Development of A Leakage Target Setting Approach For South Korea Based On Economic Level of Leakage

College of Engineering, Mathematics, and Physical Sciences
Development of a Leakage Target Setting

Approach for South Korea based on
Economic Level of Leakage
Submitted by
Euijeong Lim
to the University of Exeter as a thesis for the degree of

Master of Philosophy in Engineering
In November 2015
This thesis is available for Library use on the understanding that it is

copyright material and that no quotation from the thesis may be published
without proper acknowledgement.
I certify that all material in this thesis which is not my own work has been
identified and that no material has previously been submitted and approved
for the award of a degree by this or any other University.
Signature: …………………………………………
Abstract
ABSTRACT
Leakage has become a crucial issue that needs to be addressed effectively by the
water suppliers in terms of economic management of the water system. Leakage
management costs such as pipe replacement, pressure management, detection
and repair costs have been steadily increasing. These costs have a direct effect on
the financial performance of water suppliers. Hence, they have to continuously do
their best to reduce leakage. However, a large number of water systems are
operated by small local government operators who are not well funded and lack the
necessary expertise. Consequently, a large volume of water is being lost due to
leaving on-going leakage unrepaired. In order to resolve these problems, South
Korea has been promoting the Non-Revenue Water (NRW) reduction project of
local water supplies in which the authorized organization, specializing in water
management would operate facilities on behalf of struggling local government. K-
water, the public water company in South Korea, has been operating and
managing 22 NRW reduction projects instead of local government since 2004.
In this thesis, a target setting method based on the Economic Level of Leakage
(ELL) calculation is proposed. The methodology applied is developed specifically
for the South Korean context to select a minimum achievable level of NRW. In
addition, the thesis will examine the appropriateness of the current target within
existing financial constraints by using limited available data. This approach is
focused on the derivation of the NRW control cost curve by using the newly
developed cumulative method that minimizes data fluctuation and enhances the
cost curve reliability. This has been applied to a case study by using data collected
from the water supplier information system. The results obtained from the case
study show significant outcomes in respect of both identification of an economically
optimal target and prevention of unnecessary investment to meet this aim. This
advance in leakage management allows water suppliers to select a rational target
and manage their system economically and efficiently.
3
Acknowledgements
ACKNOWLEDGEMENTS
Two years have already passed since I started my study. If it had not been for the
help of many people at every stage, I could not have finished my study. I would like
to take this opportunity to express my appreciation to everyone who has helped
and supported me.
Firstly, I would like to express the deepest gratitude to my two supervisors,

Professor Dragan Savic and Professor Zoran Kapelan. They have given me an
opportunity to undertake my MPhil in a lovely university and have always
encouraged and advised me to accomplish my research throughout my studies.
I would like to thank all the water supply operation and maintenance department
members, especially, Director General Kyeongsik Shin, General Manager
Seungwha Jung, Manager Seonglok Do, Beomseok Kim, Kapsoon Park, Sungjae
Bang, Jihun Koh and Assistant Manager Minjae Ko for their invaluable
encouragement and support.
I would like to thank the water supply management department members,

especially, Director General Kwanglae Kim, General Manager Eunsoon Park,
Manager Byungeun Hwang. Despite having to share my workload, they reduced
my responsibilities and allowed me to focus on my studies.
In addition, I am very grateful to Manager Cheul Jang and Sinjae Lee who are
studying together in the UK. They were very supportive and gave me valuable
advice during my years of study here. I would also like to acknowledge Assistant
Manager Jongkwon Lee from the Danyang local water supply department for
providing the data used in my research.
Finally, I would like to express special thanks to my parents and parent-in-laws, my

wife Jihye, Ajin and Woojin. They have been encouraging, supporting,
understanding me throughout process. Their dedication has enabled me to
accomplish my study.
4
Table of Contents
TABLE OF CONTENTS
Abstract .................................................................................................................. 3
Acknowledgements............................................................................................... 4
Table of Contents .................................................................................................. 5
List of Figures........................................................................................................ 9
List of Tables ....................................................................................................... 10
List of Abbreviations........................................................................................... 11
List of Symbols.................................................................................................... 12
Chapter 1 Introduction ........................................................................................ 13
1.1 Background .................................................................................................. 13
1.2 Aim and Objectives of Research .................................................................. 14
1.3 Thesis structure............................................................................................ 15
Chapter 2 Literature Review ............................................................................... 17
2.1. Introduction ................................................................................................. 17
2.2. Water loss management ............................................................................. 18
2.2.1 Water loss assessment .......................................................................... 19
2.2.2 Water loss Assessment methods ........................................................... 20
2.2.2.1 Total integrated flow analysis (Top-down method) ........................ 20
2.2.2.2 Minimum night flow analysis (Bottom-up method)......................... 21
2.2.2.3 Component analysis ..................................................................... 22
2.2.3 Performance Indicators .......................................................................... 23
2.2.4 Leakage target setting ........................................................................... 27
2.2.5 Leakage detection methods ................................................................... 28
2.2.5.1 Leakage localization methods ....................................................... 28
2.2.5.2 Leakage pinpointing methods ....................................................... 29
5
Table of Contents
2.2.6 Water loss management methods ......................................................... 30
2.2.6.1 Active leakage control (ALC) ......................................................... 31
2.2.6.2 Pressure management .................................................................. 31
2.2.6.2.1 Fixed and Variable Area Discharge (FAVAD) concept ............ 32
2.2.6.2.2 Type of Pressure Management .............................................. 33
2.2.6.3 Speed and quality of repairs ......................................................... 34
2.2.6.4 Pipeline and assets management ................................................. 35
2.3. Economic Level of Leakage ........................................................................ 37
2.3.1 Definition of ELL..................................................................................... 37
2.3.2 Time frames of ELL................................................................................ 39
2.3.3 ELL target setting process and the components .................................... 39
2.3.3.1 ELL target setting process ............................................................ 39
2.3.4 Calculating Economic Level of Leakage ................................................ 40
2.3.4.1 Policy minimum ............................................................................. 40
2.3.4.2 Least cost planning approach (Option 1) ...................................... 41
2.3.4.3 Marginal cost approach (Option 2) ................................................ 42
2.3.5 Previous research on Economic Level of Leakage ................................ 46
2.4. Summary ..................................................................................................... 47
Chapter 3 Methodology For Assessing Economic Level of Non-Revenue

Water .................................................................................................................... 49
3.1. Introduction ................................................................................................. 49
3.2. Economic level of NRW target setting process ........................................... 50
3.3. Economic level of NRW Calculation ............................................................ 51
3.3.1 Deciding Target Area ............................................................................. 51
6
Table of Contents
3.3.2 Data collection ....................................................................................... 52
3.3.2.1 Policy minimum ............................................................................. 52
3.3.2.2 Current NRW ................................................................................ 53
3.3.2.3 Costs data ..................................................................................... 53
3.3.3 Policy minimum ...................................................................................... 53
3.3.4 Current NRW level ................................................................................. 54
3.3.5 Economic level of NRW calculation methods ......................................... 54
3.3.5.1 Method 1: Marginal cost approach ................................................ 55
3.3.5.2 Method 2: Cumulative cost-benefit analysis approach .................. 57
3.4. Economic Level of NRW target setting........................................................ 59
3.5. Sensitivity analysis ...................................................................................... 60
3.6. Summary ..................................................................................................... 60
Chapter 4 Case Study ......................................................................................... 61
4.1. Introduction ................................................................................................. 61
4.2. CASE 1: NRW Reducing Stage (Danyang-gun).......................................... 61
4.2.1 Description of study Area ....................................................................... 61
4.2.2 Data collection and analysis .................................................................. 65
4.2.2.1 Connection, property and population ............................................ 65
4.2.2.2 Distribution network and pressure................................................. 65
4.2.2.3 Water loss performance indicators ................................................ 66
4.2.3 Current NRW level ................................................................................. 66
4.2.4 Policy minimum ...................................................................................... 68
4.2.5 Economic level of NRW calculation ....................................................... 69
4.2.5.1 NRW cost curve ............................................................................ 69
7
Table of Contents
4.2.5.2 Reliability check ............................................................................ 73
4.2.5.3 Economic level of NRW calculation (Method 2) ............................ 73
4.3. Sensitivity analysis ...................................................................................... 80
4.4 Setting NRW target ...................................................................................... 81
4.5. Summary ..................................................................................................... 81
Chapter 5 Summary and Conclusion................................................................. 82
5.1 Thesis Summary .......................................................................................... 82
5.2 Summary of contributions ............................................................................ 83
5.3 Conclusions ................................................................................................. 84
5.4 Future work recommendations ..................................................................... 85
Appendix A .......................................................................................................... 87
Data collection form ........................................................................................... 87
Appendix B .......................................................................................................... 91
Changes in NRW ............................................................................................... 91
Bibliography ........................................................................................................ 92
8
List of Figures
LIST OF FIGURES
Figure 1 : ILIs of 34 Demand Zones in North West England (Pearson, 2002) ....... 26
Figure 2 : K-water O&M 17 cities (Environment, 2012) ......................................... 26
Figure 3 : The four pillars of a successful leakage management strategy (Force,
2003) .............................................................................................................. 30
Figure 4 : Ration of Pressures (Force and Thornton, 2003) .................................. 32
Figure 5 : Economic level of leakage calculation(Fanner et al., 2007) .................. 38
Figure 6 : ELL target setting process map (Ofwat, 2002) ...................................... 39
Figure 7 : Leakage control costs curve .................................................................. 43
Figure 8 : Economic level of NRW target setting approach ................................... 50
Figure 9 : Marginal cost curve ............................................................................... 55
Figure 10 : Cumulative cost-benefit curve A .......................................................... 58
Figure 11 : Cumulative cost-benefit curve B .......................................................... 59
Figure 12 : Location of Danyang-gun in South Korea............................................ 62
Figure 13 : Location of Danyang-gun and DMAs................................................... 63
Figure 14 : Marginal cost of NRW control curve .................................................... 70
Figure 15 : Cumulative cost curve A ...................................................................... 71
Figure 16 : Cumulative cost curve B ...................................................................... 72
Figure 17 : Cumulative cost-benefit curve A .......................................................... 78
Figure 18 : Cumulative cost-benefit curve B .......................................................... 79
Figure 19 : Economic level of NRW sensitivity test ............................................... 80
9
List of Tables
LIST OF TABLES
Table 1 : International Water Association(IWA) standard international water

balance and terminology(Trow and Farley, 2004) .......................................... 19
Table 2 : Recommended indicators for physical losses and NRW (Alegre and
Association, 2006) .......................................................................................... 24
Table 3 : Comparison of three PRV operation types (McKenzie and Wegelin 2010)
....................................................................................................................... 34
Table 4 : Water pipe replacement and rehabilitation techniques (Thornton et al.
(2008) ............................................................................................................. 36
Table 5 : Annual investment plan of Danyang Gun (K-water, 2008)...................... 44
Table 6 : Source of collected data for ELL calculation ........................................... 52
Table 7 : Year 2013 statuses of connection, property and population ................... 64
Table 8 : Annual status of connection, property and population ............................ 65
Table 9 : Annual status of distribution network ...................................................... 65
Table 10 : Annual status of water loss performance indicators ............................. 66
Table 11 : Annual water balance ........................................................................... 67
Table 12 : Unavoidable annual real losses (UARL) ............................................... 68
Table 13 : Marginal cost of NRW control ............................................................... 69
Table 14 : Cumulative cost curve .......................................................................... 71
Table 15 : Cost data .............................................................................................. 73
Table 16 : Cumulative benefit of NRW reduction................................................... 74
Table 17 : Components of marginal operating costs in 2013 ................................ 75
Table 18 : Demand related future investment........................................................ 76
Table 19 : Present value of demand related investment ....................................... 76
Table 20 : Present value of the growth in demand ................................................ 77
Table 21 : Present value of no growth in demand ................................................. 77
Table 22 : Results of sensitivity test ...................................................................... 80
10
List of Abbreviations
LIST OF ABBREVIATIONS
ALC Active Leakage Control

CARL Current Annual Volume of Real Losses
CI Cost of Intervention
CV Variable cost of Water
DMA District Metered Area
ELL Economic Level of Leakage
FAVAD Fixed and Variable Area Discharge
GIS Geographic Information System
GPR Ground Penetrating Radar
ILI Infrastructure Leakage Index
IWA International Water Association
LNC Leak Noise Correlator
NRW Non-Revenue Water
OFWAT Water Services Regulation Authority
PI Performance Indicator
PLC Passive Leakage Control
PMA Pig Mounted Acoustic
PRV Pressure Reducing Valve
RR Rate of Rise of unreported leakage
SCADA supervisory control and data acquisition
SELL Sustainable Economic Level of Leakage
TGT Tracer Gas Technique
UARL Unavoidable Annual Real Losses
WDSs Water Distribution Systems
WLSG Water Loss Specialist Group
11
List of Symbols
LIST OF SYMBOLS
Lm mains length (km)
Nc number of service connections
Lp total length of private pipe, property boundary to customer meter (km)
P average pressure (metres).
L0, L1 the leaks flow rate before and after change in pressure
P0, P1 pressures before and after change in pressure
L level of leakage(m3/connection/year)
C cost of leakage control (￡/connection/year)
La actual level of leakage for the area(m3/connection/yea)
Ca actual cost of leakage control for the area(￡/connection/year)
Lb base level of leakage for the area(m3/connection/year)
Lp passive level of leakage for the area(m3/connection/year)
R2 coefficient of determination
12
Chapter 1 Introduction
CHAPTER 1 INTRODUCTION
1.1 Background
Industrialization, environmental pollution, climate change, aging infrastructure and

increasing level of customer expectations have made huge changes in water
supply (Levin et al., 2002). The changes require various types of investments such
as reinforcement and expansion of facilities, the introduction of advanced water
treatment facilities, and strengthening risk management (i.e. climate change
resilience and strengthening preparations). This has become a serious burden on
the water suppliers economically.
According to a recent World Bank publication, the annual volume of Non-Revenue

Water (NRW) was estimated to be approximately 50 billion cubic meters globally
and the losses were equivalent to at least US $15 billion per year (Frauendorfer
and Liemberger, 2010). Similarly, a large amount of water in South Korea is
disappearing through leakage every year. The annual volume and lost revenue in
2013 were 656 million m3 and $753 million (Environment, 2014), respectively. In
spite of continuous investment and efforts to reduce NRW, its management cost
continues to increase rapidly. The costs have nearly doubled in the last 10 years
(Koo et al., 2011). However, owing to a lack of expertise and aging and
deteriorated infrastructure of local waterworks, a large volume of water is still being
lost due to leaving on-going leakage unrepaired.
In order to resolve these problems, South Korea has been promoting the NRW
reduction project of local water supplies in which an authorized organization,
specializing in water management would operate facilities on behalf of struggling
local governments. As a result, K-water, the public water company in South Korea,
has been operating and managing 22 NRW reduction projects, instead of local
governments, since 2004.
When it comes to the project target, the aim is to achieve 20% NRW within 5 years
from the beginning of each respective project, including infrastructure installations
13
and maintaining this level until the end of the project life cycle, typically 20 years
(K-water, 2014). This NRW rate has been established as a performance indicator
for a long time in South Korea. This has been the case despite the problem with
changes in the level of consumption. In addition, NRW rate does not consider the
operating environments such as finances, water use patterns and topographic
conditions of the individual areas (Koo et al., 2011).
Since 2004, identical target setting has created problems because regional
characteristics, financial conditions and water use scale were not considered.
Specifically, the efficiency of NRW reduction shows variation in NRW control cost
such as leakage repair, pipe replacement, and pressure management. Some
projects with a budget shortage may have difficulty managing their water system
for the remaining period. Therefore, it is necessary to introduce an economic
principle for achieving and maintaining the NRW target efficiently with the limited
budget to the K-water projects. This introduction of an economic framework will
allow water suppliers to manage their water system economically and efficiently.
The research carried out in this thesis addressed the issues mentioned above. The
verified method in the UK and newly developed calculation model, cumulative cost-
benefit analysis, are proposed.
1.2 Aim and Objectives of Research
The overall aim of this research is to contribute to the development of the

economic level of NRW calculation model and target setting method. The specific
objectives are:
1. To introduce and develop an economic level of NRW calculation model for

South Korean water systems. Given the condition of rising operation and
management costs in water supply systems, it is required to introduce and
develop cost-effective NRW management strategy from an economical
perspective in South Korea.
2. To identify an acceptable NRW level within the budgetary constraints. Most
water suppliers in South Korea regard NRW level of developed countries as
14
10%, and they try to achieve this as an ultimate objective. However,

different NRW levels should be set, which can be achievable and
maintainable by every water supplier.
3. To assess whether application of uniform NRW target to K-water projects is

appropriate and reasonable. Water suppliers have various operating
conditions such as financial status, labour force, infrastructure deterioration,
geographical conditions, and political and social demand.
4. To test, verify and demonstrate the applicability of developed

methodologies for economic level of NRW calculation. This economic level
of NRW calculation is a newly attempted method in South Korea. It can be
used for a range of purposes (i.e. as a system performance indicator,
budget allocations, and project target setting).
1.3 Thesis structure
The thesis is organized into five chapters showing the process and result of the
research. A specific description of each chapter is introduced below:
 Current chapter 1 presents the motivation and background. It also explains

the aims and objective of this study. Additionally, the thesis structure is
outlined.
 Chapter 2 provides a literature review and is divided into two sections:

general water loss management and Economic Level of Leakage (ELL)
concept. The former gives a wide range of overview about water loss
assessment, target setting and four water loss management methods. The
latter is focused on economic leakage management informed by previous
research.
 Chapter 3 sequentially introduces economic level of NRW calculation and

target setting procedure, required data collecting and processing, and
15
calculation methods. When it comes to methods, two methods are

presented: (1) Marginal cost Analysis and (2) Cumulative cost-benefit
analysis. The first method is generally used in the UK. The second method
is a newly developed method in this research.
 Chapter 4 shows a case study with discussed methodology from Chapter 3.

Firstly, the background of the case study area selection is presented.
Secondly, in order to decide the most suitable method, a reliability check of
the NRW control cost curve is carried out. Thirdly, the economic level of
NRW of the case study area is estimated with a chosen method. Fourthly, a
sensitivity analysis is presented with the purpose of finding the most
influential factor on the economic NRW level. Lastly, the evaluated
economic NRW level is compared with the current level and the most
efficient and economic target is suggested.
 Chapter 5 includes an overall summary of this thesis and related

contributions are briefly summarized. The conclusion of this research
emphasizes the discussed methodology in chapter 4, followed by required
future research work which could strengthen and extend the methodology.
16
Chapter 2 Literature Review
CHAPTER 2 LITERATURE REVIEW
2.1. Introduction
Leakage in water distribution systems (WDSs) represents lost water through pipes,
joints, fittings and reservoirs (Trow and Farley, 2004). These losses have been
taking place since the WDSs have been installed and they are common events
which we can notice in our everyday lives. However, recognition of leakage has
changed over time. The interruption in the water supply in the past mainly created
customer inconvenience; however, today it leads to social and economic effects
such as traffic congestion, flooding, interruptions in factory operations and
customer inconvenience. However, the complete removal of leakage in a water
distribution system is impossible and expensive (Stephens, 2003). As leakage
reduction activities follow the law of diminishing returns, there comes a tipping
point at which costs increases outweigh benefits of leakage reduction. Therefore,
finding and managing at the most economical level is required in terms of efficient
use of budgets (Pearson and Trow, 2005).
Research related to the economic efficiency of leakage has been performed since
the eighties. Particularly in the UK, economic management methods in WDSs have
been developed and all the water companies have adopted the Economic Level of
Leakage (ELL) methodology. Water Services Regulation Authority (Ofwat), which
is an organization that regulates the performance of the water companies,
evaluates their annual reports on a regular basis. With this effort, most companies
are now operating at the high efficiency level. “The optimal level of leakage is the
point at which the cost of reducing leakage is equal to the benefit gained from
further leakage reductions” (Ofwat, 2002). In other words, it is better to stop
reducing water leakage at the point where costs and benefits meet.
The focus of water supply utilities in South Korea has not been on operational and
economic management because they have concentrated on the development and
expansion of infrastructure (Koo et al., 2011). Recently, they also showed interest
in the economic leakage management by reducing both leakage and high
17
maintenance costs due to the rise in the cost of water production. However, this
work on saving costs in water companies has been performed only at a basic level.
It still requires much effort to systematize the data store and its acquisition.
The purpose of this research is to find advantages and disadvantages identified in

previous studies, especially in the UK, related to the Economic Level of Leakage
(ELL). Its aim is to develop an appropriate methodology which can be applied to
water supply systems in South Korea. This literature review was performed with
this particular aim in mind. This literature review is largely composed of two main
parts. The first part explains the overall water loss management concept and the
other presents ELL methodologies and previous research efforts.
2.2. Water loss management
Despite the global effort to reduce water loss, a large amount of water is being lost
through leakage. According to the World Bank discussion paper No.8 (2006),
32billion m³ of water disappears every year. For the efficient use of limited water
resources and its conservation, it is necessary to understand the cause of water
loss and to attempt to reduce this by using the techniques (described below in
section 2.2.6).
Farley and Liemberger (2005) stated that why water loss occurs in the WDSs
mainly occurs because of poor infrastructure, bad operation and management.
Specifically, it can happen for various reasons such as the complexity of water pipe
networks, shortage of operators and equipment, and a lack of professionalism,
insufficient repair/replacement and faulty customer meters.
Water loss has many economic and social repercussions. Firstly, water lost is
directly linked to water company profitability since it generates additional costs
such as treatment, sludge disposal and electricity. Secondly, water loss leads to
customer dissatisfaction and damages the company’s image, since disruption of
supply and low water pressure has an important effect on customer satisfaction.
Lastly, water loss can reduce available water resources (ABB Limited, 2011).
18
2.2.1 Water loss assessment
Analysis of annual water balance is very important to understand its components

and calculate its quantities, while considering optimal strategy establishment and
management (Lambert, 2003). The portion of volume of each component
compared to the system input volume represents a good measure that could be
used to decide the direction for water loss management. Therefore, the starting
point is to calculate the water balance. The components of water balance can be
seen Table 1.
Table 1 : International Water Association(IWA) standard international water

balance and terminology(Trow and Farley, 2004)
System Authorised Billed Billed Metered Consumption Revenue

Input Consumption Authorised (including water exported) Water
Volume Consumption Billed Unmetered
(corrected Consumption
for known Unbilled Unbilled Metered Non-
errors) Authorised Consumption Revenue
Consumption Unbilled Unmetered Water
Consumption (NRW)
Water Apparent Unauthorised Consumption
Losses Losses
Customer Metering
Inaccuracies
Real Losses 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
The most representative terms, which are widely used in WDS, are water losses
and NRW (Wu, 2011). Water losses consist of apparent and real losses and NRW
is made up of water losses and unbilled authorized consumption, such as public
purpose uses. NRW also expressed the difference between system input volume
and revenue water.
19
High level of water losses and NRW means huge financial losses and waste of
limited water resources. In order to prevent these losses, it is very important to
understand the reasons and factors which affect water loss and NRW. Trow and
Farley (2004) explain that the most important issue in a water loss management
strategy is deciding the leakage target and assessing current leakage level
precisely. Currently three methods are widely used by the water companies to
calculate water losses: (1) Total integrated flow analysis(Top-down), (2) Minimum
night flow analysis (Bottom-up) and. (3) Component analysis (Thornton et al.,
2008).
2.2.2 Water loss Assessment methods
Significant practical efforts to address leakage in the WDS have been made since
the Burst And Background Estimate (BABE) methodology was developed in the
mid-1990s (Lambert, 1994). In order to reduce controversy about measuring water
losses, IWA (2000) introduced standard terminology and performance measures
for water balance calculations. Unavoidable annual real losses (UARL) which is the
lowest attainable annual real losses under the current leakage control policy was
made based on the BABE concept (Lambert, 2003). This can be used for
comparisons of technical leakage performance within a water system with current
annual real losses (CARL). The specific explanation is described in Section 2.2.3.
2.2.2.1 Total integrated flow analysis (Top-down method)
A ‘best-practice’ standard water balance, as shown in Table 1, is a common

terminology originally introduced by International Water Association Task Forces
(IWATF) to unify various formats and relevant definitions. This method has been
most widely used by water companies because of its easy and simple to use
technique (Lambert, 2003). Using the data presented in Table 1(described in
section 2.2.1), water losses are easily calculated by deducting authorised
consumption comprising of billed and unbilled authorised consumption from system
input. However, as its quantity cannot be calculated directly, it can only be
assessed by evaluating other components. In this respect, it is important to note
20
that the water loss volume can be affected by errors and/or uncertainties
associated with other components (Weimer, 2001).
Generally, unbilled authorised consumption and apparent losses are considered

having high errors and uncertainty. In terms of unbilled authorised consumption,
water is used for public purposes, such as flushing distribution main, street
cleaning, fire fighting and frost protection, and is not subjected to exact metering.
According to a recent paper published by the IWA Water Loss Specialist Group
(WLSG), unbilled authorised consumption is estimated to contribute to the billed
metered consumption by 0.5%(Lambert et al., 2015). Even though this is a small
portion of water balance, the used water volume should nonetheless be recorded
for precise assessment (Weimer, 2001). With respect to apparent losses arising
from unauthorised consumption (theft and illegal use) and metering errors, these
are estimated to account for 4.5% of the System Input Volume in South Korea,
based on figures pertaining to other countries. The quantity is still significant and
must be addressed (Weimer, 2001).
However, it must be noted that obtaining information about real losses and
separating the individual components is extremely difficult, if not impossible.
Besides, since the total integrated flow method is normally used on a yearly basis,
there are some limitations to water suppliers for using this method as an alert
system, especially if the aim is to detect new leaks and bursts. For these reasons,
it is necessary to utilise it in conjunction with two additional assessment methods
described in the subsequent sections.
2.2.2.2 Minimum night flow analysis (Bottom-up method)
Minimum night flow analysis (Bottom-up method) was first developed in the UK and
is based on night flow measurements and flow-pressure relationship. Thus, real
losses are calculated by subtracting customer night time use from minimum night
flow, which is measured between 2 and 4 AM. However, this approach should also
incorporate the Night-Day-Factor to convert the results into daily volume of real
losses in order to account for diurnal variation of the pressured distribution system
21
(Wu, 2011). Nonetheless, this method is useful for checking the real losses of Top-
down calculations, whereby the gap between top-down and bottom-up should be
within 5% for precise estimation. Normally, if its gap exceeds 5%, water suppliers
are recommended to repeat the procedure. Although this is a complementary
assessment method which can supplement the top-down method, it is required to
verify average customer night use for applying to the South Korea situation. The
reason is average values for calculation of customer night use need to be adjusted
because of different per capita consumption, lpcd (litres per capita per day),
between approximately 150 lpcd in UK and approximately 330 lpcd in South Korea.
Since different water use habits and patterns exist between the two countries,
investigation of customer night use needs to be done in advance. Otherwise,
volume of leakage is likely to be overestimated.
There is another opinion about minimum night flow analysis. Handy (2011)
reported that it is difficult for a water distribution system having high water losses to
adapt this method. This is because WDSs having high water losses usually have
poor finances, inaccurate water meters and unauthorized consumption; eventually
these bring about huge and unmanageable uncertainties in water balance analysis.
2.2.2.3 Component analysis
Understanding the various components of water balance is important to remove

leakage. Component analysis is very helpful to find specific reasons for the
leakage and the volume of losses in the WDS. Lambert (2003) reported that loss
volume from each leakage location is affected by flow rate and leakage run-time
until the leaks is repaired. This is the BABE concept, which was first developed and
calibrated in order to understand the factors involved in water balance in the UK.
This method uses numbers, average flow rate and average run-times of leaks and
bursts to estimate annual real losses in each part of distribution infrastructure such
as mains lengths, number of connections, and private pipes from boundary to
meter. If all data are produced by water suppliers correctly, this method allows
water suppliers to establish the most suitable real losses reduction strategy. This is
22
significant in terms of efficient water loss management for achieving economic

level of leakage(Thornton et al., 2008).
It is very important to develop a water loss management strategy because this

makes water suppliers invest their budgets efficiently. However, in case sufficient
data is not available, many assumptions such as flow rate are needed for the
analysis. Therefore, systematic data management in analysis of water losses is
very important. The best way to evaluate water losses is the combination of three
methods (top-down, bottom-up and component analysis) in order to get reliable
results. (Wu, 2011)
2.2.3 Performance Indicators
Performance indicators (PIs) are used to compare performance across the country,
benchmark best practice for water loss management and to set targets. Until now,
when it comes to evaluate performances of NRW and water losses, the most
widely used PI are ‘percentage by volume’. It has traditionally been widely used as
a PI even though it has huge potential for misinterpretation and manipulation. The
reason is that it is affected by consumption scales and changes (McKenzie, 2002).
Alegre et al. (2000) also announced that it is not suitable for assessing operational
management of real losses. In order to resolve this matter, Alegre and Association
(2006) explained PIs based on the IWA’s Manual of Best Practice as seen in Table
2.
23
Table 2 : Recommended indicators for physical losses and NRW (Alegre and
Association, 2006)
Function Level Performance Indicator Comments
Financial: 1 Volume of NRW Can be calculated from
NRW by (Basic) [% of System Input simple water balance,
volume Volume] but not too meaningful
Operational: 1 [Litres/service Best of the simple
Physical (Basic) connection/day] ‘traditional’ performance
Losses Or indicators, useful for
[Litres/km of mains/day] target setting, limited
(only if service use for comparisons
connection density is between systems
<20km)
Operational: 2 [Litres/service Easy to calculate
Physical (Interimed) connection/day/m pressure] indicator if the ILI is not
Losses Or known yet, useful for
[Litres/km of mains comparisons between
/day/m pressure] systems
(only if service
connection density is
<20km)
Financial: 3 Value of NRW Allows different unit
NRW by cost (Detailed) [% of annual cost of costs for NRW
running system] component, good
financial indicator
Operational: 3 Infrastructure Leakage Ratio of current annual
Physical (Detailed) Index(ILI) physical losses to
Losses unavoidable annual
real losses, most
powerful indicator for
comparisons between
systems
24
From among these, the infrastructure leakage index (ILI) is considered the most
appropriate performance indicator (PI). This is a measure of how well distribution
system is managed and maintained in terms of reducing real losses. It is the ratio
of Current Annual volume of Real Losses (CARL) to Unavoidable Annual Real
Losses (UARL). This indicator has no units so it is very effective for comparing
between countries that use different units. (Wu, 2011)
The formula can be seen below:
ILI = CARL / UARL (2.1)
UARL (litres/day) = (18 x Lm + 0.8 x Nc + 25 x Lp) x P

- where Lm = mains length (km); Nc = number of service connections;
- Lp = total length of private pipe, property boundary to customer meter (km);
- P = average pressure (metres).
By the way, three coefficients used in the above UARL calculation formula are
highly uncertain and need to be developed separately for each system. Because
the three coefficients were derived from component-based approach, those values
can be varied with different system conditions such as infrastructure condition,
intensity of leakage control, and quick and effective leak/burst repair. Thus even
though above formula is being used internationally, it needs to be assured whether
the coefficients are reasonable values to each water system.
To return, ILI values close to 1.0 represents that the CARL is managing and
operating at a technical minimum. However, it does not mean that such low ILI
values are economical in leakage management. This is because ILI is a technical
performance indicator without considering the economic aspects (McKenzie, 2002).
Pearson (2002) announced that ILI of the 34 demand zones of United Utilities in
UK showed scope between 1.13 and 3.2, its average and median value were
respectively 2.38 and 2.21. This means most demand zones are well managed and
operated by water companies. Moreover, their infrastructures are maintained under
the favourable conditions. Figure 1 shows the ILI value of each demand zones.
25
Figure 1 : ILIs of 34 Demand Zones in North West England (Pearson, 2002)
In contrast, the ILI value of K-water, (the Korean Water Resources Corporation)
was much higher than the UK results. The range of ILI values are spread from 1.74
to 27.86. The average and median value was 7.99 and 6.21 (Environment, 2012) .
Specific ILI value can be seen in Figure 2.
Figure 2 : K-water O&M 17 cities (Environment, 2012)
26
There is another opinion about using ILI as a performance indicator. Liemberger et

al. (2007) pointed out that assessing ILI identifies a problem with data reliability in
developing countries. This is because they do not manage or record important data
such as network length, number of service connections and pressure and real
losses. This causes different result with the current operation environment.
However, until the present, there have been no suitable performance indicators
which can replace ILI and it is still the best indicator to describe the level of real
losses of a system.
2.2.4 Leakage target setting
It is crucial to establish an accurate water losses target along with the calculation of
the precise volume of water loss. Exact target setting allows water suppliers to use
limited water resource efficiently and it prevents unnecessary investment such as
water pipes replacement, construction of reservoirs and water treatment facilities.
PIs presented in the table 2 are good indicators as a water loss target setting.
Therefore, many water suppliers use them widely for comparing and benchmarking
their performances with others (Fanner et al., 2007).
In South Korea, the ‘percentage by volume’ method has been used as an indicator
to set a goal for a long time despite the problem with changes in the level of
consumption by customers. Their indicator is NRW rate (%) which is calculated by
dividing volume of NRW with the volume of total water produced in the treatment
facilities. In general, policy makers and water suppliers think the optimal NRW rate
is approximately 20 percent, the national average NRW rate in South Korea.
However, it does not consider various options such as financial, water use pattern
and topographic conditions of the individual areas. Recently, K-water got interested
in the ELL, which is UK’s economic framework in water loss management. The
idea is that it will enable water suppliers to manage their water loss economically.
Furthermore, it can suggest with limited budgets and current manpower structure,
which NRW level is the maximum achievable. This is first objective of this study
along with adapting the ELL methodology for the South Korean water system.
27
In the UK, the ELL methodology has been used for setting leakage targets since
1990s. ELL calculations involve finding the tipping point between the costs and
benefits of water loss management (Ofwat, 2002). Ofwat decides economic level
by using ELL methodology and manages water loss accordingly. It is based on
mega litres per day (Ml/day) units. More details about the ELL methodology will be
presented in the section 2.3.
2.2.5 Leakage detection methods
Leakage detection has been the basic approach to find and locate leaks in the
WDSs. Regardless of equipment, it can be classified with localization and
pinpointing (Pilcher, 2003, Fanner et al., 2007, Puust et al., 2010).
2.2.5.1 Leakage localization methods
Leakage localization is a procedure aimed at identifying and prioritizing areas

within the system, in order to discover leaks faster and easier (Pilcher, 2003).
Typically, step-testing, acoustic logging, ground motion sensors and ground
penetrating radars are employed (Puust et al., 2010). Step-test examines changes
in the water flow data during the period of minimum night flow by closing valves
systematically to identify suspected areas. However, owing to the difficulties
associated with planning and working at night, in the 1990s, this method was
replaced with acoustic logging (Pilcher, 2003). Elaborate equipment, such as
vibration sensors or hydrophones that can collect leak signals during night times,
used in acoustic logging approach, make leakage localization easier and simpler
than in the step-test. The data collected in this manner can be analysed statistically
using computer software. However, experts skilled in this type of analysis are
required, and the problem of background noise interfering with sound collection still
needs to be addressed (Puust et al., 2010). Lastly, ground penetrating radar (GPR)
is a non-destructive method enabling analysis of both cross-section and surface
features. Owing to the technological advances, scanning speed of up to 15-30
km/h is now possible, making this approach extremely efficient. Nevertheless, it
can yield false conclusions, when metal objects are detected or pipes are buried
28
deep in the ground (Puust et al., 2010).
2.2.5.2 Leakage pinpointing methods
Leakage pinpointing in contrast to leakage localisation is elaborate process

because this can directly affect an excavation costs and amount of labour
necessary to repair leaks. Three widely used methods are described in this
section: (a) leak noise correlators; (b) gas injection; (c) pig-mounted acoustic
sensing (Puust et al., 2010).
Leak noise correlators (LNC) are employed in order to find a leak point by
comparing the arriving time of sound from one correlator to another. The accuracy
of this method depends on pipe materials. While it can locate a leak within 1 metre
in case of metal pipes, its utility is questionable when applied to plastic pipes.
Tracer gas technique (TGT) is another method used in this field, which can detect
a leak point by using gas, such as helium or hydrogen. After injecting gas into the
pipe, operator traces its distribution using a highly sensitive gas detector. Although
this is a highly accurate method, it is very costly. Furthermore, it might adversely
affect the operating water network, as it can cause a water pipe to burst if the gas
is injected at high pressure (Puust et al., 2010).
Lastly, pig-mounted acoustic (PMA) technique is based on placing a microphone

into the main to record the leak position and noise. It is a highly accurate approach
because, by passing through the water pipe, the microphone can collect
information that can provide the exact distance and transmit the leak sound.
However, it cannot be applied to old pipes due to heavy corrosion. In addition, it
may be affected by water quality, since pigs are in direct contact with the inner pipe
(Puust et al., 2010).
29
2.2.6 Water loss management methods
After setting a specific leakage target, adopting the most suitable leakage
management methods that can accomplish the optimal level effectively and
efficiently is significant because various types of leakage occur in spite of
continuous leakage reduction activities. Moreover, the leak volume varies with the
water use patterns, geographic conditions and facilities condition. According to
IWA (2000), the following four representative leakage management methods have
been shown as most effective: (1) pressure management, (2) active leakage
control, (3) speed and quality of repairs and (4) pipeline asset management,
maintenance, and renewal. This can be seen in Figure 3.
Figure 3 : The four pillars of a successful leakage management strategy (Force,

2003)
In Figure 3, the large rectangle represents the current annual volume of real losses.
If the four methods mentioned above are not applied effectively, the volume of real
losses will slowly increase. The smaller inner rectangle - Unavoidable Annual Real
Losses (UARL) which stands for the lowest achievable level – can be obtainable
only if the water system is well managed and maintained. However, achieving this
30
level entails huge investment (Liemberger et al., 2007). The Economic level of Real
Losses is the point of balance and should be targeted by water suppliers.
2.2.6.1 Active leakage control (ALC)
Severn Trent Water (2009) stated that active leakage control process consists of
proactively seeking out or detecting leaks for repair. There are two types of
leakage control according to the response to the leak. One is passive leakage
control (PLC) and the other is active leakage control (ALC). PLC assumes that
water suppliers react only when water comes up to the surface or when their
customers complain about low pressure (Thornton et al., 2008). ALC is one of the
most rapid response leakage reduction methods today and its process consists of
leakage monitoring and regular surveys. The leakage monitoring is finding changes
of flow rate by comparing flow entering into the water supply zone to quantify
leakage and deciding the priority for the leakage survey. Regular surveys are an
activity that identifies and locates the leaks periodically through listening for leaks
on pipes, fittings, valves and water meters, and by using leakage detection
equipment (Farley and Liemberger, 2005). With the development of computational
management techniques, desk-based studies have also been made possible.
2.2.6.2 Pressure management
Pressure management is the most cost-efficient and rapid method of the leakage
reduction techniques since it is possible to decrease leakage by reducing pressure
instead of detection, repair and pipe replacement through the whole distribution
area. Moreover, this method has many advantages such as reducing the number
of leaks and bursts, extension of infrastructure life, and decreasing surge impacts
(Fanner et al., 2007). Koo et al. (2011) stated that if utility managers reduce over-
pressure by 10%, they can save as much as 10% in the volume of leakage.
Therefore, understanding the relationship between pressure and leakage is very
important in pressure management because the volume of water losses varies
along with pressure.
31
2.2.6.2.1 Fixed and Variable Area Discharge (FAVAD) concept
May (1994) define the concept of Fixed and Variable Area Discharge (FAVAD),
which is the most widely, used equation showing the relationship between pressure
and leakage. The following is the FAVAD equation.
L1  P1 
N1
 
L0  P0 
L Varies with and (2.2)
Where L0, L1 : the leaks flow rate before and after change in pressure
P0, P1 : pressures before and after change in pressure
Equation (2.2) describes how Leakage Rate varies with Pressure P to the power
N1. N1 is the controlling factor to explain different type of pipes and stature of the
distribution network (Wu, 2011). N1 value varies from 0.5 for fixed area such as
metal pipes to 2.5 for variable area such as non-metallic pipes. In the case of large
systems, which have mixed pipe material, N1 can be assumed to be 1.0 (Fanner et
al., 2007). N1 value can be decided by using pressure-leakage relationship curves.
These curves can be seen in Figure 4.
Figure 4 : Ration of Pressures (Force and Thornton, 2003)
32
This N1 value can be calculated by a pressure step test conducted at night. If there
is no specific information, N1 can be assumed to be 1.0. And IWA WLTF
developed a more empirical equation:
N1 = 1.5 – (1 – 0.65/ILI) x p/100 (2.3)
Where ILI is the Infrastructure Leakage Index and p is the percentage of detectable
leakage on rigid pipes.
2.2.6.2.2 Type of Pressure Management
Pressure management can be divided into three types: (1) Sectorization, (2) Pump
control, (3) Pressure reducing valve control.
Firstly, the most basic form of pressure management is sectorization. This method
divides large area into the small sectors naturally according to the topography,
ground level and water pressure. Otherwise, the area can be divided by installing
boundary valve artificially. This is simple and cost-effective way but not sufficient to
control pressure perfectly. It is more effective when using with the pressure
reducing valve and controller (Thornton et al., 2008).
Secondly, pump control can be a method of pressure management. This control

pump head to maintain proper water pressure at the critical point and average
zone pressure point of a pressure management area. In case of using variable
speed pump to meet a change in flow and pressure, this can lead to waste of
energy due to the frequent pump operation. This methodology needs to be used
carefully (Thornton et al., 2008).
Lastly, the most widely used method is using pressure reducing valves (PRV).
PRVs reduce or maintain the pressure from a set point to the customer
irrespectively of the upstream pressure or changes in flow rate (Wu, 2011). There
are three types of PRV operation: fixed outlet control; time-modulated control; flow
modulated control. A brief introduction of these methods can be seen in Table 3.
33
Table 3 : Comparison of three PRV operation types (McKenzie and Wegelin 2010)
Types Characteristic Advantages Disadvantages
Fixed outlet  Control Maximum  Simple control  Difficult to
pressure from  No additional control surplus
inlet point equipment pressure
 Cheap
installation
Time-modulated  Control operating  Provide further  Cannot react to
time with the reduction water
pressure changes during the off- demand change
at the critical point pick period  More expensive
than Fixed outlet
Flow-modulated  Control the flow  Most effective  The most
of the inlet point of the Three expensive
and react to the types  Require high
pressure of the level of
critical point operation skill
Each type of pressure management has own advantages and disadvantages.

McKenzie and Wegelin (2010) emphasized that operator should consider selecting
appropriate form of pressure control type rather than the valve types by comparing
variables such as budget, expected savings and operational skill in pressure
control.
2.2.6.3 Speed and quality of repairs
Volume of water losses consists of large and small losses. Large leaks like those
on water mains normally last for a short time while small leaks could last much
longer as they are difficult to detect. In order to control these leaks, it is vital to
understand three elements of leakage runtime (Thornton et al., 2008). The volume
of lost water from leaks is function of time which consists of awareness time,
location time and repair time (Trow and Farley, 2004):
34
- Awareness time: The average time from start of leak to aware of its existence.
- Location time: The average time to find the location of the leak.
- Repair time: The average time to repair the leak.
Water suppliers should reduce leakage runtime for the fast repair of leaks. In terms
of awareness time, frequent customer opinions gathering about the water pressure,
real time monitoring of district metered areas (DMA) and increased leakage
detection efforts have an effect on reducing overall awareness time. Location time
can be reduced by positive leakage detection campaign of well-trained detection
team members. At the same time, appropriate usage of leakage detection
equipment is also an important way. Another way of locating pipe bursts is using a
decision support methodology. Good quality of repairs also have an effect on
decreasing the repair time (Thornton et al., 2008). For example, when a small pin
hole happens in a cast iron pipe, the network operators have to decide how to
repair the leak and whether to replace the leaking section with new pipes or to
repair the pipe with a clamp. This decision also can reduce repair time.
2.2.6.4 Pipeline and assets management
If the target area has a high burst frequency and a high level of background
leakage, the replacement of water pipes can reduce this (Trow and Farley, 2004).
In addition, this is the most effective method to eliminate background leakage
along with pressure management. However, since water pipe replacement follows
a law of diminishing returns, making an accurate diagnosis can minimize
unnecessary pipe replacement. Currently, this has been made possible along with
the development of technology, such as internal inspection with robot or elaborate
camera (Thornton et al., 2008). Within this context, various trenchless technologies
are considered when making a decision about pipe replacement. (The advantages
and disadvantages of representative techniques are described in the Table 4)
35
Table 4 : Water pipe replacement and rehabilitation techniques (Thornton et al.

(2008)
Techniques Advantages Disadvantages
Replacement replacement  Exact replacement  Traffic congestion
 Remove background  Pedestrian
leakage inconvenience
 Improving water Quality
Slip lining  Simple and fast  Reduction of cross
construction sectional area
 Long length can be  Difficult to find leak
achieved
 Improving water quality
Pipe cracking  Fast construction  Difficult to remove
or  Improving water old pipe
pipe bursting quality
Rehabilitation Epoxy  Fast renewal without  Lining peel off
excavating service  Pipe to be clean
connections before lining
Cement  Structural  Long setting time
enhancement  Reduced hydraulic
 Flow is improved capacity
 Provide long term  Impossible for small
protection pipes
36
2.3. Economic Level of Leakage
Due to the need to extend the existing water system, to mitigate climate change
impacts and to meet customers’ expectations, management costs of water system
are rapidly increasing. In the field of leakage management, there is an increasing
need for investment to reduce leakage. This is due to the detection costs of finding
remaining smaller leakage which is much higher than for the bursts that can be
found easily. Recently, these rising costs have imposed a burden on water
suppliers in terms of the difficulty of securing a budget and achieving a low leakage
level. For this reason, water suppliers were forced to increase their interest in
economic water production and distribution. In South Korea, K-water which is the
public water company has developed an interest in economic leakage
management and tries to adapt ELL methodology.
ELL research first started in the UK in the 1980s. A number of water companies
began to use this concept while going through severe drought in 1995/1996. The
continuous effort of water companies and Ofwat regulation led to considerable
reduction in leakage levels from 5,112Ml/d in 1994-95 to 3,576Ml/d in 2005-06
(Ofwat, 2007). Currently, most UK water companies manage their leakage level
close to the ELL. In order to develop an applicable methodology for South Korea,
the following are described in this section: ELL definition, components for
calculating ELL and recent research trends on economic leakage management.
2.3.1 Definition of ELL
UKWIR (1994) used the term “Optimum level of leakage” which is “the level of
leakage where the marginal cost of active leakage control equals the marginal cost
of the leaking water”. Ofwat (2002) defined “Economic level of leakage (ELL)” as a
same concept as optimum level of leakage. That is “the point at which the cost of
reducing leakage is equal to the benefit gained from further leakage reductions”
(Ofwat, 2002). In other words, when achieving the ELL it is possible to minimize
the total costs for supplying water to the customer and to operate the water system
most efficiently (Stephens, 2003).
37
Eventually, ELL is the level that water suppliers should achieve within their
financial constraints. The concept of ELL is expressed in Figure 5.
Figure 5 : Economic level of leakage calculation(Fanner et al., 2007)
The X-axis is the level of real losses and the Y-axis is the cost of leakage control
and lost water through leakage. Annual cost of leakage reduction refers to the cost
of all leakage reduction activities which follow the law of diminishing returns. For
example, the more reduction in real losses required, the more cost compared to
the low level of real losses. If the current real losses are reduced to the background
leakage level, the costs of leakage reduction increase exponentially. Regarding the
annual cost of water lost shown in the graph, this refers to the sum of production
and distribution costs which varies according to the volume of lost water. This can
be saved directly by producing less water through leakage reduction. The overall
Economic Leakage Level is calculated by the sum of the annual cost of leakage
reduction and the annual cost of water lost. On this curve, the minimum point is the
economic leakage level. (Smout et al., 2010)
38
2.3.2 Time frames of ELL
An economic level of leakage has, according to the capital investment, both short
and long run time frame. The former only considers operational cost, without
capital expenditures. The latter evaluates life cycles and decision making. As
mentioned above (See section 2.2.6), of the four primary components of leakage
management, active leakage control and speed and quality of repairs refer to the
short run ELL. Pressure management and infrastructure that requires significant
investment decision is considered ELL in the long run (Pearson and Trow, 2005).
Recently, the upgraded concept of “Sustainable economic level of leakage (SELL)”
was introduced. SELL considers externalities such as social and environmental
cost and benefits. (Beal et al., 2012)
2.3.3 ELL target setting process and the components
2.3.3.1 ELL target setting process
Ofwat (2002) presented an ELL target setting process map. It can be seen in
Figure 6.
Figure 6 : ELL target setting process map (Ofwat, 2002)
39
The start of ELL target setting is deciding the area of analysis which requires
selecting a smaller geographical zone for more accurate results. Next, the current
policy minimum needs to be set on the basis of current leakage levels which are
estimated from the bottom-up and top-down water balances referred to in the
above chapter (See section 2.2.2). At the same time, leakage detection and repair
costs should be calculated. After this process, decision makers have to consider
new policies and technical options to ensure further reduction in leakage in the
future. Then, the relationship of leakage level and costs need to be reviewed for
greater efficiency and better financial savings. Finally, decision makers can choice
appropriate options among Option 1 and Option 2 and estimate a suitable leakage
target. (Ofwat, 2002)
2.3.4 Calculating Economic Level of Leakage
2.3.4.1 Policy minimum
Policy minimum is defined as a lowest achievable level of leakage at every DMA

with current leakage control methods, equipment and reasonable effort (Ofwat,
2002). It consists of reported leaks and un-detectable leaks under the current ALC
operation and whether theoretical or practical levels of leakage are achievable.
Therefore, deciding on the policy minimum is an important process, since it is the
lowest level which can be achieved under the current leakage control strategy. In
other words, in order to achieve a lower level than the current policy minimum,
another leakage control policy needs to be introduced. Trow (2006) also explained
that “policy minimum is a level of background losses which results from an
optimized entry and exit policy for DMA management, or exit policies for regular
survey”.
Policy minimum can be calculated as follows. First, intensive sounding and noise
logging throughout DMA should be performed and all leaks should be repaired
within the target periods. The next is analysing minimum night flow and calculates
the average minimum nightline values. At this moment, the average size of DMA
40
should be within between 900 properties and 2000 properties. Lastly, the current
policy minimum should be compared to the historical value (Ofwat, 2002).
According to the data from the Water UK leakage managers, the average
background leakage is about 51% of total leakage (Beal et al., 2012).
Estimation of policy minimum is currently impossible in South Korea, since

minimum night flow analysis has not yet been applied to the water system.
Therefore, as an alternative, indirect, comparative analysis with the UK’s average
rate of background leakage and the UARL indicator will be employed in Chapter 4.
2.3.4.2 Least cost planning approach (Option 1)
The least cost planning approach is used when setting a long-term plan (basically
25 to 30 years) for managing the supply-demand balance. This approach aims at
minimizing a net present value of the cost related to the supply-demand investment
(operating, capital, social and environmental costs). Leakage is a factor that affects
supply-demand balance. The leakage profile that leads to the lowest net present
value of costs, is precisely the economic level of leakage profile (Ofwat, 2002).
While this is similar to the method of calculating the marginal cost of water, it is a
more comprehensive concept because the least cost planning deals with all
options, related to supply-demand policy, based on supply-demand forecasts
(Ofwat, 2002, Beal et al., 2012). Furthermore, this approach gives more accurate
results compared to the marginal cost method due to applying various options.
Hence, many water companies take advantage of marginal cost approach to
establish the base line leakage level, and then use a least cost planning approach
to decide if further leakage reductions are economic (Beal et al., 2012).
However, taking into consideration of the current data level, the marginal cost
approach is more suitable for the South Korea water system. This is because many
assumptions should be made for ELL calculation due to the complexity of
assessment and data-intensive structure of the least cost planning approach.
Therefore, the marginal cost approach was used as a method for ELL calculation in
41
this study. The marginal cost approach is presented in more detail than the least
cost planning in the following section.
2.3.4.3 Marginal cost approach (Option 2)
The marginal cost approach determines the ELL as the relationship between
marginal cost of leakage control and marginal cost of water form the next resource
scheme. If the marginal cost of leakage control for saving additional water is less
than that for development of next resource to meet the demand, it will be cost
effective to increase leakage control. On the contrary to this, the new resource
should be developed. Accordingly, ELL is defined as a level which “the marginal
cost of leakage control equals the marginal cost of water from the next resource”
(Ofwat, 2002).
The marginal cost approach only considers the balance between leakage control
and resource/treatment costs, this is relatively simple and more transparent than
the least cost planning (Ofwat, 2002). For this reason, even though regulators
suggest water companies use least cost planning, significant numbers of water
companies are more likely to use a marginal cost analysis. Especially, if there is no
deficit in the water supply, 83% of the water companies without a deficit used this
marginal cost approach (Beal et al., 2012).
However, because the least cost planning deals with full modelling of the supply-
demand balance, it is recommended to develop the SELL by using a least cost
plan.
As noted above, though the least cost planning is more powerful than marginal
cost approach, the former will be used in this research. Key components of
marginal cost approach, marginal cost of leakage control and marginal cost of
water, are explained as follows:
42
1) Marginal cost of leakage control
As with the marginal cost of water, the marginal cost of leakage control is an
important factor in the ELL calculation. This is additional cost in order to reduce
leakage from the current level to a lower leakage level. The greater the leakage
reduction, the more money is required when the current level approaches the
policy minimum (UKWIR, 1994). Hence, from the economic point of view, it is
critical to forecast associated costs depending on leakage levels. There are two
methodologies for estimating the cost of leakage control, and these will now be
outlined.
The first methodology draws a regression curve using the actual investment costs.
Whilst this method is highly reliable, it requires much effort to acquire data and
process because it is based on actual operating data. After completing the data
processing, it is possible to draw the leakage control cost curve similar to the
exponential or logarithmic function shown in Figure 5.
The leakage control cost curve drawn according to the South Korea situation is a
different form of the UK’s shape. In order to compare the form of the curve, two
types of graph are shown in Figure 7.
Cost Cost
Leakage level Leakage level
(1) Typical costs curve of UK (2) Costs curve of South Korea
Figure 7 : Leakage control costs curve
43
The left graph shown in Figure 7 is the typical leakage control cost curve in UK.
The right graph which is obtained from K-water, highlights to the Annual
Investment Plan of Water Efficiency Improving Project, which is operated by K-
water in South Korea. The main reason for these differences is attributed to the
investment policy. An average of approximately 40% of investment, such as
establishing DMAs, pressure management and pipe replacement/rehabilitation, is
being made by K-water within 5 years (Koo et al., 2011). Then, during the
remaining period of the project, (2013~2027), it is anticipated that investment will
be on average 4% of the total. These projected figures can be seen in Table 5.
Table 5 : Annual investment plan of Danyang Gun (K-water, 2008)
(Unit: ￡)
2008 2009 2010 2011 2012 2008~12 2013~27

Total
1 year 2 year 3 year 4 year 5 year 5years 15years
11,813,746 751,067 1,685,485 747,538 1,033,796 715,196 4,933,082 6,880,665
- 6% 14% 6% 9% 6% 42% 58%
Therefore, the development of the cost curve derivation method, allowing for this
investment structure of the K-water project, is the crucial point for the ELL
calculation. This specific process will be addressed and explained in the Chapter 3.
The alternative method is adopting the modelling approach introduced by
UKWIR/WRc in 1994. The method is represented by following Equation (UKWIR,
1994).
{ }
Total cost = C = Ca : (2.4)
{ }
L= level of leakage, m3/connection/year

C = cost of leakage control, ￡/connection/year
La = actual level of leakage for the area, m3/connection/yea
Ca = actual cost of leakage control for the area, ￡/connection/year
44
Lb = base level of leakage for the area, m3/connection/year

Lp = passive level of leakage for the area, m3/connection/year
Once each component mentioned above is calculated, the cost curve can be easily
derived from the equation.
2) Marginal cost of water
The marginal cost of water is an additional expense which is required for water
production and distribution in each water resource zone. There are largely three
types of costs: (1) Marginal operating costs, (2) Marginal capital costs and (3)
externalities (social and environmental) costs.
Firstly, marginal operating costs are comprised of pumping costs (with the
objective of abstraction, production and distribution), treatment chemical costs,
sludge disposal costs, and raw water withdrawing or purchasing costs (Fanner et
al., 2007). These costs directly link to the cost savings when leakage is reduced
and water does not need to be produced in order to meet the increasing demand.
Secondly, marginal capital costs occur when water suppliers cannot provide
sufficient quantities of water into the supply zone. Expenses are also incurred in
the construction of new facilities, which are planned due to the increase in demand.
This marginal capital costs can be saved by deferring future capital investment
plans (Fanner et al., 2007). Lastly, social and environmental costs, (which is the
last marginal cost of water effect on leakage reduction), should be considered in
the ELL calculation. This is because changes in leakage levels influence financial
considerations. It can save money via a decrease in water abstractions and less
purchasing of bulk water. Also, it can incur expenses such as inconvenience for
pedestrians and delay of vehicles. Therefore, it is necessary to develop an ELL
calculation by considering these costs for water suppliers, customers, society and
environment (Ofwat, 2008).
Nevertheless, there is no typical guideline about social and environmental costs.

Even though an example calculation with default/typical values was introduced in
45
2012, they are still specific to each area (Beal et al., 2012). Thus, it is necessary to
apply this to the externalities of the South Korea water system, but only after
further research has been undertaken.
2.3.5 Previous research on Economic Level of Leakage
The research on economic leakage management starts from questions: How much
money should be invested for leakage reduction? Additionally according to the
investment, what benefits we can get from the investment? (Lambert et al., 2015).
Various researches on ELL have been performed in UK since first report has been
published in 1994.
The purpose of the first trial was to develop an economic model for evaluating
leakage in South Africa. This model allows water suppliers to decide the economic
ALC intervention frequency, every 6, 12 and 24 months, by comparing the costs
(McKenzie and Lambert, 2002).
Lambert and Fantozzi (2005) introduced a model which can assess leakage
intervention frequency based on regular surveys. This method requires only three
parameters for calculations: Cost of intervention (CI), Variable cost of Water (CV),
Rate of Rise of Unreported leakage (RR). These required parameters, when
calculated exactly, offer a variety of applications: operators can easily confirm
economic intervention frequency, the time of next intervention, required annual
budget and economic annual volume of unreported real losses. However, most
countries with a different data management system to the UK would have a
problem calculating these factors.
The other huge change in estimating economic level of leakage is the inclusion of
externalities such as social and environmental costs and benefits. This is the
concept of sustainable economic level of leakage (SELL). If the existing concept of
economic level of leakage is made for water companies, the SELL is optimal for
both customers and society, by considering social and environmental impacts such
as reduction abstraction charges, carbon emissions and costs of traffic disruption
46
(Ofwat, 2008). Munoz-Trochez (2012) introduced how energy externalities

associated with ALC can be estimated and included in the ELL calculation.
Islam and Babel (2012) adapted the ELL concept to find the optimal ALC costs for
the Bangkok water distribution system in Thailand. While the methodology used in
this research followed previously published equations, this research also
introduced more user friendly elements. Those included not only how to calculate
ELL quickly and estimate optimal ALC costs, but also how operating pressure and
marginal cost of water can affect the ALC cost and ELL. This is a significant
advancement for water distribution system managers who want to analyse and
operate ELL as a management tool.
Finally, Cho (2013) stated that most research on economic leakage management
have focused on individual activities such as leakage detection, pressure
management and mains renewal rather than on overall activities for leakage
reduction and setting operational targets. In terms of long term economic level of
leakage, it is difficult to make accurate economic predictions due to the change in
costs and development of leakage detection technologies. Therefore, verification of
the long term leakage targets should be checked at least every 5 years.
2.4. Summary
Large and small leaks occur continuously in any water distribution network. The
process for reducing leakage has not changed much over time, but the related
costs such as labour, material and maintaining costs have increased dramatically.
Koo et al. (2011) stated that the costs for raising the revenue water rate by one
percent have increased 1.5 times compared to the early 2000s in South Korea. In
order to use limited budgets efficiently and manage leakage systematically,
economic leakage management methods should be introduced. However, due to
the differences of data management methods and systems, it is difficult for the
South Korean water system to apply the UK’s methodology. Therefore, in order to
resolve these problems, the following two tasks are considered fundamentally
important:
47
 Based on the water loss management literature review, accurate volume of

water losses should be calculated.
 Required information and method of collecting data need to be identified for
continuous ELL analysis.
The key requirements for calculating the ELL falls mainly into three categories. The
first is estimating leakage control costs and benefits according to the leakage level.
The second is developing procedures of cost curve derivation by considering the
investment structure of the project. The last is simplifying the complex UK-based
ELL calculation process as data management systems and the variable quality
levels differ in South Korea. There is a requirement to develop a methodology that
can easily evaluate ELL with limited South Korea data:
 Exact costs and benefits resulting from leakage reduction should be worked
out.
 Cost curve derivation procedure which is suitable for the business structure
of K-water projects have to be made.
 Simple and easy methodology which can adapt with current operation data
should be included.
In the next chapter, specific methodology and process will be presented by

focusing on the difference between the UK and the South Korea.
48
Chapter 3 Methodology For Assessing Economic Level of Non-Revenue Water
CHAPTER 3 METHODOLOGY FOR ASSESSING

ECONOMIC LEVEL OF NON-REVENUE WATER
3.1. Introduction
The first local waterworks operational efficiency projects were promoted by K-water
in South Korea in 2004. Given the remaining project period (i.e., approximately 10
to 15 years), it is necessary to review the results of NRW reduction and suitability
of uniform NRW goals applied to all projects. Having has this background, this
study aims to conduct this review and suggests reasonable NRW target.
Currently, there are no serious discussions about an evaluation method related to

economical leakage management in South Korea. Only the NRW rate has been
used as an indicator to compare the level with another local waterworks’ and
national average NRW. However, this indicator is unable to take into account the
economical leakage management (Cho, 2013, Koo et al., 2011). In addition, this
indicator overlooks some variables in the network such as regional characteristics,
water use pattern and scale, and infrastructure deterioration status. These
variables can have a significant impact on the NRW level and budget security.
Meanwhile, approaching the same objectives and NRW reduction strategies,
without considering these variables, has shown a huge difference in the reduction
of NRW unit costs (i.e., more than 5 times) (K-water, 2013).
The literature review was conducted in order to understand the water loss
management concepts and economic target setting method, and guide the
development of the leakage target setting approach for South Korea. In light of this,
ELL calculation model of the UK, (reviewed in Chapter 2), is the optimal model to
examine the appropriateness of the K-water projects’ target. However, the direct
application of UK’s model to K-water projects is problematic. The main reason is
that data acquisition and processing, analytical procedures and method definitions
need to be modified to reflect the South Korea business environment.
49
In chapter 3, the applicable methodology for the South Korea context, based on the
ELL used in the UK, is developed and described. In section 3.2, the target setting
process of the economic level of NRW is illustrated. In the next section (3.3),
components of the economic level of NRW, data collecting and methods and two
approaches for economic NRW target setting are explained. In addition, section 3.4
presents how the economic level of NRW is applied to the target setting. Next,
section 3.5 describes sensitivity analysis of the key parameters. Finally,
conclusions are drawn in section 3.5.
3.2. Economic level of NRW target setting process
The economic level of NRW target setting flow chart can be seen in Figure 8.
Figure 8 : Economic level of NRW target setting approach
50
The first step is to decide a target area having at least 5 year operating data of
DMAs, due to the acquisition of minimum available data. Next, the current level of
NRW and policy minimum need to be estimated which is referred to in the above
chapter (See section 2.2.2). The UARL would be employed instead of minimum
achievable background leakage. At the same time, costs data should be collected
from the company information system. After this process, the NRW control cost
curve would be developed by using two approaches respectively (Ofwat, 2002).
After that, it is possible to determine the most reliable approach for this study by
comparing derived cost curves fitting. The method showing highest coefficient of
determination R2 is used for economic level of NRW calculation. The other method
having low coefficient of determination R2 will not be considered in this thesis.
Then, the economic NRW level is estimated by using the selected method.
Through this process, decision makers can identify economic level and decide a
project target, by considering these financial constraints, social and political
request.
3.3. Economic level of NRW Calculation
3.3.1 Deciding Target Area
Ofwat (2007) recommended that the analysis period for calculating ELL should be
a minimum of five to ten years. This means the target area should have operation
data during the same period. These data must be automatically and periodically
collected by the supervisory control and data acquisition (SCADA) system. Taking
into account the above explanation, the current 12 of 20 projects operated by K-
water can be selected for target area as only 12 has established DMAs and has
been monitored for at least 5 years after the installation of DMAs. In order to
decide a suitable target area, the following various factors are comprehensively
considered: 1) the scale of water supply, 2) presence of own water resource and
bulk water imports, 3) completion of DMA installation and water pipe replacement,
and finally 4) sufficient operating data. Having accounted for these factors, a
project is selected and analysed as a target area since it have managed operating
51
data systematically and have shown outstanding data accuracy due to the recent
establishment of a Geographic Information System (GIS).
3.3.2 Data collection
Munoz-Trochez (2012) stated that quality of data have a major impact on the
reliability of ELL calculation. All the data used in this research are collected based
on various statistics and information systems. The sources of data are shown in
Table 6 and details about required data can be found in Appendix A.
Table 6 : Source of collected data for ELL calculation

Sources Managing Main contents
body
Water Country, Local Water supply status, water facilities information,
statistics government revenue water rate etc.
Annual report K-water Water supply status, water facilities information,
operation data etc.
Information K-water NRW rate, accounting, customer service, water
system pipe network operation system etc.
Technical K-water Annual investment plan of development of DMA
Report and water distribution system maintenance
Future Local Water facility expansion plan including pumping
maintenance Government and booster stations, water treatment, service
strategy reservoirs and water pipe expansion
3.3.2.1 Policy minimum
Policy minimum data consists of the number of connections and properties, the
length of water mains and communication pipes with average pressure of each
DMA, and the performance indicator values. Connection and property information
can be collected from the Customer Information System. Length of water pipes and
pressure values are collected by DMA units and those values are recorded at one-
minute intervals from the SCADA system. Lastly, performance indicator values can
be assessed through various calculations, such as CARL, the length of the water
pipes, the number of connections and the average pressure. In this study, UARL
divided by connection numbers, will be used as a policy minimum.
52
3.3.2.2 Current NRW
NRW data, estimated by total integrated flow analysis, can be obtained from
annual and monthly reports that are submitted to the general work system by local
office managers of waterworks systems. This consists of two types of data
according to the size of the target area. The two forms of data collection can be
seen in Appendix A.
3.3.2.3 Costs data
There are largely two types of costs data, NRW control costs and cost of lost water.
The NRW control costs are classified as expenses such as pipe replacement,
installing DMAs, valves and pressure equipment, repair costs and leakage
detection. These data are recorded on an individual basis by K-water accounting
information system and can be collected in a simple and easy manner. Specific
contents are described in Appendix A.
The cost of lost water is a function of marginal costs (i.e. production, distribution
and capital investment associated with future upgrades to meet the increasing
demand) and level of NRW. The more NRW increases, the cost of lost water is
higher. On the other hand, the cost of lost water can be regarded as a benefit due
to the amount of water saving as a result of NRW reduction by the water utilities.
That is, the benefit is the reduced costs associated with the volume of water
reduced by NRW control. Therefore, the costs of production and distribution, such
as bulk water purchase costs, pumping or boosting costs, and water and sludge
treatment costs, usage charges of dam water, expansion of pipe diameter and
construction of reservoir, can be an important component the in economic level of
NRW calculation. These data will be linked to the benefit calculation, can be
collected through national and local government statistics, annual reports and
various information systems.
3.3.3 Policy minimum
The Policy minimum is the lowest achievable level of leakage which is calculated
53
through the night flow analysis and it is called background leakage (Ofwat, 2002).
Moreover, it is impossible to estimate this under the South Korea data
management system because there is no standard for estimating night use. Even
though it is possible to estimate with the UK’s average night use values, it is
necessary to verify the UK’s night use value by comparing the result of total
integrated approach. Mckenzie (1999) presented a simple calculation process,
which uses default loss parameters derived from night flow analysis. This makes it
possible to calculate background leakage very quickly. However, when considering
the different water supply environment and status of infrastructure, the reliability of
value needs to be checked. Therefore, as an alternative, the UARL indicator will be
employed as a policy minimum in Chapter 4 (Beal et al., 2012).
3.3.4 Current NRW level
NRW is defined as the difference between the system input volume and the
amount of authorised consumption (Pilcher, 2003). The system input volume is
managed by water suppliers in real-time. It is easy to acquire the volume of system
input data. When it comes to authorized consumption, water meter reading is
carried out at every month and the accurate volume of authorized consumption is
estimated. This NRW can be calculated by IWA standard international water
balance (described in section 2.2.1)
3.3.5 Economic level of NRW calculation methods
Once current NRW and policy minimum are estimated, economic level of NRW can
be calculated by analysing relationship between costs and benefits. In this thesis,
two types of methods are employed to estimate the optimal level of NRW by
comparing costs and benefits; (1) Marginal cost analysis and (2) Cumulative costs-
benefits analysis. The first method is generally used in the UK (UKWIR, 1994). The
second method is the newly developed method specifically suited for South Korean
situation. All methods will be applied to a selected case study area. In order to
select the most appropriate methodology for the South Korea business
environment, a comparative analysis based on the reliability of each cost curve will
54
be performed. After that, the economic level of NRW will be estimated by the
preferred method.
3.3.5.1 Method 1: Marginal cost approach
This approach finds an intersection point between the marginal (unit) cost of NRW
control and the marginal cost of water. Both the marginal (unit) cost of NRW control
and the marginal cost of water curve can be drawn, as in Figure. 9. In this graph,
the intersection point of both curves is the most economic NRW level (UKWIR,
1994) .
Figure 9 : Marginal cost curve
1) Marginal cost of NRW control
The marginal (unit) cost of NRW control means additional cost for further 1m³ of
NRW reduction. This cost can be calculated by dividing the annual volume of NRW
reduction over the previous year into the annual cost of NRW control, which
consists of various activities such as water pipe replacement/rehabilitation,
pressure management, water meter replacement, and leakage detection/ repair
(UKWIR, 1994).
55
2) Marginal cost of water
The marginal cost of water can be estimated by adding marginal operating costs,
marginal capital costs, and social and environmental costs (UKWIR, 1994).
However, the social and environmental costs will not be included in this research
due to the difficulty and uncertainty in exact calculation. The marginal operating
costs are based on production and distribution costs such as power, chemical, bulk
purchase, and abstraction. The marginal capital costs can be affected by NRW
reduction. A reduction in the level of NRW may allow to change the size of a
project or to postpone its plan. These two costs are explained as follows.
(1) Marginal operating costs can be calculated by dividing annual variable costs
such as bulk water purchase, pumping and pump boosting, water treatment,
sludge treatment and usage charges of dam water into total annual volume of
water provided into the distribution network (UKWIR, 1994). This can be estimated
by Equation (3.1).
Marginal operating costs =
+ (3.1)
(2) Marginal capital costs occur when capital expenditure is decreased or

postponed due to the NRW reduction. Therefore, all the items which can be
affected by NRW reduction such as water resource works, treatment works,
pumping and boosting stations, service reservoirs and distribution main
replacement and rehabilitation should be considered (UKWIR, 1994).
UKWIR (1994) introduced two methods: The method 1 is dividing the present value
of the planned investment programme by the present value of the growth in water
demand over the same period. The present value of the growth in water demand is
difference between demand considering increase rate in water demand of the
56
future and demand without increase rate. The present value of demand related
investment is changed future capital cost by applying discount rate to planned
investment according to the future maintenance strategy. The method 2 is dividing
the difference in the present value of investment expenditures by the present value
of the demand reduction. The difference between method 1 and method 2
considers the changes of planned investment projects according to the leakage
reduction.
It is important to note that both methods can be adapted under the assumption that
future demand increases during the project period. Since expecting the exact
timing of capital expenditure is not easy under the current K-water business
environment and data management system, marginal capital costs would be
calculated by method 1. If the demand decreases or maintains constant level, the
marginal capital costs will not be adapted. In this study, marginal capital costs can
be calculate with the following equation (3.2)
Marginal capital costs
= (3.2)
3.3.5.2 Method 2: Cumulative cost-benefit analysis approach
Unlike the previous method, the cumulative cost-benefit method is a newly

developed approach taking into account the particular operating conditions in
South Korea. The economic level of NRW can be identified by analysing the
relationship between the cumulative cost of NRW control and the cumulative
benefit of NRW reduction. Based on the data collected from the water supplier
billing system, the cumulative costs of NRW control are estimated. The cumulative
benefits are represented as the aggregated value of annual benefits of NRW
reduction. The annual benefits of NRW reduction are calculated by multiplying both
volume of NRW reduction over previous and marginal cost of water. With this data,
the cumulative benefits of NRW reduction can be estimated in the same way as the
cumulative costs of NRW reduction. The cumulative cost curve can be identified by
57
using both data worked out through the above process. The cumulative cost and
benefit curves are illustrated by two forms in Figures 10 and 11 against the
cumulative volume of NRW reduction and the annual level of NRW per connection,
respectively.
In Figure 10, the cumulative benefit curve (the red line) shows that the cumulative
benefit increases over time due to the continuous NRW reduction. However, the
rate of increase diminishes over time. On the contrary, the slope of cost curve, (the
blue line), shows an initial linear increase before growing exponentially beyond the
economic point. Once the graph is developed, the intersection point of both curves
is the most economically optimal NRW level. Though Figure 10 suggests an
optimum cumulative volume of NRW, it has a disadvantage in indicating the best
NRW level represented by “m3/connection/year”. Therefore, the cumulative cost-
NRW level curve is used in conjunction with Figure 10 to help calculate the optimal
NRW level. This is illustrated in Figure 11.
Cumulative cost of NRW control

NRW control/reduction (￡)
Cumulative cost/benefit of
Cumulative benefit of NRW reduction
Economic level of NRW
Cumulative volume of NRW reduction (m3)
Figure 10 : Cumulative cost-benefit curve A
58
NRW control/reduction (￡)

Cumulative cost of NRW control

2013
2011 2010
2009
2012
2008
Economic level of NRW
Level of NRW (m3/connection/year)
Figure 11 : Cumulative cost-benefit curve B
In Figure 11, when NRW level moves from high(right) to low(left), the cumulative
benefit of NRW reduction increases only slightly because of the difficulty
associated with reducing NRW to a low level. In other words, the cumulative
benefit of NRW reduction increases over time due to the continuous NRW
reduction. However, the rate of increase diminishes over time. On the contrary, the
cumulative cost of NRW control rises rapidly. Due to all these reasons leakage
reduction activities follow the law of diminishing returns.
3.4. Economic Level of NRW target setting
The long-term goal of the South Korean government, in line with public opinion, is
to achieve 20% NRW. Most of the NRW reduction projects, which have promoted
by K-water, are designed to achieve that level. Under the current status, though
estimated NRW from this study is the most economically appropriate target, it is
difficult to modify the current NRW target, in a short time. Therefore, in this study,
the estimated economic NRW level through the discussed approaches would be
suggested as an optimal target by comparing both current and historical NRW
levels.
59
3.5. Sensitivity analysis
The economic NRW level is determined using a variety of variables that are subject
to uncertainty. The representative variables applied to this study are volume of
NRW, NRW control costs, marginal cost of water, and number of connections. The
sensitivity analysis will be performed to investigate which factor has the most
influence on and how a change in the parameter causes a change in economic
level of NRW value. The sensitivity will be tested by applying +/-5%, +/-10%, and
+/-15% to each variable.
3.6. Summary
The Economic Level of Leakage calculation model developed in the UK has

provided water suppliers with economically useful information for the operation and
maintenance of WDSs. This chapter has presented the economic level of NRW
calculation methodology with detailed reference to the UK model. In this chapter,
two methods for calculating the Economic level of NRW were discussed: (1)
Marginal cost analysis and (2) Cumulative costs-benefits analysis.
The first approach is well known method in the UK. The second Cumulative cost-
benefits approach is newly developed method in the economic level of NRW
calculation. It uses both cumulative costs and benefits for the economic level of
NRW calculation. This is advantageous for minimizing data fluctuations resulting
from an application to the project having a short period of operating data.
In Chapter 4, two methods will be applied to a case study involving K-water

projects. In order to select the most appropriate methodology for South Korea
business environment, a comparative analysis of the reliability of each cost curve
will be made. After that, the economic level of NRW will be estimated by the
chosen methodology.
60
Chapter 4 Case Study
CHAPTER 4 CASE STUDY
4.1. Introduction
In this chapter, two methods for estimating economic level of NRW are applied to a
K-water case study. The project area has just completed infrastructure
improvement such as DMA instalment, water pipe replacement and rehabilitation.
The case study presented in this chapter employed empirical data collected from
the project management system. The objective of this chapter is to find an optimal
NRW level based on the real data. Then, through comparison with the current
project target, a reasonable NRW target will be decided.
This chapter is outlined as follows. After this introduction, Sections 4.2 and 4.3
present how the economic NRW level is estimated with the most appropriate
method. Then a sensitivity check in section 4.4 was made for identifying the most
influential factors affecting optimal NRW levels. Finally, Section 4.5 includes the
chapter summary and presents conclusions.
4.2. CASE 1: NRW Reducing Stage (Danyang-gun)
4.2.1 Description of study Area
The Danyang-gun water system has been operated by K-water since they were
contracted to operate and manage the system in 2008. One of the main goals is to
achieve 20% NRW rate by 2014 starting from 52% NRW in 2008. In 2013, the
recorded NRW rate was recorded 21%.
Danyang-gun is a small city which is located in the north-east of South Korea

covering a total area of 780.65 km². The location of the case study area is shown in
the Figure 12.
By the end of 2013, out of a total population of 31,390, 22,433 people, (71.5%) use
treated water from the K-water. The remainder of the population is using a small-
scale water supply system based on ground water (Ministry of Environment, 2013).
61
Three water treatment facilities have been providing water to this case study area.
The average volume of this supplied water was 8,149m3/day. Recently, it has been
increasing due to the attraction of businesses as well as tourists.
Figure 12 : Location of Danyang-gun in South Korea
62
The Danyang-gun DMAs were allocated by region. The location and name of DMA
are described in Figure 13.
The rest of the Danyang-gun area, except for Danyangeup and Maepoeup is small
and geographically far apart. The status of each DMA is explained in Table 7.
Figure 13 : Location of Danyang-gun and DMAs
63
Table 7 : Year 2013 statuses of connection, property and population

Large Medium Small Number of Average Water pipe length (Km)
DMA DMA DMA connections water
Total Mains Communication
(-) use
pipe
(m3/d)
Total 6,414 8,149 348.5 255.4 93.1
DY DY DY1 1,456 2013 48.8 35.6 13.2
DY2 587 725 21.1 16.8 4.3

DY3 65 973 3.0 2.6 0.4
DY4 1,073 850 48.0 30.5 17.5
MP MP1 41 60 6.1 4.7 1.4

MP2 1,041 984 27.0 17.9 9.1
MP3 373 961 37.8 31 6.8
RC RC1 169 77 14.5 9.8 4.7

RC2 554 899 50.8 40.9 9.9
RC3 351 154 34.4 25.1 9.3
YC YC YC1 552 351 41.7 29.8 11.9

YC2 3 34
JS JS JS 149 68 15.3 10.7 4.6
64
4.2.2 Data collection and analysis
4.2.2.1 Connection, property and population
Connections, property and population were obtained from the K-water billing
system and Danyang-gun statistical yearbooks. Table 8 shows that they have been
increasing steadily due to the invigoration of tourist industry.
Table 8 : Annual status of connection, property and population

Year 2008 2009 2010 2011 2012 2013
Connections 5,315 5,376 5,454 5,778 6,089 6,107
Property 13,378 13,665 14,037 14,080 14,100 14,268
Population 20,017 20,331 21,269 22,210 22,241 22,433
4.2.2.2 Distribution network and pressure
The length of pipes obtained from GIS and statistical yearbooks is summarized in
Table 9. There seems to be an error in water mains and communication pipe
length in between 2010 and 2011. A large number of pipe information was
corrected in 2011 through the total inspection while the local office was conducting
a research about the WDSs maintenance plan. There was no average pressure
data about the whole area but average pressure of each DMA has operated
between 35metres to 45metres, hence, the average pressure was assumed to be
40 metres
Table 9 : Annual status of distribution network

Year 2008 2009 2010 2011 2012 2013
Total pipe length (km) 201 216 220 271 289 348
Water main (km) 93 103 107 200 216 255
Communication pipe (km) 108 113 113 71 73 93
65
4.2.2.3 Water loss performance indicators
Three types of performance indicators are calculated from Tables 8, 9, 11 and the
results are presented in Table 10. One of performance indicators, ILI was
estimated by dividing CARL into UARL. For example, CARL in 2008 was 2,259,146
m3/year and UARL was 88,460 m3/year. Both figures are shown in Table 11 and 12.
Therefore, the estimated value is 25.54. As it can be seen from the latter table, this
case study area has significantly reduced real losses due to intensive infrastructure
investment in the time period shown.
Table 10 : Annual status of water loss performance indicators

Water loss PI
2008 2009 2010 2011 2012 2013
Name / Year
m3/km/year 24,292 11,504 8,271 3,886 3,461 2,502
m3/km/day 66.6 31.5 22.7 10.6 9.5 6.9
m3/km/hour 2.8 1.3 0.9 0.4 0.4 0.3
m3/connection/year 425 220 162 134 123 104
m3/connection/day 1.16 0.6 0.44 0.37 0.34 0.29
m3/connection/hr 0.05 0.03 0.02 0.02 0.01 0.01
l/connection/hr 48.3 25 18.3 15.4 14.2 12.1
(ILI) 25.54 12.9 9.43 6.36 5.75 4.54
4.2.3 Current NRW level
The annual water balance of Danyang-gun between 2008 and 2013 is presented in
Table 11. The table shows that the volume of NRW has fallen by more than a third
from 2,823,933m3/year in 2008 to 813,111m3/year in 2013 due to the intensive
operational and capital investments for NRW reduction. The CARL values were
estimated by deducting apparent losses from water losses. Table 11 also shows
that 109,713 m3 has decreased between 2012 and 2013.
66
Table 11 : Annual water balance
Total Volume ( m3/year)

Component
2008* 2009* 2010 2011 2012 2013
System Input Volume 5,524,642 4,122,296 3,994,187 3,695,966 3,862,353 3,868,393
1. Revenue Water 2,700,709 2,641,109 2,850,823 2,733,323 2,937,848 3,055,282
- Build Authorized Consumption(a+b) - - 2,850,823 2,733,323 2,937,848 3,055,282
(a) Billed Metered Consumption - - 2,850,823 2,733,323 2,936,561 3,055,282
(b) Billed Unmetered Consumption - - - - 1,287 -
2. Non-Revenue Water(NRW) 2,823,933 1,481,187 1,143,364 962,643 924,505 813,111
- Unbilled Authorized Consumption(c+d) - - 78,637 19,186 3,022 1,071
(c) Unbilled Metered Consumption - - - - - 49
(d) Unbilled Unmetered Consumption** - - 78,637 19,186 3,022 1,022
3. Water Losses - - 1,064,727 943,457 921,483 812,040
- Apparent Losses (e+f) - - 179,741 166,317 173,806 174,076
(e) Unauthorized Consumption - - - - -
(f) Customer Metering Inaccuracies*** - - 179,741 166,317 173,806 174,076
- Current Annual Real Losses 2,259,146 1,184,950 884,986 777,140 747,677 637,964
* Notes: there were no data about specific components in 2008 and 2009.
** Unbilled unmetered consumption is the sum of items such as fire fighting, flushing of mains and sewers, street
cleaning, and frost protection.
*** Customer metering inaccuracies are estimated to be 4.5% of the System Input Volume according to the guidance
of annual water balance analysis in South Korea.
67
4.2.4 Policy minimum
The policy minimum is the lowest achievable theoretical level of leakage through
the active leakage control. It can be derived from minimum night flow analysis
(Utilities, 2010). However, because of inaccuracy with the customer night use
measurement, it is impossible to calculate policy minimum exactly. Alternatively,
the UARL is used as a policy minimum and it was calculated by using equation (2-
1). The results are shown in Table 12.
ILI = CARL / UARL (2.1)
UARL (litres/day) = (18 x Lm + 0.8 x Nc + 25 x Lp) x P

- where Lm = mains length (km); Nc = number of service connections;
- Lp = total length of private pipe, property boundary to customer meter (km);
- P = average pressure (metres).
Table 12 : Unavoidable annual real losses (UARL)

Item 2008 2009 2010 2011 2012 2013
Water main(km) 93 103 107 200 216 255
Connections 5,315 5,376 5,454 5,778 6,089 6,107
length of private pipe(km) 5.32 5.38 5.45 5.78 6.09 6.11
Average pressure(metres) 40 40 40 40 40 40
UARL(m3/year) 88,460 91,822 93,813 122,156 130,107 144,271
UARL(m3/connection/year) 17 17 17 21 21 23
In Table 12, the UARL is increasing steadily year on year after the first three years
where it was static. This is because the UARL is affected by the increase in various
parameter values (i.e., water mains lengths, number of connections, private pipe
lengths and average pressure). Especially, there was a rapid growth in the length
of water mains and connections from 2011 to 2013. The other significant cause is
that pipe properties were redefined in 2011 through total inspection. A large
number of pipes that were classed as communication pipes were changed into
water mains.
68
4.2.5 Economic level of NRW calculation
4.2.5.1 NRW cost curve
A cost curve is a key factor in the calculation of the economic level of NRW. This is
because it enables water suppliers to predict future NRW control costs. A derived
cost curve is subject to large uncertainty if there is not enough reliable data. In this
section, through the comparison of each cost curve, the most suitable approach
between the two will be selected and the most economically efficient level of NRW
will be calculated using the chosen method.
1) Marginal cost of NRW control curve (Method 1)
The marginal costs of NRW control was calculated by dividing the annual costs of
NRW control into the changes in NRW. The specific annual costs of NRW control
are shown in Table 16 and the changes in NRW are explained in Appendix B. The
marginal costs of NRW control curve can be drawn by using both the marginal
costs of NRW control and the NRW per connections. Both the estimated values are
presented in Table 13 and the marginal cost of NRW control curve is illustrated in
Figure 14.
Table 13 : Marginal cost of NRW control
Annual costs of Marginal costs

End of Changes in
NRW control of NRW control NRW/connections
Year year NRW NRW(A)
(B) (B/A)
m3/year m3 ￡￡/ m3 m3/connection/year
2008 2,823,933 - 125,500 - 531
2009 1,481,187 1,355,325 412000 0.3 276
2010 1,143,364 354,816 886,500 2.5 210
2011 962,643 234,522 314,000 1.3 167
2012 924,505 86,670 653,000 7.5 152
2013 813,111 152,225 1,142,500 7.5 127
69
Marginal cost of NRW control curve

NRW control (￡/㎥) 10
Marginal cost of
9
8
7 R² = 0.7786
6
5
4
3
2
1
-
- 50 100 150 200 250 300
NRW level per connection(㎥/connection/year)
Figure 14 : Marginal cost of NRW control curve
2) Cumulative cost curve (Method 2)
The cumulative cost curve can be drawn into two ways according to the different X-
axis values. X-axis of the first graph used the cumulative volume of the NRW
reduction, and alternatively, the level of NRW per connection was employed.
Commonly, cumulative cost of NRW reduction was used as a Y-axis. The
components of cumulative cost curve are presented in Table 14. The Changes in
NRW was estimated according to the Appendix B and the cumulative volume of
NRW reduction was calculated by adding changes in NRW year by year. The
NRW/connection was estimated by dividing annual NRW into number of
connections. The cumulative costs of NRW control is shown in table 15. The
cumulative cost curve A and B are illustrated in Figure 15 and 16.
70
Table 14 : Cumulative cost curve

Cumulative
Cumulative
Volume of
Year Changes in NRW NRW/connection costs of
NRW
NRW control
reduction
m3 m3 m3/connection/year ￡1,000,000
2008 - - 531 0.126
2009 1,355,325 1,355,325 276 0.538
2010 354,816 1,710,141 210 1.424
2011 234,522 1,944,663 167 1.738
2012 86,670 2,031,333 152 2.391
2013 152,225 2,183,558 127 3.534
Cumulative NRW control cost curve A

5
Cumulative costs of NRW control (￡m)
1
R² = 0.9737
-
- 500,000 1,000,000 1,500,000 2,000,000 2,500,000
Cumulative volume of NRW reduction(㎥)
Figure 15 : Cumulative cost curve A
71
Cumulative NRW control cost curve B

5
Cumulative costs of NRW control (￡m)
1
R² = 0.9619
-
- 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
Figure 16 : Cumulative cost curve B
72
4.2.5.2 Reliability check
Among the two cost curves, the cumulative cost curve showed the best fit to data.
Therefore, in this case study, the most economical level of NRW is estimated by
the cumulative cost-benefit analysis (method 2).
4.2.5.3 Economic level of NRW calculation (Method 2)
1) Costs data
The NRW control costs in Danyang-gun between 2008 and 2013 are presented in
Table 15. The Tabe 15 shows the annual costs for NRW reduction. Danyang-gun
has tried to reduce NRW by using various methods which are shown in Table 15.
Especially, Danyang-gun has focused on replacing water pipe and rehabilitation.
Table 15 : Cost data

Year 2008 2009 2010 2011 2012 2013
Cumulative
125,500 537,500 1,424,000 1,738,000 2,391,000 3,533,500
Total costs (￡)
Cost of NRW
125,500 412,000 886,500 314,000 653,000 1,142,500
reduction (￡)
Pipe replacement
1,000 152,500 400,500 77,500 460,000 939,500
and rehabilitation
Old and faulty
valve - 17,000 8,500 - 2,500 12,500
replacement
Establishment of
- - 228,500 30,000 2,000 -
DMAs
Water meter
33,500 55,500 49,500 26,500 8,000 14,500
replacement
Leakage Repair 30,500 100,000 106,500 90,500 88,000 79,000
Leakage
60,500 87,000 93,000 89,500 92,500 97,000
detection
73
2) Benefit data
The cumulative benefit of NRW reduction can be calculated from multiplying both
the cumulative volume of NRW reduction and the marginal cost of water. The
cumulative benefit of NRW reduction was expressed in Table 16. The marginal
cost of water calculation is described successively.
Table 16 : Cumulative benefit of NRW reduction

Year Cumulative Volume Marginal cost of Cumulative benefit of
of NRW reduction water NRW reduction
(A) (B) (A x B)
m3 (￡) (￡1,000)
2008 - - -
2009 1,355,325 1.51 2,045
2010 1,710,141 1.51 2,581
2011 1,944,663 1.51 2,935
2012 2,031,333 1.51 3,066
2013 2,147,024 1.51 3,240
(1) Marginal cost of water
The marginal cost of water can be estimated as the sum of marginal operating
costs plus marginal capital costs. It has been calculated using the 2013 data. The
marginal cost of water was estimated to be ￡1.51/m3 through the following
calculation.
① Marginal operating costs
The components of marginal operating costs calculation is presented in Table 17.

The marginal operating costs has been calculated through the Equation (3-1)
described in section 3.3.5.2 and its costs were ￡0.08/ m3.
74
Table 17 : Components of marginal operating costs in 2013

Components Unit Value
A Annual volume of production m3/year 3,868,393
B Annual volume of dam water m3/year 525,915
C Annual volume of purchased bulk water m3/year -
D Pumping or Boosting costs ￡ 174,782
E Water treatment costs ￡ 20,473
F Sludge treatment costs ￡ 3,187
G Usage charges of Dam water ￡ 13,227
H Bulk water purchase costs ￡ -
Marginal operating costs = + + = ￡0.08/ m3 (4-1)
② Marginal capital costs
The marginal capital costs are calculated by dividing the sum of the present value
of demand related investment into the present value of the growth in water demand
over the time period. The discount rate for converting future investment into the
present value of investment was assumed to be 5.2% according to the social
discount rate (5.2~6.5%) of the Korea Development Institute report published in
South Korea. The estimated value, (￡9,782,499 and volume 6,827,836 m3),
through the following successive sections were used and its costs were estimated
to be 1.43￡/ m3 through Equation (3-2), as described in section 3.3.5.2.
(a) Sum of present value of demand related investment

=
(b) Present value of the growth in water demand
9,782,499
=￡1.43/ m3
6,827,836 (4-2)
75
(a) Present value of demand related investment
Demand related to future investment and the present value of demand related
future investment is given in Tables 18 and 19. The present value of investment,
￡9,782,499, in table 19 was used as a denominator in marginal capital costs
calculation.
Table 18 : Demand related future investment

Items Total 2015 2020 2025
Total(￡) 14,158,000 3,423,500 5,392,000 5,342,500
Reservoirs(￡) 94,500 - - 94,500
Expansion of water main(￡) 12,797,000 3,181,000 4,953,000 4,663,000
Boosting station(￡) 369,000 20,500 100,000 248,500
Design costs(￡) 709,000 175,000 268,000 266,000
Supervision charge(￡) 188,500 47,000 71,000 70,500
Table 19 : Present value of demand related investment

- Total 2015 2020 2025
Total investment(￡) 14,158,000 3,423,500 5,392,000 5,342,500
Present value of investment(￡) 9,782,499 3,093,420 3,781,288 2,907,741
(b) Present value of the growth in water demand
The present value of the growth in demand and the present value of no growth in
demand are displayed in Table 20 and 21.
76
Table 20 : Present value of the growth in demand

Item Total 2013 2014 2015 2016 2017 2018
Demand(m3/d) 284,023 19,705 20,065 20,425 20,858 21,291 21,725
Demand(m3/year) 103,668,395 7,192,325 7,323,725 7,455,125 7,613,243 7,771,361 7,929,479
Present value of Demand(m3/year) 77,054,542 7,192,325 6,961,716 6,736,331 6,539,168 6,345,036 6,154,120
Item 2019 2020 2021 2022 2023 2024 2025
Demand(m3/d) 22,158 22,591 22,741 22,891 23,041 23,191 23,341
Demand(m3/year) 8,087,597 8,245,715 8,300,465 8,355,215 8,409,965 8,464,715 8,519,465
Table 21 : Present value of no growth in demand

Item Total 2013 2014 2015 2016 2017 2018
Demand(m3/d) 256,165 19,705 19,705 19,705 19,705 19,705 19,705
Demand(m3/year) 93,500,225 7,192,325 7,192,325 7,192,325 7,192,325 7,192,325 7,192,325
Item 2019 2020 2021 2022 2023 2024 2025
Demand(m3/d) 19,705 19,705 19,705 19,705 19,705 19,705 19,705
Demand(m3/year) 7,192,325 7,192,325 7,192,325 7,192,325 7,192,325 7,192,325 7,192,325
77
The present value of the growth in water demand over the time period can be
calculated by deducting the present value of no growth in demand (70,226,706
m3/year) from the present value of demand, (77,054,542 m3/year). The difference
between these two values, (6,827,836 m3/year), was used as the numerator.
3) Cumulative cost-benefit curve
The economic level of NRW can be identified by adding cumulative benefit of NRW
reduction curve which comes from table 15 and 16 into the cumulative cost curve.
These data used in this analysis are real operational data collected and stored by
K-Water. The two type of graph are illustrated in Figure 17 and 18.
Cumulative cost-benefit curve A

4
NRW control/reduction (￡1,000,000)
2
Cumulate benefit of NRW reduction
Cumulate cost of NRW control Cumulative economic
1 volume of
NRW = 2.2M㎥
-
- 500,000 1,000,000 1,500,000 2,000,000 2,500,000
Cumulative volume of NRW reduction(㎥)
Figure 17 : Cumulative cost-benefit curve A
78
Cumulative cost-benefit curve B

4
Cumulative cost/benefit of NRW
control/reduction (￡M)
Current NRW Cumulative cost of NRW control

3
=127 ㎥/connection/year
R² = 0.9746
Policy minimum
1 R² = 0.9619
Economic NRW
-
- 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300
Figure 18 : Cumulative cost-benefit curve B
4) Optimal NRW level
The optimal NRW level can be identified by finding a point where the two curves
meet. The calculated economic cumulative volume of NRW and the economic level
of NRW are is 2.2 Mm3/total and 132m3/connection/year respectively. In this
research, 132 m3/connection/year was used for convenience.
79
4.3. Sensitivity analysis
Sensitivity analysis has been carried out to examine which factors would have the
most impact on economic level of NRW level and how far the economic NRW level
was changed. The results are summarized in Table 22 and shown in Figure 19.
Table 22 : Results of sensitivity test

Sensitivity Parameters Economic level of NRW (m3/connection/year)
Value change -15% -10% -5% 0% 5% 10% 15%
Volume of NRW(m3/year) 145 140 136 132 128 125 121
Marginal cost of water(￡/ m3) 145 140 136 132 128 125 121
NRW control costs(￡M/year) 120 124 128 132 135 140 143
Connections 140 137 134 132 129 126 124
Figure 19 : Economic level of NRW sensitivity test
From the above results, both volume of NRW and the marginal cost of water were
reduced by as much as 4m3/connection/year at a steady rate when changes in
parameters are applied. The NRW control costs, meanwhile, were increased at a
80
similar rate to the volume of NRW and the marginal cost of water. The number of
connections had a minimal impact on the level of NRW. Even though there is no
significant governing factor affecting the economic level of NRW, this analysis
demonstrates that economic NRW target can be set within the calculated limits
actively or passively according to the financial conditions.
4.4 Setting NRW target
The optimal NRW levels were calculated by the cumulative cost-benefit analysis.
The optimal level is 132m3/connection/year. It should be noted the recent NRW
level was recorded at 127m3/connection/year by the end of 2013. It is illustrated
with red line in Figure 17. This means the study area has already achieved the
desirable NRW level. Therefore, it is recommended to maintain current NRW level
and simultaneously to set this level as the optimal NRW level for this study area.
4.5. Summary
This chapter has presented the economic level of NRW calculation by applying the
methodology developed in Chapter 3 to a case study. The cumulative cost-benefit
analysis was selected as an evaluation method by cost curve reliability check. The
estimated economic NRW level by cumulative cost-benefit analysis was 132
m3/conn/year, which is beyond the current NRW level, 127m3/connection/year.
Thus, the NRW target of this case study can be set with maintaining current level
but this needs to be reviewed when enough data is secured.
The sensitivity analysis attempted to identify the dominant factor and how far the
economic NRW level changed. The results obtained by sensitivity analysis showed
that all the parameters can affect to the economic NRW level to a similar extent,
approximately ±4m3/connection/year. Although it was impossible to identify the
most influential factor, both lower and upper limits of the economic NRW level were
determined
81
Chapter 5 Summary and Conclusion
CHAPTER 5 SUMMARY AND CONCLUSION
5.1 Thesis Summary
What is the economically acceptable NRW level? Why do all the K-water projects
have an identical NRW target in South Korea? This study has been conducted to
address these questions. NRW reduction projects, which are being promoted by K-
water, are to achieve 20% NRW within a short period (i.e. 5 or within 7 years).
Over the last 11 years, there has been no discussion about whether 20% NRW is
an economically reasonable and practical target, even though all the projects have
shown different efficiency in NRW reduction.
Thus, this study was aimed at evaluating the current NRW level by developing a
model for the calculation of the economic level of NRW and setting an economic
NRW target. With this background, the literature review was conducted on both
water loss management and Economic level of Leakage principles. Through the
literature review, two methodologies were proposed; (1) Marginal cost analysis,
and (2) Cumulative cost-benefit analysis. The first approach is commonly used
method in the UK. The second method is newly developed allowing for the K-water
business environment.
It should be noted that there was a difference in the way these methods have been
applied. Methods 1 has basically been applied under steady state conditions (the
state of maintaining NRW at a given level) using traditional active leakage control
methods (i.e. monitoring, detecting, locating and pinpointing for repair). However,
in this research, all the NRW control activities such as DMA installation, pipe
replacement, and pressure management were considered at the same time. In that
respect, a possibility of the application of those methods was examined through
cost curve comparison in Chapter 4. Meanwhile, the method 2 was intended for
integrating all the activities for NRW reduction. This is due to the fact that all the
strategies for NRW reduction need to be adopted simultaneously to achieve 20%
NRW within short period (i.e. 5years or 7years). This research was carried out
based on this background.
82
Through the comparison of cost curve fitting in the Chapter 4, it was possible to
determine the most suitable method. A case study area was determined to apply
the chosen methodology from Chapter 3. The area, Danyang-gun, is a small city in
South Korea with a population of approximately 30,000 and has managed its
operation data by relating it to the NRW reduction systematically since 2008. In this
research, the economic level of NRW was calculated through the cumulative cost-
benefit analysis.
In the next stages, a sensitivity analysis was carried out to identify the most
influential factor affecting the NRW level. This was followed by investigating, how
changes in each component value from -15% to 15% can affect NRW level. Four
main factors: volume of NRW, marginal cost of water, NRW control costs and
number of connections were tested.
Finally, the estimated economic NRW level was compared with the current level to
decide a reasonable target for the case study area. Since this study area has
already achieved the desirable NRW level, it was recommended to maintain the
current NRW level and simultaneously to set this level as the most optimal NRW
level of this study area.
5.2 Summary of contributions
Economic leakage management is common in the UK but it has not been

discussed nor investigated in South Korea until now. This study contributes to
developing a change in attitude about economic NRW management in South
Korea. The contributions of this study are described as follows:
 The newly developed methodology, cumulative cost-benefit analysis, enables

water suppliers to evaluate economic NRW or leakage level with high reliability.
Since this method uses cumulative values, it is less sensitive to data
fluctuations. This leads to an acquisition of reliable results. Especially, it is
useful to water systems that have a short data gathering period and rapidly
83
growing systems where the number of connected customers changes

substantially each year.
 This study suggests two types of calculation models for identifying economic
NRW level. It is possible to apply them to various projects that could have a
different business environment. Through the comparison with applied
approaches, the most appropriate method can be selected and it can suggest
the optimal NRW level. This allows water suppliers to select the most
appropriate method and to estimate the most reliable NRW level for their
system.
 Finally, this research contributes to managing water systems the most

economically and efficiently by preventing overinvestment and focusing on
optimal system operation. Ultimately, this would be beneficial to both water
suppliers and customer in terms of budget savings, restraint of water rates and
improvement of customer satisfaction.
5.3 Conclusions
The conclusions of this study are as follows:
Firstly, the Economic Level of NRW model developed and applied in this study
shows the applicability to the South Korean water system as illustrated by the case
study results. Because the new method uses cumulative data, data fluctuations
due to the data records have less of an effect on the results. On the other hand,
the UK model showed low reliability in cost prediction. The reason comes from the
different investment methods for NRW control in the UK and South Korea. Water
companies in the UK, because they have been managed economically and
optimally, have maintained low costs to control NRW. In contrast, in the same 5-
year period, South Korea has seen increased investment of about 40% of the
overall project management costs resulting in systems that now have NRW level
below the optimal one.
84
Secondly, the sensitivity analysis attempted to identify the dominant factor for NRW
and how far the economic NRW level changed over time. The results obtained by
sensitivity analysis showed that all the parameters can affect the economic NRW
level to a similar extent, approximately ±4m3/connection/year. Through the
sensitivity analysis, both lower and upper limits of the economic NRW level were
determined.
Thirdly, the comparison between the current project target and the calculated
economic NRW level was made through this research. In many cases, optimal
NRW level is lower than the current NRW level. However, the result of this study
shows a similar level of the NRW target to the currently achieved. Hence, the
relationship between the new model result and the current target needs to be
confirmed by applying the methodology to other projects.
Lastly, another potential issue identified in this research is the difference between
modelled economic NRW level and planned target of operation and management
(O&M) contract. When the calculated economic level of NRW is much higher than
the planned target or vice versa, a thoroughgoing review is needed whether current
contract has to be changed or maintained in respect of cost-effective. This was not
discussed in this study since this needs to be checked from an integrated point of
view such as political, economic, social, technical, legislative and environmental
factors.
5.4 Future work recommendations
The current study has attempted to apply a methodology commonly used in the UK
and develop a new methodology for the South Korea water system. Since a large
part of this study has been focused on evaluating the economic level of NRW,
other related questions could not be dealt with fully due to the limitation of time.
Specific recommendations for future research are described as follows:
1) It will be necessary to further test and validate the presented methods by

applying it to various systems and operation conditions.
85
2) A method for systematic NRW related data collection and processing

should be developed. This is because the calculation of economic level of
NRW is data intensive and time consuming work.
3) The economic level of NRW can be an operational performance indicator.

Although the economic NRW level was derived from the whole city level,
without dividing into each DMA, if specific data for economic NRW
calculation is available for small DMAs, it would be possible to use it as a
performance indicator. This needs to be checked in the future.
4) Estimation standard for external social and environmental costs of NRW

should be established. Recently, the averseness has increased that NRW
has social and environmental dimensions. Although the externalities cannot
greatly affect the economic level, this still needs to be considered.
5) In order to set up a long term NRW management plan, least cost planning
method should be considered. The NRW management in the least cost
planning is only a component for the supply/demand balance. This means
the NRW management plan is treated more widely than the current
marginal cost method. Furthermore, the least cost planning method
contains a variety of factors influencing the supply/demand.
6) It is required to develop a NRW trend prediction method. Due to the short

operation time, it was impossible to make a prediction of future trends in this
study. The NRW reduction projects promoted by K-water are conducted
under the fixed business time period. If a reasonable target or required time
to meet the target become accurately estimated, it will be possible to
allocate an exact budget and manage it to meet the target.
86
Appendix A
APPENDIX A
Data collection form
A.1 Components of economic level of NRW calculation

Table A- 1. Components of economic level of NRW calculation
Basic data NRW data
Number of DMAs Annual Total Integrated Flow
Number of Connections -result
System input volume
Number of Properties - Billed authorized consumption
Length of pipe - Unbilled authorized
Pressure -consumption
Apparent losses
Performance indicators
Data Costs data Benefits data

Collection Pipe rehabilitation, replacement Bulk water purchase costs
Pressure management Pumping or Boosting costs
Valve install/replacement Water and Sludge treatment
Water meter replacement costs charges of Dam water
Usage
Establishment of DMAs Expansion of pipe diameter
Own detection team operation Construction of reservoir
Leakage repair costs Construction of pumping station
All sorts of technical service Construction of water treatment
facilities
A.2 Basic data
Table A- 2. Data form of Connection number, Property and Population
Year 2008 2009 2010 2011 2012 2013
Number of Connections
Number of Property
Population
87
Appendix A
Table A- 3. Data form of distribution network information.

Year 2008 2009 2010 2011 2012 2013
Total pipe length (km)
Water main (km)
Communication pipe(km)
Pressure(kgf/cm²)
Table A- 4. Performance indicators

Name 2008 2009 2010 2011 2012 2013
m3/km/year
m3/km/day
m3/km/hour
m3/connection/year
m3/connection/day
m3/connection/hr
l/connection/hr
Current Annual Real Losses (m3/year)
Unavoidable Annual Real Losses (m3/year)
Infrastructure Leakage Index (ILI)
88
Appendix A
A.3 NRW data

Table A- 5. Form of Total Integrated flow (Company or Water Resource Zone level)
Component Volume(m3/year)
System Input Volume
1. Revenue Water
(1) Build Authorized Consumption
(a) Billed Metered Consumption(including water exported)
(b) Billed Unmetered Consumption
2. Non-Revenue Water(NRW)
(2) Unbilled Authorized Consumption
(c) Unbilled Metered Consumption
(d) Unbilled Unmetered Consumption
(3) Apparent Losses
(e) Unauthorized Consumption
(f) Customer Metering Inaccuracies*
(4) Real Losses
Table A- 6. Form of Total Integrated flow (Small DMA level)

Component Volume(m3/year)
System Input Volume
Revenue Water
Non-Revenue water
Revenue water rate (%)
Non-revenue water rate (%)
89
Appendix A
A.4 Costs data

Table A- 7. Direct and Indirect costs for NRW reduction
Type Activities Contents
Direct Pipe rehabilitation, Old and deteriorated water pipe
costs Replacement replacement
Small-scale pipe replacement for
leakage repair
Pressure management Pressure reducing valve instalment
Water meter replacement Water meter replacement
Detection team operating cost Salary, insurance, vehicle, fuels,
(including detection service) Equipment costs
Detection service costs
Leakage repair costs Leak repair costs
Indirect Valve instalment /replacement Old, non-working and leaking valve
costs replacement
Establishment of DMAs DMAs establishment costs
(including SCADA system costs)
All sorts of technical service Design and research service costs
A.4 Benefits data

Table A- 8. Benefit from NRW reduction
Type Cost items Contents
Operating Decrease in imported water from
Water purchase
Costs other water suppliers
Electric costs reduction for
Pumping and boosting abstracting, producing and
distributing water
Chemical materials reduction for
Water and sludge treatment water treatment and sludge
disposal
Decrease in dam water
Usage of dam water
abstraction charge
Capital Deferment or cancellation of
New resources development
Costs resource development
Deferment of pipe replacement or
Water pipe replacement
Quantity reduction
Deferment or cancellation of
Service reservoir reservoir construction due to
demand decrease
Deferment or cancellation of
Water pipe diameter enlargement
Water pipe enlargement
90
Appendix B
APPENDIX B
Changes in NRW
The “Changes in NRW” is a crucial factor not to be overlooked in the benefit

calculation. Since cumulative benefit of NRW reduction is estimated by multiplying
both the volume of NRW reduction over last year and the marginal cost of water. If
the total volume of NRW is increased compared to the year before, due to the rise
in connection numbers, the volume of NRW reduction, (compared to the previous
year), would have minus value. This is the case even if NRW per connection has
shrunk considerably because the rise in connections does not take this into
account. Therefore, it is necessary to present a possible way to take into account
increase in connection. Following Equation B-1 presents a possible way of
factoring in this increase in connection:
Changes in NRW = (1) × (2)

(1) Connection number of last year
(2) NRW per connection of last year –this year (B-1)
ex) Changes in NRW in 2009 = (NRW/connection in 2008 – NRW/connection in

2009) × connection numbers in 2008
= (531-276) × 5,315 = 1,355,325
Table B- 1. Changes in NRW
Year End of year Changes in Connection NRW/connection

NRW NRW numbers
m3/year m3 - m3/connection/year
2008 2,823,933 5,315 531
2009 1,481,187 1,355,325 5,376 276
2010 1,143,364 354,816 5,454 210
2011 962,643 234,522 5,778 167
2012 924,505 86,670 6,089 152
2013 813,111 152,225 6,414 127
91
Bibliography
BIBLIOGRAPHY
Paper presented by the candidate
LIM, E., SAVIC, D. & KAPELAN, Z. 2015. Development of a Leakage Target

Setting Approach for South Korea Based on Economic Level of Leakage. Procedia
Engineering, 119, 120-129.
92
References
ALEGRE, H. & ASSOCIATION, I. W. 2006. Performance indicators for water

supply services, IWA publishing.
BEAL, S., TROW, S. & HAYWOOD SMITH, B. 2012. Review of the calculation of
sustainable economic level of leakage and its integration with water
resource management planning.
CHO, H. 2013. Study for Assessing Economic Level of Leakage(ELL) and Setting
Management Target. 1-83.
ENVIRONMENT, M. O. 2012. Water Statistics 2011.
ENVIRONMENT, M. O. 2014. Water Statistics 2013.
FANNER, P., WORKS, W. & ASSOCIATION, A. W. W. 2007. Leakage
management technologies, Awwa Research Foundation.
FARLEY, M. & LIEMBERGER, R. 2005. Developing a non-revenue water reduction
strategy: planning and implementing the strategy. Water Science &
Technology: Water Supply, 5, 41-50.
FORCE, I. W. L. T. 2003. Practical Approach Initiatives to Water Loss Reduction.
Water 21.
FORCE, I. W. L. T. & THORNTON, J. 2003. Managing leakage by managing
pressure: a practical approach. Water 21.
FRAUENDORFER, R. & LIEMBERGER, R. 2010. The issues and challenges of
reducing non-revenue water.
HANDY, S. N., MANURUNG 2011. Bottom-up Water Balance Estimation
Applications and limitations.
ISLAM, M. S. & BABEL, M. S. 2012. Economic analysis of leakage in the Bangkok
water distribution system. Journal of Water Resources Planning and
Management, 139, 209-216.
K-WATER 2008. Business plan of operational efficiency project of Danyang-Gun
local water supplies
K-WATER 2013. 2013 NRW analysis report.
K-WATER 2014. Local Water Supply Services Handbook.
93
KOO, J., CHOI, T., PARK, S. & KIM, J. 2011. A Basic Study for Establishment of
Setting Objective for Leakage Management of Water Pipe Network. 223.
LAMBERT, A. 1994. Accounting for losses: The bursts and background concept.
Water and Environment Journal, 8, 205-214.
LAMBERT, A. 2003. Assessing non-revenue water and its components: a practical
approach. Water21, pp50-51
LAMBERT, A. & FANTOZZI, M. 2005. Recent advances in calculating economic
intervention frequency for active leakage control, and implications for
calculation of economic leakage levels. Water Science & Technology: Water
Supply, 5, 263-271.
LAMBERT, A., MERKS, C. & TROW, S. 2015. Good practices in leakage
management: lessons from Europe.
LEVIN, R. B., EPSTEIN, P. R., FORD, T. E., HARRINGTON, W., OLSON, E. &
REICHARD, E. G. 2002. US drinking water challenges in the twenty-first
century. Environmental Health Perspectives, 110, 43.
LIEMBERGER, R., BROTHERS, K., LAMBERT, A., MCKENZIE, R., RIZZO, A. &
WALDRON, T. Water loss performance indicators. International Conference
on the WaterLoss, IWA, Bucharest, 2007. 148-160.
MCKENZIE, A. O. L. A. D. R. D. 2002. Practical Experience in using the
Infrastructure Leakage Index.
MCKENZIE, R. 1999. Development of a standardized approach to evaluate burst
and background losses in potable water distribution systems: SANFLOW
User Guide. South African Water Research Commission, Report TT 109/99.
ISBN 1-86845490-8.
MCKENZIE, R. & LAMBERT, A. 2002. ECONOLEAK.
MCKENZIE, R. & WEGELIN, W. Scope for pressure management in South Africa.
Proceedings of the IWA International Specialised Conference ‘Water Loss
2010, 2010.
MUNOZ-TROCHEZ, C. 2012. Inclusion of energy externalities in the economic
level of leakage (ELL) model. © Camilo Munoz-Trochez.
OFWAT 2002. Best Practice Principles in the Economic Level of Leakage
Calculation. 158.
94
OFWAT 2007. Alternative approaches to leakage target setting.
OFWAT 2008. Providing Best Practice Guidance on the Inclusion of Externalities in
the ELL calculation.
PEARSON, D. 2002. Testing the UARL & ILI approach using a large UK Data Set.
IWA Conference ‘Leakage Management - A Pratical Approach’.
PEARSON, D. & TROW, S. Calculating the Economic Levels of Leakage.
Leakage 2005 Conference Proceedings, 2005.
PILCHER, R. 2003. Leak detection practices and techniques: a practical approach.
Water 21.
PUUST, R., KAPELAN, Z., SAVIC, D. & KOPPEL, T. 2010. A review of methods
for leakage management in pipe networks. Urban Water Journal, 7, 25-45.
SMOUT, I. K., KAYAGA, S. & MUNOZ-TROCHEZ, C. 2010. Adapting the
economic level of leakage concept to include carbon emissions, and
application with limited data.
STEPHENS, I. 2003. Regulating Economic Levels of Leakage in England and
Wales. Office of water.
THORNTON, J., STURM, R. & KUNKEL, G. 2008. Water loss control, McGraw Hill
Professional.
TROW, S. 2006. Economic Level of Leakage (ELL) - Approaches to Estimating
and Achieving ELL.
TROW, S. & FARLEY, M. 2004. Developing a strategy for leakage management in
water distribution systems. Water Science & Technology: Water Supply, 4,
149-168.
UKWIR 1994. Managing Leakage : Setting Economic Leakage Targets (Report C).
UTILITIES, U. 2010. Business Plan 2010-2015.
WEIMER, D. 2001. Water loss management and techniques. German National.
WU, Z. Y. 2011. Water loss reduction, Bentley Institute Press.
95
96

Development of A Leakage Target Setting Approach For South Korea Based On Economic Level of Leakage

Uploaded by

Document Informationclick to expand document information43098372

Document Informationclick to expand document information

Copyright:

Available Formats

Development of A Leakage Target Setting Approach For South Korea Based On Economic Level of Leakage

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Development of A Leakage Target Setting Approach For South Korea Based On Economic Level of Leakage

Uploaded by

Copyright:

Available Formats

College of Engineering, Mathematics, and Physical Sciences

Development of a Leakage Target Setting

to the University of Exeter as a thesis for the degree of

This thesis is available for Library use on the understanding that it is

Firstly, I would like to express the deepest gratitude to my two supervisors,

I would like to thank the water supply management department members,

Finally, I would like to express special thanks to my parents and parent-in-laws, my

Table of Contents .................................................................................................. 5

List of Tables ....................................................................................................... 10

Chapter 1 Introduction ........................................................................................ 13

1.1 Background .................................................................................................. 13

1.2 Aim and Objectives of Research .................................................................. 14

1.3 Thesis structure............................................................................................ 15

Chapter 2 Literature Review ............................................................................... 17

2.1. Introduction ................................................................................................. 17

2.2. Water loss management ............................................................................. 18

2.2.1 Water loss assessment .......................................................................... 19

2.2.2 Water loss Assessment methods ........................................................... 20

2.2.2.1 Total integrated flow analysis (Top-down method) ........................ 20

2.2.2.2 Minimum night flow analysis (Bottom-up method)......................... 21

2.2.2.3 Component analysis ..................................................................... 22

2.2.3 Performance Indicators .......................................................................... 23

2.2.4 Leakage target setting ........................................................................... 27

2.2.5 Leakage detection methods ................................................................... 28

2.2.5.1 Leakage localization methods ....................................................... 28

2.2.5.2 Leakage pinpointing methods ....................................................... 29

2.2.6 Water loss management methods ......................................................... 30

2.2.6.1 Active leakage control (ALC) ......................................................... 31

2.2.6.2 Pressure management .................................................................. 31

2.2.6.2.1 Fixed and Variable Area Discharge (FAVAD) concept ............ 32

2.2.6.2.2 Type of Pressure Management .............................................. 33

2.2.6.3 Speed and quality of repairs ......................................................... 34

2.2.6.4 Pipeline and assets management ................................................. 35

2.3. Economic Level of Leakage ........................................................................ 37

2.3.1 Definition of ELL..................................................................................... 37

2.3.2 Time frames of ELL................................................................................ 39

2.3.3 ELL target setting process and the components .................................... 39

2.3.3.1 ELL target setting process ............................................................ 39

2.3.4 Calculating Economic Level of Leakage ................................................ 40

2.3.4.1 Policy minimum ............................................................................. 40

2.3.4.2 Least cost planning approach (Option 1) ...................................... 41

2.3.4.3 Marginal cost approach (Option 2) ................................................ 42

2.3.5 Previous research on Economic Level of Leakage ................................ 46

2.4. Summary ..................................................................................................... 47

Chapter 3 Methodology For Assessing Economic Level of Non-Revenue

3.1. Introduction ................................................................................................. 49

3.2. Economic level of NRW target setting process ........................................... 50

3.3. Economic level of NRW Calculation ............................................................ 51

3.3.1 Deciding Target Area ............................................................................. 51

3.3.2 Data collection ....................................................................................... 52

3.3.2.1 Policy minimum ............................................................................. 52

3.3.2.2 Current NRW ................................................................................ 53

3.3.2.3 Costs data ..................................................................................... 53

3.3.3 Policy minimum ...................................................................................... 53

3.3.4 Current NRW level ................................................................................. 54

3.3.5 Economic level of NRW calculation methods ......................................... 54

3.3.5.1 Method 1: Marginal cost approach ................................................ 55