more promising are those that combine
hydraulic modelling software, GIS and SCADA system into one package (Tabesh and
Delavar, 2003). With advanced SCADA systems and large asset, customer and
maintenance databases, water service providers are facing the challenge of efficiently
extracting useful information from data. Data mining techniques can be used for different
purposes. For example, artificial neural network (ANN) models can be used for demand
forecasting (Bougadis et al., 2005) and for scanning large amounts of data like operational
variable and historical records to identify a failure event (Mounce and Machell, 2006;
Aksela et al., 2009; Mounce et al., 2007; 2008; 2009) or to estimate failure patterns. A
sample of other active control techniques applied onto real data is presented in Table 6.
There are many other methodologies for active leakage control that are not so commonly
used and therefore not discussed here. For a list of different active leakage technologies
please see Puust (2007). In general, available active leakage control techniques are either
expensive (and time consuming) or have a long leak detection and location time. Active
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
26
leakage control techniques are used regularly to survey the system for leaks and hence to
reduce the time elapsed between the burst occurrence and its repair thus reducing the
number of potential customer complaints. The main drawback of the active leakage
control is that it is labour intensive and expensive.
Active pressure management has been called a well-proven method that has an
effect on the whole network or pressure zone. Previously it has been shown that leakage
is tightly coupled with network pressure (Eq. 1). Therefore when overall pressure is
reduced, the same happens to leakage. One should still be aware that in such conditions
the leak detection itself is quite challenging because of a reduced leak flow. Pressure
reduction in water distribution systems is normally achieved through pressure reducing
valves (see Figure 14). The objective of pressure reduction is to ensure the target pressure
at any given zone/area/node satisfies the customers. When pressure reduction is made
dynamically over a period of time, some computer algorithm/program can definitely
make this step easier. For example genetic algorithms are used for that purpose in Reis et
al. (1997). There are many other optimisation techniques available in the literature that
can achieve this but their mathematical advances are out of scope in this review.

6. Conclusions and Future Work Recommendations

*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
27
Leakage assessment, detection and control methods have come a long way since their
introduction in the mid 1950s. Based on the review completed and presented here, the
following main conclusions and future work recommendations are made.
Bottom-up leakage assessment methods are still preferred to top-down approaches
despite the fact that they are much more data hungry and time consuming to use. The
main value of top-down approaches is seen in making fast system-level leakage
assessments but also in verifying/controlling the results obtained by using bottom-up
methods. A certain novel value is seen in integrating these methods, especially bottom-up
methods, with pressure-driven hydraulic models of these systems (e.g., see Giustolisi et
al. 2008). Finally, both approaches are expected to benefit in the future from explicit
uncertainty analysis used to characterise and quantify the major sources of errors
involved in the leakage assessment process. This should be made possible with the
constant increase in better yet cheaper computational power available.
When it comes to leakage detection methods, significant advances have been
made in the past in both equipment-based and numerical models. The hardware based
methods (e.g. leak noise correlators) still remain superior in terms of detection accuracy
but also remain much more expensive to use than the numerical models. Further
developments of the promising equipment-based leak detection methods are envisaged
(e.g. pig-mounted acoustic sensing devices and/or ground penetrating radars).
With regard to the use of various transient based methods for leakage detection it
should be noted that these methods had limited success so far, typically in simpler pipe
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
28
systems only. It is envisaged that transient simulation models need to be developed
further before they can be utilised for leakage detection and assessment in more complex
pipe systems.
The further development of other numerical (i.e. non-equipment) based methods
is envisaged, especially the on-line type methods for real-time detection and diagnosis of
leaks caused by pipe bursts in networks. This should be made possible by the latest
developments in the pressure and flow sensor technology which should enable water
companies to install larger number of more accurate and cheaper devices in the near
future. The latest advancements made in the development of water quality (e.g. turbidity)
sensors could be potentially utilised too, through additional information available (e.g.
turbidity tends to increase significantly during pipe burst events). The most promising
techniques in the context of on-line models include various Artificial Intelligence
techniques, e.g. artificial neural networks for pressure/flow signal forecasting, wavelets
for signal de-noising and fuzzy sets and Bayesian networks for improved inference
analysis. Note that the successful development of the above real-time models will enable
merging the leak detection and assessment techniques, pressure-driven hydraulic solvers
and active leakage control methods.
Finally, an integral part of the above should be the development of novel
sampling design methods for locating pressure and flow sensors in pipe networks so that
better detection and diagnosis results can be obtained for both background and especially
burst related leaks.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
29

Acknowledgements
The first author would like to acknowledge the financial support from the Estonian Science Foundation
(ETF7646).

References
Ainola, L., Koppel, T., Tiiter, K. and Vassiljev, A., 2000. Water network model
calibration based on grouping pipes with similar leakage and roughness estimates. In:
Proceedings of Joint Conference on Water Resources Engineering and Water
Resources Planning & Management. Minneapolis, USA.
(AHL) Authority of the House of Lords., 2006. Water Management. In: 8
th
Report of
Session 2005-06. London : Science and Technology Committee.
Aksela, K., Aksela, M. and Vahala, R., 2009. Leakage detection in a real distribution
network using a SOM. Urban Water Journal, 6(4), 279-289.
Alonso, J. M., Alvarruiz F., Guerrero, D., Hernandez, V., Ruiz, P. A., Vidal, A. M.,
Martinez, F., Vercher, J. and Ulanicki, B., 2000. Parallel computing in water network
analysis and leakage minimization. Journal of Water Resources Planning and
Management, 126(4), 251-260.
Araujo, L. S., Coelho, S. T. and Ramos, H. M., 2003a. Estimation of distributed pressure-
dependent leakage and consumer demand in water supply networks. In: Proceedings of
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
30
the International Conference on Advances in Water Supply Management. London :
Balkema.
Beck, S. B. M., Curren, M. D., Sims, N. D. and Stanway, R., 2005a. Pipeline network
features and leak detection by cross-correlation analysis of reflected waves. Journal of
Hydraulic Engineering, 131(8), 715-723.
Beck, S. B. M., Bagshaw, N. M., and Yates, J. R., 2005b. Explicit equations for leak rates
through narrow cracks. International Journal of Pressure Vessels and Piping, 82, 565-
570.
Behzadian, K., Kapelan, Z., Savic, D. A. and Ardeshir, A., 2009. Stochastic sampling
design using Multiobjective Genetic Algorithm and Adaptive Neural Networks.
Environmental Modelling and Software, 24(4), 530-541.
Beuken, R. H. S., Lavooij, C. S. W., Bosch, A. and Schaap, P. G., 2006. Low leakage in
the Netherlands confirmed. In: Proceedings of the 8th Annual Water Distribution
Systems Analysis Symposium. Cincinnati, USA.
Bougadis, J., Adamowski, K. and Diduch, R., 2005. Short-term municipal water demand
forecasting. Hydrol. Process. 19, 137-148.
Bracken, M. and Hunaidi, O., 2005. Practical aspects of acoustical leak location on
plastic and large diameter pipe. In: Leakage 2005 Conference Proceedings. Halifax
,Canada.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
31
Brunone, B. and Ferrante, M., 1999. On leak detection in single pipes using unsteady-
state tests. In: Proceedings of the IASTED International Conference: Modeling and
Simulation (MS ’99). Philadelphia, Pennsylvania, USA.
Brunone, B., 1999. Transient test-based technique for leak detection in outfall pipes.
ASCE Journal of Water Resources Planning and Management, 125(5), 302-306.
Brunone, B. and Ferrante, M., 2001. Detecting leaks in pressurized pipes by means of
transients. Journal of Hydraulic Research, 39(5), 539-547.
Brunone, B. and Ferrante, M., 2004. Pressure wave as a tool for leak detection in closed
conduits. Urban Water Journal, 1(2), 145-155.
Buchberger, S. G. and Nadimpalli, G., 2004. Leak estimation in water distribution
systems by statistical analysis of flow readings. Journal of Water Resources Planning
and Management, 130(4), 321-329.
Carpentier, P. and Cohen, G., 1991. State estimation and leak detection in water
distribution networks. Civil Engineering Syst., 8, 247-257.
Cascetta, F. and Vigo, P., 1992. Location and assessment of water leakage. Measurement
+ Control, 25, 269-275.
Cassa, A. M., Van Zyl, J. E. and Laubscher, R. F., 2005. A numerical investigation into
the behaviour of leak openings in uPVC pipes under pressure. In: Proceedings of
CCWI2005, Water Management for the 21st Century. Exeter, UK
Chastian-Howley, A., 2005. Transmission main leakage: how to reduce the risk of a
catastrophic failure. . In: Leakage 2005 Conference Proceedings. Halifax ,Canada.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
32
Chen, L.-C., 1995. Pipe network transient analysis – the forward and inverse problems.
Thesis (PhD). Cornell University.
Cist, D. B. and Schutz, A. E., 2001. State of the art for pipe & leak detection. Report by
Geophysical Survey Systems, Inc.
Coetzer, A. J., Van Zyl, J. E. and Clayton, C. R. I., 2006. An experimental investigation
into the turbulent-flow hydraulics of small circular holes in plastic pipes. In:
Proceedings of the 8th Annual Water Distribution Systems Analysis Symposium.
Cincinnati, USA.
Colombo, A. F. and Karney B. W., 2002. Energy and costs of leaky pipes: Toward
comprehensive picture. Journal of Water Resources Planning and Management,
128(6), 441-450.
Covas, D. and Ramos, H., 1999. Leakage detection in single pipelines using pressure
wave behaviour. In: Proceedings of Water Industry Systems: modelling and
optimization applications. Baldock : Research Studies Press Ltd.
Covas, D., Stoianov, I., Butler, D., Maksimovic, Ć., Graham, N. and Ramos, H., 2001a.
Leakage detection in pipeline systems by inverse transient analysis – from theory to
practice. In: Proceedings of Water Software Systems: theory and applications.
Leicester, UK.
Covas, D. and Ramos, H., 2001b. Hydraulic transients used for leakage detection in water
distribution systems. In: Proceedings of the 4th Conference on Water Pipeline Systems:
Managing Pipeline Assets in an Evolving Market. York, UK
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
33
Covas, D., Graham, N., Maksimović, Ć., Ramos, H., Kapelan, Z., Savic, D. and Walters,
G., 2003. An assessment of the application of inverse transient analysis for leak
detection: Part II – Collection and application of experimental data. In: Proceedings of
the International Conference on Advances in Water Supply Management. London :
Balkema.
Covas, D., Ramos, H. and de Almeida, A. B., 2005a. Standing wave difference method
for leak detection in pipeline systems. Journal of Hydraulic Engineering, 131(12),
1106-1116.
Covas, D., Ramos, H., Young, A., Graham, N. and Maksimovic, C., 2005b. Uncertainties
of leak detection by means of hydraulic transients: from the lab to the field. In:
Proceedings of CCWI2005, Water Management for the 21st Century. Exeter, UK.
Covas, D., Ramos, H., Lopes, N. and Almeida, A. B., 2006a. Water pipe system
diagnosis by transient pressure signals. In: Proceedings of the 8th Annual Water
Distribution Systems Analysis Symposium. Cincinnati, USA.
Covas, D., Jacob, A. and Ramos, H., 2006b. Bottom-up analysis for assessing water
losses: A case study. In: Proceedings of the 8th Annual Water Distribution Systems
Analysis Symposium. Cincinnati, USA.
Deagle, G., Green, A., Scrivener, J. and Tansley, N., 2007. Advanced tools for burst
location. Water Management Challenges in Global Change. Ulanicki et al. (eds),
Taylor and Francis, 307-312.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
34
Dias, N., Covas, D. and Ramos, H., 2005. Leakage reduction and customer level of
service improvement by pressure management. In: Proceedings of CCWI2005, Water
Management for the 21st Century. Exeter, UK.
Dunlop, E. J., 1999. A dynamic hydraulic model for water distribution network
simulation. In: Proceedings of Water Industry Systems: modelling and optimization
applications. Baldock : Research Studies Press Ltd.
EPRI, 1989. Condenser leak detection by using SF
6
as a tracer gas. Electric Power
Research Institute, Technical Report RP1689.
EPSRC Engineering Programme Network in Trenchless Technology, 2002. Underground
mapping, pipeline location technology and condition assessment. Birmingham :
Infrastructure Engineering and Management Research Centre.
Farley, M. and Trow, S., 2003. Losses in water distribution networks. London : IWA
Publishing.
Farley, M., 2008. Finding ‘difficult’ leaks. International Water Association Specialist
Group – Efficient Operation and Management. Available from: http://www.iwaom.org
[Accessed 21 January 2009].
Ferrante, M., Brunone, B. and Rossetti, A. G., 2001a. Harmonic analysis of pressure
signal during transients for leak detection in pressurized pipes. In: Proceedings ofthe
4th International Conference on Water Pipeline Systems. York, UK : BHR Group
Limited.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
35
Ferrante, M., Brunone, B. and Rossetti, A. G., 2001b. Leak detection in pressurized pipes
by means of wavelet analysis. In: Proceedings of the 4th International Conference on
Water Pipeline Systems. York, UK : BHR Group Limited.
Ferrante, M. and Brunone, B., 2003a. Pipe system diagnosis and leak detection by
unsteady-state tests. 1. Harmonic analysis. Advances in Water Resources, 26, 95–105.
Ferrante, M. and Brunone, B., 2003b. Pipe system diagnosis and leak detection by
unsteady-state tests. 2. Wavelet analysis. Advances in Water Resources, 26, 107–116.
Ferrante, M., Brunone, B., Meniconi, S. and Almadori, C., 2005. Wavelet analysis of
numerical pressure signals for leak monitoring. In: Proceedings of CCWI2005, Water
Management for the 21st Century. Exeter, UK.
Field, D. B. and Ratcliffe, B., 1978. Location of leaks in pressurised pipelines using
sulphur hexafluoride as a tracer. Water Research Centre, Technical Report 80.
Fletcher, R., 2008. SmartBall™ – a new approach in pipeline leak detection. In:
Proceedings of the International Pipeline Conference. Calgary, Alberta, Canada.
Gao, Y., Brennan, M. J., Joseph, P. F., Muggleton, J. M., and Hunaidi, O., 2004. A model
of the correlation function of leak noise in buried plastic pipes. Journal of Sound and
Vibration, 277, 133-148.
Gao, Y., Brennan, M. J., Joseph, P. F., Muggleton, J. M., and Hunaidi, O., 2005. On the
selection of acoustic/vibration sensors for leak detection in plastic water pipes. Journal
of Sound and Vibration, 283, 927-941.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
36
Gao, Y., Brennan, M. J. and Joseph, P. F., 2006. A comparison of time delay estimators
for the detection of leak noise signals in plastic water distribution pipes. Journal of
Sound and Vibration, 292, 552–570.
Garcia, V. J., Cabrera, E. and Cabrera (Jr.), E., 2006. The minimum night flow method
revisited. In: Proceedings of the 8th Annual Water Distribution Systems Analysis
Symposium. Cincinnati, USA.
Garzon-Contreras, F. and Thornton, J., 2006. Pressure-leakage relationships in urban
water distribution systems. In: Proceedings of the 8th Annual Water Distribution
Systems Analysis Symposium. Cincinnati, USA.
Germanopoulos, G., 1985. A technical note on the inclusion of pressure dependent
demand and leakage terms in water supply network models. Civ. Engng Syst., 2, 171-
179.
Germanopoulos, G., 1995. Valve control regulation for reducing leakage. Improving
efficiency and reliability in water distribution systems. Dordrecht : Kluwer Academic
Publishers.
Giustolisi, O., Savic, D.A. and Kapelan, Z., 2008. Pressure-driven demand and leakage
simulation for water distribution networks. ASCE Journal of Hydraulic Engineering,
134(5), 626-635.
Goodwin, S. J., 1980. The results of the experimental programme on leakage and leakage
control. Water Research Centre, Technical Report 154.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
37
Greyvenstein, B. and Van Zyl, J. E., 2005. An experimental investigation into the
pressure leakage relationship of some failed water pipes. In: Leakage 2005 Conference
Proceedings. Halifax ,Canada.
Grunwell, D. and Ratcliffe, B., 1981. Location of underground leaks using the leak noise
correlator. Water Research Centre, Technical Report 157.
Hernandez, V., Martinez, F., Vidal, A. M., Alonso, J. M., Alvarruiz, F., Guerrero, D.,
Ruiz, P. A. and Vercher, J., 1999. HIPERWATER: A high performance computing
EPANET-based demonstrator for water network simulation and leakage minimisation.
In: Proceedings of Water Industry Systems: modelling and optimization applications.
Baldock : Research Studies Press Ltd.
Hesssel, G., Schmitt, W., Van Der Vorst, K. and Weiss, F.-P., 1999. A neural network
approach for acoustic leak monitoring in the VVER-440 pressure vessel head. Progress
in Nuclear Energy, 34(3), 173-189.
Hope, W., 1892. The waste of water in public supplies, and its prevention. Proc. Instn.
Civ. Engrs. Mun. Engr., (republished 1996), 115, 68-75.
Hough, J. E., 1988. Leak testing of pipelines uses pressures and acoustic velocity. Oil &
Gas Journal, 21, 35-41.
Hunaidi, O., 1998. Ground-penetrating radar for detection of leaks in buried plastic water
distribution pipes. In: Proceedings of 7
th
International Conference on Ground-
Penetrating Radar. Lawrence, Kansas, USA.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
38
Hunaidi, O., and Chu, W. T., 1999. Acoustical characteristics of leak signals in plastic
water distribution pipes. Applied Acoustics, 58, 235-254.
Hunaidi, O., Chu, W., Wang, A. and Guan, W., 2000. Detecting leaks in plastic pipes.
Journal AWWA, 92(2), 82-94.
Hunaidi, O., Bracken, M., Gambino, T. and Fricke, C., 2004. Acoustic methods for
locating leaks in municipal water pipe networks. In: International Conference on Water
Demand Management. Dead Sea, Jordan.
Jönsson, L., 1995. Leak detection in pipelines using hydraulic transients – laboratory
measurements. Department of Water Resources Engineering, University of Lund,
Sweden.
Jönsson, L., 2003. Hydraulic transients in a pipeline with a leak. VA-FORSK Rapport 28.
Kapelan, Z., Savic, D. A. and Walters, G. A., 2000. Inverse transient analysis in pipe
networks for leakage detection and roughness calibration. In: Proceedings of
International Symposium: Water Network Modelling for Optimal Design and
Management. Exeter, UK.
Kapelan, Z., 2002. Calibration of WDS hydraulic models. Thesis (PhD). University of
Exeter.
Kapelan, Z. S., Savic, D. A. and Walters, G. A., 2003a. A hybrid inverse transient model
for leakage detection and roughness calibration in pipe networks. Journal of Hydraulic
Research, 41(5), 481-492.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
39
Kapelan, Z., Savic, D., Walters, G. A., Covas, D., Graham, I. N. and Maksimović, Ć.,
2003b. An assessment of the application of inverse transient analysis for leak detection:
Part I – Theoretical considerations. In: Proceedings of the International Conference on
Advances in Water Supply Management. London : Balkema.
Kapelan, Z., Savic, D. A. and Walters G. A., 2003c. Multi-objective sampling design for
water distribution model calibration. ASCE Journal of Water Resources Planning and
Management, 129(6), 466-479.
Kapelan, Z., Savic, D. A. and Walters, G. A., 2004. Incorporation of prior information on
parameters in inverse transient analysis for leak detection and roughness calibration.
Urban Water Journal, 1(2), 129- 143.
Kapelan, Z., Savic, D. and Walters, G., 2005. Optimal sampling design methodologies for
water distribution model calibration. Journal of Hydraulic Engineering, 131(3), 190-
200.
Kim, S. H., 2005. Extensive development of leak detection algorithm by impulse
response method. Journal of Hydraulic Engineering, 131(3), 201-208.
Lambert A. and Hirner, W., 2000. Losses from water supply systems: Standard
terminology and recommended performance measures. Available from:
www.iwahq.org [Accessed 21 August 2009].
Lambert, A., 2001. What do we know about pressure:leakage relationships in distribution
systems? In: Proceedings of IWA Conference ’System Approach to leakage control and
water distribution system management. Brno, Czech Rebublic.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
40
Lambert, A., 2002. Water losses management and techniques. Water Science and
Technology: Water Supply, 2 (4), 1-20.
Lambert, A., 2003. Assessing non-revenue water and its components: a practical
approach. Water 21, August, 50-51.
Lee, P. J., Vitkovský, J. P., Lambert, M. F., Simpson, A. R. and Liggett, J. A., 2003.
Frequency response coding for the location of leaks in single pipeline systems. In:
Proceedings of PEDS 2003 (Pumps, Electromechanical Devices and Systems Applied
to Urban Water Management). Valencia, Spain.
Lee, P. J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R. and Liggett J. A., 2005a.
Frequency domain analysis for detecting pipeline leaks. Journal of Hydraulic
Engineering, 131(7), 596-604.
Lee, P.J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R. and Liggett J., 2005b. Leak
location using the pattern of the frequency response diagram in pipelines: a numerical
study. Journal of Sound and Vibration, 284, 1051-1073.
Lee, P. J., Lambert, M. F., Simpson, A. R., Vitkovský, J. P. and Liggett, J., 2006.
Experimental verification of the frequency response method for pipeline leak detection.
Journal of Hydraulic Research, 44(5), 693-707.
Liemberger, R. and Farley, M., 2004. Developing a non-revenue water reduction strategy
Part 1: Investigating and assessing water losses. In: Proceeding of IWA WWC 2004
Conference. Marrakech, Morocco.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
41
Liggett, A., 1993. Network monitoring and the algorithmic location of leaks under steady
and unsteady conditions. In: Water supply systems: State of the art and future trends.
Southampton : Computational Mechanics.
Liggett, J. A., and Chen, L.-C., 1994. Inverse transient analysis in pipe networks. ASCE
Journal of Hydraulic Engineering, 120(8), 934-955.
Liou, J. C. P., 1994. Mass imbalance error of waterhammer equations and leak detection.
Journal of Fluids Engineering. 116, 103-109.
Liou, C. P., 1998. Pipeline leak detection by impulse response extraction. Journal of
Fluids Engineering, 120, 833-838.
Liou, J. C. P., Hall, R. J. and McMahon, M. C., 2003. Hazardous liquid leak detection
techniques and processes. General Physics Corporation, Report No. DTRS56-02-D-
70037-01.
Lockwood, A., Murray, T., Stuart, G. and Scudder, L., 2003. A study of geophysical
methods for water leak location. In: Proceedings of PEDS 2003 (Pumps,
Electromechanical Devices and Systems Applied to Urban Water Management).
Valencia, Spain.
Makar, J. M. and Chagnon, N., 1999. Inspecting systems for leaks, pits, and corrosion.
Journal of American Water Works Association, 91(7), 36-46.
Mamlook, R. and Al-Jayyous, O., 2003. Fuzzy sets analysis for leak detection in
infrastructure systems: a proposed methodology. Clean Techn Environ Policy, 6, 26-31.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
42
Martinez, F., Signes, M., Savall, R., Andrés, M. and Ponz, R., 1999. Construction and use
of a dynamic simulation model for the Valencia metropolitan water supply and
distribution network. In: Proceedings of Water Industry Systems: modelling and
optimization applications. Baldock : Research Studies Press Ltd.
May, J., 1994. Pressure dependent leakage. World Water and Environmental
Engineering, October, 10.
McInnis, D. and Karney, B. W., 1995. Transients in distribution networks: field tests and
demand models. Journal of Hydraulic Engineering, 121 (3), 218-231.
McNulty, J. G., 2001. An acoustic-based system for detecting, locating and sizing leaks
in water pipelines. In: Proceedings of the 4th International Conference on Water
Pipeline Systems: Managing Pipeline Assets in an Evolving Market. York, UK.
Mergelas, B. and Henrich, G., 2005. Leak locating method for precommissioned
transmission pipelines: North American case studies. In: Leakage 2005 Conference
Proceedings. Halifax, Canada.
Miller, R.K., Pollock, A.A., Watts, D.J., Carlyle, J.M., Tafuri, A.N. and Yezzi Jr, J.J.,
1999. A reference standard for the development of acoustic emission pipeline leak
detection techniques. NDT&E International, 32, 1–8.
Misiunas, D., Vítkovský, J., Olsson, G., Simpson, A. R. and Lambert, M. F., 2003.
Pipeline burst detection and location using a continuous monitoring technique. In:
Proceedings of the International Conference on Advances in Water Supply
Management. London : Balkema.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
43
Misiunas, D., Vítkovský, J., Olsson, G., Simpson, A. and Lambert, M., 2005a. Pipeline
break detection using pressure transient monitoring. Journal of Water Resources
Planning and Management, 131(4), 316-325.
Misiunas, D., Vítkovský, J., Olsson, G., Lambert, M. and Simpson, A., 2005b. Failure
monitoring in water distribution networks. In: Proceedings of the 2nd IWA Conference
on Instrumentation, Control and Automation. Busan, Korea.
Misiunas, D., Lambert, M. and Simpson, A., 2006. Transient-based periodical pipeline
leak diagnosis. In: Proceedings of the 8th Annual Water Distribution Systems Analysis
Symposium. Cincinnati, USA.
Morris Jr., R. E., 1967. Principal causes and remedies of water main breaks. In:
Southwest Section Meeting, Austin, Texas, Oct. 19, 782-798.
Mounce, S. R. and Machell, J., 2006. Burst detection using hydraulic data from water
distribution systems with artificial neural networks. Urban Water Journal, 3(1), 21-31.
Mounce, S. R., Boxall, J. B. and Machell, J., 2007. An artifical neural network/fuzzy
logic system for DMA flow meter data analysis providing burst identification and size
estimation. Water Management Challenges in Global Change. Ulanicki et al. (eds),
Taylor and Francis, 313-320.
Mounce, S. R., Boxall, J. B. and Machell, J., 2008. Online application of ANN and fuzzy
logic system for burst detection. In: Proceedings of the 8th Annual Water Distribution
Systems Analysis Symposium. Kruger National Park, South Africa.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
44
Mounce, S. R., Boxall, J. and Machell, J., 2009. Development and verification of an
online artificial intelligence system for detection of bursts and other abnormal flows.
Journal of Water Resources Planning and Management, In Press.
Moyer, E., Male, J. W., Moore, C. and Hock, G., 1983. The economics of leak detection
and repair – a case study. American Water Works Association, January, 29-35.
Mpesha, W., Gassman, S. L. and Chaudhry, M. H., 2001. Leak detection in pipes by
frequency response method. Journal of Hydraulic Engineering, 127(2), 134-147.
Mpesha, W., Chaudhry, M. H. and Gassman, S. L., 2002. Leak detection in pipes by
frequency response method using a step excitation. Journal of Hydraulic Research,
40(1), 55-62.
Muggleton, J. M., Brennan, M. J. and Linford, P. W., 2004. Axisymmetric wave
propagation in fluid-filled pipes: wavenumber measurements in vacuo and buried pipes.
Journal of Sound and Vibration, 270, 171-190.
Muggleton, J. M. and Brennan, M. J., 2004. Leak noise propagation and attenuation in
submerged plastic water pipes. Journal of Sound and Vibration, 278, 527-537.
Muggleton, J. M. and Brennan, M. J., 2005. Axisymmetric wave propagation in buried,
fluid-filled pipes: effects of wall discontinuities. Journal of Sound and Vibration, 281,
849-867.
Muggleton, J. M., Brennan, M. J., Pinnington, R. J. and Gao, Y., 2006. A novel sensor
for measuring the acoustic pressure in buried plastic water pipes. Journal of Sound and
Vibration, 295, 1085-1098.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
45
Mukhopadhyay, S. and Srivastava, G. P., 2000. Characterisation of metal loss defects
from magnetic flux leakage signals with discrete wavelet transform. NDT&E
International, 33, 57-65.
Nicholas, R. E., 1990. Leak detection on pipelines in unsteady flow. In: Forum on
unsteady flow. In: 1990 winter annual meeting of ASME. New York, USA.
Noack, C. and Ulanicki, B., 2006. Modelling of soil diffusibility on leakage
characteristics of buried pipes. In: Proceedings of the 8th Annual Water Distribution
Systems Analysis Symposium. Cincinnati, USA.
O’Brien, E., Murray, T. and McDonald, A., 2003. Detecting leaks from water pipes at a
test facility using ground-penetrating radar. In: Proceedings of PEDS 2003 (Pumps,
Electromechanical Devices and Systems Applied to Urban Water Management).
Valencia, Spain.
O’Day, D. K., 1982. Organizing and analyzing leak and break data for making main
replacement decisions. AWWA Journal of the American Water Works Association,
November, 589-594.
Ofwat, 2008. International comparison – leakage. Available from:
http://www.ofwat.gov.uk/regulating/reporting/rpt_int_08leakageintro [Accessed 25
February 2009].
Pilcher, R., 2003. Leak detection practices and techniques: a practical approach. Water
21, December, 44-45.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
46
Pilcher, R., Hamilton, S., Chapman, H., Ristovski, B. and Strapely, S., 2007. Leak
location and repair guidance notes. In: International Water Association. Water Loss
Task Forces: Specialist Group Efficient Operation and Management. Bucharest,
Romania.
Poulakis Z., Valougeorgis, D. and Papadimitriou, C., 2003. Leakage detection in water
pipe networks using a Bayesian probabilistic framework. Probabilistic Engineering
Mechanics 18, 2003, pp. 315–327.
Pudar, R. S. and Liggett, J. A., 1992. Leaks in pipe networks. ASCE Journal of Hydraulic
Engineering, 118(7), 1031-1046.
Puust, R., Kapelan, Z., Savic, D. and Koppel, T., 2006. Probabilistic leak detection in
pipe networks using the SCEM-UA algorithm. In: Proceedings of the 8th Annual Water
Distribution Systems Analysis Symposium. Cincinnati, USA.
Puust, R., 2007. Probabilistic leak detection in pipe networks using the SCEM-UA.
Thesis (PhD). Tallinn University of Technology.
Rahman, S., Ghadiali, N., Wilkowski, M. and Paul, D., 1997. A computer model for
probabilistic leak-rate analysis of nuclear piping and piping welds. Int. J. Pres. Ves. &
Piping, 70, 209-221.
Rajtar, J. and Muthiah, R., 1997. Pipeline leak detection system for oil and gas flowlines.
Journal of Manufacturing Science and Engineering, Transactions of ASME, 19(1), 105-
109.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
47
Ramos, H., Reis, C., Ferreira, C., Falcao, C. and Covas D., 2001. Leakage control policy
within operating management tools. In: Proceedings of Water Software Systems: theory
and applications. Leicester, UK.
Reis, L. F. R., Porto, R. M. and Chaudhry, F. H., 1997. Optimal location of control valves
in water supply networks by genetic algorithm. Journal of Water Resources Planning
and Management, 123(6), 317-326.
Reis, L. F. R. and Chaudhry, F. H., 1999. Hydraulic characteristics of pressure reducing
valves for maximum reduction of leakage in water supply networks. In: Proceedings of
Water Industry Systems: modelling and optimization applications. Baldock : Research
Studies Press Ltd.
Rossman, L. A., 2000. Epanet2 User Manual. Washington, USA : US EPA.
Rougier, J., 2005. Probabilistic leak detection in pipelines using the mass imbalance
approach. Journal of Hydraulic Research, 43(5), 556-566.
Saldarriaga, J. G., Fuentes, D. A. A. and Galvis, L. F. C., 2006. Implementation of the
hydraulic transient and steady oscillatory flow with genetic algorithms for leakage
detection in real water distribution networks. In: Proceedings of the 8th Annual Water
Distribution Systems Analysis Symposium. Cincinnati, USA.
Savic, D. A., Kapelan, Z. and Jonkergouw, P., 2009. Quo vadis water distribution model
calibration? Urban Water Journal, 6(1), 3-22.
Scott, S. L. and Barrufet, M. A., 2003. Worldwide assessment of industry leak detection
capabilities for single & multiphase pipelines. Project report prepared for the minerals
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
48
management service under the MMS/OTRC cooperative research agreement, 1435-01-
99-CA-31003, Task Order 18133, August 6.
Shimanskiy, S., Iljima, T. and Naoi, Y., 2003. Development of microphone leak detection
technology on Fugen NPP. Progress in Nuclear Energy, 43(1-4), 357-364.
Simpson, A. R. and Vitkovský, J. P., 1997. A review of pipe calibration and leak
detection methodologies for water distribution networks. In: Water in the Balance:
Proceedings of the Australian Water & Wastewater Association 17th Federal
Convention. Melbourne, Australia.
Soares, A. K., Covas, D. I. C. and Reis, L. F. R., 2007. Inverse transient analysis for leak
detection in a PVC pipe network. In: Proceedings of CCWI 2007 and SUWM 2007
Conference: Water Management Challenges in Global Change. Leicester, UK.
Stathis, J. A. and Loganathan, G. V., 1999. Analysis of pressure-dependent leakage in
water distribution systems. Analysis of pressure-dependent leakage in water distribution
systems. In: Preparing for the 21
st
Century, 29
th
Annual Water Resources Planning and
Management Conference. Tempe, Arizona, USA.
Stenberg, R., 1982. Leak detection in water supply systems. Stockholm, Sweden : The
Swedish Water and Waste Water Works Association.
Stephens, M., Lambert, M. F., Simpson, A. R., Vitkovský, J. P. and Nixon, J., 2004. Field
tests for leakage, air pocket, and discrete blockage detection using inverse transient
analysis in water distribution pipes. In: Proceedings of the 6th Water Distribution
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
49
Systems Analysis Symposium, World Water and Environmental Resources Congress,
ASCE. Salt Lake City, USA.
Stoianov, I., Karney, B., Covas, D., Maksimović, Ć. and Graham, N., 2001. Wavelet
processing of transient signals for pipeline leak location and quantification. In:
Proceedings of Water Software Systems: theory and applications. Leicester, UK.
Tabesh, M. and Delavar, M. R., 2003. Application of integrated GIS and hydraulic
models for unaccounted for water studies in water distribution systems. In: Proceedings
of the International Conference on Advances in Water Supply Management. London :
Balkema.
Thornton, J., 2002. Water loss control manual. New York : McGraw-Hill.
Thornton, J. and Lambert, A., 2005. Progress in practical prediction of pressure: leakage,
pressure: burst frequency and pressure: consumption relationships. In: Leakage 2005
Conference Proceedings. Halifax, Canada.
Todini, E., 2003. A more realistic approach to the ‘extended period simulation’ of water
distribution networks. In: Proceedings of the International Conference on Advances in
Water Supply Management. London : Balkema.
Tripartite Group, 2002. Best practice principles in the economic level of leakage
calculation. Tripartite Group. Available from: http://www.ofwat.gov.uk [Accessed 22
January 2009].
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
50
Tucciarelli, T., Criminisi, A. and Termini, D., 1999. Leak analysis in pipeline systems by
means of optimal valve regulation. ASCE Journal of Hydraulic Engineering, 125(3),
277-285.
Tuhovčák, L., Svoboda, M., Sviták, Z. and Tothova, K., 2005. The technical audit of
water distribution network using the different leakage indicators. In: Leakage 2005
Conference Proceedings. Halifax, Canada.
TWDB, 2005. Water loss manual. Austin, Texas, USA : Texas Water Development
Board
TWDB, 2007. Water audit reporting form. Texas Water Development Board. Available
from: http://www.twdb.state.tx.us/assistance/conservation/Municipal/Water_Audit/
HB3338ReportingForm.pdf [Accessed 3 April 2007].
Ulanicka, K., Bounds, P., Ulanicki, B. and Rance, J., 2001. Pressure control of a large
scale water distribution network with interacting water sources – a case study. In:
Proceedings of Water Software Systems: theory and applications. Leicester, UK.
Ulanicki, B., Bounds, P. L. M., Rance, J. P. and Reynolds, L., 2000. Open and closed
loop pressure control for leakage reduction. Urban Water, 2, 105-114.
Vairavamoorthy K. and Lumbers, J., 1998. Leakage reduction in water distribution
systems: optimal valve control. Journal of Hydraulic Engineering, 124(1), 1146-1154.
Van Zyl, J. E. and Clayton, C. R. I., 2005. The effect of pressure on leakage in water
distribution systems. In: Proceedings of CCWI2005, Water Management for the 21st
Century. Exeter, UK.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
51
Vela, A. F., Espert, V. and Fuertes, V. S., 1995. General overview of unaccounted for
water in water distribution systems. Improving efficiency and reliability in water
distribution systems. Dordrecht : Kluwer Academic Publishers.
Verde, C., 2005. Accommodation of multi-leak location in a pipeline. Control
Engineering Practice, 13, 1071-1078.
Vitkovský, J. P. and Simpson, A. R., 1997. Calibration and leak detection in pipe
networks using inverse transient analysis and genetic algorithms. Research Report No.
R 157 : Department of Civil and Environmental Engineering, The University of
Adelaide.
Vítkovský, J. P., Simpson, A. R. and Lambert, M., 2000. Leak detection and calibration
using transients and genetic algorithms. Journal of Water Resources Planning and
Management, 126(4), 262-265.
Vitkovský, J. P., Lambert, M. F., Simpson, A. R. and Wang, X. J., 2001. An experimental
verification of the inverse transient technique. In: Conference on Civil Engineering
Hydraulics. Hobart, Australia : Institution of Engineers.
Vitkovský, J. P., Liggett, J. A., Simpson A. R. and Lambert, M. F., 2003a. Optimal
measurement site locations for inverse transient analysis in pipe networks. Journal of
Water Resources Planning and Management, 129(6), 480-492.
Vitkovský, J. P., Lee, P. J., Stephens, M. L., Lambert, M. F. and Simpson, A. R., 2003b.
Leak and blockage detection in pipelines via an impulse response method. In:
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
52
Proceedings of PEDS 2003 (Pumps, Electromechanical Devices and Systems Applied
to Urban Water Management). Valencia, Spain.
Vitkovský, J. P., Liggett, J. A., Simpson, A. R. and Lambert, M. F., 2003c. Optimal
measurement site locations for inverse transient analysis in pipe networks. Journal of
Water Resources Planning and Management, 129(6), 480-492.
Vitkovský, J. P., Lambert, M. F., Simpson, A. R. and Liggett, J. A.., 2007. Experimental
observation and analysis of inverse transients for pipeline leak detection. Journal of
Water Resources Planning and Management, 133(6), 519-530.
Walski, T., Bezts, W. and Posluszny, E. T., 2006. Modeling leakage reduction through
pressure control. Journal AWWA, April, 147-155.
Wang, X.-J., Lambert, M. F., Simpson, A. R., Liggett, J. A. and Vitkovský, J. P., 2002.
Leak detection in pipelines using the damping of fluid transients. Journal of Hydraulic
Engineering, 128(7), 697-711.
Wang, X.-J., Simpson, A. R., Lambert, M. F., Vitkovský, J. P. and Liggett, J. A., 2003.
Simulation of transients in a looped laboratory pipe network. In: Proceedings of PEDS
2003 (Pumps, Electromechanical Devices and Systems Applied to Urban Water
Management). Valencia, Spain.
Wang, X.-J., Simpson, A. R. and Lambert, M. F., 2006. An analytical solution for the
transients in a pipeline with a variable boundary condition: leak detection in pipe
networks using coded transients. In: Proceedings of the 8th Annual Water Distribution
Systems Analysis Symposium. Cincinnati, USA.
*Corresponding author. Tel.: ++37-2620-2552; fax: +37-2620-2550.
E-mail address: raido@staff.ttu.ee (R. Puust).
53
WATERPIPE, 2009. Global change and ecosystemsec research project WATERPIPE.
Available from: http://www.waterpipe-eu.org [Accessed 21 August 2009].
Wiggert, D. C., 1968. Unsteady flows in lines with distributed leakage. ASCE Journal of
the Hydraulics Division, 94, 143-161.
Willems, H. and Barbian, O. A., 1998. Operational experience with inline ultrasonic
crack inspection of German crude oil pipelines. In: Proceedings of the 7
th
European
Conference on NDT. Copenhagen, Denmark.
Wu, Z. Y. and Sage, P., 2007. Pressure dependent demand optimization for leakage
detection in water distribution systems. Water Management Challenges in Global
Change. Ulanicki et al. (eds), Taylor and Francis, 353-361.
Zecchin, A.C., White, L. B., Lambert, M. F. and Simpson, A. R., 2005. Frequency-
domain hypothesis testing approach to leak detection in a single fluid line. In:
Proceedings of CCWI2005, Water Management for the 21st Century. Exeter, UK.
Zecchin, A. C., Lambert, M. F., Simpson, A. R. and White, L. B., 2006. Laplace-domain
comparison of linear models of a reservoir-pipe-valve system with a leak. In:
Proceedings of the 8th Annual Water Distribution Systems Analysis Symposium.
Cincinnati, USA.


List of Figures

Figure 1. 24h leakage modelling based on minimum night flow measurement. [Adapted
from Liemberger and Farley (2004, Figure 2)].

Figure 2. Pressure time-history at the leak during transient due to the closure of the end
valve. Even though the transient attenuates quickly, the risk of compromising water
quality exists (where pressure falls below tank level, h
0
and the ambient pressure external
to the leak, p
ext
is higher than the internal pipe pressure at the leak, p
e
). [Adapted from
Brunone and Ferrante (2004, Figure 1)].

Figure 3. Transient analysis for leak detection. Head variations measured at the same
nodes for no-leak case studies (a, b) and for leak case studies (c, d). Considerable head
damping can be seen when leaks exist in the system. [Adapted from Vitkovský et al. (2001,
Figures 3, 4)].

Figure 4. Impulse response functions for the non-leaking (a) and leaking (b) cases. The
first leak-induced reflection for the leaking case determines the correct location and size
of the leak. The secondary reflection is negligible compared to the main leak reflection.
[Adapted from Vitkovský et al. (2003b, Figure 4)].



Figure 5. Fourier series analysis of transients measured from pipeline without leak (a) and
with leak (b). By analysing the damping of harmonic components the leak can be
identified. Through the ratios of leak damping rates the leak location can be calculated.
[Adapted from Wang et al. (2002, Figure 9)].

Figure 6. Continuous wavelet transform (CWT) and discrete wavelet transform (DWT)
for no-leak (a, b) and for leak (c, d) case studies. It is possible to show the presence of the
leak of diameter equal to 1.49 mm in both cases. It is observed that chains of maxima
appear in figures (c) and (d) in correspondence with the instants t = 1.23 s and t = 3.23 s,
which are not present in figures (a) and (b). [Adapted from Ferrante et al. (2005, Figures 2 -
5)].

Figure 7. Impact of changing leak size and position on the frequency response diagram
extracted at the inline valve at downstream boundary: (a) leak at 700 m, C
d
b = 0.00014
m
2
; (b) C
d
b = 0.00028 m
2
; (c) leak at 1400 m, C
d
b = 0.00014 m
2
; (d) no leak. [Adapted
from Lee et al. (2005a, Figure 6)].

Figure 8. Results of acoustic noise loggers in two consecutive days (after repairs): "line
with crosses" – noise amplitude and "line with dots" – noise dispersion. Leakage situation
corresponds to "line with crosses" above "line with dots ". [Adapted from Covas et al.
(2006b, Figures 8)].



Figure 9. Left image: Continuous-wave radar principle shown on the left image. Right
image: GPR image showing water accumulation from a leak. [Adapted from Farley (2008,
Figure 1)]. Single frequency f
0
transmitted by radar is received back from moving targets
slightly different frequency f
0
+ ∆f. By rejecting f
0
, only moving objects (such as leaking
water) are detected.

Figure 10. Several leak detection methods by historical appearance. [Adapted from
Pilcher et al. (2007, Figure 17)].

Figure 11. Recording of a typical leak response when tracer gas technique is used
[Adapted from EPRI (1989, Figure 2)]. Response time depends on gas that is used and a
magnitude of the response depends on a gas volume that was injected.

Figure 12. Typical economic level of leakage (ELL) analysis. Economically feasible
leakage level compared with net present value (NPV) costs. There are also some level of
leakages (background losses) that are not possible to eliminate at all. [Adapted from
Tripartite Group (2002, Figure 4.1)].

Figure 13. On-line leak analysis. Comparison of pressure traces measured with and
without leakage (a). The change in difference between the two traces (a) indicates the
presence of a leak. The actual difference between measured pressures can be analysed to
get better resolution as shown in (b). [Adapted from Misiunas et al. (2006, Figures 9, 10)].



Figure 14. Rate of leakage (with and without optimised valve control i.e. pressure
regulation). Lines 'OBJ1' and 'OBJ2' indicates different optimisation model types.
[Adapted from Vairavamoorthy and Lumbers (1998, Figure 6)].


List of Tables

Table 1. IWA standard for international water balance and terminology.

Table 2. Recommended indicators for real losses and non-revenue water. [Adapted from
Liemberger, R. and Farley, M. (2004, Figure 4)].

Table 3. Summary of exponents b derived from field tests. [Adapted from Garzon-
Contreras and Thornton (2006, Table 1)].

Table 4. Efficiency ranges of various leak detection methodologies using pressure
transients.

Table 5. Efficiency ranges of various leak localisation and pinpointing techniques.

Table 6. A snapshot of various leakage control techniques applied to real data.




Figure 1. 24h leakage modelling based on minimum night flow measurement. [Adapted
from Liemberger, R. and Farley, M. (2004, Figure 2)].



Figure 2. Pressure time-history at the leak during transient due to the closure of the end
valve. Even though the transient attenuates quickly, the risk of compromising water
quality exists (where pressure falls below tank level, h
0
and the ambient pressure external
to the leak, p
ext
is higher than the internal pipe pressure at the leak, p
e
). [Adapted from
Brunone and Ferrante (2004, Figure 1)].



Figure 3. Transient analysis for leak detection. Head variations measured at the same
nodes for no-leak case studies (a, b) and for leak case studies (c, d). Considerable head
damping can be seen when leaks exist in the system. [Adapted from Vitkovský et al. (2001,
Figures 3, 4)].



Figure 4. Impulse response functions for the non-leaking (a) and leaking (b) cases. The
first leak-induced reflection for the leaking case determines the correct location and size
of the leak. The secondary reflection is negligible compared to the main leak reflection.
[Adapted from Vitkovský et al. (2003b, Figure 4)].



Figure 5. Fourier series analysis of transients measured from pipeline without leak (a) and
with leak (b). By analysing the damping of harmonic components the leak can be
identified. Through the ratios of leak damping rates the leak location can be calculated.
[Adapted from Wang et al. (2002, Figure 9)].



Figure 6. Continuous wavelet transform (CWT) and discrete wavelet transform (DWT)
for no-leak (a, b) and for leak (c, d) case studies. It is possible to show the presence of the
leak of diameter equal to 1.49 mm in both cases. It is observed that chains of maxima
appear in figures (c) and (d) in correspondence with the instants t = 1.23 s and t = 3.23 s,
which are not present in figures (a) and (b). [Adapted from Ferrante et al. (2005, Figures 2 -
5)].



Figure 7. Impact of changing leak size and position on the frequency response diagram
extracted at the inline valve at downstream boundary: (a) leak at 700 m, C
d
b = 0.00014
m
2
; (b) C
d
b = 0.00028 m
2
; (c) leak at 1400 m, C
d
b = 0.00014 m
2
; (d) no leak. [Adapted
from Lee et al. (2005a, Figure 6)].



Figure 8. Results of acoustic noise loggers in two consecutive days (after repairs): "line
with crosses" – noise amplitude and "line with dots" – noise dispersion. Leakage situation
corresponds to "line with crosses" above "line with dots ". [Adapted from Covas et al.
(2006b, Figures 8)].



Figure 9. Left image: Continuous-wave radar principle shown on the left image. Right
image: GPR image showing water accumulation from a leak. [Adapted from Farley (2008,
Figure 1)]. Single frequency f
0
transmitted by radar is received back from moving targets
slightly different frequency f
0
+ ∆f. By rejecting f
0
, only moving objects (such as leaking
water) are detected.


Figure 10. Several leak detection methods by historical appearance. [Adapted from
Pilcher et al. (2007, Figure 17)].




Figure 11. Recording of a typical leak response when tracer gas technique is used
[Adapted from EPRI (1989, Figure 2)]. Response time depends on gas that is used and a
magnitude of the response depends on a gas volume that was injected.




Figure 12. Typical economic level of leakage (ELL) analysis. Economically feasible
leakage level compared with net present value (NPV) costs. There are also some level of
leakages (background losses) that are not possible to eliminate at all. [Adapted from
Tripartite Group (2002, Figure 4.1)].



Figure 13. On-line leak analysis. Comparison of pressure traces measured with and
without leakage (a). The change in difference between the two traces (a) indicates the
presence of a leak. The actual difference between measured pressures can be analysed to
get better resolution as shown in (b). [Adapted from Misiunas et al. (2006, Figures 9, 10)].



Figure 14. Rate of leakage (with and without optimised valve control i.e. pressure
regulation). Lines 'OBJ1' and 'OBJ2' indicates different optimisation model types.
[Adapted from Vairavamoorthy and Lumbers (1998, Figure 6)].






Table 1. IWA standard for international water balance and terminology.
Billed metered consumption (including water
exported)
Billed unmetered consumption
Unbilled metered consumption
Unbilled unmetered consumption
Unauthorized consumption
Customer metering inaccuracies
Leakage on transmission and/or distribution mains
Leakage and overflows at utility's storage tanks
Leakage on service connections up to point of
customer metering
System
input
volume
Authorized
consumption
Water losses
Revenue
water
Non-
revenue
water
Billed
Unbilled
Apparent losses
Real losses



Table 2. Recommended indicators for real losses and non-revenue water. [Adapted from
Liemberger, R. and Farley, M. (2004, Figure 4)].




Table 3. Summary of exponents b derived from field tests. [Adapted from Garzon-
Contreras and Thornton (2006, Table 1)].
Country
Number of
zones tested
Range of
exponents b
Average
exponent b
United Kingdom (1970's)
17 0.70 to 1.68 1.13
Japan (1979) 20 0.63 to 2.12 1.15
Brazil (1998) 13 0.52 to 2.79 1.15
United Kingdom (2003)
75 0.36 to 2.95 1.01
Cyprus (2005) 15 0.64 to 2.83 1.47
Brazil (2006) 17 0.73 to 2.42 1.4
Totals 157 0.36 to 2.95 1.14



Table 4. Efficiency ranges of various leak detection methodologies using pressure
transients.
LRM Laboratory 135 m 0.04 l/s 1.9 m 44 m Jönsson (2003)
ITA Real 5'936 m 3 l/s 50 m 13 m Covas et al. (2005b)
ITA Real network 1 l/s 4.85% * 7 m Saldarriaga et al. 2006
IRA Numerical 20'000 m 10 l/s 2000 m 2 m Liou (1998)
TDM Laboratory 37.2 m 0.01 l/s 0.38 m 2 m Wang et al. (2002)
FRM Numerical 2'000 m 4.73 l/s < 500 m 26 m Lee et al. (2005a)
Case study
Transient
methodology
Reference
Inspection
range (pipeline
length)
Transient
pressure wave
height
Detectable
leak size
Location
precision

* - Leak size error
Note: This table is for general guidance only based on data that is available from given references. It may
not reflect the best solution for that particular technology that is currently available.


Table 5. Efficiency ranges of various leak localisation and pinpointing techniques.
* - depends on inspection range



Table 6. A snapshot of various leakage control techniques applied to real data.
Technique*
Hydraulic
model
Network size
Burst detection
size
Leak size
error
Detection
time
Reference
SE Yes DMA 8.3 - 83 l/s n/a n/a Carpentier et al. (1991)
GIS Yes
DMA (1533
properties)
18.1 l/s 10% n/a Tabesh and Delavar (2003)
SA No 200 homes 0.063 l/s n/a n/a
Buchberger and Nadimpalli
(2004)
ANN No DMA 5 l/s 10% 2.5h
MNFA No DMA 5 l/s 20% n/a
Mounce et al. (2007)

(1)
SE – state esitmation; SA – statistical analysis; MNFA – minimum night flow analysis