Water Supply Systems 1996

NATO ASI Series
Advanced Science Institutes Series

A series presenting the results of activities sponsored by the NA TO Science Committee,
which aims at the dissemination of advanced scientific and technological knowledge,
with a view to strengthening links between scientific communities.
The Series is published by an international board of publishers in conjunction with the

NATO Scientific Affairs Division
A Life Sciences Plenum Publishing Corporation

B Physics London and New York
C Mathematical and Physical Sciences Kluwer Academic Publishers
D Behavioural and Social Sciences Dordrecht, Boston and London
E Applied Sciences
F Computer and Systems Sciences Springer-Verlag
G Ecological Sciences Berlin Heidelberg New York
H Cell Biology London Paris Tokyo Hong Kong
I Global Environmental Change Barcelona Budapest
PARTNERSHIP SUB-SERIES
1. Disarmament Technologies Kluwer Academic Publishers
2. Environment Springer-Verlag/Kluwer Academic Publishers
3. High Technology Kluwer Academic Publishers
4. Science and Technology Policy Kluwer Academic Publishers
5. Computer Networking Kluwer Academic Publishers
The Partnership Sub-Series incorporates activities undertaken in collaboration with

NATO's Cooperation Partners, the countries of the CIS and Central and Eastern
Europe, in Priority Areas of concern to those countries.
NATo-peo DATABASE
The electronic index to the NATO ASI Series provides full bibliographical references
(with keywords and/or abstracts) to about 50000 contributions from international
scientists published in all sections of the NATO ASI Series. Access to the NATO-PCO
DATABASE compiled by the NATO Publication Coordination Office is possible in two
ways:
- via online FILE 128 (NATO-PCO DATABASE) hosted by ESRIN,
Via Galileo Galilei, 1-00044 Frascati, Italy.
- via CD-ROM "NATO Science & Technology Disk" with user-friendly retrieval software
in English, French and German (© WTV GmbH and DATAWARE Technologies Inc.
1992).
The CD-ROM can be ordered through any member of the Board of Publishers or
through NATO-PCO, Overijse, Belgium.
2. Environment - Vol. 15
Springer
Berlin
Heidelberg
New York
Barcelona
Budapest
Hong Kong
London
Milan
Paris
Santa Clara
Singapore
Tokyo
Water Supply Systems
New Technologies
Edited by
Cedo Maksimovi6
IRTCUW, Institute of Hydraulic Engineering
P.O. Box 895
11000 Belgrade, Yugoslavia
Francesco Calomino
Dipartimento di Difesa del Suolo
Universita della Calabria
1-87040 Montalto Uffugo (Cs), Italy
John Snoxell
Wessex Water, Wessex House
Passage street, B22 OJQ Bristol, UK
Springer
Published in cooperation with NATO Scientific Affairs Division
Proceedings of the NATO Advanced Study Institute "New Technologies
for Large Water Supply Projects", held in Varna, Bulgaria,
October 24 - November 4, 1994
Library of Congress Cataloging-in-Publication Data
Water supply systems: new technologies / edited by Cedo Maksimovic.

Francesco Calomino. John Snoxell.
p. cm. -- (NATO ASI series. Partnership sub-series 2.
Environment; vol. 15)
Includes bibliographical references and index.
1. Water-supply--Management--Technological innovations.
I. Maksimovic. C. II. Calomino. Francesco. 1949- III. Snoxell.
John. 1948- IV. Series.
TD353.W39 1996
628.1--dc20 96-32374
CIP
ISBN-13: 978-3-642-64739-0 e-ISBN-13: 978-3-642-61187-2

001: 10.1007/978-3-642-61187-2
This work is subject to copyright. All rights are reserved, whether the whole or part of the material
is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other way, and storage in data banks.
Duplication of this publication or parts thereof is permitted only under the provisions of the German
Copyright Law of September 9, 1965, in its current version, and permission for use must always
be obtained from Springer-Verlag. Violations are liable for prosecution under the German Copy-
right Law.
© Springer-Verlag Berlin Heidelberg 1996
Softcover reprint of the hardcover 1st edition 1996
Typesetting: Camera-ready by authors/editors
SPIN: 10470207 31/3137 -5 4 3 210 - Printed on acid-free paper
New Technologies for Water Supply Systems
Preface
Technology now affects almost every aspect of Water Supply Management, Operation,
Planning and Design; the speed of development means that assessing what is "new" is
sometimes difficult. Old ideas can now be applied because of new technology; technology is
now revealing problems that were unnoticed 10 years ago. Some emerging technologies
promise much but are still underdeveloped for use in real world conditions, while we should
always remember that "new" technology depends upon the state of development in respective
countries, a point which is particularly relevant to the NATO Advanced Study Institute, for
which this book has been produced.
Thus our objective in producing the book has been to highlight, in a wide range of
technical areas, where and how technology is being applied, what is "new" and what the
limitations of these technologies are in the real world. We have also tried to provide an
European and American perspective where possible to illustrate how problems are tackled in
different cultural environments. It is probably true that "technology" is also somewhat
dependent upon the political, economic and organisational climate in different countries and we
have included a chapter covering these aspects.
Water Demand in most countries has shown a steady, relentless increase in the past and
this has put increasing pressure on water resources globally. Chapter I, Water Supply Systems
on the Threshold of the 21st Century, looks at water demand and some new concepts in
meeting these demands: separate systems for potable and agricultural use,differing qualities of
water for differing uses, protecting water sources and transmission lines, and waste water
reuse, a subject which is coming under increasing scrutiny.
Increasingly, water is being moved over very long distances to balance areas of surplus
and deficit, both in space and time, and Chapter n, Water Supply Conveyance and Control,
looks at some examples of these systems and the associated problems. Firstly, traditional and
new technologies in pipeline construction and structural requirements are considered and
secondly two such systems, Brussels in Belgium, and Wessex in the UK, are looked at in detail
to illustrate the kinds oftechnology now in use to monitor and control these large networks.
VI
Over recent years new technology has affected water treatment quite radically. Firstly,
new analytical techniques have revealed many complex organic pollutants to be present in raw
and treated water (some formed as by-products of treatment), and some are associated with
diseases such as cancer. The response to this has been a tightening of standards for a wide
range of such parameters and, in consequence, major investment in new and innovative water
treatment technologies.
Chapter III, New Technologies for Water Treatment, looks at these problems,
providing an European and American perspective, and goes on to look in detail at the
innovative approach adopted by Amsterdam Water Supply, in treating raw water derived from
the river Rhine. An American view is provided which looks at technologies emerging in the
United States, with particular reference to health effects, costs of different technologies, and
research being conducted at Colorado' State University.
Disinfection using chlorine has been used almost universally for decades but it is only in
recent years that the presence of disinfection by-products, and their potential effects as
carcinogens, has been recognised. This has led to new technologies for controlling and
removing such complex chemicals, and research into alternatives. Chapter IV, Emerging
Problems and Technologies in Water Treatment, looks at these technologies and also looks at
new approaches to optimising older technologies for water clarification, coagulation and
flocculation.
There have been numerous advances in the management and operation of existing
systems and chapter V, Diagnosis and Rehabilitation of Existing Systems, looks at these
technologies. Flow measurement, water balancing, network modelling, data logging, leak
detection and pressure control have all seen major advances in the last decade along with new
techniques for pipe rehabilitation. The current situation in the UK is examined in detail with
some examples from around the world to provide a global perspective.
Network Analysis, or computer modelling of supply systems, is an old technique which
is' how finding new life because of advances in computer technology and' advances in data
management. Chapter vi, The Application of Modelling in the Management of Water Supply
Systems, looks at the numerical methods underlying this technique and a new approach to the
computing aspects. It also looks at the way that network 'modelling is now being integrated
with telemetry and GIS to provide a live, up-to-date, dynamic model of the network which can
be used as a real operational management tool, as well as;in its more traditional role in planning
and design.
In' some special conditions, transient flows giving rise to water hammer can be very
destructive if not controlled properly and Chapter VII, Transient Flows, looks at some of the
technologies nOw being employed to examine and analyse these phenomena, including new
numerical methods and computing techniques.
VII
Geographical Information Systems (GIS) are relatively new computer systems now
extensively used in the management, operation and analysis of water supply networks. Chapter
VIII, The Application of GIS and Remote Sensing, looks at the theoretical basis of such
systems, and how they can be used for spatial analysis and as digital maps for the management
of underground assets in a water company environment. The potential of remote sensing for
data capture and leak detection is also covered.
The social, economic and political factors affecting water supply systems vary
enormously between countries and often have a profound effect on the management, operation
and technological status of the network. Chapter IX, Institutional, Social and Economic
Aspects, concludes the book by looking at some of these frameworks and the way in which
such issues as charging, privatisation, levels of service, management structures and other issues
are dealt with using some European examples.
The scope of this book is very wide, and indicates that "new technology" now affects
almost every aspect of water supply. From resource management, conveyance, treatment,
hydroinformatics, pumping, to telemetry, data logging and GIS, the modem water supply
network now encompasses technologies from every part ofthe spectrum. The challenge is to
learn how to apply and manage these technologies and to spread this knowledge throughout
the global world of water.
This book 'and the ASI (Advanced Study Institute of the NATO Science series of
Advanced Science institutes) from which it is derived, has made a contribution to this process
and has helped in spreading this understanding of new technology between the participating
nations.
This is a reflection of NATO's objectives with the ASI programme - "to promote the
dissemination of scientific knowledge, to impart experiences and foster professional contacts
among scientists" - and it is to be hoped that this book will prove to be a worthy record of the
ASI in Bulgaria 1994, and the hard work and commitment of all those who took part.
Cedo Maksimovic
Francesco CaIomino
JohnSnoxeU
April, 1996
Acknowledgements
The Organizers wish to thank the following companies for their co-sponsorship of the
NATO ASI in Bulgaria, 1994:
• BERMAD Control Valves, ISRAEL

• BIWATER Industries, UK
• DEGREMONT ITAUA, ITALY
• l. T. T. Flygt, UK
The Editors would also like to express their thanks and appreciation to the individual
authors, chapter coordinators and all those who assisted in the preparation of this book.
Special thanks to prof. Vujica Yevjevich for his invaluable suggestions in preparation
and ASI organization and running. We are also indebted to prof. Paolo Veltri who helped us a
lot in the final phase of preparation of the text for publishing.
Finally, special thanks are due to Ms Zorica Todorovic, B.Sc. for her outstanding work
and assistance in transforming the individual contributions into a consistent and high-quality
volume.
The Editors
April,1996
List of Contributors
A. Graveland
Daniel L. 8rumagne
Amsterdam Water Supply
Compagnie Interconununale
Voge1enzangseweg 21
Bruuxelloise des Eaux
Rue aux Laines 70, I ()()() Bruxelles
2114 BA Vogelenzang
BELGIUM THE NElHERLANDS
Vladimir Havlik
Francesco Calomino
Dipartimento di Difesa del Suolo Czech Technical University of Prague
Thakurova 7, 16629 Prague 6
87040 Montalto Uffugo (Cs) CZECH REPUBLIC
ITALY
Jan Peter van der Hoek
Tory L. Champlin
Department of Civil Engineering Vogelenzangseweg 21
2114 BA Vogelenzang
Colorado State University
Fort Collins, Colorado THE NElHERLANDS
USA
Peter David Hedges
Thomas William Charnock Department of Civil Engineering
Aston University, Aston Triangle
Department of Civil Engineering
Birmingham B4 7ET
Birmingham B4 7ET UK
UK
Petr Ingeduld
LisaM. Dunn Czech Technical University of Prague
Department of Civil Engineering Thakurova 7, 16629 Prague 6
Colorado State University CZECH REPUBLIC
Fort Collins, Colorado
Marko Ivetie
Institute of Hydraulic Engineering
JohnElgy
Department of Civil Engineering Faculty of Civil Engineering
University of Belgrade
Birmingham B4 7ET P.O. Box 895, II 000 Belgrade
YUGOSLAVIA
UK
x
Pavel Jirou§ek Thomas G. Sanders

Charles University Department of Civil Engineering
Prague Colorado State University
CZECH REPUBLIC Fort Collins, Colorado 80523
USA
Peter KaIinkov
Higher Institute of Architecture John Snoxell
and Civil Engineering, WESSEX WATER
I Christo Smimenski, 1421 Sofia WESSEX House, Pas$8ge str.
BULGARIA BS2 OJQ Bristol
UK
Jean Louis Monzo
Higher Institute of Architecture Stanislav Vanecek
and Civil Engineering, Czech Technical University of Prague
I Christo Smimenski, 1421 Sofia Thakurova 7, 16629 p.rague 6
BULGARIA CZECH REPUBLIC
tedo Maksimovic Adrian Verwey

IRTCUW, Insitute of Hydraulic Engineering IHE, Delft
Faculty of Civil Engineering P.O. Box 3015,2601 DA Delft
University of Belgrade THE NETHERLANDS
P.O. Box 895, II 000 Belgrade
YUGOSLAVIA Paolo Veltri
Mario J. N. Neves Universitil della Calabria
Universidade do Porto 87040 Montalto Uffug~ (Cs)
Rua Jer6nimo Mendon~ 65 ITALY
4200 Porto Codex
PORTUGAL Radisav D. Vidic
Department of Civil and Environmental
Duian Obradovic Engineering
WINS Ltd. University of Pittsburgh
Tower House, Parkstone Road 949 Benedum Hall, Pittsburgh PA 15261
Poole, Dorset BHI5 2DH USA
UK
Vujica Yevjevich
Dusan Prodanovic Department of Civil Engineering
IRTCUW, Insitute of Hydraulic Engineering Colorado State University
Faculty of Civil Engineering Fort Collins, Colorado 80523
University of Belgrade USA
YUGOSLAVIA
EvienZeman
Czech Technical University of Prague
CZECH REPUBLIC
Table of Contents
CHAPTER I:
WATER SUPPLY SYSTEMS ON THE THRESHOLD
OF THE 21 sT CENTURY
Introductory Statement ................................................................................................... 3
1. Urban Water Demand ................................................................................................................ 7

Thomas G. Sanders, Vujica Yevjevich
2. Water Demand Has a Threshold ............................................................................................... 19

Paolo Veltri
3. Availability and Selection of Sources of Water for Water Supply Systems ................................. 25
Vujica Yevjevich, Thomas G. Sanders
4. New Concepts for Water Supply Systems: What Kind of Water

for What Purpose in the Future ............................................................................................... .43
VUjica Yevjevich
5. Pollution Control for Urban Water Supply Systems .................................................................. 59

Thomas G. Sanders, Vujica Yevjevich
6. Wastewater Reuse - a Worldwide Issue ..................................................................................... 71

Thomas G. Sanders, Lisa M Dunn, Vujica Yevjevich
CHAPTER II:
WATER SUPPL YCONVEYANCE AND CONTROL
Introductory Statement .................................................................................................. 87
1. Construction Methods and Pipe Structural Requirements .......................................................... 89

Francesco Calomino, Paolo Veltri
2. Water Transfer Within Brussels Water Services ................................................................................ 111

Daniel L. Brumagne
3. Transfer ofWaier and Data Management in Wessex Water - UK ..................................................... 127

John Snoxell
XII
4. Control Valves - Principles of Operation ................................................................................. 145

Marko Ivetic
CHAPTER III:
NEW TECHNOLOGIES FOR WATER TREATMENT
Introductory Statement ................................................................................................ 163
1. New Technologies for Water Treatment: General Aspects ................................................................. 165

J.P van der Hoek, A. Graveland
2. New Technologies for Water Treatment: Case Studies ....................................................................... 185

J.P. van der Hoek, A. Graveland
3. Treatment of Drinking Water in the United States on the Threshold of the 21 st Century .......... 211
Thomas G. Sanders, Tory L. Champlin
CHAPTER IV:
EMERGING PROBLEMS AND TECHNOLOGIES IN WATER TREATMENT
1. Control of Disinfection By-Products in Drinking Water: Regulations and Costs ....................... 259
Radisav D. Vidic
2. Control of Disinfection By-Products in Drinking Water: Case Studies of Alternative ............... 275
Disinfection Technologies
Radisav D. Vidic
3. Technological Aspects of Advanced Processes of Coagulation and

Flocculation in Water Treatment ............................................................................................ 293
Peter Kalinkov, Jean Louis Monzo
4. Study of Floc-Blanket Clarifier, Combined with Lamella Settlement

under the Condition of Prolonged Coagulation ........................................................................ 307
Peter Kalinkov, Jean Louis Monzo
CHAPTER V:
DIAGNOSIS AND REHABILITATION OF EXISTING SYSTEMS

XIII
1. Measurement Methods for Diagnosis and Rehabilitation of Existing Systems .......................... 327
Cedo Maksimovic
2. Diagnosis and Rehabilitation of Existing Systems .............................................................................. 349

John Snoxell
CHAPTER VI:
THE APPLICATION OF MODELLING IN THE MANAGEMENT
OF WATER SUPPLY SYSTEMS
1. Water Supply Networl<: Modelling and its Application in Design,

Planning and Operational Management .............................................................................................. 371
DuSan Obradovic, John Snoxell
2. Object-Oriented Hydroinfonnatics Tools for Water Distribution Analysis .............................. .401

Petr Ingeduld. Vladimir Havlik. Stanis/av Vanecek, Pavel Jirou§ek
CHAPTER VII:
TRANSIENT FLOWS
1. Pumping Stations and Transient Flow Analysis ...................................................................... .423

Marko Ivetic
2. Object-Oriented Tools for Transient Flow Analysis ................................................................ .451

Petr Ingeduld. EvZen Zeman. Adrian Verwey
CHAPTER VIII:
USE OF GIS AND REMOTE SENSING IN WATER SUPPLY
1. Relevant GIS and Remote Sensing Technology: Theoretical Considerations ............................ .467
John Elgy. Thomas William Chamock. Peter David Hedges
2. Use of GIS and Remote Sensing in Water Supply: Applications ............................................. .491
John Elgy. Thomas William Chamock, Peter David Hedges
3. Application of GIS in Analysis and Design of Water Supply Systems ...................................... 527
Cedo Maksimovic. Dusan Prodanovic
XIV
4. Geographical Infonnation Systems and Digital Mapping for Water Supply Inventory .............. 537
Thomas William Charnock, John Elgy, John Snoxell
CHAPTER IX:
INSTITUTIONAL SOCIAL AND ECONOMIC ASPECTS
I. Historical and Institutional Aspects ......................................................................................... 559

Paolo Veltri
2. Management Options .............................................................................................................. 565

Paolo Veltri
3. Quality Management in Water Supply .................................................................................... 573

Mario J. N. Neves
4. Examples of Existing Management Systems in Europe ............................................................ 589

Paolo Veltri
5. Institutional Framework for Training of Specialists for Water Supply Systems:

The Status and Problems to be Solved .................................................................................... 597
Cedo Maksimovic
Index.......................................................................................................................................... 607
Chapter I
Water Supply Systems on the Threshold

of the 21st Century
Chapter I
Water Supply Systems on the Threshold

of the 21st Century
INTRODUCTORY STATEMENT
The development and expansion of water supply systems are continuing problems for
the highly industrialized countries as well as the transition countries on the threshold of the
21 st century. Although the magnitude and the complexities of the problems will vary from
country to country, the fundamental issues of demand, planning concepts, water sources,
pollution control and wastewater reuse are universal. In this Chapter each one of these issues
will be discussed.
The past and present demand of water systems in Europe and the United States are
presented and reviewed. Water demands from as early as the 1900's in the US are presented as
well as more recent demands illustrating the increasing variability of demand making the
prediction of future demand, which is the basis for design, extremely difficult. In addition,
many ofthe European countries' water demand are reviewed as well. The annual, seasonal,
monthly, weekly, daily and hourly demands experienced in the US are summarized. As
discussed, each has a value and is used in the design of water systems in the US. It is
interesting to note that there appears to be an increasing trend in domestic or urban water use
in the US, using data from New York City, Dallas, Texas and Los Angeles, California and the
consumption per capita has increased. In Italy, the increase in consumption per capita has been
small through the period 1975 to 1984. In Europe, the average domestic use of water ranged
from a little over 110 lip/day (liters/person/day) in Belgium to 250 lip/day for Switzerland
4
while in the US, the average urban use varied from 130 Vp/day to much greater than 800
Vp/day with a national average of628 Vp/day.
As a result of the worldwide problem of diminishing sources of fresh water, the
availability and selection of sources of water for future water supply systems is increasingly
important and are summarized and discussed. Surface water sources include: springs, lakes and
reservoirs, ponds, streams, estuaries wetlands and swamps. Groundwater sources include:
alluvial aquifers, rock fissure water systems, karstic water systems, volcanic rock groundwater,
artificially recharged groundwater and juvenile groundwater aquifers. Brackish and sea water
sources are also mentioned. Criteria needed for selection of sources for water supply to be
covered are, source characteristics, regional characteristics, water rights and markets, stage
development and economics.
The Section of the Chapter covering new concepts for water supply systems reviews
the existing and new concepts and investigates the issues associated with the safety, technology
and economics of interconnecting adjacent water supply systems. New water quality standards
and the concept of producing various qualities of waters for various uses is also discussed. The
uses include: boiler feed water, drinking water, human hygiene and recreational water,
irrigation water, industrial including mining, smelting, and cooling water and hydropower and
navigation water. Each requires a specific water quality, from distilled water for boilers to
water free from floating material for navigation. Other water uses such as wildlife protection to
wastewater dilution will have specific water quality requirements as well. Although not a new
concept, the idea of a dual water system is discussed also.
The protection of water supplies from pollution is a problem of continuous concern
particularly as countries become more industrialized. The concept of the multiple barrier
system which is the principle most water supply systems in the US is based upon is discussed.
The protection in the multiple barrier system varies from restricting access to the source
watershed and collection systems to the technology based water pollution control program
promulgated by the federal government and enforced by the individual states. In many areas of
the United States, physical access to source water supplies is restricted. This includes not only
the watershed but also transmission works and storage structures. Land use restrictions and
prohibition of specific pesticides use have been imposed and improved water quality has
resulted. T.he water quality legislation since 1965 when a water quality based pollution
abatement program was first introduced and in 1972 when the technology based program was
instituted with a goal of "zero discharge" and all rivers feasible and swimmable has endeavored
to improve water supply quality by diminishing pollution sources. The foundation of the 1972
legislation, the National Pollution Discharge Elimination System (NPDES), Best Available
Technology (BAT), and the Best Management Practices (BMP) are discussed. This focus on
pollution sources will continue with the emphasis in the future on nonpoint sources. In
5
addition, there has been a renewed interest in new technologies in water treatment as a result
of the presence of trace compounds in drinking water supplies and the continuous vulnerability
of the water supplies to infectious agents. The advanced and new technologies are discussed in
greater detail in Chapter m.
And finally, the use of wastewater as a viable source for urban water supplies is
discussed in this Chapter. Since history was first written, man has been reluctant to use his
wastewater, usually disposing of it in the nearest river, lake or ocean. And although it was not
until the mid 1800's that the relationship between disease and contaminated drinking water was
established, the reluctance in using it today continues even though the technology exists to
treat wastewater to a quality exceeding existing drinking water standards. This is due in part to
the concept of a multiple barrier system discussed in the Chapter where the source water for
drinking should be unpolluted and, in particular, unpolluted by man's waste, and thus has
prevented widespread consideration of wastewater as future water supply source. However,
things are changing as the sources of new water supplies are decreasing and the cost of treating
brackish and sea waters continues to increase. Examples of existing uses of wastewater and the
costs are summarized.
1. Urban Water Demand
Thomas G. Sanders
Vujica Yevjevich
Fort Collins, Colorado 80523
USA
1.1. INTRODUCTION
Water demand generally referred to as future water requirements will be a function of

the demanding entity. The demand characteristics such as trending, jumps, periodicity and
stochasticity are different for an urban demand, irrigation and agriculture demand, industrial
and hydropower demands. The variation in demand can be large as well. For example, urban
water demand will vary hourly, daily, weekly, seasonally and annually. Whereas, irrigation
demand in the Western United States, where appropriative water rights exists (first in time,
first in right), although quite seasonal is much more predictable as most water rights specify a
specific amount of water delivered during a specified period of time. If the demand exceeds the
supply, the water rights having the highest priority will be supplied first and those rights having
a lower priority may not receive any water at all.
Urban water demand which is probably the most complex of all the demands as it is a
combination of domestic, public, commercial, industrial, thermal power and even irrigation
demand, especially during the summer months, will be the focus of this Section.
Annual water demand is an indicator of the size and characteristics of the population
served. The location of cities to a large extent have been a function of the annual
characteristics of the nearby water resource; this availability of supply leads to the demand. As
the population increases and water uses diversify, estimating demand becomes more complex.
The factors that affect demand include not only the population demographics but also
availability, climate, economic development, technological development, water pricing,
possibly water metering and the level of embracing the environmental ethic of conserving
resources to name a few.
NATO ASI Series. Partnership Sub-Series, 2.Environment- Vol. 15
Water Supply Systems - New Technologies
Edited by C. Maksimovic, F. Calomino, and J. Snoxell
8
Engineers use the annual demand statistics to determine the size of the water
infrastructure to meet these demands. Components of the infrastructure would include
transmission works, reservoirs and treatment facilities. Seasonal demand affects the operational
schedule of the infrastructure: for instance, empty reservoirs in the Fall and full reservoirs in
the Spring, use of less than full capacity of the treatment system during the winter but full
capacity utilized in the summer. Monthly and weekly demands are associated more with the
seasonal fluctuations. Variations in daily demand will affect production as water demand
decreases on the weekends. And finally, hourly demand fluctuates sufficiently that pumping
stations and the capacity and configuration of the pumps are sized and designed on the
maximum hourly flow.
1.2. DEMAND VARIATIONS
1.2.1. Annual
The annual water demand in a survey of 40 cities in the United States varied a great
deal in 1890. A high of 186 gallons per capita per day, gpcd, (704 Up/day) was recorded in
Buffalo, New York and a low of 13.8 gpcd (52 Up/day) was recorded in Chippewa Falls,
Wisconsin (Table 1.).
Fifteen years later the daily per capita use in Buffalo was 324 gallons (1226 Up/day)
and in Chippewa Falls 100 gpcd (378 Up/day) indicating a 74% and 624% increase in demand
respectively in 15 years. During this same peri<;>d of time, the per capita use of water in
Rochester, New York increased from 66 (250 Up/day) to 88 gpcd (333 Up/day), a 33%
increase (Table 2.). In 1980, the consumption of water in Rochester increased to 451 Up/day, a
583% increase from 1890. Of the 22 cities surveyed in 1980, Rochester had the seventh lowest
consumption while Passiac, New Jersey had the highest at 807 Up/day and St. Paul, Minnesota
had the lowest at 277 Up/day. Recently, in Fort Collins, Colorado where much of the municipal
water is used for lawn irrigation in the summer time, the daily per capita rate was 187 gallons
(708 Up/day).
While the aforementioned data indicates trending, there is variation in the annual
demands as well. With reference to Figure 1., the variation in annual demand can be
substantial. Even as early as the late 1800's, the annual demand could shift over 100% in
several years. For example, the demand in Detroit, Michigan went from 120 gpcd (454 Up/day)
to over 200 gpcd (Up/day) and back to 125 (Up/day) in a period of20 years.
9
Table 1. Consumption of water in american cities and towns in 1890 and 1905
(Tumeaure and Russell,1908)
City Population Population Percent Consump. Percent Consump.

per Tap of Taps per of Taps per
Metered Inhabitant Metered Inhabitant
Daily Daily
1900 1890 1890 1890 1905 1905

Chicago 1698.600 7.1 2.5 140 ... 200
Philadelphia 1293.700 6.1 0.3 132 1.0 230
StLuis 575.200 1l.8 8.2 72 7.8 92
Boston 560.900 6.6 5.0 80 5.0 151
Cleveland 381.800 8.7 5.8 103 68 137
Buffalo 352.400 6.3 0.2 186 3 324
San Francisco 342.800 9.9 41.4 61 21 96
Cincinnati 325.900 . 8.5 4.1 112 12 130
Detroit 285.700 5.1 2.1 161 9 188
Milwaukee 285.300 11.1 31.9 110 80 91
Louisville 204.730 11.9 5.9 74 8 81
Minneapolis 202.720 16.5 6.3 75 47 76
Providence 175.600 9.4 62.4 48 86 68
Indianopolis 169.160 35.6 7.6 71 10 82
Kansas City 163.750 ... ... ... 38 73
St. Paul 163.065 12.7 4.2 60 38 56
Rochester 162.600 5.4 11.4 66 41 88
Toledo 131.820 18.6 9.4 72 70 75
Columbus, O. 125.560 1l.5 6.4 78 76 110
Worchester, Mass. 118.420 8.9 89.4 59 95 75 I
Fall River,.Mass. 104.860 14.9 74.6 29 97 42 II
Memphis, Tenn. 102.320 11.9 3.7 124 20 100
Lowell, Mass. 94.970 9.2 22.9 66 69 58
Atlanta, Ga. 89.870 20.0 89.6 36 100 65
Dayton, Ohio 85.333 20.0 3.8 47 70 70
Nashville Tenn. 80.870 14.9 0.8 146 52 148
Camden, N. J. 75.940 ... ... 131 3 155 I
Yonkers, N. Y. 47.930 12.0 82. 68 99 115 I
Newton, Mass. 33.587 5.5 67. 40 86 58
Aurora, Ill. 24.147 8.2 19.3 40.7 36 56 I
Madison, Wis. 19.164 11.0 31.0 40 97 46 I
Ashland, Wis. 13.074 .
9.9
. _--,--
2.8 90 ... 81
Champaign&Urbana, Ill. 14.826 7.3 2.5 I 43 I ... I 45 I
Chippewa Falls, Wis. 8.094 7.4 6.6 13.8 ... 100 I

Middleborough, Mass. 6.885 11.7 24.0 21 47 38
Beloit, Wis. 10.436 10.2 10.0 64 ... 130
Foxborough, Mass. 3.266 8.7 34.0 44.0 46 63
Clinton, Ill. 4.452 4.1 3.0 27.0 1.1 99
Shenandoah, Ia. 3.573 15.5 15.5 39 ... 35
Melrose, Mass. 12.962 4.2 1.7 71 3 112
10
With reference to Figure 2., the annual demand of New York City, increased from
approximately 125 billion gallons per year (473 billion liters per year) in 1898 to 450 billion
gallons per year (2044 billion liters per year) in 1968, a 260010 increase.
Viessman and Hammer (1985) report an average water consumption in the United
States of 628 Up/day. As recently as 1991, the average per capita demand reported in the
United States varied from 130 to 2000 Up/day (McGhee, 1991).
Table 2. Recorded rates of consumption in some american cities (McGhee, 1991)
City Average daily per Maximum l-day Maximum in

capita consumption, consumption in a proportion to
1* 3-year period, I * avenll!:e, %
Rochester, N.Y. 451 637 141
Syracuse, N.Y. 728 917 126
Hartford, Conn. 671 889 132
Albany, N.Y. 671 860 128
EI Paso, Tex. 447 739 165
Portland, Me. 572 773 135
Camden, N.J. 641 963 150
Albuquerque, N.M. 402 766 190
Winston-Salem, N.C. 447 580 130
Waterloo, Iowa 383 625 163
Passaic, N.J. 807 1016 126
South Gate, Calif. 550 891 162
Fort Smith, Ark. 474 652 138
Poughkeepsie, N.Y., 569 728 128
Tyler,-Tex. 371 743 200
Monroe La. 584 875 150
Spartanburg, S.C. 754 955 126
Pomona, Calif. 629 1092 173
St. Cloud, Minn. 277 701 254
Salina, Kan. 603 1357 225
Alliance, Ohio 796 III tV I
Ashtabula, Ohio 766 985 129

* Gallons = liters x 0.264
11
250 250
/
II"
. ...-
200 ~ / 200
1\
\ A
f---f-~~ \ I /
!150 r-----W
\ J..
I
!"Ii.'
III
...
/ ;::. (J
8- L- ~ ~J
::::00- .gl-
1"
.j. y
',..r IP'"
........
8 100 :fhA- ,
./ iiIfr .~L ,fi.
J a<' 1,\
vI 'il'v :':"'~
~ ~~ -
~ I
.J
..."j-.;., ./
"
1./
to
,L.lA
..,. "'~
\
~gO'"
-
-~
I' ~ 'C\\ .
50
L- ~'dence ~.£~..,.- V' 50
~~
o 0
1875 '80 '85 '90 '95 1900 'OS 1875 '80 '85 '90 '95 1900 '0.5
Figure 1. Variation of yearly rates of consumption (Tumeaure and Russell,1908)
500000
450000
r
;g 400000 I ./ ~I
II V
'-I
.9
i 350000 I'\. ~ '\
} v,"
.§ I
11 300000
(I
.S 250000
~
1
J
V ~
200000
~ 150000
/
:::;l 100000
50000
o
o 5 10 15 20 25 30 35 40 45 50 60 65 70 75 80
Time in years
Figure 2. Annual water use for New York City, New York, for 1898-1968 (Salas and Yejevich,1972)
12
1.2.2. Seasonal
Seasonal water demand is characteristic of most municipal water supplies in the United
States. With reference to Figure 3., during the period of 1950-1969 in Dallas, Texas in addition
to the obvious trending of the mean, there is an appreciable seasonality of the monthly demand
indicated by the annual peaks and troughs of the water demand. The fluctuations are not only
occurring every year but they also appear to be increasing indicating a trending
(nonstationarity) of the variance as well. A similar trending of the mean and variance can be
seen in the water demand for Los Angeles, California for the time period of 1940 to 1965
(Figure 4.). The demand increased from about 4 billion gallons per year (15 billion liters per
year) in 1940 to over 10 billion gallons per year (38 billion liters per year) in 1965 and seasonal
fluctuations of a little more than 2 billion gallons per year (8 billion liters per year) and 4 billion
gallons per year (15 billion liters per year) respectively.
10000
9000 -
8000 -
5'"
i 7000 -
5
] 6000 -
.9
~ sooo
,
-
='
~
1;j
~ 4000 r-
...=
~ 3000 r- ~\ ~
'j ~ ~
2000
1000
~~ IW U "
'" ~~ "
I I I I I I I I I I I I
0 20 40 60 so 100 120 140 160 ISO 200 220 240 260
TIme in months
Figure 3. Monthly water use for Dallas, Texas, for 1950-1969 (Salas and Yevjevich,1972)
13
12000 , - - - - - - - - - - - - - - - - - - - ,
] 10000 f-
-;
-
bI)
.s~ 8000 f-
.5
1!l
~
::;
6000 -
1;j
~
~ 4000 -
2000 - -
,I I I I I I
o 50 100 150 200 250 300 400
Time in months
Figure 4. Monthly water use for Los Angeles, California, for 1940-1965 (Salas and Yevjevich, 1972)
1.2.3. Monthly and Weekly
Monthly and weekly demands will be related to seasonal variations and not as
important as the seasonal or daily fluctuations. However, weekly maximum have varied
between 119 % and 167% of the average daily consumption (Tumeaure and Russell, 1908).
Tchobanoglous and Schroeder (1985) report daily averages in the maximum month to be
typically 120% of the annual average and the daily average of the maximum week to be 140%
of the annual mean.
1.2.4. Daily
Daily demands vary depending on the day of the week. Usually, the demands on
Saturday and Sunday are less than during the week. An exception would be in areas similar to
Fort Collins, Colorado in a summer without appreciable rainfall, the demand for irrigating the
lawns may be greater than the demand during the week. Tchobanoglous and Schroeder (1985)
report the maximum day to be typically 180% of the annual average.
14
1.2.5. Hourly
The hourly water demands show the greatest fluctuations because of the
anthropogenetic diurnal cycle. Other factors which affect the hourly demand are seasonal and
antecedent precipitation. As can be seen in Figure 5., the demand in gallons per dwelling unit
approaches zero around 4 am and peaks at lOam and 8 pm at approximately 400 gallons
(1514 liters) per day per dwelling during a typical winter day. A maximum winter day was
found to be over 2000 gallons (7571 liters) per dwelling per day. A similar fluctuation occurs
during a summer day (90 0 F), with and without precipitation (Figure 6.). A low demand of
approximately 50 gallons ( 189 liters) per day per dwelling and two peaks of approximately 350
gallons (1325 liters) per dwelling on days with rain were realized while demands as high as
1560 gallons (5905 liters) per dwelling on days without rain were realized. Tchobanoglous and
Schroeder, 1985, report the maximum hour to be typically 270% of the annual average.
Typical maximum day ~

2,000 .------.-------.----r-----.----+--f-:-----,
1,600 1----+----1----+------11--+---+11----1
1bO
:§ 1,2001----+----+---1----+-+---++----1
!Iil
Q.
~ 800~--4---_+---~--1---~--~
1;l
::s
!i
~
4001----+----±~~-~~--+-----I--~
4 8 12 4 8 12
A.M. P.M.
Time of day
Figure 5. Daily water-use patterns, maximum day and winter day (from Residential water-use research
project, Johns Hopkins University and Federal Housing Administration)
(Viessman and Hammer, 1985)
15
1400
L
1200
J
90°F day without rain
~ -
~ 1000
.§
bI)
.9
==.,[$
.8.
800
....
lrL
"CI
"CI
~ 600
,..1;l -
.,
'0$ 90"F day with-...rain
::t '-
C:r -
400
r-
L.r.- Il-, 'l
200 -
L
L
I--'
o
~
12 4 8 12 4 8 12
A.M. P.M.
Time of day
Figure 6. Daily water-use patterns in R-6 area, maximum day and minimum day (from Residential
water-use research project, Johns Hopkins University and Federal Housing Administration)
(Viessman and Hammer, 1985)
1.3. ANALYSES
In order to estimate urban water demand, the trends, periodicities and stochasticity of
the data should be understood and then modeled. In a study of four cities in the United States
and one city in Canada, Salas and Yevjevich, 1972, determined the polynomial trend model,
the dependence model and the stochastic model for urban water use for each city. With
reference to Tables 3. and 4., the annual trends were modeled with a quadratic with the
exception of Colorado Springs, Colorado which had a cubic trend model. Fort Collins,
16
Colorado and Colorado Springs did not require dependence models as the residuals after
removing the trends were independent. However, the remaining cities' residuals showed
dependence requiring 1st and 3rd order autoregressive models. The independent residuals of
each city was further analyzed and it was found that all could be modeled using a normal or in
the case of Montreal, Canada, a lognormal model.
Table 3. Polynomial regression coefficients for the trend in annual urban water use
(Salas and Yevjevich,1972)
Name Unit Order of Rei ;ression coefficients

trend A B C D
Fort Collins, Colo. m.g. Quadratic 1,172.420 - 7.245 1.573
Colo. Springs, Colo. m.g. Cubic 3,532.943 -282.570 25.118 -0.276
Baltimore Maryland m.g. Quadratic 13,088.346 302.060 7.244
Baltimore, Maryland g.p.c.d. Quadratic 91.120 1.368 - 0.0071
New York New York m.g. Quadratic 101,737.57 8187.10 -49.577
New York, New York g.p.c.d. Quadratic 104.389 0.844 0.0031
Montreal, Canada m.g. Quadratic 44,792.895 1135.30 17.344
Montreal, Canada g.p.c.d. Quadratic 114.120 - 0.2569 0.0267
Table 4. Dependence model and distribution of the independent stochastic component for annual urban
water use (Salas and Yevjevich, 1972)
Dependence model Distribution

Name Unit Type Parameters Type Parameters *
pi p2 p3 A B C
Fort Collins, m.gc Normal 0.0 204.29
Colo.
Colorado m.g. Normal 0.0 419.99
Springs, Colo.
Baltimore, m.g. First 0.825 Normal 0.0 0.563
Maryland
Baltimore, g.p.c.d. First 0.766 Normal 0.0 0.633
MarYland
New York, m.g. First 0.681 Normal 0.0 0.729
New York
New York, g.p.c.d. First 0.493 {,u11l".. v.v v.o'v - ~J'T
New York
Montreal, m.g. Third 0.629 0.318 -0.23 Lognormal-3 2.019 0.084 -4.90
Canada
Montreal, g.p.c.d. Third 0.725 0.413 -0.067 Lognormal-3 1.593 0.116
Canada
* for Normal: A = Mean for Lognormal-3: A - Mean ofln <e-C)
B = Standard deviation B - Standard deviation
C - Lower boundary
17
Further analyses of the Fort Collins' weekly data yielded that the trend of the data
accounted for 41.9 % of the variance while the periodicities accounted for 40.9% of the
variance as well and the stochastic component accounted for 17.20 % of the variance (Salas
and Yevjevich, 1972). Analyses of urban demand data for other cities will probably yield
similar results pointing out the necessity of taking into account the trends, periodicity and
stochasticity of the data when determining urban water demand.
1.4. CONCLUSIONS
Urban water demand has had and continues to have large variations as a result of many
factors. These factors include population growth, water availability, climate, economic
development, technological development, water pricing, water law (appropriative vs. riparian),
and possibly water metering. As a result, urban water demand exhibits trends, periodicities and
stochasticities each responsible for a proportion of the variance. Therefore, in order to be able
to develop realistic and useful estimates of urban water demand, procedures and
methodologies must be able to quantify and take into account these factors of urban water
demand which are responsible for the variances.
1.5. REFERENCES
McGhee T.J. (1991) Water Supply and Sewerage. 6th edition, McGraw-Hili, New York.
Salas J.D., Yevjevich V. (1972) Stochastic Structure of Water Use Time Series. Hydrology
Papers, Number 72, Colorado State University.
Tchobanoglous G., Schroeder E.D. (1985) Water Quality, Characteristics, Modeling,
Modification. Addison-Wesley Publishing Company, Reading, Massachusetts.
Turneaure F.E., Russell H.R. (1908) Public Water Supplies. 2nd edition, John Wiley & Sons
Inc., New York.
Viessman Jr. W., Hammer MJ. (1985) Water Supply and Pollution Control. 4th edition,
Harper & Row, Publishers, New York.
2. Water Demand Has a Threshold
Paolo Veltri
87040 Montalto Uffugo (Cs)
ITALY
2.1. INTRODUCTION
The limits to sustainable development in the supply and use of freshwater resources
have been reached during the last decades in many parts of Europe: in spite of limited growth
in specific domestic consumption in the most industrialised countries, many factors show that
in the same field a lot of countries will demand in the next years more and more freshwater
resources. As a matter of fact it is believed that all the countries with improving life conditions
will use more water consumption facilities and, in addition, an increase in water demand for
irrigation and industries has to be expected. These tendencies are however in conflict with the
inadequacy of water resources in terms of both quality and quantity: this is true both in the
Southern regions of Europe, and in the most industrialized parts of the Central and Northern
Europe in terms of water resource degradation. The misuse and overuse of the resources have
resulted in long term damage such as groundwater contamination, both by polluting
concentrated sources, widespread agricultural pollution, heavy pollution of rivers and estuaries
and salt water intrusion into groundwater.
2.2. WATER DEMAND - EUROPEAN EXPERIENCE
All freshwater demands interfere with the reasonable maintainance of the right
ecological equilibrium: the terms of a sustainable development demand more cautious use of
the water.
NATO ASI Series, Partnership Sub-Series, 2. Environment - Vol. 15
Edited by C. Maksimovil:, F. Calomino, and J. Snoxell
© Springer-Vedag Berlin Heidelberg 1996
20
In the recent SAST Project n.6 of the UE (Schmitz, 1994) the following key figures are
given for UE countries:
• domestic water consumption 10 %

• industry (processing and cooling) water consumption 56 %
• agricultural water consumption 33 %
• others 1%
Groundwater resources represent 60010 of sources in public water supply system;

several member states depend on other countries for more than 50% of delivered water.
This global scenario must be verified case by case, in fact in Italy the 1990 updating of
the National Water Conference (1990) provided a different estimate:
• civil water consumption 15 %

• industry water consumption 24%
• agricultural water consumption 60%
• others 1%
Consumption value, disaggregated according to consumer category, have been also

quoted in Great Britain (Archibald and others, 1984) as follows:
• category of consumer (consumption) M1/day % of total

• households 6.000 54
• industry 3.050 27
• commerce and service industries 1.600 14
• agriculture 500 5
• miscellaneous 50 <1
As regards individual consumption, Archibald and others found an evident relationship

between this consumption and the number of householders. The range is from 140 Vp/day for
single person apartments, to 64 Vp/day in the case of more than seven inhabitants.
In order to justifY the different terminology in UE and Italian scenarios it is useful to
refer (Veltri and Maiolo, 1994) to a civil consumption scheme (Figure 1.), in which losses are
also quoted in evaluating the volumes to be delivered at the starting point of the drinking water
supply system. The term domestic is intended as that necessary for drinking, washing, cooking
and sanitary purposes.
21
The most relevant problem in assessing drinking water demand is that of finding the
necessary requirement, avoiding misuse and limiting losses both by end-users and water distri-
bution companies. In this respect it is possible to make analytical evaluations of water demand
in a civil context, even if this one is dependent on many factors, among which climate, water
consumption customs, continuous and unlimited resource availability, and last, but not least,
the water price and the adopted tariff system. Severe measures and controls of the delivered
water volumes and high water prices become the most significant factors discouraging water
misuse and loss. Figure 2. (Veltri, 1994) shows the average domestic consumption in several
countries, in which the average tariff (Figure 3.) plays an evident role. In any case comparisons
are difficult because some tariffs include both drinking and dirty water charges.
PROVIDED VOLUMES
MAIN LOSSES
NOT MEASURED VOLUMES 14---------1
Figure 1. Civil consumption scheme

22
The trend of domestic and total civil consumptions in the case of Italy is also shown in
Figure 4. (Maione and Orsi, 1990). It seems reasonable to assume that in the countries in
which new economic management entities are going to operate, and consequently higher water
charges will be applied, domestic consumptions will decrease.
300
c 250
mr.-
.2
~>; 200
.."
" co
cc:
8 ~ 150
.!l 8.
)~ 100
I I I I« I
50
0
:r: en ~ Z ....I ....I w z :r: u. o III
~
u o Z u::
Countries
Figure 2. Average domestic consumption in several countries, year 1986 (Veltri, 1994)
• I
• N
• __NL
-
•
_lIiIiiiil:"·~_
GB
-
'"' OJ
• B
0.2 0.7 1.2 1.7 2.2 2.7

Average tariff (DMARKlm',
Figure 3. Domestic consumption versus average tariff in some european countries

23
~r-----------fi·~~
;;;~~;;~~t-----------------------------,
350 .Tclal cMI
300
.~ ~ 250
15."
E~ 200
~ ~
8! 150
0_
100
50
o
1975 1980 1984
Year
Figure 4. Trend of domestic and total civil consumption in Italy (Maione and Orsi, 1990)
Based on a study of civil consumptions of many Italian centres a very simple formula
(Veltri and Maiolo, 1994) can be roughly used among the many existing criteria to forecast the
highest daily consumption:
in which: (Vp/day) is the specific civil consumption,

(Vp/day) is the specific domestic consumption,
Co (Vp/day) is the specific collective consumption,
P (Vp/day) are specific losses.
In the formula, Cd is weakly dependent on the city scale (it seems reasonable to assume
that it has almost the same value both in large and small towns); even if its value is affected by
the factors already highlighted, first of all water prices, it does not go below a reasonable
physiological treshold, which could represent the individual requirement (for water). Cc is
instead wrongly dependent on town dimension, in terms of public and collective water
demanding services; it is possible both to determine it from historical consumption data and to
assign an adequate value in terms of percentage of C,. Finally P is the losses term and it must
be assigned as accepted and not completely eliminable (it must be assumed to be not greater
than 10% ofC,).
24
If the water demand for collective usages is not available and a reasonable value of
specific losses is tolerated, it can be assumed that these are a percentage, a and ~ respectively,
of CI; by re-writing the formula, CI becomes:
in which Cd can be assumed according to either existing data or on the basis of analytical
summation of required usages, whereas a and ~ can be evalueted as follows: a between 0.10,
for small towns, and 0.30, for large cities (and even higher for the largest ones), ~ not higher
than 0.10.
2.3. REFERENCES
Archibald G.G., Kitson T., Waters B.D., Watson T.J. (1984) Consumer Requirements. In
Water Practise Manual, vol. 4, Water Distribution Systems. T.W. Brandon editor,
IWES, London, England.
Conferenza Nazionale delle Acque (1990) I problemi delle acque in Italia. Aggiomamento al
1989 dei risultati della Conferenza Nazionale delle Acque, Edizioni Agrico\e, Bologna.
Maione U., Orsi E. (1990) Consumi idrici delle realta urbane. Convegno: La conoscenza dei
consumi per una migliore gestione delle infi'astrutture acquedottistiche, Sorrento,
aprile, CDEN, Napoli.
Schmitz B. (1994) Research and Technological Development for the Supply and Use of
Freshwater Resources. SAST Project n. 6, EUR 15645 EN, Brussels.
Veltri P. (1994) Analisi dei dati storici disponibili. In: Rapporto sui Consumi Idropotabili in
Italia sui finire degli anni 80. A cura di P. Bertola e L. Taglialatela, Quademo n. 1,
CUEN, Napoli.
Veltri P., Maiolo M. (1994) Consumi annui dei centri urbani e turistici. In: Rapporto sui
Consumi Idropotabili in Italia sui finire degli anni 80, a cura di P. Bertola e L.
Taglialatela, Quademo n. 1, CUEN, Napoli.
3. Availability and Selection of Sources of Water for
Vujica Yevjevich
Thomas G. Sanders
USA
3.1. INTRODUCTION
Sources of water are systematically enumerated by the five major categories: (1)
surface, (2) underground, (3) atmospheric, (4) recycled, and (5) brackish or saline waters.
Each category of sources of water contains two or more specific kinds of water, with brief
description of each. A total of 21 kinds of sources of water are identified. Similarly, criteria
needed for selection of source(s) of water for the water supply systems are divided into the five
major categories. They are: (1) criteria which characterize the source(s) of water and water
itself, (2) regional characteristics of sources as related to the water supply systems, (3) water
rights and market for water, (4) development of the system in stages, and (5) economics of
water supply systems criteria of selection of water source(s). Each of these categories of
criteria contains two or more kinds of criteria, with their brief description. A total of 15 criteria
are identified. The principles of the application of criteria to selection of the source(s) of water
for water supply systems are reviewed in the third part of the text. The last part of this text
treats briefly the comprehensive and integrated development, conservation, control and
protection of all regional sources of water.
This text on the availability and selection of sources of water for the future water
supply systems covers first a section on the systematic listing of categories and kinds of water
sources, with a brief description of major properties of each category, and of each kind of
sources of water in a category. Examples of the categories of sources of water are: surface,
underground, atmospheric, recycled and saline waters. Examples of the kinds of sources of
water, in the category of the surface sources of water, are: springs, lakes, streams, reservoirs,
NATO AS! Series, Partnership Sub·Series, 2. Environment - VoL IS
26
ponds, wetlands, etc. The third section treats criteria for the selection of sources of water in
planning new or expanding the existing water supply systems. The fourth section covers the
principles of selection of sources, including the future factors that may influence that selection.
The fifth and final section is devoted to regional integration of all sources of water into a
comprehensive solution of water supply problems.
3.2. CATEGORIES AND KINDS OF SOURCES OF WATER, AND THEIR

PROPERTIES
3.2.1. Surface Sources of Water
The surface sources of water are characterized by an easy access for tapping them for
the water supply systems. Their pollution is easy also either by slow processes and accidents in
nature or by various human activities. The following kinds of surface sources of water are (see
the book by O. Starosolszky, Surface Water Hydrology, 1987):
3.2.1.1.Springs. Springs are natural outlets of different types of underground water

systems. They occur when the piezometric subsurface water table intersects the ground
surface. The average quantity of water and the time variation in discharge and quality of this
source of water depend highly on the type of underground system, especially for the springs of
karstified water-bearing formations. In most cases, springs are easily tapped for water supply
systems. The large majority of cities in ancient times were supplied by spring waters. The best
examples are the cities of the ancient civilizations around the Mediterranean Sea. Quantity and
quality of spring waters may change with time due to various natural or anthropogenic
change-inducing factors. The quality of spring water is usually suitable for drinking purposes,
provided it comes from the unpolluted underground water systems. Springs in most cases
should be of a high water quality. Unfortunately, pollution of underground water systems and
water itself leads also to pollution of springs, as the causal consequence.
3.2.1.2. Lakes. Lakes are likely to be the source of water with the second highest
quality of water for water supply systems, behind the springs. Under the natural conditions,
without the anthropogenic pollution, many lake waters are used by the native populations for
drinking without treatment. Their water budget, and level and volume of discharge fluctuations
determine the quantitative capacity for their use in water supply systems. Lake water is easy to
27
tap, and also in most cases easy to treat to drinking water standards. The trends in population
increase often lead to dense settlements around lakes with subsequent deterioration of the
quality of water, making more expensive water treatment necessary. To get acquainted with
the properties of lakes as the source of surface water for water supply systems, any good book
on limnology, the science oflakes, would be sufficient.
3.2.1.3. Ponds. Ponds are small lakes or small reservoirs, either natural or man-made
bodies of water, which usually serve a particular purpose (such as fish production, water
supply, irrigation, recreation, etc.). Water quantity and quality of ponds are limited, and in
many cases their waters need treatment to be a source of drinking water.
3.2.1.4. Streams (rivers). Streams are currently the most available source of surface
waters. Because of the large fluctuation of discharge with time, the tapping of water from
small rivers for water supply systems most often requires the building of reservoir(s) for
storage of water. This is most certainly true as the water diversion points for water supply
systems are very often located at the upper regions of the catchment where the fluctuation of
flow is the greatest. Because of pollution by the natural processes as well as by the
anthropogenic sources, river waters, nearly without exception, require treatment for drinking
water purposes. Any classical book on rivers and river waters (potamology) will enable the
reader to obtain enough information on them for the purpose of tapping this most frequent
surface water source.
3.2.1.5. Estuaries. The most downstream sections of rivers, at their entrances into the
seas or large level fluctuating salt lakes, may be mixed with waters of those receiving bodies of
water. In the form of estuaries they also represent a potential source of surface water. The
central problem becomes, how to avoid the mixing between the fresh and saline waters, which
produces a brackish wate~ unsuitable for drinking water purposes without desalination. By
proper hydraulic barriers, it is possible to avoid that mixing. The tidal wave propagation
upstream along the estuaries may also be eliminated by a hydraulic barrier in the form of a
weir. An example of such a structure is in the estuary of the Thames River downstream from
London. Estuaries may be transformed into reservoirs for water supply by similar weir barriers
or by dams.
3.2.1.6. Reservoirs. More than 250 thousand of all types and sizes of reservoirs exist
already in the world, of which about 55.000 are large and medium size reservoirs (according to
the International Commission on Large Dams), and the rest are small and very small reservoirs
and ponds. Many of them serve as the source of surface water for water supply systems.
28
Reservoirs have been and will continue to be one of the most important sources of water in the
foreseeable future. Settlement of suspended river sediment and temperature stratification of
water in the deeper reservoirs as well as the mineralization differences, make reservoir water
usually easier to treat for drinking purposes than the river water, particularly if the water is
drawn from the epilimnion (the surface layer of a stratified lake or reservoir). However, if the
water is drawn from the hypolimnion (the bottom layer), as is the case for many reservoirs in
the United States of America, the reverse may be true. The water may be highly mineralized,
turbid and without oxygen, necessitating treatment prior to most uses.
3.2.1.7. Wetlands and Swamps. Ifwetlands and swamps are permitted to be partially or
fully drained, they may be utilized as surface sources for water supply systems. Usually, the
preservation of wetlands stemming from ecological concerns, or their use for tertiary treatment
of wastes, may forbid the use for any other water resources purpose. They may be used for
smaller water supplies, especially if the maintenance of wetlands does not require all the
available water. The problem of fluctuating water quality will require treatment if water is used
for drinking purposes.
3.2.2. Subterranean or Underground Waters
Usually, these waters are called groundwater. The products of atmospheric

precipitation, which infiltrate into the different types of geological formations, create
underground water systems. The quality of these waters are characteristic of the geologic
setting and may require treatment before use in water supply systems. The exception to this
rule is the juvenile groundwater, retained in the soil and rock formations during their formation
(this water when used is classified as "mining water. ") The example of mining of this type of
juvenile water is the large water supply system in Libya, where such an aquifer in the desert is
tapped and water is brought a long distance to the coastal region, however not as a renewable
water resource. In general, the underground water resources can be of very good quality as
long as there are few sources of pollution. The costs of pumping water in many cases prohibit
the use for some price-sensitive water supply purposes. The variations in these underground
water systems are many (see the book by G. Kovacs and Associates, Subterranean Hydrology,
1981), and their kinds are:
3.2.2.1. Alluvial Aquifers. The permeable deposits of silt, sand and gravel, with an
adequate water supply, form most of the groundwater systems known as alluvial aquifers. They
are found everywhere but particularly along, on the sides and beneath the rivers as well as
29
along the lakes and seas. Their waters are usually clean and easy to tap for water supply. They
are available directly for drinking purposes, provided that anthropogenic pollution does nor
occur requiring water treatment before use.
3.2.2.2. Rock Fissures Water Systems. Many rock formations, either granular or solid,
have connected faults and fissures, generated by the orographic movements and processes, so
that they permit infiltration, accumulation and movement of water. These sources of
underground water are not simple to tap, except under particularly convenient circumstances.
Some of the best such systems are sandstones, with primary or granular porosity, and
faults/fissures system of porosity. This latter system permits a fast movement of water in or out
of the system, by slowly recharging or draining the significant primary, granular porosity.
3.2.2.3. Karstic Water Systems. The dissolution of carbonate rocks (limestones and
dolomites) and gypsum creates not only the specific surface forms of karstic type formations,
but also the systems of underground channels and caverns, thus creating the subterranean
system of channels with free turbulent flow or turbulent flow under pressure, or both. (see the
book by P. Milanovic, Karst Hydrogeology, 1981). These groundwater systems usually
terminate either with a spring or with many outlets of several springs. The karstic waters may
be of drinking quality during low flows, but often polluted during highest flows. This property
requires treatment of water during the high flows at the springs. In other cases, channels are
the part of the karstic aquifer, needing water pumping to tap it for water supply systems.
3.2.2.4. Volcanic Rocks Groundwater. The geologically recent volcanic rocks outflows
have created the underground porosity and channels (often called tubes), which are in many
aspects hydraulically and hydrologically similar to the karstic systems. Water from these rocks
are of much lower pH than water from the karstic systems which tend to be alkaline (high pH).
Water supply systems of the Hawaiian islands are the best examples ofthese systems, where
the horizontal drainage galleries produce substantial flow discharges. This drainage system by
constructing the long horizontal galleries are often referred to as the Maui type (after the
Hawaiian Maui island where it serves as the water supply system).
3.2.2.5. Groundwater Drawn in Directly From Surface Water. In many cases, the
groundwater tapped from aquifers are nothing more than the surface water (of rivers, lakes,
ponds, reservoirs), drawn in by operating production wells at relatively small distances from
the surface water sources. This source of water may then be considered as a combined
surface/subsurface source of water. The conventional surface water treatment costs can be thus
30
decreased substantially since the water is filtered through the bank alluvial or other porous
media.
3.2.2.6. Artificially Recharged Groundwater. The underground water which is

artificially recharged with surface waters using recharge wells, galleries, reservoirs, ponds or
irrigation system of ditches, sprinklers and land inundations, is also to be considered a
combined surface/subsurface source of water for water supply systems. The storage of water,
purification by filtration through porous media and sometimes the prevention of saltwater
intrusion are the reasons for recharge.
3.2.2.7. Juvenile Groundwater Aquifers. These aquifers are very old waters, locked
into the subterranean porous media during its formation. This water most often is highly
mineralized, which requires either treatment (usually desalination) or mixing with low
mineralized vadose waters in order to be tapped and used for drinking water purposes. This
water is most often used only for the limited time supplies, matching the life of water supply
user's needs with the durability of the available quantity of juvenile water source. As already
mentioned, an example of the use of this water is the tapping of the juvenile water in Sahara by
Libya.
3.2.3. Atmospheric Waters
The atmospheric waters are tapped directly, either as precipitation (rainfall and/or
snowfall) or as condensation of moisture from the humid air at the water collecting natural or
artificial barriers. These waters are low in dissolved solids but carry air-borne pollutants.
Atmospheric water may have low pHs, and is a source of usually small quantities of water. It is
stored in the specially built tanks, most often in the arid and semi-arid regions of the world and
on waterless islands.
3.2.3.1. Precipitation Collected on Impervious Surfaces. Precipitation on roofs, paved

collecting surfaces or other impervious surfaces are collected and stored. This water is
relatively aggressive, because of the lack of dissolved minerals. Many islands, especially those
composed ofkarstified limestones and dolomites, or the highly pervious volcanic rocks, in arid
and semi-arid regions, like at some Greek and Caribbean islands, are forced to use this type of
collection. This water is often a cheaper solution than bringing water by boat-cisterns. A good
example is at Gibraltar, where water was collected on a paved inclined surface in the times of
conflict between England and Spain for the ownership and control of the "Rock." The way
31
how this water is kept in storage, sometimes with crushed rocks (mainly carbonates), enables
minerals to be picked up as dissolved solids. Before water is used for drinking and cooking
purposes, it is filtered or boiled.
3.2.3.2. Directly Condensed Atmospheric Water. It is known that in the arid or semi-
arid regions the passage of the moist air over the cold elevated land forms often leaves
condensed water. This can be artificially enhanced, and small quantities of drinking water
obtained.
3.2.4. Recycled Wastewaters
Wastewaters from urban and rural communities, industries, mining and other uses may
be recycled as new sources of water. Different kinds of these waters are:
3.2.4.1. Recycled Urban and Rural Community Wastewaters. The modem technology
of treatment of wastewaters can produce clean waters that even drinking water quality could
be obtained. Psychologically, the solutions are such that the treated wastewater is traded with
the other drinkable waters. This may be the primary water source when no other source is
available. However, much effort must be taken as this source of drinking water has a high
potential to propagate, and transmit, the communicative diseases such as cholera, dysentery
and hepatitis. Therefore, as long as there are other, safer, sources of water, recycled
wastewaters should not be a source of drinking water. It is no longer a question of treatment
technology, it is a question now of reliability and psychology.
3.2.4.2. Industrial Wastewaters. The wastewater from industry may be recycled either
to be reused for the same purpose, or it can be treated to such a level that it can be used even
as drinking water. It will depend on which industrial purpose was originally served and how
expensive is treatment for recycling the water to serve the same or a new purpose. Cooling
waters may be so little changed from the original quality, that minimal recycling treatment
would be necessary for drinking water purposes, if the original water used for cooling was of a
good quality.
3.2.4.3. Mining Related Wastewaters. Basically three sources of contaminated water

come from the mines and mining processes. They are: (1) drainage water from the mine
galleries and shafts, either from the old, abandoned mines or from those in operation; (2) ore
processing waters; and (3) waters leaching pollutants from the mine tailings. All of these
32
sources can pollute the natural water courses and treatment or other expensive solutions
become necessary. Unfortunately, the costs to clean up these waters may exceed the value of
the mined element. This cost certainly appears to be the case for the costs to clean up the
impacts of the summit gold mines in Colorado. The value of the gold was estimated
approximately at 90 million dollars and the cost of cleanups in excess of a 100 million dollars
(McWhorter, 1992).
3.2.5. Brackish and Saline Waters
Close to seas, in estuaries, salt lakes, or even in some groundwaters, one finds waters
which contain too much salt or even some other undesirable minerals, that they can not be used
for drinking purposes without treatment. They are:
3.2.5.1. Brackish Waters. These waters are often encountered in springs close to the
sea in karstic and recent volcanic formations, in estuaries, close to salt lakes, and at the lower
layers of some aquifers. The cost of desalination is a function of concentration of dissolved
solids to be removed. Whenever one has to select between the desalination of brackish and
saline water, the former usually is a cheaper solution.
3.2.5.2. Sea Water. This source of water is inexhaustible for the coastal regions. The
major problem is the high cost of converting these waters to become drinking water. All
depends on the cost of energy. Many examples of saline water conversion plants, on
continents, islands and ships, testify that in the absence of other cheaper drinking water
alternatives, the conversion of saline water to drinking water becomes the right and sometimes
the only solution. Some conversion plants in the Persian Gulf Region are of the capacity of
several cubic meters per second. The major value of saline water conversion may be on the
ships that can move to and be used temporarily when conventional water sources are not
available, during and after floods, hurricanes, fires or earthquakes.
3.3. CRITERIA FOR SELECTION OF SOURCES OF WATER FOR WATER

SUPPLY SYSTEMS
The selection of water source(s) for a water supply system depends on the criteria
chosen for that selection. They are grouped in five categories which are:
33
(1) criteria which characterize the source(s) of water and water itself,
(2) regional geographic properties of the relationship between the water supply system and
the source of water (such as the distance, terrain characteristics, geomorphological and
geological feasibilities for creating storage reservoirs, etc.),
(3) water rights and the availability of market for these rights,
(4) feasibility for development of the system in stages as the function of the quantity and
quality of water in a changing water demand,
(5) economics of water supply criteria.
These five categories may be then classified in kinds of criteria, and discussed
appropriately.
3.3.1. Criteria which Characterize the Source(s) of Water and Water Itself
The major kinds of this category of criteria for the selection of the source(s) of water,
and related to the source itself, are:
3.3.1.1. Availability of Water. This criterion means how much and of what quality of
water each considered source has. The parameters of variables describing water of the source,
such as the mean discharge, the standard deviation, the periodicity and randomness describing
functions and parameters, the time variation functions (models) and their parameters, etc., all
are related to the availability of the water sources and may be used to describe these kinds of
criteria. For example, the mean annual discharge may determine the size of the population to
be served, while the standard deviation of annual flows is related to the storage capacity
necessary to meet the maximal needs of the served population. The same measures apply to
quality determining the amount of necessary water treatment. Simply, the state of these
hydrologic or hydrogeologic characteristics make up the science of water quality hydrology.
3.3.1.2. General Longevity of Water Supply Sources. This criterion requires that very
small chances exist for water of a source to be diverted or completely lost either by a natural
process or man-made actions. It is known from the historical events that earthquakes have
completely diverted water to go other directions, and sometimes created the new springes) in
34
large karstic formations. Even the upstream uses of water by humans often imperil the quality
of water supply downstream. In discussion of water conservation needs and measures, often
one could hear that "upstream water conservation may mean downstream drought." Here one
must distinguish and investigate both the natural processes of sudden or slow change, which
are not under human control, and those changes that are man-induced and man-controlled. In
fact, that is the major research challenge of the next century to distinguish those processes and
results that are man-induced and thus controllable and those processes and results that are not.
3.3.1.3. Potential for an Easy Water Pollution. It is known that various sources of
water for water supply systems have different potentials for pollution of water and the system's
environment. Examples of an easy water quality deterioration are many, since all sources of
water can be polluted. All surface and underground sources of water may be polluted. Karstic
water sources are among those which may be most easily polluted with difficulties to clean
them evident. The long history of attempts and costs of cleaning the polluted groundwaters
and their systems in the United States testifies that polluting water is easy, fast and long-
lasting, while cleaning the water environments is a long-range, difficult and costly endeavor.
Estimates as high as exceeding 300 billion dollars may be required to clean up known
contaminated groundwater sources in the USA (McWhorter, personal communication, 1994).
3.3.1.4. Access to the Source for Water Tapping Purposes. This kind of criteria is
basically reflected in the cost and safety of access to the source and the easy way of which
water can be tapped without the risk of either losing it or polluting it.
3.3.2. Regional Characteristics of Water Sources as Related to Water Supply Systems
Both the quantity and quality of water can be understood in many cases as a resource
of the region. Obviously, waters near the ocean and industrial centers will have different quality
characteristics and will have different economical values. The regional category of criteria
relates basically to the distance of the source of water from the center of water demand, to the
potential construction of the seasonal flow regulation along the connecting line between the
source and the center of water demand, as well as the local potential disasters threatening the
interruption of water supply. The kinds of these criteria are:
3.3.2.1. Distance of Source of Water From the Center of Water Consumption. The
distance of the source of water from the center of water use of a water supply system is one of
the major regional characteristics, which determines the cost and the reliability of connecting
35
the source to the system. Since the long distance requires the main conduit(s) to traverse
valleys and other geographic features, which may be disrupted or even destroyed during
floods, earthquakes and other natural or man-made disasters, the safety of connecting lines
becomes an important kind of criteria in selecting the source(s) of water.
3.3.2.2. Availability of Natural or Man-Made Water Storage Capacities. A

geologically sound and spatially attractive conditions for seasonal water flow regulation along
the connecting line from the water source to the water supply system, either existing or to be
constructed, represents an asset for sources of water. The natural assets may be existing lakes
and aquifers. The man-created assets may be reservoirs and ponds.
3.3.2.3. Availability of Elevated Terrain in the Vicinity or Within the Water

Distribution Network of Water Supply Systems. Each water supply system, which is fed with
nearly constant or low varying daily inflow from the source(s), requires daily water regulation
because of highly variable water consumption within 24 hours of operation. Any convenient
elevation of terrain enabling a relatively economical construction of daily regulation
reservoir(s), instead of building towers with tanks all over the plain terrains, is an asset to the
system. This is especially important if reservoirs also have the role of weekly flow regulation,
namely to store the surplus inflow water during lower water consumption over the weekends,
and to increase the corresponding average water supply during the working days.
3.3.2.4. Non-Aggressive Soils in Which the Main Conduits are Buried The connecting
conduit lines from the source of water to the water supply system's distribution network should
not be laid down in a highly aggressive soil, because of the corrosion of conduits or the
increased cost in special protection of conduits by the exterior coatings.
3.3.3. Water Rights and Availability of Market for Water
Water rights, their distribution and length of ownership will always play an important
role in the development of water supply systems. Obviously, to commit large sums of capital,
development system managers must have access and ownership for long periods of time. In
Colorado and much of the West in the USA, once obtained, the right lasts forever, while in
other regions of the world they are for a given time period. In either case the guarantee of
access is vital and must be institutionalized. Legal and market aspects of water supply systems
also impose some criteria in selecting a water source for that system. These kinds of criteria
are:
36
3.3.3.1. Availability of Water Rights and Their Protection. For a secure future water
supply, there should exist a governmental set up under which the owner or operator of any
water supply system must have assurance of their water rights of at least 50-100 years in
advance, with the guarantees for the future renewal or extension of those rights. If those rights
are marketable, they may be purchased or sold as economics and reliability of a water supply
system may require.
3.3.3.2. Availability of Market for Water. If there is a market for water, namely that
holders of water rights (public, semi-public and private institutions, or individuals) have the
right to sell partially or fully waters under those rights, it would insure that the highest need
users of water can apply the market mechanism and economics to procure the water. When
Thornton (a suburb of Denver, Colorado, USA) buys a ranch, not to raise cattle or cultivate
crops, but to procure ranch water rights for this city's water supply system, this becomes a
market vehicle to assure waters to the most important users. Unfortunately, having a free
market system for water using means that agricultural use cannot compete with the many other
uses. For example, an industry would have no problem paying $1000's for a unit of water (in
this case an acre-feet, or about 1200 cubic meters) while a farmer would have great difficulty
paying several hundred dollars for the same unit.
3.3.3.3. Availability of Water Market Among Countries. It is very delicate problem

when water from a water rich country has to be administratively decided to be diverted to a
water deficient country. However, if water is considered the market commodity, like
electricity, to be sold to neighbors for a price, often it would be easier to find a solution. If the
USA asks Canada to permit the diversion of water flowing into the Hudson Bay to the Great
Plains of the US West, it would be difficult to find a solution. But, if somebody offers to buy
that water, at a price per cubic meter or otherwise, definitely the Canadians would be at least
interested to listen. This will be analogous to the selling of electric power, like the Canadian
province of Quebec selling electric power to New England of the USA.
3.3.4. Development of Water Supply Systems in Stages
The development of water supply systems could usually be traced to minimal cost of
new water, either as the system extension by using the surplus water of the already used water
source, or by tapping the new source(s) of water. This extension most often starts with the
nearby surface waters, then with subsurface waters, interbasin transfer waters, and finally
recycling wastewaters. Most sources of water for the major urban areas are nearby surface and
37
interbasin surface waters, while many smaller urban areas as well as rural areas use the nearby
surface and subsurface sources of water. Recycling wastewaters is becoming a very viable
option, especially in water-short California, where recycled waters are used for recharging
aquifers and for irrigating crops that are processed before being consumed. With the advanced
technology of recycling and recharging wastewaters, and their reliability, the re-use of waters
will become more widespread as it is the last economically feasible option as a source of new
water in many regions of the world. A feasibility to expand the water supply systems in stages,
as well as the procuring of sufficient reserve water sources for the future, require some criteria
to be considered in selecting a water supply system alternative. These kinds of criteria are:
3.3.4.1. Feasibility of Development of Water Supply Systems in Stages. Two factors

affect the future growth in water demand, the increase of population to be supplied with water
by that system, and the increase in standard of living of water users. This is usually translated
into a larger daily average consumption of water per inhabitant of an increased population.
Therefore, each system must be planned to be developed in stages, with each new stage also
improving the performance of the previously developed stages whenever possible or feasible.
3.3.4.2. Procuring Water for Future Demand by New Water Rights. The wise policy in
planning future developments of water supply systems in stages is to insure that sufficient
sources of water for future demand will be available either through additional water rights, or
by the purchase of water already allocated to other water users or purposes. The projected
demand for water of 25-75 years, regardless of the low accuracy of that projection, will well
serve the policy of insuring the future system's expansion. The city of Fort Collins, Colorado,
USA, has accumulated enough water rights to meet the projected water demands in 2020.
Other cities like Thornton, Colorado, mentioned previously, are securing water rights and
designing water systems to meet their water demands at the tum of the century.
3.3.5. Economics of Water Supply Systems
Economics of a water supply system is important source of criteria in selecting the best
alternative of that system. However, as households and other users of water may tolerate a
relatively high price of drinking water, in comparison with the economics of the other purposes
in developing water resources, the water supply systems are less affected by its economics than
other water resources purposes, so the other criteria may have the same or a much higher
weight that in the case of other water resources development purposes. The economic kinds of
criteria are:
38
3.3.5.1. Measures of Economic Efficiency. The price of delivery of a gallon or cubic

meter of water to users is the major measure of economic efficiency, especially when compared
with the same prices of other water supply systems. The major measure of economic efficiency
of the other water resources purposes (irrigation, hydropower production, navigation, flood
control, recreation, etc), are most often measured by the benefit/cost ratio. This index of the
purpose of water supply systems is not applicable, since drinking water unit price can be high
and still people would accept it, if there were no other more attractive alternative. This unit
price of water is usually adjusted for inflation with the time. The price serves as one of the
criteria in selecting the best water supply system in the case of several available alternatives.
3.3.5.2. Upper Limit Cost of Unit Water Volume as a Criterion. Nearly any water
price will be paid by people for drinking purpose if there is no other source of water. So, in
planning water supply systems there is usually an upper price limit which in each case should
not be exceeded. One type of these limits is the cost of good bottled drinking water which will
be surely available on the market if that limit is exceeded. The other limit may be economically
imposed, since the high cost of utilities in general decreases the standard of living in
communities, and the future growth of communities may be affected. The job creating
businesses may be shying away from communities with the exorbitant prices of utilities in
general, and of water in particular. The third factor which imposes a kind of upper price limit
on the unit volume of water is the tendency by the lowest economic strata of people to tap any
available but cheaper local source of water, usually water found in communities themselves.
Such cases are the excavating or drilling a well close to rivers or underlying aquifer, with that
water being in principle of lesser quality, and less controlled by professionals, than water from
the well designed, constructed, operated and maintained water supply system.
3.4. SELECTION OF SOURCE(S) OF WATER FOR A WATER SUPPLY SYSTEM
3.4.1. Use of Single or Multiple Selection Criteria
It is simpler to use only one criterion in selection of an alternative for a water supply
system, than to embark on all intricacies of application of the multiple criteria of selection. To
do that, all other criteria beyond the selected single criterion should be reduced to their effects
on this selected criterion. In many water resources plannings and designs, the single criterion is
usually monetary, say in the form of benefit/cost ratio or the difference between the expected
39
annual benefit and the expected annual cost, in the case the average annual values of benefit
and cost are being compared and used in decision making.
Many criteria described in Section 3. of this text may be partially or fully expressed in
the form of their influence on the cost of water supply system, or as the contributor to the cost
of unit volume of water delivered by the system. For instance, criteria of the impact of distance
of water source from the center of water consumption, as well as the cost of seasonal, weekly
or daily flow regulation either at the source itself, or on the line connecting the source and the
system, or in the area of water consumption, can be expressed as the contribution to the cost.
Monetary effects of some other criteria of Section 3. may be only approximately estimated.
The monetary effects of the criteria related to the quality of water may be also determined by
using the cost of treating water to the given water quality demand standards.
However, some criteria, especially related to availability and risk, may not be accurately
applied through their monetary effects. Their influences on the selection of a water supply
system alternative may be only feasible through professional judgements. An example would be
the selection of a water supply source which can be controlled both in water quantity and
quality by another country, but is much cheaper than an alternative.
3.4.2. Selection of Water Source or Several Water Sources
The single economic criterion, used for the selection of the alternative in cases where
the planning of water supply system comes out with two or more alternatives, may decide in
the selection of its source(s) of water. The minimum cost of delivered water of the given
quality, per unit volume, rather than the application of all other criteria, may help decide on the
final alternative. However, if the other criteria, especially the availability of water, reliability of
supply, risk of major accidents and damages by disasters, are considered important, the
professional judgement and the best outlook for the system's maintenance, operation, extension
and future renovation may prevail in the decision, especially if the cost difference is not
substantial, say not more than 5-20 percent of the cost of the cheapest alternative.
If the multiple criteria are used in selecting the best alternative, criteria may be
compared by giving each a weight in the analysis of each alternative. The maximum total
weight may give the preference to that alternative. Or, the other, already classical multiple
criteria approaches of decision making may be used. The differences in the quality of supplied
water between alternatives may be weighed heavily in comparison toward the minimum cost
criterion.
40
3.S. COMPREHENSIVE AND INTEGRATED PLANNING FOR DEVELOPMENT OF

ALL SOURCES OF WATER
The comprehensive water resources planning for utilization of all water resources in a
region includes all water sources above described by using all the above outlined criteria. In
that case the present and future water supply systems must compete for the available water
with all other sources of water, however giving always some kind of preferences to the
drinking water purpose. In most cases, but particularly in the developed countries the volume
of drinking water is only a small percentage of all available sources of water (less than 1% of
per capita water use in the United States of America). If it doesn't exceed a couple of percent
of all regionally available waters, except in arid and some semi-arid region, drinking water has
always preference.
The integrated water resources planning means the inclusion of all the interests and
purposes in the development of water resources in the region. In that regard, the answer to the
question "what kind of water to what purpose" becomes crucial in this comprehensive and
integrated planning of water resources development, conservation, control and protection. This
approach to regional planning makes water supply systems only a subsystem of a system of
regional water resources. This fact often does not sit well with the leaders of the urban water
supply systems, which by the importance of drinking water always tend to be autonomous and
independent of the solutions proposed by the plan for all other water resources purposes.
Though there is something vital in this attitude, economics, safety and quality requirements of
water supply may be resolved adequately by the regional comprehensive and integrated water
resources planning and resulting schemes.
The important aspects of this approach are:
(1) The planning of water resources development has passed many phases in the last
100-150 years. From a concept of single structure for a single water resources purpose
and a single source of water, the planning has evolved to the multi-structure,
multi-purpose and multi-source of water,
(2) The planning is more and more regional or basin-wide, so that the selection of the areal
horizons become unavoidable, such as dividing the areas of the physical limits in
planning, then the adjacent areas to them with which there will be an exchange or only
one way transfers of water, and finally still a wider area which affects the plans by the
influence of joint demand and the fulfillment of that demand,
41
(3) Planning has several time horizons to look at, such as the time horizon of the economic
life of a system, the time horizon of its physical life, horizon of its life of obsolescence,
and eventually the horizon imposed by the eventual climatic changes.
3.6. CONCLUSIONS
The following three major conclusions may be drawn from the above discussions:
(1) The selection of the source, or the sources of water for the urban or rural water supply
systems is very complex process, because the quality of water for drinking purposes
imposes several strict criteria to fulfill, that many water sources can not meet
economically,
(2) With the above classification of five categories, and within them of 21 kinds of sources
of water, a systematic approach to the investigation of all sources of water, and the
eventual combination of those sources for water supplies, are indicated, in a
comprehensive and integrated water resources planning for their development,
conservation, control and protection,
(3) With the above classification offive categories, and within them of 15 kinds of criteria
for judging the suitability of the sources of water and their selection for the water
supply systems is facilitated, especially in the complex regional planning of water
resources schemes and in the allocation of water rights for the required quantities and
qualities of the available sources of water.
3.7. REFERENCES
Kovacs G. & Assoc. (1981) Subterranean Hydrology. WRP, 970 pages.

McWhorter D. (1992) Profession of Agricultural and Chemical Engineering. Report, Colorado
State University, Fort Collins, Colorado, USA.
McWhorther D. (1994) personal communication.
Milanovic T.P. (1981) Karst Hydrogeology. WRP, 434 pages.
Starosolszky O. (1987) Applied Surface Hydrology. WRP, 821 pages.
4. New Concepts for Water Supply Systems
* What Kind of Water for What Purpose in the Future *
Vujica Yevjevich
USA
4.1. INTRODUCTION
The most basic concepts in planning and design of water supply systems are already
well shaped and currently widely used. The unique system provides the drinking quality water,
regardless that many users of high quantity water would tolerate water of lesser quality.
Sometimes users of large quantities of non-drinking water quality may construct a separate
system, as an economical solution. Regardless of the availability of advanced technology,
research and development in many areas promise the planning and design of better water
supply systems in the future. These research and development areas are: safety, technology and
economics of interconnecting the adjacent water supply systems; planning systems renovation
in advance; design of new water quality standards based on advanced research; role of water
quality and composition on human health and nutrition; use of new materials in construction
and operation of systems; 'incrustation, erosion and corrosion in water supply systems; new
materials for pipes and conduits; etc. In deciding what kind of water for what purpose may be
used, distinction should be made between using the natural untreated good quality waters and
the treated waters to various degrees of quality.
4.2. CONCEPTS FOR DESIGN AND PLANNING OF WATER SUPPLY SYSTEMS
4.2.1. Main Existing Concepts
The main concepts which are currently applied in design and planning of water supply
systems are:
NATO AS! Series, Partnership Sub-Series, 2.Environment - Vol. 15
44
(1) Water demand, both by its quantity and quality requirements, is determined as
accurately as the available information permits. This is done either for a new or an
expanding water supply system. It is usually stated in the form of parameters of the
required water demand variables, as these parameters change with time,
(2) All sources of water within, or in the vicinity of, the area covered by the water supply
system are considered and investigated as the potential sources of water in planning a
system or in a system's expansion,
(3) Criteria are selected and applied for the evaluation of the alternatives of water supply in
planning the new or expanding an existing water supply system,
(4) Major criteria applied to water supply systems are: the availability of water, its
reliability of supply (time of access, assured water rights, etc.), water quality, cost of
the unit of water supplied by the system, and the future potential for system expansion,
(5 Safety of the system as a whole against various types of water pollution as well as
natural or man-induced disasters, and the related overall water quality, must be
acceptable to both the decision makers and the water users,
(6) Treatment of water to meet the quality requirements, the pollution-free transport and
storage of treated water, and the distribution network are designed and planned
according to the prevailing technologies and the best engineering practice at the time of
decision making.
4.2.2. Composition a~d Characteristics by Quality and Quantity of Water Users in

Most water supply systems, especially the urban ones, have a variety of water users,
often with requirements of quite different water quantity and quality characteristics. The major
concept used for the contemporaneous water supply systems is that water treatment must meet
the criteria of the user of the system with the highest water quality requirement. Because
drinking water is usually the use with the most stringent water quality requirements, all other
users which can tolerate the lesser water quality are generally supplied with, and use, this
drinking quality water.
45
As an example, watering of lawns, shrubbery and trees around houses, in parks and
around the public buildings and facilities in most urban water supply systems uses a much
better, treated water than is necessary for such irrigation. And, this is often the user of the
largest quantity of water of the system. This concept, that all water users of a system, receiving
water of the user with the highest quality requirements, may be designated as the unique
network water supply system. This means there is only one water distribution network of
conduits and pipes, conveying the same quality of water for all users, regardless of their
differences in required quantity and quality of water. This unique system imposes a possible
waste of resources in overtreating water for many large urban or rural water users. Also, due
to the uncertain quality of an untreated or insufficiently treated water, which may come into
contact with the general population, the water supply system decision maker, trying to avoid
public controversies, tends to prefer a single, high quality water supply system.
A case of departure from the rule in planning the general unique water supply system is
when a large water user finds it economically and safety-wise acceptable to have a separate
water supply system in a part of area covered by the unique water supply system. Such an
example is in Fort Collins, Colorado, USA and the Colorado State University. The city is
supplied with treated water from the Horsetooth Reservoir. However, irrigation of the large
area of the Campus of the University is supplied directly from an untreated reservoir water,
because it was a cheaper solution to have a separate system for that use. It was, therefore,
considered that little risk for water-generated health problems existed using the reservoir water
for watering of Campus lawns. This situation is not more likely to be a risk than is recreation
(boating, swimming, picnicing, fishing) at the same reservoir. The city, however, often uses
treated water for irrigation.
4.2.3. Ongoing Technology and its Potential for Improvements
In principle, one rarely encounters ongoing technologies of the complex infrastructural

systems, like the large urban water supply systems, that research and development could not
improve on the existing or create new technologies, which can successfully replace the existing
ones. So, both the new concepts and the new technologies are worthy searching for in order to
improve the water supply systems of the 21st century.
The goals of this research and development are versatile. They are:
(1) To increase the availability of water sources for a system,

46
(2) To make reliability of supply as high as the needs and economic criteria would require
or permit,
(3) To continually increase work on water quality research and development, especially on
the effects of water quality on humans, with new findings expected,
(4) To decrease the costs of water of the prescribed drinking water standards,
(5) To increase the potential for relatively economical expansion of existing water supply
systems,
(6) To better plan and design how to overhaul and innovate the aged water supply systems
in the future,
(7) To improve the monitoring of water supply processes and better maintain water supply
subsystems; and similar goals.
4.3. FUTURE ALTERNATIVES FOR WATER SUPPLY SYSTEMS
4.3.1. Basic Dilemma in Selecting the Type of Water Supply System
The application of the principle that each water user in a water supply system should
obtain the quality of water which meets only the necessary quality requirement of that user,
means that the unique. system of water distribution network must be changed, modified or
abandoned. The alternatives are either two or more combined water supply systems, each
serving the particular user, or a group of users, with the required water quality standard. The
highest quality water user will be in the system supplying the drinking and cooking water for
the population. The lowest water quality user happens to be usually the bulk large quantity
water uses, such as cooling water, water for irrigation, water for washing of streets, and other
public or private large quantity water using facilities. However, it must be noted that many
industrial processes (food processing industry included) may require higher quality water than
drinking water.
In the middle of the quality scale will be water quality needed for the basic, human
hygienic needs. In a two-pronged water supply system, water needed for hygienic purposes
47
(bathing, washing, toilet, flushing, etc.) will be served by the bulk water of lower quality,
provided this quality is sufficient for those purposes. However, drinking water quality will be
used if the quality of the bulk water supply is not acceptable for human hygienic purposes.
The immediate reaction of planners, designers and decision makers to such a multiple
network water supply system alternatives would be that the increased cost in the double or
multiple pipe networks will have to be compared with, and be smaller than, the additional cost
of treating all water to the drinking water standards in the unique network alternative. This
comparison by the economic analysis can be easily done since the cosst of treatment of water
and the cost of construction of networks are readily available.
4.3.2. Alternatives for Future Water Supply Systems
The investigation of new alternatives of water supply systems with two or more
separate pipe networks, one for each major, particular user, or the groups of users, faces
several dilemmas. For example, should the best drinking water be of the same quality for all
people? It is known that people have different acidities of the stomach juices (hypertonic,
neutral, hypotonic, or of several more classes instead of only these three classes). Some may
prefer pH of the water that best corresponds to their health and stomach tolerance, and all of
them can not satisfy their preferences with the same drinking water quality at the same time.
Only the industrially processed waters may produce several desired types of water, each of a
different pH and different mineral composition, to be sold in food and/or beverage markets like
the varieties of food selection that are offered currently. The alternatives of the future water
supply systems, especially of the large and medium size systems, are:
4.3.2.1. Unique System. That means to continue planning the unique water supply
systems, for which all water is treated to the prevailing drinking water standards.
4.3.2.2. Variety of Drinking and Cooking Water Provided at the Market. These types
of water are provided by the market oriented to production of drinking water of various
compositions and pH values, which can be regularly purchased in the food and beverages
stores, and then to use the unique water supply system, with the quality of treated water to
satisfy the people's needs for human hygienic purposes.
4.3.2.3. Dual Water Supply System. The dual system of water distribution networks
will mean that two pipe networks are laid into the same ditch or gallery, with the smaller size
pipes of the network serving the drinking, cooking and hygienic human purposes, while the
48
larger size pipe network would serve all the other purposes, especially the users of the large
quantities of water, with water treated only to be disinfected (safe from bacteria) and sterile
standards (safe from viruses and protoviruses).
4.3.2.4. Multiple Network Water Supply Systems. The multiple network water supply
systems will have three or more networks of pipes. Apart from the network for the users of
water of drinking, cooking and hygienic quality standards, the urban irrigation, various
industrial water users, plant cooling and other specific purpose uses, would have their
particular networks or the individual network of pipes, while all other remaining uses as a
group will form the separate pipe network. Then all these waters of different quality would
each meet the water quality requirement of the use with the most stringent standards in each of
the separate groups of uses.
4.3.3. Specific Topics for Investigation, which may Improve Future Water Supply
Systems
There are several aspects of water supply systems which deserve continual active
research and development, as well as a better attention and efforts in operation and
maintenance than was done in the past and at present. They are:
4.3.3.1. Interconnection of Adjacent Water Supply Systems. Many large water supply
systems are often close by distance, or the medium or small systems are located in between the
two or more large systems. With the relatively short conduit interconnections, systems may be
connected in such a way that water can be exchanged or shifted between the presently separate
systems. This is not only important for the potential matching of surplus and deficit of water
which occur in the interconnected systems, but also for the solution of the problems of drought
deficit. This is analogous to the interconnected electric power grid systems. The more distant
the geographic centers of water of the corresponding sources of water (rainfall, surface and
subsurface sources of water), the smaller the probability that drought will cover both system
sources with the same proportional deficit. This concept should be compared with the
alternative concept of solving the drought problems by large reservoir capacities in both
systems for the seasonal and/or interannual flow regulation. An example of such
interconnection of the large systems may be the relatively densely populated urban areas
between Boston and Washington D.C. in the United States. Many separate large and medium
water supply systems between these two cities (including New York City, Philadelphia and
Baltimore) can be easily interconnected for the benefit of all of them. A major problem still
49
remains when one system produces a better quality water than another, even though both may
meet the existing EPA (US Environmental Protection Agency) drinking water standards. An
example of this is the quality of water of the city of Fort Collins, Colorado, and a water
company nearby. The latter wants to be interconnected but the city of Fort Collins does not
want the connection since the other water company does not produce as good a quality of
water as the city does. However, if safety of water supply in emergencies and in droughts are
the major reason for interconnection, the mixing of different quality of drinking water may be
temporarily accepted.
4.3.3.2. Planning in Advance of System Renovation when Aged, or Deteriorated, or

Obsolete. The ongoing renovation or overhauling of many large water supply systems in the
world, especially those with deteriorated pipe networks, pose a simple question, namely how
many of the existing water supply systems in their original planning and design have taken into
account the fact that these systems will definitely age and will need renovation, or also when
obsolete, be replaced? Therefore, one can justifY future studies, with research and
development, of these problems in order to incorporate into planning and design of the new or
renovated systems also how to fight most economically, the aging process and when
unavoidable to renovate or replace the systems.
4.3.3.3. Definition and New Standards for Water Quality. Regardless of the
tremendous efforts to define water quality in general, and for the specific users in particular,
the fact is that several hundred chemical compounds and dozens of biological species may be
found in water. Apart from individual effects of these compounds and species, there are also
many synergistic effects by which two or more compound or species have a larger effects on
the water consumers than the sum of the effects of the individual compounds or species. As an
example, recent research has shown a link of cancer in the stomach and intestines to the
consumption of chlorinated water making a need for a better monitoring or disinfection of
water, which do not contribute to increases of cancer (Canter, 1992). A huge potential is
available· for the systematic and long-lasting research and development, uncovering many
synergistic aspects of water quality. This may change the ongoing definitions of water quality
for the water supply systems and lead to the new, significantly better water quality standards.
4.3.3.4. Water in Human Health and Nutrition. It is well known that quality of water
affects human health. Otherwise, there will be much less international and national efforts in
the world, to supply water for drinking purposes, especially in developing countries of large
populations. Less is known on how qUality of water affects nutrition of people. Definitely,
minerals and eventually benign biological matter, brought by water into the human bodies,
50
must play some role in nutritional aspects, if not also in human health. There is a speculative
theory, that water from rocks of the subsurface water systems brings with it some very rare
elements, which serve as catalyzers (enzymes) in biological processes, especially in
reproduction of cells. Waters from alluvial aquifers have less of those rare elements because
the river water which deposited sediments of alluvial aquifers have washed out those elements
into the sea. Research and development on the role of quality of water on human health and
nutrition, likely will be a subject of many projects and new research approaches in the future.
Maybe, that human longevity partially depends on these quality variables.
4.3.3.5. Use of New Materials in Construction of Water Supply Systems and in Their
Operation. Materials used for water supply networks, pumps, water treatment structures and
materials used in treating water, may affect the quality of water, sometimes decisively. The best
example is the use of lead and copper pipes in water systems, even for the short connections
and the ends of pipes bringing drinking water into households. Even a speculative theory goes
with it, namely that the Roman Empire may have partially fallen due to the use of lead pipes.
During that empire, water in water supply systems was mainly brought from springs to cities by
the free-flowing water aqueducts. Only rich and politically leading and influential people could
have afforded in houses and restricted public bath houses the internal water supply system
under pressure with lead pipes As a poison, lead is a slow destructor of human tissues.
Regardless of the speculative character of this theory, the fact that the lead and copper pipes
are being removed from water supply systems in many countries, as dangerous material,
testifies that the theory may have some validity. Also, the use of the chlorine in treating the
drinking waters is not without controversies, because of the creation of its compounds which
are related to cancer increases.
4.3.3.6. Incrustation, Erosion and Corrosion. Many water supply subsystems,

especially their pipe networks, may exhibit three phenomena: (1) internal incrustation,
particularly those which deposit calcium carbonates at the internal walls of pipes, with a
decrease of effective flow area and an increase of roughness and head losses within the main
conduits and pipe networks; (2) erosion of the interior pipe surface when the untreated water
carries fine sediments, especially the fine quartz grains; and (3) corrosion of pipes either from
outside or inside if water or soil are chemically corrosive, with this phenomenon most often
occurring on outside rather than inside of the pipes. Research and development need to help in
a better understanding of these three phenomena which decrease longevity of pipes and
networks in water supply systems and decrease capacity for conveying waters. New pipe
materials or new protection coatings, both inside and outside, may prolong the life of the
systems and make their maintenance and operation more economical.
51
4.3.3.7. New Pipes and Main Conduits. The science of behavior of materials under
various conditions is progressing rapidly by the end of the 20th century. It is likely that this
progress will be maintained also in the first part of the 21st century. Therefore, special research
and development on the new pipes and new conduits may help both the new and the renovating
water supply systems. The stress in these special investigations should be on both the
advantages but also the eventual disadvantages of any new material, being it in plastics,
laminated materials or others, or insuring the anti-erosion, anti-corrosion and even eventually
anti-incrustation properties of materials or coatings.
4.4. WHAT KIND OF WATER FOR WHAT PURPOSE IN THE FUTURE
In the treatment of the topic what kind of water to be allocated or used for what
purpose in the future, a distinction may be beneficially made between the natural untreated
waters of satisfactory qualities, and the treated waters to various degrees of quality, even
pushing the treatment to the point of "manufacturing of special types of quality water for
drinking purposes" by the industrial processes.
4.4.1. Natural Untreated Waters of Good Quality
The basic premise which must be made in using natural untreated waters for any
purpose is that natural waters may be and will be protected from a substantial deterioration by
pollution. Or, if these waters become polluted, that there will be the means and administrative
or legal vehicles to force P911uters to return the quality of these waters to the pristine natural
original conditions.
In matching the users' demand for water quality and the quality of natural waters,
historical developments and the recent experience may be the most useful guides in this task.
The best quality water in nature, originally being with the springs and groundwater aquifers
(used by excavating or drilling wells and galleries such as the ancient and even recent
Persion-invented kanats), should be reserved primarily for drinking purposes. They need safe
tapping, protected from any kind of local pollution. Because originally the quantity of these
waters often exceeded the needs for drinking purposes, many spring and aquifer waters have
been used for irrigation in agriculture, or for industrial, cooling and mining processing and
washing purposes. With the increase of population and their standard of living, with both
52
factors requiring more water of the best quality, pressures are usually exercised to use more
spring and aquifer waters for drinking purposes, taking them away from agriculture, industry,
cooling and mining purposes. Wastewaters from urban and rural communities, return flows
from irrigation, wastewaters from industry, mining, cooling and other uses, a\I may be changed
or deteriorated by use, either physically, chemically or biologically, or by any combination of
them. This fact usually and in general threatens the quality of all spring and aquifer waters.
The recent experience has shown also that protection of the quality of naturally very
clean drinking water is not an easy assignment. The Rocky Mountains "clear waters",
advertised by a beer producer which uses those waters in beer production, may be now
questioned. The fumes from copper smelters of Arizona, fumes from the coal-fired power
plants of the Southwest and West of the USA, fumes from automobiles of Southern California,
are all carried by the southwesterly and westerly winds over the Rocky Mountains. As these
westerlies produce 6-15 snow or rainfall storms over the Rockies during the winter season, a
lot of those air-borne fume materials are deposited over the Rockies. In some way, one can tell
"good-by to the clear Rocky Mountains waters." Combined with the pollution by the mines of
small rivers in the Rocky Mountains region (old mines drainage, leaching from tailings, and
processing of ores of the ongoing mines), it is easy to demonstrate that the preservation of the
quality of natural pristine waters for drinking purposes is a Sisyphusian task in the modem
intensive agricultural, versatile industrial and broad mining activities.
The altered waters of springs and aquifers may become unusable for drinking purposes
by the broad and versatile sources of pollution, especially by the chemical pollution. That
pollution may not impede the use of those changed waters for agricultural or urban irrigation,
for industries and mining operations, which tolerate the less quality waters than required by the
drinking water purpose. If those spring and aquifer waters are first used in mining or in
industry, and if their wastewaters become further polluted by these uses, they may not be
suitable even for agriculture irrigation. Therefore, the priorities in using the present or
previously pristine natural waters should be:
(1) for drinking water purposes,
(2) for special industries requiring water of very good quality,
(3) for agriculture irrigation or irrigation in urban areas, and finally,
(4) for many types of general industrial and mining water uses, which
tolerate water oflesser quality.
53
The difficulties in carrying out this priority sequence arise from the facts that major
mining water users are in mountains where mines and pristine waters are located. Industries
and urban water users are often located in the wider downstream valleys, and irrigation users in
the plains, usually in the lowest sections of rivers. This sequence of the location of water users
does not follow, or coincide with, the sequence of the quality of water along rivers. Ideally, the
pristine waters should be used first for urban and rural water supply systems as well as for the
special industrial purposes requiring waters of a particular good quality, followed by
agricultural and urban irrigation, and finally by the industry and mining tolerating in their use
the waters oflesser quality (washing, cooling, ore separating, etc).
4.4.2. Altered Waters after Various Uses
Once water is used, it is usually either discharged into the rivers and underground, to
be taken again later downstream by the next user, or it will be passed to the next user directly
from the outlet of the previous user. Ifwater is discharged into the rivers to flow freely along a
river section and for a time, the natural purification capacity of rivers or underground water
systems can be utilized to improve on water quality. Unfortunately, due to the limited power of
this natural cleansing, either biological through the natural oxidation process of biological
matter, or chemical by dilution of their concentration in river waters, the ability of many rivers
has been exceeded many times over. This has caused massive pollution control problems
especially during the low flows and the highest yearly temperatures. Also, it created the
awareness and policies that rivers should not be considered in the future as organic and
chemical waste transport and disposal sinks. If water goes directly from a user to the next one,
planners and designers must cope with the altered water quality and changes in its quantity
made by the previous user, when that water does not meet the water quality/quantity criteria of
the next user. All this creates the interdependence among water users in a river system. The
users can be a group or individual uses, cities, or even countries. Thus a chain of sequential
uses results, which require most often water treatment between these uses.
A classical example is the Ruhr Region of Germany, with the chain of uses and
treatments of water in between them, so that symbolically people used to say "that the Ruhr
water passes through seven stomachs." However, because of the inflows of both natural and
the used water in between the intakes of water for these users, water for any user is more a mix
of the used and new fresh waters than only the upstream wastewater. Thus water is not only
easier to treat, but also psychologically it does not imply that only the upstream wastewater is
being used downstream after treatment.
54
4.4.3. Matching Kinds of Water and Water Use Purposes
When one considers what kind of water should be used for what purpose, first it is
necessary to take into account the original characteristics of the source of water, before any
anthropogenic changes in water quantity and quality have been made. Second, one must
investigate whether the natural water source may be economically returned to its original
pristine state. Third, if water quality may be returned to its original pristine state, and
eventually also improved, the question to answer is whether the new characteristics of that
source of water may be effectively protected from various sources of pollution in the future.
The major sources of water meeting these properties are springs, aquifers and lake and
reservoir surface waters. All other sources, such as direct use of the atmospheric water,
recycled water, conversion of brackish and saline waters, may be considered as special cases to
be considered if the major sources of surface and subsurface waters are no longer available.
4.4.3.1. Drinking Water. The original supply of ancient human communities has been
first oriented to the use of water of perennial springs. They required aqueducts, usually with
the free surface water flow, from the spring(s) to the water fountains inside the community.
When low flows of these springs no longer satisfied the need for water, the next source and the
priority in tapping water sources were the aquifers. They required usually the digging of a well
or several wells, and then the power to lift water from the well (human, animal or wind lifting
powers). In modem times, when the spring and aquifer sources of water have been already
tapped for water supply systems, the attention was passed to the lake and reservoir surface
waters. This latter alternative in most cases required treatment of water to be bacteriologically
disinfected and free from viruses. In recent times, and due to an easy pollution of many types
of springs and aquifers, the attention was paid to lake and reservoir waters. The special value
was for the upper most reservoirs of smaller rivers, where no intensive agriculture, water
polluting industries and mines existed. Grazing and silvaculture, with timber production, and
the eventual recreation (wilderness areas, national forests, national and local parks and
protected areas), enable an effective protection of water from high chemical pollution (with the
exception of air borne pollutants), so that the necessary, classical water treatment is relatively
inexpensive. When none of the above sources of water could satisfy the new water demands,
especially in semiarid and arid regions, the drinking water may be directly collected on paved
surfaces from the rain, of by converting brackish and/or saline water into the fresh water. The
recycled wastewater usually is not directly used for drinking purposes. If that water is
recharged into an aquifer, it may be pumped later and used as a source of water (with the new
eventual treatment) for drinking purposes.
55
4.4.3.2. Water for Human Hygienic Needs and for Recreation. Since these uses of
water have a less stringent water quality requirements than the drinking water, many surface
and underground sources of water may satisfY these quality requirements, usually with the
minimal treatment (say, by chlorinating such waters). Everybody knows that the swimming
pool water is most often chlorinated to meet the minimal quality standard for such water use.
Surplus spring and aquifer waters are directly used without treatment, even in the case of
karstic springs which may carry turbid waters in floods. Lakes and reservoirs are the main
sources of such quality waters.
4.4.3.3. Water for Irrigation. Nearly all sources of water in nature, which by its quality
is acceptable for irrigation purposes, have been used one way or another. When springs are
used for drinking purposes, if any water is left unused in the growing season, usually that
surplus of water goes to irrigation. The same is valid for aquifers. However, water of rivers,
lakes and large aquifers is predominantly used for irrigation, where irrigation is needed. From
all the water resources development purposes irrigation requires the largest storage capacities
in relation to the available mean annual water volume, because the best use of water in
irrigation is by storing surplus water in wet seasons and wet years, and then using that stored
water in dry seasons and dry years. The largest irrigation schemes in the world are developed
by constructing large reservoirs. Among others, the US West is one of the best examples of
those large irrigation schemes, not to mention also such large schemes in China, Russia, India
and all over the world. Similarly, huge aquifers like the Ogallala formation in the US West feed
irrigation projects in many states through millions of wells. As an example, only the state of
Nebraska has nearly half a million of irrigation wells pumping water from this huge aquifer, so
that the difficult problem of aquifer recharge looms at the horizon (Bittinger and Green, 1980).
Contemporaneous agriculture uses lots of chemicals. Among them the nitrates are at the basis
of artificial fertilizers. If irrigation is carried out over the aquifers, two modem phenomena
occur. The lower layers of aquifer water are loaded with nitrates. By pumping water from
these layers for irrigation, nitrates are recovered. The second positive aspect of this type of
conditions, is that the use of irrigation water distribution network in non-growing season,
aquifers may be easily and most economically recharged. Thus aquifers are not only used as
sources of irrigation water, but also as storage capacity and the recovery offertilizers.
Recycled water, or treated urban wastewaters have been used for irrigation of fast
growing trees for paper industry, since one can control easily in such cases the contact of
humans with these irrigation waters. In the watering of lawns, shrubbery and trees in urban
areas, usually that water is of the quality of drinking water, so that health of the population is
protected. If such irrigation uses a special independent network, the water usually must be
disinfected.
56
4.4.3.4. Industrial Water Supplies. There is no uniformity in the industrial and mining
water quality requirements. From food and pharmaceutical industries, which require drinking
water quality, and sometimes even still better water quality which fits the need of a specific
industrial process, all the way up to the washing water which doesn't even need to be
desinfected and/or made sterile. The quality of water then should meet the needs of the
particular industries, mines and smelters. Some industries may use waters of different qualities,
the best for the specific process and the least quality water for simple washing. Close to the
seas and salt lakes, the limited quantity of good quality fresh water may be used only for the
industrial process, which can not use the lesser quality water, while the brackish and/or saline
waters are used for the purposes which do not require the more expensive, quality water for
such uses as washing floors, trucks, cars and machines. Water used for cooling purposes may
be loaded with minerals, if its temperature is sufficiently low to meet the requirement of the
cooling purpose. In classical thermal and in nuclear power plants the condensation process
water should be as little mineralized as the aggressiveness of that water permits, while the
cooling tower or flow-through cooling waters may be of a much lower quality as long as they
accomplish the cooling purpose.
4.4.3.5. Water for Hydropower and Navigation. Because hydroelectric power plants
use the energy head in passing water through the plant, the consumption of water
(evaporation, leakages, losses, etc.) is relatively minimal. The only requirement on water
quality is that it is not too aggressive for the structures through which water passes, so that
they do not highly and rapidly corrode, erode or incrustate. A similar situation may be with
navigation, where the quality requirements are such that water has no floating debris and
should have sufficient navigable depths, or and sufficient volume of water to fill locks in the
passage of ships and barges from lower to higher levels by using locks.
4.4.4. Six Basic Types of Water Uses
In summary, there are six basic types of water uses:
(1) Destilled water for use in boilers and uses in special industrial processes,
(2) Uses needing drinking quality water,
(3) Human hygienic and recreation water uses,

57
(4) Irrigation in agriculture and in urban areas,
(5) Industrial, mining, smeltering, cooling purposes uses, and
(6) Hydropower and navigation water uses tolerating the relatively low quality waters.
These basic uses may satisfY the water demands by quite different qualities of water,
which depends on the character of the user's water quality requirements. All other water uses
(wildlife protection, fish production, low flow augmentation, wastewater dilution, etc.) may be
satisfied by one or more types ofthe above six water quality potential uses.
4.5. CONCLUSIONS
The following major conclusions of the above discussion are:
(1) The unique water supply system forces the treatment of water quality to the standards
of drinking water, though the majority by the volume of the users may tolerate the
lesser quality waters,
(2) The future investigations should resolve the dilemma whether the multiple network
water supply systems have any safety and economic advantages, which may lead to
different treatment levels of the multiple network water supply systems,
(3) There are many areas, with some discussed above in this text, which deserve the
repeated research and development attempts, especially as it concerns the human health
and nutrition related to the drinking water quality.
4.6. REFERENCES
Bittinger W. M., Green E.B. (1980) You Never Miss the Water Til (The Ogallala Story). Water
Resources Publications, 116 pages.
Canter K.P. (1992) Cancer Risks Associated with Chlorination By-products. Water Review
Technical Report, a publication of the Water Quality Research Council, USA, Vol.
10, No.2.
5. Pollution Control for Urban Water Supply
Systems
Thomas G. Sanders
Vujica Yevjevich
USA
5.1. INTRODUCTION
The protection and prevention of contamination of an urban water supply system is a

complex process which can be as simple as fencing off and preventing access to the
contributing water catchment and as complicated as adding slow acting disinfectants in the
treated water to protect it from contamination while it is being delivered through the
distribution system. There are many other control strategies in fighting the contamination of
water in an urban water supply, forming the multiple barrier system within that supply
system.
The concept of the multiple barrier system is the cornerstone of the environmental
engineering approach to producing safe drinking water. In order to protect the public water
supply many measures are undertaken to eliminate, prevent and/or destroy pathogenic
organisms. Each measure in itself could eliminate the pathogenic organisms but maintaining the
multiple barrier system approach assures a high reliability of the entire system that the
pathogenic organisms will be eliminated.
This strategy is the reason why the preferred source of the water is one where man's
access is restricted, through the unit processes treating the water to remove the pathogenic
organisms as well. In fact, the multiple barrier system reduces significantly the probability that
something may go wrong in the prevention and protection of the drinking water against
contamination, with all the consequences that the unremoved pathogenic organisms may
produce on human health. Urban water supply systems can be divided into the following four
componential parts (called here either the components or the subsystems) as related to the
NATO AS! Series, Partnership Sub-Series, 2.Environment- Vol. IS
Edited by C. Maksimovif, F. Calomino, and J. Snoxell
60
protection and prevention of the contamination of the systems' waters: (1) collection (and/or
tapping); (2) transmission (or transportation); (3) treatment; and (4) distribution.
The collection component or subsystem of a water supply system can be visualized as
the catchment upstream from the point of diversion and/or the point of intake structures, used
to collect, divert and/or control the proper inflow of the water into the man-made water supply
system.
The other type of visualization are the reservoirs, as man-made structures, used to
collect and store the water, as well as places where the water intake structures may be or often
are located. The transmission or transportation component of a water supply system includes
the main conduits, canals, aqueducts, tunnels, siphons and other water conveyance structures,
which connect the diversion or intake structure(s) of the main source(s) of the water to the
treatment plants or further to the local storage tanks or reservoirs for the treated water within
the major water consumption areas. The basic concern is to prevent that no uncontrolled water
inflow would penetrate the transmission conveyance line. The treatment works component or
subsystem may only be the disinfection, or the conventional treatment of chemical coagulation,
flocculation, and sedimentation followed by filtration and disinfection, or the conventional
treatment followed by carbon absorption, reverse osmosis, etc. The final product of this
subsystem will be the drinking water meeting the water quality standards selected by all the
parties responsible for the public safety of the delivered water to consumers.
The distribution component of urban water supply systems is composed of a series of
interconnected pipes which transmit the water from the treatment plant(s) or from the storage
of treated water to points of use at a prescribed lowest water pressure. Most often, the
distribution system of pipes uses the closed loops of pipes, so that each user of water may
receive water from two directions, and that the local service interruption of a pipe does not
interrupt also the water supply. Depending on the availability of head, standpipes or reservoirs
located at elevated positions in the urban area or water pumping stations may be required to
maintain pressure in th~ system and therefore would be also part of the distribution component.
5.2. PROTECTING SYSTEM'S WATER COLLECTION COMPONENT
5.2.1. Access to Catchment and Reservoirs
The protection of the collection component of the source of water in small water
supply systems from contamination is carried out either by limiting public access to the
61
catchment and reservoirs, or by permitting that access but in a regulated way. In the first case,
only those people may be permitted access who are connected with the operation of reservoirs
and the water supply system, and those who have the property or land utilization rights within
the catchment. Those permits, however, must be clearly defined and regulated, so that the
offenders may be prosecuted for violations. In the second case, which do not restrict public
access, restrictions may be used for both motors on the reservoirs (sometimes permitting only
the electrical, water non-polluting motors, operated on batteries), and allowing the use of
sailboats, rowboats and canoes.
Whether the river, lake or reservoir is a source for water supply or not, disposal of all
human wastes is usually prohibited. If wastewaters occur in the collection area, a strict
regulation on the degree of their treatment before they are released into the catchment or
reservoir must be enforced. Similarly, any other point or non-point source of contaminants
must be identified and strictly controlled.
As the catchments get bigger and the restrictions for public access become unfeasible,
even to reservoirs, the alternative is the restriction of land use. This could include the
prohibition of the use of septic tank systems and require vaults for human wastes. Restrictions
to allow only certain land uses is a viable alternative. An example would be to restrict or
prohibit the use of pesticides, herbicides, insecticides and most of the artificial fertilizers thus
discouraging intensive farming, or farming at all, in the catchment. If restricting land use is not
feasible, aggressive point source and non-point source control of pollutants is the only
alternative.
5.2.2. Control of Point Sources of Contaminants
Since 1972 a US national policy of zero wastewater discharge has been pursued to
eliminate point source discharges from municipal wastewater treatment plants and from
industries, without permits. This is a technology based program using the best available
technology (BAT's) for each discharger. Enforced by using the NPDES permits (National
Pollution Discharge Elimination System permit), each point discharger is required to meet
the permit's discharge requirements which will change as best technology changes. The
individual states of the USA are charged with the enforcement of this policy and to periodically
review and reissue new permits. The previous US legislation of 1965 was a pollution control
program based upon a water quality based program. Each navigable stream in the USA was
classified according to existing use and water quality. The idea was to not allow any river to
get worse (antidegradation policy) with the eventual improvement and protect the water
quality in courts. Unfortunately, it did not work as it was based upon legal enforcement of
62
stream standards where the source of the violations could not, in most cases, be proven in
courts.
5.2.3. Control of Non-Point Sources of Contaminants
Presently the focus of the US national policy on pollution control is directed toward
non-point sources of pollutants. Primarily, looking at urban non-point sources at present, the
program will eventually expand to non-point sources of contaminants from agriculture,
silvaculture, pastures, roads and highways, airports, etc. Developing the concept of the best
management practices, BMP's, which are analogous to the BAT's for point sources of
pollution, will be the thrust of research and development of technologies for the next decades.
Unfortunately, the BAT's are being recommended and may eventually be required without
knowing whether they work or will have any kind of impact in the reduction of pollution. An
example is the recommendation to treat surface runoff by using detention ponds, followed by
filtering it through sand to remove phosphorus. The detention ponds will trap the phosphorus
absorbed to sediment particles, which could be as high as 70% of the phosphorus load but
there is no indication that sand and smaller particles sediment will remove significant amounts
of dissolved phosphorus (Mestas, 1994).
Some control of quality of the water may be accomplished by using the proper types of
intake structures at reservoirs. The multiple outlet water intake structures at reservoirs permit
the drawing of water from the particular layers of stratified reservoirs and lakes, as the best
management of water quality may require.
Thus, some reservoir inflow waters can stay longer in the storage giving them
opportunity to eventually deposit with the suspended sediment as much phosphorus as
particles can accommodate and store at the bottom. This is the case especially with the mud
lake at the lowest part of reservoirs, close to the dam sites. These mud deposits may be washed
out during floods, provided downstream conditions permit and there is the corresponding
bottom outlet(s) at dams.
5.2.4. Protecting Water from Pollution at Intake Structures
Water intake structures are located at the transition between the collection and
transmission components of water supply systems. The tapping of the spring and well waters
requires special attention because of the potential of penetration of polluted waters. The most
sensitive structure at the spring intakes is the spill-over part, due to the fact of the most
63
frequent condition that only part of spring discharge is taken for the water supply system,
producing an overflow of water at the spring and flow its natural stream or go underground.
Spillways at such intake structures are a sensitive part of the intake, since animals and
impurities may penetrate the intakes through the spillway or other surplus water outlets. In
case of water wells, the penetration of polluted surface water along the well casings must be
often specially treated.
5.2.5 Protecting Underground Water Collection Component from Pollution
It is much more complex and expensive to protect the underground water systems from
pollution than it is the case with the surface waters. Special measures and technologies may be
required to protect the karst or highly permeable lava deposit water systems from
contamination, than is the case with the alluvial aquifers, though these aquifers may also be
easily polluted, especially if they underlay the permeable and agriculture-used soils. Special
techniques are needed and are available for building on, and using, the karst surfaces for
various purposes. Urban and industrial areas require protection measures from an easy
infiltration and/or sinking of the polluted water into karst formation.
A similar case may be with the water systems in the recent lava deposits (recent in the
geological sense). These measures require that the permeable surfaces are paved over, or made
impervious by other methods, urban and industrial drainage water carefully collected, and
wastewaters of both treated to the point that they may be permitted either to be disposed in the
sinks or permeable terrains, or if not sufficiently treated, to be disposed outside of these
geological formations (Turk, 1975).
The protection of alluvial aquifers from contamination requires a careful control of the
aquifer's levels in order to avoid drawing in the polluted surface waters, basically through the
large side infiltrations. Th,e agricultural pollution due to infiltration of various chemicals is part
of the non-point sources of pollution.
For the aquifers under the cultivated land, often nitrates and pesticides collect in the
lower layers of the aquifer. If the aquifer is used for irrigation, a method of solving the problem
of this pollution is to pump water for irrigation purpose from the deep, the high nitrate
concentration water layer, and not only to protect it from the expansion of nitrate loaded water
layers but also to recover fertilizer.
64
5.3. PROTECTING SYSTEM'S WATER TRANSMISSION COMPONENT
One should distinguish between the protection of the water from contamination of an
open and a closed conveyance water transmission. A closed conveyance may work either with
the free surface flow or under pressure. The open conveyances are: natural channels, canals
and open aqueducts. The free surface and/or under pressure flow conveyances are tunnels and
closed aqueducts. Flow under pressure only is usually in the conduits. They may be on the
surface and easily accessible, or they may be underground or buried into the soil, therefore less
accessible.
5.3.1. Protecting Open Conveyances from Water Pollution
If the reservoir water is released into its natural channel, and the diversion or intake
structure is located sufficiently downstream from the reservoir outlet, the protection of water
from contamination will depend on the water users in the catchment between the reservoir and
the diversion or intake point. This case is often encountered in mountains, where the optimal
points of dam and diversion locations do not coincide. There are many such examples in the
Rocky Mountains in Colorado. Simply, the problem of protection from pollution of the water
becomes the same as it is for the collection component of the system. The only difference is
that the water at the diversion point is a mixture of the reservoir stored and quality transformed
water and the natural or man-affected waters inflowing into the stream in between.
Canals are the special case, very often when water is conveyed for two or more
purposes, say for irrigation, hydropower, navigation, water supply, etc. Practically the diverted
river water, liberated from floating and bedload materials, flows through the canal to the
diversion or intake structure for the water supply system. Transmission canals are generally
restricted to access to prevent contamination, if water is predominantly used for drinking water
supply. In some areas of the high water power heads with relatively small discharge, water of
the hydroelectric power plants can be also used for drinking water supply, after the last power
plant. In this case, headrace and tailrace canals of hydroelectric plants are strictly restricted to
access, or are covered.
Protection of open aqueducts is simple, because it is easy to completely restrict access
to everybody except the maintenance crew. If the aqueduct carries clean, spring drinking
water, usually it is covered to avoid the wind blown dust to reach the water.
65
5.3.2. Protecting Oosed Conveyances from Pollution and Interruption
Two dangers threaten the closed transmission conveyances in operation, the

penetration of contaminated water and the interruption of the line, meaning the out-of-
operation state of the system. Rarely in the medium and large water supply systems is the
safety secondary conveyance line economically feasible. The easiest way to protect a closed
conveyance from direct pollution is to build a tunnel, to 'lay down a good conduit, or to cover
an aqueduct. And it is done in such a way that protection from penetration of any outside,
potential polluting water is well assured, except in case of disasters (such as earthquakes,
avalanches, landslides, floods, wars, revolutions, etc.).
Protection from interruptions is done by the professional investigation of the terrain
over and in which a line is positioned. Geotechnical and geological factors, as well as
hydrological (traversing the stream and creek valleys) are most important, since they are
determining the stability of any such conveyance structure. In some cases, interruptions may be
also connected with the potential pollution.
5.3.3. Monitoring Conveyances for Safety against Pollution and Interruption
It is necessary to properly design a monitoring system for both the protection of

conveyances from penetration of polluted waters as well as for location of potential sources of
interruptions (a weak point in the line, a slow evolving erosion or corrosion of conduits).
Either the pollution or interruption of water supply by the transmission line represents major
accidents in the system's operation. Monitoring this component by a well designed plan and
program may be paid otT very many times over since either of the two types of accidents may
be very costly, highly inconveniencing water users.
5.4. PROTECTING SYSTEM'S WATER TREATMENT COMPONENT
5.4.1. Types of Water Treatment and Protection of Quality
The treatment could be as simple as just disinfecting water using chlorination or more
sophisticated approach using chemical coagulation, flocculation and sedimentation, followed
by filtration and disinfection. Each unit process is part of the multiple barrier system by having
66
the particular capability of removing pathogenic organisms. The addition of chemical

flocculants deactivates bacteria due to the pH change and the actual entrapping and
precipitation out of solution. The filtration, be it rapid sand filtration or slow sand filtration,
physically removes these organisms. It should be noted, however, that this process is much
more complicated than just physical removal. Research and practice have shown that without
chemical pretreatment, rapid sand filtration is not that effective in removing some organisms,
such as Giardia lamblia cysts and Cryptosporidium oocysts.
The final process of disinfection will deactivate any remaining bacteria and higher
concentrations and longer detention times then are used in conventional treatment, apparently
deactivates viruses as well. Chlorine, when added with ammonia, forms various species of
chloramines (combined chlorine, NH2Cl, NHCIz, NCb) which are bactericidal but tend to be
longer lasting than free chlorine (hypochlorous acid-chlorate ion, HOCl-OCf) and will remain
active for several days in the distribution subsystem prior to consumption. If sufficient amounts
of chlorine is added with ammonia to drive the reaction to completion creating NCb, a gas, the
process is called breakpoint chlorination and can be effective for denitrification and destroying
some of the other organic compounds, but unfortunately creates trihalomethanes (THM's),
known precursors to cancer.
5.4.2. Potential for Increasing Future Protection
Future processes to be incorporated into the conventional water treatment train, which
can remove chlorinated hydrocarbons, dissolved inorganics and possibly viruses include:
granular activated carbon, reverse osmosis, ion exchange and ultra filtration. Other
disinfectants besides chlorination are being considered as well since they have the capability of
not forming THM's. They are ozone and chlorine dioxide (Sanders and Champlin, 1994).
Unfortunately the use of chlorination and alum (aluminum sulfate), the most widely used
chemicals in the water treatment industry for disinfection and flocculation, respectively, are
being. questioned due to known or speculated health consequences. There are indications that
drinking a chlorinated water for many years increases an individual's chances of getting bladder
cancer 9% and rectal cancer 15% versus individuals drinking unchlorinated water (American
Journal of Public Health, 1992). And there has been some speculations that the consumption of
drinking waters treated with alum is related to memory loss.
In order to provide more protection for future drinking water, the U. S. Environmental
Protection Agency (EPA) has proposed the Information Collection Rule (ICR) in February
1994.
67
The rule requires utilities to collect information which will become a data base for the
Enhanced Surface Water Treatment Rule (ESWTR) and the stage Two
DisinfectantslDisinfection By-Products Rule (D / DBPR). The study is nationwide,
primarily to quantify the occurrence of contaminants and their removal, and will lead to
regulatory controls for microbiological and disinfection by-products contamination (Black and
Veatch,1994).
Besides keeping records on the disinfection by-products, ICR mandates routine
analyses for Giardia lamdia, Chryptosporidium, and culturable viruses. The expected cost to
implement ICR for a typical water treatment plant is estimated to be $145,500 (Black and
Veatch,1994).
5.4.3. Monitoring Treatment Processes
Real time monitoring of the treatment processes using turbidity, pH, and residual
chlorine, and effluent monitoring for coliform bacteria at the treatment plant and predetermined
sites throughout the distribution system maintains the best quality control. The statistics-based
monitoring programs for measuring the bacterial quality of the drinking water are related to the
size of the population served and are the last barrier before water enters the storage or
distribution system to be consumed. If viable organisms are found in the tests, immediate
action is taken to find and eliminate the origin and cause ofthe problem.
5.5. PROTECTING SYSTEM's DISTRIBUTION COMPONENT
5.5.1. Storage of Treated Water and its Protection before it is Released into Pipe
Network
The storage of treated water in tanks or reservoirs, for matching daily and/or weekly
water supply and water demand, between the treatment plant and the beginning of distribution
network, is not only needed for this regulation of flow but also for maintaining the pressure in
the distribution network. In recent years there has been a tendency in water supply systems to
cover these storage facilities. This is done primarily to eliminate problems with atmospheric
deposition but also to prevent public access to the already treated water. These reservoirs also
help for gases, especially those linked to chlorination, to leave water and removed from the
68
tanks or reservoirs. Attention should be paid in design of reservoirs for all water to slowly
move through reservoirs most of the time, so that no eventual solids precipitate and
accumulate in dead angles of some types of water supply reservoirs.
5.5.2. Protecting Distribution Subsystem from Contamination
In many instances the storage of treated water in reservoirs at elevated locations

provides sufficient head to maintain at least 40 psi (pounds per square inch) throughout the
entire distribution subsystem. The minimum of 40 psi is primarily needed for fire fighting
purposes but it also lessons the probability that very low or negative (sub-atmospheric)
pressures occur in the distribution subsystem where untreated water may enter the subsystem.
This can be a real problem with most distribution components of water supply systems during a
fire. Once a large fire truck is connected to the distribution subsystem, very low or even
negative pressures in the subsystem can occur allowing untreated water to enter into the
distribution network.
An example of this occurring happened at Holy Cross College, USA. It was a hepatitis
outbreak which infected the entire Holy Cross football team. In epidemiological analysis the
outbreak was traced back to a sprinkling irrigation system cross connection (a physical
connection of a potable water supply to a non-potable water supply) which during a large fire
nearby lost pressure which became negative and sucked water in having urine from a boy
carrying hepatitis. This was, by the way, one of the first proven cases that a hepatitis virus can
survive and be transmitted in water.
The cross connections in the distribution subsystem are a continuing problem and is
addressed by local authorities. Most water supply systems in the USA have laws or ordinances
that ban cross connections. If they do occur, one way valves and pressure relief valves are
required. All faucets for sinks and bathtubs are well above them so that if the sinks or tubs fill
up and overflow, there will be no contact with the faucets. Although all faucets, which are
threaded, permitting a hose to be connected, should have a one-way pressure relief valve, but
in practice many do not (or do not function properly), constitute continuous potential problems
with cross connections. Ironically, hospitals are among the major sources of this type of cross
connection.
A recent phenomenon, a rapid proliferation of home underground sprinkler systems in
the USA has created many new potential cross connections. Many of the cities where this is
occurring have been slow to pass ordinances and/or enforce proper designs which required
one-way valve protection.
69
Another design in the distribution subsystem to protect the quality of the water is the
absence of dead ends of pipes where water can stand for many days before being used. Most
distribution subsystems are interconnected in the sense that water can come from two
directions (no pipe dead ends). This concept of design also allows a more even distribution of
network pressures, and decreases the number of users losing access to water when a pipe has
to be repaired.
Because there is an increasing fear of the general public that their drinking water from
the tap is not safe, there is a new product on the market that connects directly to the faucet
usually consisting of activated carbon, that is finding wide acceptance. It may help in the
reduction of trace organics and particles. But if not properly used and replaced periodically, it
constitutes an increased health menace if a break through occurs.
5.5.3. Aging and Corrosion of Buried Pipes as Related to Water Pollution
Buried pipes if not corrosive from inside is more likely to be so from outside, whenever
pipes are buried into acid soils. Besides the problems may arise with leaking joints for whatever
reasons. Leaking pipes buried into less permeable soils may maintain relatively large pressure
outside the pipe in comparison with the inner water pressure. Because of the regular
fluctuation of pressure at leaking points, water flow out and in the pipes frequently. These
aspects are very important especially if polluted water infiltrates the soils in which pipes are
buried. This is an occurrence in many old distribution networks, requiring precise pipe
discharge measurments and monitoring, or even uncovering pipes at test sites for checking
their conditions.
5.6. CONCLUSIONS
The above discussion of the safety of water supply systems against pollution of the
water leads to the following conclusions:
(I) Security against contamination of the water within the water supply systems is highly
enhanced by a strict application of the multiple barriers system concept,
70
(2) Systematic investigations of problems with the water contamination inside the water
supply systems and the studies of measures for counteracting these dangers, will be
much more successful if they are done by the four major system's components,
collection, transmission, treatment and distribution,
(3) Major measures in the collection subsystem are those which limit human access to
collecting areas, and when it can not be done fully, to regulate access and control both
the point and non-point sources of pollution,
(4) Major measures in protecting transmission subsystem from water contamination are
related to intake structures, non permitting access to uncovered conveyances, and by
avoiding interruptions of water conveyance lines,
(5) Principal measures in protecting water treatment subsystem from contamination is to

follow the principle of multiple barriers system and to apply the relatively rapidly
evolving, new water treatment technologies,
(6) Major actions in protecting water distribution subsystem from contamination to insure
the safe water storage of already treated water and to maintain the minimum prescribed
water pressure in the network of pipes all the time.
5.7. REFERENCES
American Journal of Public Health (1992) V. 82, p. 955.

Black & Veatch (1994) SWDA Information Collection Rule. June.
Mestas P. (1994) The Efficiency of Dissolved Phosphorus Removal from Urban Runoff Sand
. Filtration. Master's thesis in partial fulfillment of the requirements for Master of Science
of Civil Engineering at Colorado State University, Fort CoIlins, Colorado, USA.
Sanders G.T., Champlin T.L. (1994) Treatment of Drinking Water in the United States of
America on the Treshold of the 21st Century. NATO Advanced Study Institute, Varna,
Bulgaria, October 24 - November 4,1994.
Turk L.J. (1975) Predicting the Environmental Impact of Urban Development in a Karst Area.
Karst Hydrology and Water Resources, V.Yevjevich (ed.), WRP, Vol. 2, p. 681-702.
6. Wastewater Reuse - a Worldwide Issue
Thomas G. Sanders
Lisa M. Dunn
Vujica Yevjevich
USA
6.1. INTRODUCTION
Water resource shortages are a problem that are plaguing the world. Wastewater
recycling and reuse may be the solution to this worldwide problem. Recycled and properly
treated wastewater is a safe alternative source for irrigation, industry, groundwater recharge,
and miscellaneous subpotable uses. The biggest problem with wastewater recycling and reuse
is the concern for potential health risks. Currently, there are wastewater recycling and reuse
projects in Singapore, Israel, Colorado, California, Florida as well as others. While California
leads the way in developing regulations for wastewater recycling and reuse and the US EPA
has developed a manual on water reuse criteria, much debate exists regarding regulations, uses
and appropriate levels.
Three fifths of the earth is covered by water. While this may appear to be an
abundance, most of this water is not available for most uses. The oceans account for 97.2% of
the earth's water (Warner, 1994). Because of the high salinity, this water is of no use for many
water users. The remaining 2.8% of the water consists of surface, subsurface and atmospheric
water. Planning and managing this 2.8% has become a great challenge. Since 1900, the United
States' population has increased by 200%. Yet, the per capita water usage has increased by
800% (Arber, 1991). This type of increase in demand through out the world is forcing
countries to develop new water sources, while protecting the existing ones, as well as
implementing conservation measures.
The continental United States receive an average of 762 mm of rain annually.
Approximately two thirds of this water is lost through evaporation (Asano et aI., 1991).
NATO AS! Series, Partnership Sub-Series, 2. Environment - Vol. 15

Edited by C. Maksirnovic, F. Calomino. and 1. Snoxell
© Springer· Verlag Berlin Heidelberg 1996
72
Agricultural activities and steam electric plants use over 75% of all the fresh water withdrawn
in the United States (Asano et aI., 1991). Currently there are water shortages in the Rio
Grande, the Texas-Gulf, Upper and Lower Colorado, Missouri, and California (Asano et aI.,
1991). Most of these regions are using water in excess of their presently sustained renewable
water resources. Others are entirely dependent on mining groundwater.
Water resource shortages are not limited to the United States; they are a world-wide
problem. In Jordan, water is the limiting factor for industrial and agri,culturaI growth and
development (Gur and AI Salem, 1992). Not only is water scarce, it also needs to be
transported long distances. As a result, their groundwater resources have been utilized to the
extent of overpumping, leading to a deterioration in the quality.
Israel is also faced with a severe water shortage and P9llution threat to its water
resources. In order to service its concentrated urban population and its large agricultural
demands, the groundwater resources are currently being overused beyond their natural
recharge (Kuttab, 1993).
Singapore is faced with not having enough water for its growing population. The main
island of Singapore, with a total area of 620 km2, is home to 2.69 million people. The annual
precipitation is 2400 mm. However, 1600 mm/year is lost in the form of evaporation (Shelef,
1991).
In California, a five year drought has cut agricultural water supplies by 75% (Asano,
1992). Under these circumstances, using potable water for landscape irrigation appears
unreasonable and wasteful.
6.2. NEW SOURCES OF WATER
With the population continuing to grow, the water demand is exceeding the supply
capabilities in many places. As a result, it is necessary to find and develop new sources of
water.
Wastewater recycling and reuse may be a solution to this world-wide problem, Arid
and semi-arid areas can augment 15-20% of the water supply through wastewater reuse
(Shahalam, 1989).
This wastewater is best used, but not limited to agricultural purposes. Besides
providing moisture and nutrients to the soil, the reuse of this water would reduce wastewater
pollution and disposal costs.
73
6.3. RECYCLED WASTEWATER USES
6.3.1. Irrigation
Agricultural and landscape irrigation are the largest users, current and future, of
recycled wastewater. Some potential constraints to wastewater reuse are the effect of the
water quality on soils and crops, public health concerns, and the marketability of the crops
(Asano et al., 1991). Yet, many times the need for additional water far outweighs the potential
negative effects. The recent drought in California has increased the incentives for using
recycled water because many cities had passed drought control ordinances that prohibited the
use of potable waters for construction, dust control and compaction. Also recycled water is to
be used on landscapes if it were available (Newby and Hough, 1994). In Santa Clara,
California, recycled water was utilized for golf course irrigation. Not only did this conserve
water, it additionally saved money. The reused water rates were set at 90% of the retail rate
(Smith and Walker, 1991). Reuse of wastewater in Florida for agriculture irrigation is
promoted as a means of decreasing the nutrient loads into adjacent streams and reducing
treatment costs. The treatment requirements are increased if the wastewater is discharged into
streams when it cannot be used for agriculture (Wright, 1994). The country of Jordan
supplements its irrigation supply, approximately 4 MCMlyr, by using wastewater for producing
fodder, cereals, fruit trees, and vegetables which are cooked before being eaten (Gur and AI
Salem, 1992).
6.3.2. Industry
A second category for reused wastewater is industry, where it is used primarily for
cooling and processing. In Singapore, industrial demand is increasing as a result of restriction
of potable water to industrial users and a preferential water tariff system to encourage the
industry to use recycled industrial wastewater. Reused water costs industry SO. 241m3 in
Singapore, while potable water costs SI.l/m3. In addition, to cover sewer costs in Singapore, a
minimum of SSO/month is charged to industrial users (Chin and Ong, 1992). The program is
working so well, current demand for industrial water has exceeded the capacity of the Jurong
Industrial Water Works. Expansion is being done to increase yield by 20,000m3/day. The
ultimate capacity of the plant is 91,000m3/day (Chin and Ong, 1992). Potential constraints to
industrial use are public health concerns, and scaling and corrosion of the pipes (Asano et al.,
1991).
74
6.3.3. Groundwater Recharge
Another way of using wastewater involves groundwater recharge. For example, in

Jordan, groundwater recharge uses wastewater from cesspools and etlluent infiltrating through
the layers of ground, ultimately reaching and recharging groundwater aquifers (Gur and AI
Salem, 1992). Potential constraints include toxicological effects of chemicals in the
wastewater, dissolved solids and pathogens (Asano et al., 1991). For many years, recharge of
treated wastewater has been used in Southern California to prevent saltwater intrusion.
6.3.4. Miscellaneous Subpotable Uses
The last category is miscellaneous subpotable uses. These include recreational lakes,
aquaculture, and toilet flushing. "Recreational water used for body contact sports must be
aesthetically acceptable, must not contain substances that are toxic upon ingestion or irritating
to the skin, and must be reasonably free from pathogenic organisms" (Culp and Culp, 1971).
This category accounts for less than 5% of total wastewater reuse. Potential constraints to this
category are mostly regarding public health concerns (Asano et al., 1991).
6.4. HEALTH RISKS
Common to all the categories of wastewater reuse is the concern for potential health
risks. Most of the potential health risks associated with the reuse of wastewater are due to
pathogenic organisms in wastewater. These pathogens can be spread three different ways
(Shuval, 1991).
(1) Through wastewater contamination of food, mainly eaten uncooked to the public at
large,
(2) To the farm workers who have direct contact with the wastewater,
(3) To population groups living near wastewater irrigated fields possibly by aerosolized
pathogens from sprinklers irrigation or by direct contact of children playing in or near
wastewater irrigated fields.
75
There are several factors that increase the relative effect of various groups of pathogens
to cause infection in humans (Shuval, 1991a).
(1) Long persistence in the environment,
(2) Low minimal ineffective dose,
(3) Short or no immunity,
(4) Minimal concurrent transmission through other routes such as food, water and poor
personal and domestic hygiene,
(5) Long latent period and/or soil development stage required.
Based on these criteria, the worm (helminth) pathogen will be the most effectively
transmitted through irrigation of wastewater. Enteric virus disease should be least effective.
What does all this mean to people? How does this effect people's health and activities?
Depending on the method of transmission, different diseases can be linked to raw or poorly
treated wastewater (Shuval, 1991a):
(1) for the general public, eating salads or vegetable crops irrigated with raw wastewater:
ascariasis, trichuriases, cholera, tapeworm,
(2) for wastewater irrigation workers: ancylostomiasis (hookworm), ascariasis, cholera,
(3) for the general public nearby wastewater irrigation: there is little or no transmission
even when sprinklers are used.
6.5. TREATMENT PROCESSES
In order to protect public health, while making use of this new water source, it is
necessary to utilize various treatment processes. Many different agencies have studied the use
of wastewater reuse and the safety of the treatment process, each with different results and
76
limits. See Table 3. for regulations. In order to understand the different treatments, it is
necessary to understand the processes.
6.5.1. Primary Treatment
The available treatment processes can be classified into three group, primary, secondary
and tertiary. Primary treatment produces the lowest quality effluent. It is composed of physical
processes, primarily sedimentation followed by disinfection with chlorine (Shahalam, 1989).
Primary processes are septic tanks, Imhoff tanks, primary sedimentation, racks, screens.
Primary treatment is typically capable of removing 35% Biochemical Oxygen Demand (BOD),
10% total coliform bacteria, 35% fecal coliform bacteria, and 55% suspended solids
(Shahalam, 1989).
6.5.2. Secondary Treatment
Primary treatment is followed by secondary treatment which is a biologically based

process or its equivalent. Its main purpose is to remove dissolved organic materials (Shahalam,
1989).
I = Extended aeration system 4 = Waste stabilization ponds
2 = Conventional activated sludge 5 = Land treatment
3 = Facultative aerated lagoons 6 = Rapid infiltration system
100 ~------~~=-----------------~6
3
2
Stabilization S. 20 Da s
Pond Stages Detention Detention
World Bank
Figure 1. Generalized removal curves for BOD, helminth eggs, excreted bacteria, and viruses in waste
stabilization ponds at temperatures above 200 C (Shuval, 1991a)
77
Table 1. Performances of secondary processes treating domestic sewage
ITEM 1 2 3 4 5 6
PERFORMANCE
BOD Removal % 90-95 85-93 70-90 70-90 80-90 99
-~
• • (1"1"- T 15-:"~ ;tn..4(l _. - .AJ.l:.'in 80-Qn Rn
Nutrient removal %P 10-20 30-45 - 20-60 80-99 80
Suspended solids % 88 88 50 88 92 -
There are many systems that perform secondary treatment. They include oxidation
ponds, stabilization ponds, oxidation ditches, aerated lagoons, trickling filters, and activated
sludge processes. The removal of BOD, nutrients and suspended solids as well as pathogenic
organisms vary widely depending on the system used and its detention times (Table 1. and
Figure 1.).
6.5.3. Tertiary Treatment
Tertiary treatment is made up of generally more chemically based processes and is

usually preceded by primary and secondary treatment. Examples of tertiary treatment processes
include nitrification denitrification, phosphorus precipitation, activated carbon adsorption,
ionexchange, reverse osmosis, desalination, coagulation, sedimentation and filtration.
What treatment is the best and most cost effective will be the source of considerable
research in the future. The Sanitation District of Los Angeles County's Pomoana Research
Facility conducted a study in 1976-1977 to compare virus removal capacity of an alternative
tertiary process, contact filtration, with the full treatment process. The full treatment process
includes primary and secondary treatment followed by polymer and alum addition, flocculation,
sedimentation, granular medium filtration and disinfection. This is the most stringent treatment
process specified by criteria. The alternative process was similar to the full process except it
did not include the flocculation and clarification steps. Full treatment is an expensive process
due to the high costs associated with coagulant chemicals, sludge handling, and tertiary tanks
(Asano, 1992). The study found that with high chlorine residuals (10 mg/l) following both
processes, no difference was seen in the overall removal or inactivation of the seeded
poliovirus. With low residuals (5 mg/l), a slight difference was observed; the full treatment was
better (Asano,1992). The total annual wastewater reclamation cost ranges from
$200/acft($0.16/m3) to($1300/acft($1.05/m3). The ratio of full treatment to direct or contact
78
filtration is 2.0 to 2.4 for capital costs, 3.9 to 5.6 for operations and management, and 2.4 to
2.9 for total life cycle cost. Although the treatment costs may be high, the distribution system,
for reclaimed water is the principle costs for reuse projects (Asano, 1992).
6.6. EXISTING PROJECTS
To offset the rapid increase of potable water demand in Singapore, the Jurong
Industrial Water Works (JIWW) uses treated effluent from the Ulu Pandan Sewage Treatment
Works. The JIWW has a design capacity of 45.000 m3/day (Chin and Ong, 1992). Singapore
has six treatment works, which treat more than 700.000 m3/day. The JIWW treatment process
includes screening and pre chlorination, chemical clarification, rapid gravel filtration, aeration,
and postchlorination. Effluent water quality for treated water is shown in Table 2. The
estimated cost of treating water is $1.76/m3 prod. water (US) (Chin and Ong, 1992)
Table 2. Water quality (Chin and Ong, 1992)
IVariables 1-3 mgll
I
M~'!-
~uspended sohds 1.5-3.5 mgll
In rural areas the question of how to provide sewage drainage/treatment facilities is as

important as how to make water available. Save the ChildrenlUSA has utilized a wastewater
treatment/reuse system, Subsurface Drainage Technique (SDT) (Kuttab, 1993). The system is
being used in Israeli occupied villages on the West Bank and in the Gaza Strip. The system can
be used by any household with enough land in the immediate vicinity. The wastewater is
drained from the house in sewer lines to a watertight septic tank, where the sewage undergoes
treatment utilizing sedimentation and stabilization by anaerobic bacteria. The sewage
undergoes secondary treatment in the drainage field. The total cost for the system and labor is
about $500 (US) (Kuttab, 1993).
In Colorado, Denver's Potable Water Reuse Demonstration project created and
provided answers on product quality, public health, technological and economic feasibility,
conservation and consumer and regulatory agency acceptance. In 1984, the Potable Water
79
Reuse Demonstration Plant began operation. The plant used unchlorinated secondary effluent
from the metropolitan Denver Wastewater Treatment Facility as raw water supply (Rogers and
Lauer, 1992). The plant processes included high pH lime treatment, single or second stage
recarbonation, pressured filters, selective ion exchange for ammonia removal, two stages of
activated carbon adsorption, ozonation, reverse osmosis, air stripping, and chlorine dioxide
disinfection. The plant is designed with multiple barrier protection. For example, viruses and
bacteria can be deactivated with high pH lime clarification, ultraviolet irradiation, reverse
osmosis/ultrafiltration, ozonation, and chlorinelchloramination. The ultimate goal in design was
to eliminate unreliable or unnecessary treatment steps without jeopardizing water quality or
consumer safety (Rogers and Lauer, 1992). In all, the plant has shown that the advanced
wastewater treatment technology can produce an effluent with water quality exceeding that of
approved drinking water supplies.
flocculation
unchlorinated
secondary effiuent
no I. water
no. 2 water
pump station
Figure 2. Process train for the Denver potable water reuse plant (Rogers and Lauer, 1992)
6.7. REGULATIONS
With the worldwide importance of wastewater reuse, how are regulations determined?
Currently there are no worldwide regulations. In fact, there are no regulations specific to the
80
United States of America although USEP A has issued guidelines (EPA, 1992) for water reuse
for states that have no standards or are revising their existing standards.
Regulations are set by the standard of living and the acceptance of the people in the
area. For example, the World Health Organization report offers a minimum guideline for
developing countries. In contrast, the State of California has one of the most stringent criteria,
which requires extensive wastewater treatment, frequent waterquality monitoring, and strict
use-area controls.
California lead the way by developing the first wastewater reuse regulations in 1918.
The regulations were formed out of the growing belief that a new science of hygiene and
sanitation could create an "antiseptic", bacteria free environment, that would protect the public
from infectious diseases (Table 3.). These limits were set despite the lack of epidemiological
evidence and were based on the safe drinking water level (2.2 coliformsll00ml), justitying it
must be safe for irrigation and thereby conforming to the "zero risk" attitude of society
(Shuval, 1991b).
California's regulations were later revised to include standards for landscapes and
recreational impoundments. See Table 4. for California's current regulations.
While California set the pace for wastewater regulation, many have disagreed with their
"zero risk" attitude. The strictness of the codes makes it economically unfeasible for
developing and transition countries. In 1973, the World Health Organization (WHO)
developed their own regulations for developing countries. Their regulations called for primary
and secondary treatment and disinfection before use on raw edible crops (Table 3.).
In July of 1985, WHO, along with the World Bank, UNEP, UNDP and the
International Reference Center of Waste Disposal reviewed and set new guidelines, the
Engelberg Report. (Table 3.). The new WHO guidelines are easily achieved at a low cost
through the use of robust stabilization pond systems, which are well suited for developing
countries. These requirements can also be met with other processes for areas not suited for
stabilization ponds (Culp and Culp, 1971).
Israel is faced with a severe water shortage and a pollution threat to its diminishing
water resources. In addition, it has a concentrated urban population and large agricultural
demands. Their groundwater resources are currently overused beyond the natural recharge.
Wastewater reuse is a solution to their problems. In 1973 the Israeli Minister of Health
appointed an advisory committee on quality requirements for wastewater reuse in Israeli
(Shelef, 1991). See Table 3. for the regulations. In 1978 those regulations were still guidelines
for wastewater reuse in Israel.
In Jordan, water is both scarce and far from the people. As the population grows, the
demand for water increases, making- wastewater reuse vital. In 1982 the first wastewater reuse
legislation was adopted, Martial Law No.2. This brief legislation defined types of crops that
81
could be irrigated with treated emuent. It did not allow any use of treated wastewater on
cereals and vegetables (Gur and AI Salem, 1991). Martial Law No. 2 was revised in September
of 1989 (Table 3.).
Table 3. CA 1918, WHO, Engelberg Report Regulations (Shuval, 1991b)
REGULATIONS
CA 1918 WHO Engelberg Report Israel Jordan
(WHO)
Raw Crops 2.2/100ml <100/I00ml geometric mean 12/100 mI NA*

in 80% 1000/100 mI 80%
of samples
~ I helminth
eAAII
Processed 231l00ml 250/100ml
Produce
Cereal crops, <= I helminth <10001l00
industrial crops, eggll mI
fodders, trees
Landscape: <= 1O001l 00 mI 2.21l00 mI <200/100
public, parks 50% mI
~ I helminth
egg/!
Sprots fields <= I 0001l 00 mI 2501100 mI <2001l00
mI
<= I helminth
egg/l
Lakes 23/100ml
Rec. Lakes 2.t/100ml
NA* - Not allowed
Colorado has always enjoyed the luxury of having fresh, clean snowmelt from the
mountains to irrigate their crops. With the increasing population, and growing need for potable
water, Colorado is turning towards wastewater reuse for crop irrigation. The Northglenn
Colorado Water Reuse Program utilizes this concept. The city Northglenn Colorado borrows
the farmer's irrigation water for domestic purposes. The water is then returned along with 30%
makeup water plus a 10% bonus (Sanders, 1981). See Table 5. for Colorado's regulations.
Unfortunately, Colorado has very little control over wastewater reuse. As long as the sewage
emuent is in a natural watercourse, even only for a few meters, it can be used to irrigate any
crop, even raw edibles.
82
Table 4. California Regulations (Shuval, 1991b)
CALIFORNIA REGULATIONS
Application Minimum Characteristics
Primaty Secondary and Secondary Coliform
disinfected coagulated MPN/l00ml
filtered and Median
disinfected (daily sampling)
Irrigation X NR
Fodder CI'OJ)S X NR
Fiber X NR
-
Seed crops X
Produce eaten raw,
surface irrigated X ..2.2
..
I
."" "
Produce eaten raw,

spray irrigated ...
Processed Produce,
spray irrigated X 23
Landscapes:
golf course,
cementeries,
freeways X 23
Landscapes:
parks,
playgrounds,
schoolyards X 2.2
Recreational
Impoundments
No public contact X 23
Boating and
fishing only X 2.2
Body contact
Table 5. The Code of Colorado Regulations (1994)
-- ._. .. _.. COLORADO REGULATIONS

Parameter Parameter Limitation Instantaneous J
7-Day 30-Day Maximum I
Average Average
BOD5 45 mg/l 30 mg/l NA
Suspended solids 45 mg/l 30 mg/l NA
CBOD5 40 mg.ll 25 mg/l NA
Residual chlorine 0.5 mgII
single grab
pH 6.0-9.0 mgII
single gr.\b
Oil and grease 10 mgII
single grab
83
6.8. CONCLUSIONS
The need for new sources of water supplies in many regions of the world is increasing.
Aggravated by recent droughts and the increasing demands, reuse of wastewaters for
nonpotable uses appears to be a viable source of new water in developed, transitional and
developing countries. However, reusing or reclaiming wastewater should only be used for
nonpotable uses. Much more needs to be done in research on the health based risks and criteria
and technology reliability before potable uses can be safely met. Using recycled wastewater for
only nonpotable uses will increase available water supplies and may, in fact, decrease costs of
wastewater treatment since less treatment is required when compared to treatment costs for
potable waters.
6.9. REFERENCES
Arber H., Richard P. (1991) Wastewater Reuse: Thoroughly Examined. Water Engineering &
Management, 138: 5, pp. 36-38.
Asano T. (1992) Evolution of Tertiar Treatment Requirements in California. Water
Environment and Technology, Feb., pp. 36-41.
Asano T., Tchobanoglous G. (1991) The Role of Wastewater Reclamation and Reuse in the
USA. Water Science Technology, 23: pp. 2049-2059.
Chin K.K., Ong S.L. (1992) Experience of Non-Potable Reuse of Wastewaters. Water Science
Technology, 26: 78, pp. 15651571.
The Code of Colorado Regulations (1994) 5 CCR 10023, 17CR 6,6- 94,10.1.4.
Culp R.L., Culp G.L. (1971) Advanced Wastewater Treatment. Van Nostrand Company, New
York.
Gur A., AI Salem S.S. (1992) Potential and Present Wastewater Reuse in Jordan. Water
Science Technology, Vol. 26: 78, pp. 1573-1581.
Kuttab A. (1993) Wastewater TreatmentlReuse in Rural Areas. Water Science Technology,
27: 9, pp. 125-130.
Newby J.P., Hough S.G. (1994) Urban Retrofit on a Grand Scale. Journal of Water
Environment & Technology, Vol. 6, No. 11, Nov.
Rogers S.E., Lauer W.C. (1992) Denver's Demonstration of Potable Water Reuse: Water
Quality and Health Effects Testing. Water Science Technology, 26: 7-8, pp. 1555-
1564.
Sanders T.G. (1981) Recycling Wastewater in Colorado. The Issues, the Problems and the
Solutions. Industrial Wastex, Proceedings on the Thirteenth mid-Atlantic Conference,
June 29-30, 1981, pp. 588-595.
Shahalam A.B.M. (1989) Wastewater Eft1uent vs. Safety In Its Reuse: State of the Art. The
Journal of Environmental Sciences, Sept./Oct., pp. 35-41.
84
ShelefG. (1991) The Role of Wastewater Reuse in Water Resources Management in Israel. In
Water Resources Management on Israel, 23: pp. 2081-2089.
Shuval H.I. (1991) Health Guidelines and Standards for Wastewater Reuse in Agriculture
Historical Perspectives. Water Science Technology, 23: pp. 2073-2080.
Shuval H.I. (1991) The Development of Health Guidelines for Wastewater Reclamation. Water
Science Technology, 24: 7, pp. 149-155.
Smith R.G., Walker M.R (1991) Water Reclamation and Reuse. Research Journal WPCF, 63:
4, pp. 428-430.
United States Environmental Protection Agency (1992) Guidelines for Water Reuse.
Warner J.W. (1994) Class notes. Colorado State University, Fort Collins, Colorado USA
Wright RR (1994) The Best of Both Worlds. Journal of Water Environment & Technology,
Vol. 6, No. 11, Nov.
Chapter II
Water Supply Conveyance and Control

Chapter II
Water Supply Conveyance and Control
fflTRODUCTORYSTATEMENT
Worldwide demand for water has increased almost continuously in the recent past and
this has increased the pressure on available resources and has led to increasingly innovative
methods for moving and controlling water over very long distances.
Examples of such schemes include the Eastern Cross Harbour Main in Hong Kong
which involved the construction of major transfer mains across a harbour and innovative
methods of pipeline construction, installation and corrosion protection.
The Az Zour Water Distribution scheme in Kuwait involves the treatment and
distribution of up to 720 MIld of desalinated water over a distance of 100 km, with particular
problems of maintaining disinfectant, controlling corrosion and avoiding taste problems.
The London Ring main is an example of a major scheme carried out below one of the
world's major capital cities with enormous potential problems of disruption to existing
servIces.
This Chapter looks at some of these problems and how technology is employed to
provide solutions. Firstly traditional and new technologies in pipeline construction are
examined and then two examples of long distance water conveyance systems, in Brussels,
Belgium, and Wessex Water, Great Britain, are used to illustrate some of the engineering
techniques and information technology now being used to understand and control such large
integrated networks.
1. Construction Methods and Pipe Structural
Requirements
Francesco Calomino
Paolo Veltri
87040 Montalto Utfugo (CS)
ITALY
1.1. INTRODUCTION
In pipeline construction, the use of new technologies is by far less impressive than in
pipeline appraisal and design.
This is true if one considers that every innovation, before it becomes widespread, needs
to be tested for years devoted to practice, documentation of results, discussions, publicity, and
official acknowledgement.
Moreover, if on one hand it is difficult to convince engineers of the advantages of using
new technologies in their projects, on the other hand it is even more difficult to convince
contractors which have to make practical use of these technologies.
Among several aspects involved in pipeline construction, this Section will deal with
current pipe production and main laying technologies. Computational aspects will be briefly
considered for the assessment of buried pipe structural requirements, highlighting those that
could bring real changes in perspective.
NATO ASI Series. Partnership Sub-Series. 2. Environment - Vol. 15

Edited by C. Maksimovic. F. Calomino. and J. SnoxeII
@Springer-VerlagBerlin Heidelberg 1996
90
1.2. PIPE MATERIALS
1.2.1. General
Several materials are in use today for pipe manufacture; some of them are classical
materials, used for decades in this field: ductile iron, steel, asbestos cement, prestressed
concrete. Plastic materials like PVC and PE are more common today: for relatively smaIl
diameters and low pressures, they have completely substituted the former ones in most
countries.
The use of plastics is more and more important for internal lining and external
protection, in connection with other materials, and the union constitutes really new composite
pipe materials.
1.2.2. Ductile Iron
These pipes should be mentioned before all others. Pipes made up of its ancestor, cast
iron, were used since the XVI century in Europe. Today, ductile iron is the dominant material
for pipelines with 250 to 800 mm diam (Osborne, 1994).
In the past, spun iron pipes were obtained by a centrifugation process of an alloy of
Iron and Carbonium (Twort et aI., 1974, The Pipeline Industries Guild, 1994).
With the addition of a small quantity of Magnesium, and using centrifugation, the
Ductile Iron (DI) is today manufactured, whose internal structure is characterised by the
presence of rounded crystals (spheroidal graphite iron). Fragility of the external surface is
prevented by means of a thermic process, giving at the same time a ferritic structure to the
material.
ISO 2531 gives the guidelines for ductile iron pipes, fittings and accessories for
presstH"e pipelines (Brandon, 1984).
The pipes are internally lined by a spun cement mortar layer; externally, they can be
painted with zinc and a second bitumen layer. An external polyethylene sleeving can be used
before laying the pipes in particularly aggressive soils.
Currently, DI pipes on the market have diameters from 60 mm to 2000 mm (of which,
of course, the largest ones are produced to order), and are 6 to 8.3 m long.
Wall thickness should be in agreement with ISO 2531.
Working pressures are relatively high, ranging from 20 to 40 bar.
91
The pipes are nonnally provided with socket-and-spigot joints, although a number of
proprietary types of joints are known and can be used in different situations; a rubber ring
usually makes the joints watertight. A small angular deviation (up to 5° for smaller diameters)
may be allowed by some joints.
Internal lining and external coating, in addition to electrical discontinuity at the joints,
are usually good measures against corrosion. The wall thickness is able to provide high
pressure perfonnance; the weight of the pipes is, on the other hand, high, and the transport and
laying costs are usually high. The need for thrust blocks at angular deviation points is also to
be taken into account, when not using specific joints.
1.2.3. Steel
Steel pipes have been used since last century, in preference them to cast iron pipes for
their lightness and improved resistance, due to the low carbon content. The most widespread
manufacturing procedure consisted in piercing a bar of hot steel and then obtaining the right
dimensions by rolling it (Mannesmann).
Nowadays there are high-strength weldable steels suitable for pipe manufacture (The
Pipeline Industries Guild, 1994), and a steel pipe can be obtained by bending a steel plate and
welding the two edges, or by welding a spirally bent plate (Twort et a1., 1974, The Pipeline
Industries Guild, 1994).
These are usually referred to as "welded" or "seamless" pipes.
Steel is the dominant material for pipelines installed over 800 mm in diam in the UK
(Osborne, 1994).
ISO R559 gives the standard for steel pipes. Fe 360, Fe 420 and Fe 510 are the types
of steel in use.
Pipes in current production range from 40 to 500 mm in diam. for longitudinally
welded pipes, to 900 mm for seamless pipes. Different wall thickness are produced for the
same nominal diameter, according to the working pressures, which gives the designer a larger
choice. Working pressures can be as high as 150 bar.
The first steel pipes had run-lead joints to be used with socket-and-spigot. Today joints
are nonnally prepared for welding; spheric, cylindric and butt welded joints are the most
common types.
The disadvantage of steel is corrosion; protection against corrosion is provided by
means of external glass fiber and bitumen wrapping; internal bitumen lining is in many cases
insufficient and active (cathodic) protection has to be chosen (Brandon, 1984).
92
Polypropylene for external pipe protection can be used; epoxy paint up to thickness of
0.3 mm can be used for internal protection.
In case of severe corrosion problems, clad steel pipes may be used; these consist of an
outer body of lower grade steel with a liner of a corrosion resistant steel alloy (The Pipeline
Industries Guild, 1994).
1.2.4. Prestressed Concrete
For diameters larger than 800 mm prestressed concrete pipes have been largely used,
for the last 50 years, on the grounds of cost (McGhee, 1991).
Pipes are manufactured by winding a spiral of steel wire around a core, made up of a
reinforced concrete or a steel cylinder. The concrete core is longitudinally prestressed by
means of steel bars, vibration and centrifugation are employed to produce the cores. After
stressing the wire, the external concrete layer is completed. Very large diameters can be
produced this way, up to 3800 mm; working pressures usually range from 4 to 12 bar, up to 30
bar in some cases.
Standards for these pipes are given by BS 4625 and ANSI!AWWA C301-92 and C304-
92 (Brandon, 1984).
The joints are of the socket-and-spigot type, with the use of a rubber ring. For steel-
core pipes, the steel cylinders may be butt-welded.
The pipes must be filled and pressurized for a few days before testing to allow
absorption of water by the pipe wall (The Pipeline Industries Guild, 1994).
1.2.5. Plastics
Many kinds of plastic pipes are today on the market: Polyvinyl chloride (PVC),
Polyethylene (PE), Polypropylene (PP), Acrilonitrile-butadiene-styrene (ABS), etc. Thermoset
plastics are cross-linked polymers; while thermoplastics consist of long chains of molecules.
The first material shows greater ductility and capacity to withstand strain. Petroff (1993)
reports the mechanical characteristics of plastic materials. The most common materials for
pipeline construction are discussed below.
1.2.5.1. PVc. PVC, a blend of polyvinyl chloride polymers, was introduced as a pipe
material in the 60's. For this purpose, no plasticizers are added to the material, so it is also
called unplasticized PVC (uPVC) (Brown, 1994).
93
The pipes are manufactured by means of an extrusion process. Current diameters range
from 60 mm to 600 mm; the pipes are usually 6 m long. According to working pressures, the
pipes are usually produced in 3 classes: PN 9 bar, PN 12 bar, PN 15 bar.
Guidelines for pipe characteristics are given by ISO Rl165 (Brandon, 1984).
Joints are of the socket-and-spigot type, with a rubber sealing ring, or welded by the
use of a sealing cement.
One of the disadvantages of UPVC is the tendency to produce local breaks traveling
through the pipeline (Brandon, 1984).
A new material derived from PVC is MOPVC (Molecular Oriented PVC), obtained by
expanding a UPVC pipe by internal pressure. In this way a re-orientation of the molecular
structure is achieved, and the pipe working pressure is increased significantly. Moreover, the
material shows a good resistance to break traveling (Brown, 1994).
Pipe diameters range from 90 to 315 mm; working pressure classes are 8 bar, 12.5 bar,
16 bar.
The latest finding in the field seems to be MPVC (modified PVC); these pipes are made
up of a plastic alloy of great toughness and durability (blend of chlorinated PE and traditional
PVC).
The pipe length is 6 m, the diameters range between 110 mm and 630 mm, working
pressure classes are 10 bar and 16 bar.
1.2.5.2. Polyethylene. Polyethylene can be manufactured as LDPE (Low Density

Polyethylene), MDPE (Medium Density Polyethylene) and HDPE (High Density Polyethylene)
(Beech, 1994).
LDPE pipes are used - in water supply systems - only for the very small diameters.
MDPE is the most largely diffused, and it is reported to be the dominant pipe in the UK for
diameters up to 250 mm (Osborne, 1994).
Due to the lower strength, a MDPE pipe has a wall thickness 2-3 times greater than an
equivalent PVC pipe. The advantage of MDPE is the resistance to long term break
propagation.
Current diameters range between 60 and 600 mm, working pressure classes are 6 bar
and 9 bar.
The jointing system can be butt-fusion jointing and electro-fusion jointing, both
performed by means of the appropriate fusion equipment.
HDPE is a stiffer material, able to withstand larger pressures, but with a tendency to
rapid crack propagation. Nominal diameters usually range between 100 and 300 mm. The
94
pipes can be rolled in coils (up to 180 mm) in lengths up to 150 m. Working pressures are 6
bar, 10 bar and 12 bar.
1.2.5.3. GRP. Glass reinforced plastic (GRP) pipes are made up of polyester laced with
glass fibers brought about in successive layers by centrifugation, the remaining space being
filled by resin and fillers and, in the reinforced-plastic-matrix pipe, sand. The sand is
incorporated into the wall structure to increase the stiflhess of the pipes, making it possible to
handle, transport and instal them.
Joints are of the spigot-and-socket type with sealing ring.
BS 5480 gives guidelines for GRP pipes (Brandon, 1984).
The advantages of these pipes are the very high resistance to corrosion, smooth internal
surface, easy transport and laying.
The pipes have lengths up to 6 m, with diameters up to 2000 mm.
1.2.6. Mixed-Structure
In this category there are pipes made up of a matrix including different fibers; these are
asbestos cement and diffused reinforcement concrete.
Asbestos cement was for over 50 years one of the most popular pipe materials; pipe
manufacture was firstly developed in Italy around the 20's (The Pipeline Industries Guild,
1994), the matrix is Portland cement, including asbestos fibers. Pipes are manufactured in
diameters up to 900 mm; ISO R160 establishes the standard for asbestos cement pipes
(Brandon, 1984).
Recent studies demonstrated that asbestos is carcinogenic when the fibers are inhaled,
implying that such eff~t could affect also fibers possibly released from pipes (McGhee, 1991).
In fact, asbestos cement is now not in widespread use; in Italy, a 1992 law states that
asbestos cement could no longer to be used as a construction material.
Diffused reinforced concrete is one of the most interesting findings in the reinforced
concrete technology, when reduced thickness and weight are needed. This composite material
benefits from a larger specific surface at the contact between concrete and reinforcement, to
increase the pipe strength.
The pipes are manufactured by winding tiny steel wires around a rotating cylinder,
where concrete layers are formed. A longitudinal reinforcement also made up of steel wires is
placed between successive concrete layers.
95
Pipes are up to 6 m long, with diameters from 400 mm to 1200 mm. Working pressures
are up to 15 - 20 bar.
As for prestressed concrete pipes, joints are of the socket-and-spigot type with sealing
rings.
Advantages of such pipes are the relatively reduced waIl thickness, the resistance to
internal and external corrosion, the smooth internal surface.
1.3. CONSTRUCTION METHODS
1.3.1. General
Almost all of the length ofa pipeline is made up of buried pipes, although external parts
may exist, at crossings etc., and in some urban conditions the distribution network, as well as
other services, can be put into underground tunnels. The following sections relate mainly to
buried pipelines and their construction.
Current construction methods make use of the appropriate machinery for different
operations. As this changes more rapidly than the pipe materials and characteristics, the
pipeline installation practice is subjected to rapid improvements.
Moreover, increasing interest is nowadays devoted to new - trenchless - techniques.
In the following sections traditional and trenchless construction procedures will be separately
treated.
1.3.2. Traditional Methods
Traditional construction operations consist of preliminary operations, trench

excavation, bedding, pipe lying, backfilling (Brandon, 1984, ASCE, 1994).
1.3.2.1. Preliminary Operations. Preliminary operations are topographic operations

and site preparation. As to topographic operations, the current practice for pressure pipelines
consists in laying the pipes at a given depth under the soil surface, without following a given
grade, although it is necessary to be aware of the elevation of each point in order to obtain a
correct position for air valves and washouts. Site preparation may involve removal of
96
vegetation and the opening of a track for the excavation machines. It is always important to
locate exactly the services interfering with the pipeline under construction, for example gas
pipelines, underground electrical and communication cables, sewer pipes.
1.3.2.2. Trench Excavation. Trench excavation is usually performed by excavators;

these machines are suitable for widths over 50 cm and depths down to 5-6 m. SmaIl trenches in
urban centres require very smaIl excavators with reduced bucket width, but in many cases hand
excavation is the only solution. The usual trench depth is the one leaving 1-1.5 m cover on top
of the pipe. The bottom width is equal to the pipe diameter plus 50-60 cm; in general it is the
minimum which will allow laying and jointing operations. The trench upper width should be
established taking into account the bottom width, the nature of the soil and its angle of repose.
Excavation of vertical trenches is often possible without any sheeting and bracing. In
unstable soil it is necessary to support excavations, and for this purpose different methods are
used, by means of timber or steel shields (McGhee, 1991).
Trench side slopes, as suggested by the Occupational Safety and Health Administration
(O.S.H.A.) and other suggestions concerning trench escavation, are reported by Brewer and
Hurd (1993).
1.3.2.3. Bedding. The trench bottom has to be accurately prepared to support the pipe.
Foundation can be requested in order to maintain a stable trench bottom. Bedding will
influence the distribution of reaction against the bottom of the pipe. Usually, a suitable bed
granular material, sand or gravel,not self-compacting, 10-20 cm deep, is set on the trench
bottom; in some cases a concrete bed is judged to be more suitable, depending upon the pipe
and backfill characteristics. ANSI A 21-50 1981 (AWWA C 150-1981) shows standard pipe
bedding conditions (ASCE, 1994).
Gravel or sand beddings are suggested for D.I, steel, GRP, UPVC, PE pipes; concrete
bedding may be suggested for concrete pipes and for large diameter plastic pipes.
1.3.2.4. Backfilling. Terminology adopted by ASCE (1994) for backfilling of pipes will
be useful (Figure 1.); foundation and bedding are described in the above paragraph. Haunching
will influence the performance of the pipe; its compaction is essential for flexible pipes,
whereas for rigid pipes it will ensure better distribution of forces on the pipe wall; this
operation is often performed manually. Initial backfill has the function of protecting the pipe
from damage by subsequent backfill. It should consist of a suitable granular material. Final
97
backfill is important for trenches in traffic areas, where it is usually made of material that is
easily densified; otherwise it can consist of excavated material, which should be lightly com-
pacted and left mounded over the trench.
Final backfill
Initial backfill
Haunching
Bedding
Foundation
Figure 1. Trench cross section and backfilling
Brewer (1990) and Brewer and Hurd (1993) emphasize the utility of using Controlled
Low Strength Material, Controlled Density Fill (CLSM-CDF) in filling trenches under road
pavements. This material is nothing else than a sort of concrete, with a low content of cement.
The advantage of using this mixture is the possibility of easily filling the trench directly
from the concrete truck and the fact that compaction is not required. Like concrete, the
mixture dries achieving cohesion; up to 70 kglcm2 of compressive stress is the value at 28 days.
1.3.3. Trenchless Installations
Trenchless construction techniques have become more and more frequent in

underground services like water, sewerage, gas, electricity, communications in urban
environment.
The use of these techniques for pipelines and distribution networks is limited, when
compared to sewers: the reason for this is economic, because the pipes are usually laid under a
one-meter cover, and pipes for trenchless installation need to be specially manufactured.
Trenchless techniques are on the other hand especially suitable in case of railway or motorway
crossings (White and Tregoing, 1992).
98
Two main techniques exist: boring and jacking (ASCE, 1994).

Boring is perfonned by means of special machines, that are able to bore a hole into the
ground, more or less horizontally, working from a pit; the pipe is then pushed into the hole
(van Kesteren, 1992).
Excavation is perfonned by mechanical, hydraulic, or compressed-air techniques (for
very small pipes of up to 60 mm). The excavated ground is taken away from the hole by means
of mechanical devices, or mixed with water and evacuated through slurry pipes; when
compressed air is used for boring, vacuum techniques may be used. Temporary pipes may be
employed, equipped with internal cables and conduits, and then removed and substituted by the
final pipes. This technique is referred to as microtunneling (Orchard, 1992).
Jacking is perfonned by means of jacks pushing a steel or reinforced concrete pipe
through the ground. Behind the jack, strong backstops distribute the load to the soil. The pipe
head has a cutter or a shield, and the excavated soil is evacuated through the pipe (Orchard,
1992, Stevens and Girka, 1992).
Two main possibilities exist today: the first one consists in the use of a steel or
reinforced concrete pipe as a housing, into which the pipeline is inserted; the second one is the
boring or jacking of the final pipe directly. In the latter case, pipes are especially manufactured
to have good resistance to axial thrust; their joints must be able to transmit this force (Henry,
1992).
A new microtunneling technique has been recently presented by Najafi et al. (1993).
The primary thrust force is transmitted through a steel linear casing running inside the pipe. An
air-inflatable internal gripper enables PVC or other flexible pipes to be jacked to long
distances. The frictional resistance force is transmitted through the gripper mechanism which
binds the internal surface of the pipe to the linear casing, requiring less axial compressive
forces than with conventional microtunneling.
1.4. STRUCTURAL REQUIREMENTS
1.4.1. General
Structural perfonnance evaluation is today requested by many national regulations for

pipeline construction, and this aspect has become nowadays more and more important
nowadays with the increasing use of flexible pipes (Katona, 1993).
99
Conditions to be taken into account are internal pressure and external loads, together
or separately according to the mechanical characteristics of the pipes.
Stress analysis or indirect methods, the latter based on the experimental strength of the
pipe, can be carried out.
1.4.2. Internal Pressure
The m(;del is that of the thin elastic ring embedded into an elastic environment;
traditional stress analysis can be easily developed for internal pressure (prevost, 1994).
Internal pressure p, according to the membrane theory, will result only in stresses
parallel to the surface of the membrane (membrane stresses); hoop stress O'} and longitudinal
stress 0'2 are given by, as is well known:
0'
It'
=r£" cr2 =rR.
2t
where r is the pipe radius and t is the wall thickness (McGhee, 1991).
It is worth noting that classical results for maximum pressure values derived from
classical water-hammer analysis are today under question in many cases. Indeed, in the
classical water-hammer theory, the structure is assumed to be rigidly supported; wall thickness
and elasticity affect only the celerity.
On the other hand, as pointed out by Kruisbrink and Heinsbroek (1992), pipe wall
inertia effects related to pressure changes must be taken into account, as well as pipe motion
caused by friction and elbows, tee and end effects; all these influence structural conditions.
A study carried out by the above Authors by means of the Fluid Structure Interaction
(FSI) model, and supported by laboratory experiments on PVC pipes, shows that the classical
values of maximum water-hammer pressures can be exceeded by more than 30010 and the
frequency of pressure waves is increased.
Moreover, the presence of air pockets can either damp or amplify the pressure peaks,
depending on the character of the transient (Larsen and Burrows, 1992). In special cases
severe overload and failure of the pipeline may occur.
100
1.4.3. External Loads
External loads are earth load (dead load) and superimposed load (live load). Pipe
weight and water weight are usually taken into account, although in many cases they can be
neglected. Thermic and seismic actions should be correctly considered.
When subject to external loads, the pipes may show different behaviour: if E' and E
are the elasticity moduli of soil backfill and pipe material, r is the pipe radius and I the inertia of
the longitudinal section oflength 1 (1=t3112, where t is the pipe wall thickness), when the value
ofE' is less than 250 Ell?, the pipe should be considered rigid. In general, metallic (steel and
D.I.) and plastic pipes result to be flexible.
For rigid pipes, evaluation of earth loads in trench conditions (Figure 2.) is possible by
the classical Marston formula (ASCE, 1994):
where w is the load per unit length of pipe,

B is the width of trench at the top of the pipe,
y is the specific weight of backfill,
C is a dimensionless coefficient.
Iw
(0
B
Figure 2. Variables for load computation

101
In practice, the coefficient C takes into account the actions between the backfill and the
sides of the trench, and it is a function of the ratio HIB and of soil characteristics. The
coefficient C is evaluated by:
1- e(-2 km' ~)
C= 2km'
where k=tan2(7tf4 - cp/2), (cp is the angle of internal friction of backfill and m' = tane =
coefficient of internal friction between backfill and sides of the trench).
Values of coefficient C are well-documented, such as that shown by ASCE 1994.

For shallow trenches of depth H, it can be assumed that C = HIB; then
W=yHB
that is, the pipe will cany the load of the solid backfill above its crown. Conservative design
practice is based on this principle.
In case of flexible pipes, the soil at the sides will compact and the pipe will deform
more than the soil; in this case the backfill at the sides of the pipe will cany a part of the load
and the load carried by the pipe itself will be reduced.
The Marston formula in this case will become (ASCE, 1994):
W=CyBD
where D is the pipe diameter and C is the trench load coefficient.
The expressions for this coefficient takes into account soil friction on the sides of the
trench; but there are opinions that this cannot always be relied upon (Stephenson and King
1994). For shallow trenches, assuming C=HIB, is
W=yHD
Appropriate coefficients are developed for jacked pipes, including cohesion of soil
(ASCE, 1994) and for embankment conditions.
102
Superimposed loads are evaluated according to Boussinesq theory, and taking into
account road loads for appropriate traffic conditions.
It is frequent practice to use concrete slabs as a pavement foundation, when a pipeline
is located under a road. In this case, the pressure p on the pipe crown is reduced to a small
fraction of the traffic pressure. For complete information, see ASCE, 1994.
1.4.4. Rigid Pipes Computation
For rigid pipe design, two methods are today in use: the Stress analysis method and the
Load Factor Method.
For rigid pipes, the stress analysis method is based on the superposition of effects
(Figure 3.); the moments and stresses in different points of the section are computed;
eventually the maximum stress will be evaluated. Usually, this concerns the invert section of
the pipe. Maximum stress should be not greater than the ratio of the modulus of rupture of the
pipe material and a safety factor.
Figure 3. Direct method for rigid pipes (pressure distribution)
The indirect method or Load Factor method is based on three edge bearing tests
(Figure 4.), performed at the manufacturing plant. The field strength will be equal to the three
edge bearing strength times the load factor or bedding factor Bf.
This factor is defined as the ratio of the strength of a pipe under any stated conditions
ofloading and bedding, W, to its strength measured by three-edge test, S (ASCE, 1994):
103
W
B f =-
S
Values for this factor were determined by the classical experiences of Spangler at Iowa
State College.
The required three-edge bearing strength for a given bedding factor, R, can be
evaluated taking into account a safety factor Sf:
+
Figure 4. Indirect methjod for rigid pipes (Scheme of three-edge bearing test)
Although the usual procedure in the direct method consists in superposition of stresses
computed by external loads and internal pressure, there is evidence that the pipe strength is less
when the pipe is pressurized.
This effect is consequent to the fact that a pipe under a vertical load is subject to a
deflection, the section passing from a circular to an elliptic shape; the added internal pressure
tends to magnifY this deformation, creating stresses larger than the ones computed by
superposition of effects.
Schlick reported the following experimental equation
104
Wp =~1- P
W Po
where p is the internal pressure
po is the bursting pressure with no external load
W is the pipe strength with no pressure
Wp is the pipe strength when internal pressure p is applied.
The equation was demonstrated to apply to cast iron, asbestos cement, plain and
prestressed concrete pipe.
Since the indirect method can correctly take into account the internal pressure action, it
is today preferred to the direct method (prevost, 1994). Traditional and new methods for
reinforced concrete design are reviewed by McGrath (1993). The Author shows the pressure
distributions obtained by traditional methods as well as those obtained by means of Finite
Elements.
1.4.5. Flexible Pipes Computation
Flexible pipes defonn under external loads, and a part of the trench load is carried by
the side fill; the rest of the load is carried by the pipe, that bends and arches under its action.
This means that the vertical load will partly defonn the elastic ring, showing bending
actions in its wall; and it will be partly transferred to the side fill, giving a compressive action
to the ring longitudinal section (Gabriel, 1990, Hashash and Selig, 1990).
The combined effect in terms of deflection is usually accounted for by means of the
classical Spangler fonnula.
This fonnula gives the horizontal pipe deflection x (Figure 5.) as a function of soil and
pipe characteristics:
D,Kb W
X= EI
3+ 0,061E'
r
where is a deflection lag factor, consequent to the time consolidation

characteristics of the soil; suggested values 1.25-2.5.
~ is a bedding factor
W is the load per unit length
r is the pipe radius
E is the pipe modulus of elasticity
105
I is the pipe wall moment of inertia per unit length

EO is the modulus of soil reaction.
The bedding factor I<J, depends upon the width of the pipe bedding.
Values of soil reaction modulus EO were experimentally found by Howard (1977)
depending on the backfill characteristics and compaction. In reality, these values come from
back-calculation using Spangler formula, but they can not be directly measured.
.".
;",.. -------
I
I ........
.,.
,
\
I I \
-- ------~------~--
I I ,
\ I
\ I
\
, /
I
''''''.......: ~,.//
Figure 5. Flexible pipe deflection
More recently, Hartley and Duncan (1987) found a dependency of the modulus EO on
the cover depth and published a table of values updating the original Howard table. The same
Authors show an interesting relationship between EO and the Young modulus of the soil E..
It should be noted that the modulus of elasticity of plastic materials tends to reduce in
orders of magnitude with time; it is then necessary to take into account its final value.
The theoretical background of the Spangler formula is today under question. Gabriel
(1990, 1993) points out that at least two points should be criticized: first, the ring compression
stiffiless is important for plastic pipes and is to be considered in addition to ring bending
stiffiles; and second, the composite stiffiless of pipe-soil coupling should be obtained in a more
effective way than by simple addition.
Currently, the pipe computation is performed by assessing the value of vertical
deflection y equal to horizontal deflection x (for small values).
Moreover, when dealing with large diameter pipes, the term EI/r may be much smaller
than 0.061 EO. It is recommended that the first should never be less than 10-15% of the second
one, so that the pipe can have sufficient local strength in bending and thrust to develop and
utilize passive resistance of the soil (ASCE, 1994).
106
The flexible pipe computation is based on deflection. Limits to vertical deflection are
suggested for different pipe materials. It is important to point out that superposition of all
loads, including internal pressure, is the rule, except for material not sensitive to strain (steel,
d.i., pvc); in these pipes, the computation can be carried out separately, for external loads and
for internal pressure (prevost, 1994). Anyway, as clearly pointed out by Petroff (1993), waIl
strains should be evaluated, especially for strain sensitive materials, and compared with
expected design strains. To evaluate strains, the current formula is
where f is a shape factor (Ciaponi and Papiri, 1995).
Buckling is an extreme phenomenon occurring when vertical deflection is critical; in

this case, the sides of the pipe pull inward, the lateral soil support is largely reduced and this
brings to the rapid collapse of the pipe (prevost, 1994; Petroff, 1993). Kienow and Prevost
(1989) have shown the possibility that, in the case of low pipe stiffuess, installation and
compaction can result in irregular pipe deformation (squaring), reducing the resistance of the
supposed elliptically deformed pipe and leading to its collapse. Some manuals of practice
suggest how to evaluate the maximum buckling pressure (Ciaponi and Papiri, 1995).
Today, there are models in which soil-pipe interactions are evaluated by means of finite
elements. Pipe structural behaviour is also considered by more refined structural models, based
on finite elements analysis. Zaman and Laguros (1990), among others, show a finite element
procedure for studying both the deformations (strains) and stresses in a buried pipe and
surrounding soil medium. Pipe material and soil medium are treated as elastic bodies and the
interface elements are assumed to exhibit non-linearity.
The study includes the analysis of the interface between the pipe and the surrounding
soil. It is recognized that the interface conditions can noticeably influence the values and
distribution of normal and shear stresses in the pipe and in the surrounding soil.
It is to be said that, at the moment, current design computation of both rigid and
flexible pipes is carried out by classical methods. F.E., although a promising tool, is considered
in many cases too sophisticated for the design scope, because the in situ soils are frequently not
adequately characterized to take advantage of the precision of Finite Element analysis
(McGrath, 1993).
Although Spangler type consideration of soil-pipe interaction is questionable, as
discussed above, simple and effective developments recently shown by Stephenson and King
(1994) lead to a validation of this approach.
107
1.5. CONCLUSIONS
The Section highlights some classical as well some innovative aspects in pipeline
construction.
New materials for lining and coating of traditionally manufactured pipes are examined
and progresses in plastic pipe production are discussed.
As to pipe installation, the most common recommendations concerning buried pipes are
taken into account; trenchless techniques are considered as an interesting development.
As to pipe structural requirements, buried pipe analysis is discussed on the basis of
classical assumptions, pointing out the necessity of more studies on pipe-soil interaction, soil
pressure distribution and real pipe deformation.
1.6. REFERENCES
Beech S.H. (1994) Polyethylene Materials for Pipeline Systems. 2nd International Conference
on Water Pipeline Systems. Edited By D.S. Miller. BHR Group Publication n. 10,
MEP, London.
Brandon T.W. (1984) The Institution of Water Engineers and Scientists (lWES). Water
Practice Manuals. 4., Water Distribution Systems.
Brewer W.E. (1990) The Design and Construction of Culverts Using Controlled Low Strength
Material-Controlled Density Fill (CLSM-CDF) Backfill. Structural Performance of
Flexible Pipes. Edited By S.M. Sargand, G.F. Mitchell and 10. Hurd, Balkema,
Rotterdam.
Brewer W.H., Hurd J.O. (1993) Controlled Low Strengthh Materials - Controlled Density Fill
(CLSM-CDF) as a Backfill Around Flexible Structures. Structural Performance of
Pipes 1993. Edited by M. Sargand, G.F. Mitchell & 10. Hurd, Balkema, Rotterdam.
Brown IE. (1994) Developments in Thermoplastics Pressure Pipe Systems - Molecular
Oriented PVC Pipe Systems. 2nd International Conference on Water Pipeline Systems.
Edited By D.S. Miller. BHR Group Publication n. 10, MEP, London.
Ciaponi and Papiri (1995) Verifica statica delle condotte a pelo libero 0 in pressione. Sistemi di
drenaggio urbano, manuale di Progettazione, costruzione e gestione. Centro Studi
Deflussi Urbani, Politecnico di Milano, I ediz. (draft) (in Italian).
Gabriel L.G. (1990) Keynote Address: Pipe Deflections- A Redeemable Assett. Structural
Performance of Flexible Pipes. Edited By S.M. Sargand, G.F. Mitchell and 10. Hurd,
Balkema, Rotterdam.
Gabriel L.H. (1993) Checking the Premise-Deflections of Flexible Pipe are Predictable.
Structural Perfomance of Pipes 93. Edited by M. Sargand, G.F. Mitchell and 10.
Hurd, Balkema, Rotterdam.
108
Hashash N., Selig E.T. (1990) Analysis of the Performance of a Buried High Density
Polyethylene Pipe. Structural Performance of Flexible Pipes. Edited By S.M. Sargand,
G.F. Mitchell and J.O. Hurd, Balkema, Rotterdam.
Henry Ph. (1992) Pipes for Microtunnels: Suitability of Use. No Trenches in Town. No-Dig 92
Paris. Edited by 1.P. Henry & M. Mermet. Balkema, Rotterdam.
Howard A (1977) Modulus of Soil Reaction Values for Buried Flexible Pipe. Journal of the
Geotechnical Engineering Division, ASCE, January.
Katona M.G. (1993) On the Analysis of Buried Conduits- Past, Present and Future. Structural
Perfomance of Pipes 93. Edited by M. Sargand, G.F. Mitchell and J.O. Hurd, Balkema,
Rotterdam.
Kienow KW., Prevost RC. (1989) Pipe/Soil Stifihess Ratio Effect on Flexible Pipe Buckling
Threshold. Journal of Transportation Engineering, ASCE, vol. 115, March.
Kruibrinsk AC.H., Heinsbroeck AG.T.1. (1992) Fluid-Structure Interaction in Non-Rigid
Pipeline Systems - Large Scale Validation Tests. Pipeline Systems. Edited by B.
Coulbeck and E. Evans, Kluver Academic Publishers, Dordrecht.
Larsen T., Burrows R (1992) Measurement and Computations of Transients in Pumped Sewer
Plastic Mains. Pipeline Systems. Edited by B. Coulbeck and E. Evans, Kluver
Academic Publishers, Dordrecht.
McGhee T.J. (1991) Water Supply and Sewerage. McGraw-Hill International Editions. Civil
Engineering Series. .
McGrath T.1. (1993) Design of Reinforced Concrete Pipe - A Review of Traditional and
Current Methods. Structural Perfomance of Pipes 93. Edited by M. Sargand, G.F.
Mitchell and 1.0. Hurd, Balkema, Rotterdam.
Najafi M., !seley D.T., Pumphrey N.D.Jr, Nishida H. (1993) 'LLB', an Innovative
Microtunneling Propulsion System. Structural Perfomance of Pipes 93. Edited by M.
Sargand, G.F. Mitchell and 1.0. Hurd, Balkema, Rotterdam.
Orchard S. (1992) New Developments mMicrotunnelling Techniques. No Trenches in Town.
No-Dig 92 Paris. Edited by 1.P. Henry & M. Mermet. Balkema, Rotterdam.
Osborne 1.S. (1994). North West Water and the Use of Plastics Pipes for Larger Diameter
Mains. 2nd International Conference on Water Pipeline Systems. Edited By D.S.
Miller. BHR Group Publication n. 10, MEP, London.
Petroff L.J. (1993) Ring Bending Stiffuess and the Design of Flexible Pipe. Structural
Perfomance of Pipes 93. Edited by M. Sargand, G.F. Mitchell and 1.0. Hurd, Balkema,
Rotterdam.
(The) Pipeline Industries Guild (1994) Pipelines: All You Wanted to Know...
Prevost C. (1994) Buried Pipe Structural Design. 2nd International Conference on Water
Pipeline Systems. Edited By D.S. Miller. BHR Group Publication n. 10, MEP, London.
Stephenson D., King. A (1994) Pipe-Soil Interaction with Flexible Pipe. 2nd International
Conference on Water Pipeline Systems. Edited By D.S. Miller. BHR Group Publication
n. 10, MEP, London.
Stevens W.W., Girka A (1992) Ductile Iron Pipe for Trenchless Techniques. No Trenches in
Town. No-Dig 92 Paris. Edited by J.P. Henry & M. Mermet. Balkema, Rotterdam.
Twort AC., HoatherRC., LawF.M. (1974) Water Supply. Edward Arnold Ltd, London.
van Kesteren W.G.M., Bisshop F. (1992) The Excavation Behaviour of Soils in Trenchless
Tunnelling Techniques. No Trenches in Town. No-Dig 92 Paris. Edited by 1.P. Henry
& M. Mermet. Balkema, Rotterdam.
109
Water Environment Federation - American Society of Civil Engineers (1994) Design and
Construction of Urban Stonnwater Management Systems. ASCE Manuals and Reports
of Engineering Practice n.o. 77. WEF Manual of Practice FD-20.
White H., Tregoing C. (1992) Microtunnelling Advances in UK. No Trenches in Town. No-
Dig 92 Paris. Edited by J.P. Henry & M. Mennet. Balkema, Rotterdam.
Zaman M., Laguros J.G. (1990) Predicting the Behavior of Buried Pipes Using a Mixed Finite
Element Approach. Structural Performance of Flexible Pipes. Edited By S.M. Sargand,
G.F. Mitchell and 1.0. Hurd, BaIkema, Rotterdam.
2. Water Transfer Within Brussels Water Services
Daniel L. Brumagne
Compagnie Intercommunale
Bruxelloise des Eaux
Rue aux Laines 70, 1000 Bruxelles
BELGIUM
2.1. INTRODUCTION
In common with many large city water departments, the Brussels Water Service,
Compagnie Intercommunale Bruxelloise D'Eaux (CffiE) now uses water sources a considerable
distance from the point of demand; in this case some 90 Jan from Brussels, capital of Belgium
(Figure 1.).
o ZAVENfEM
BRUXELLES
MONS - ADDUCTION
o CAPTAGE
.tt. NOFlJD
1:1 RE'lFRVOIR
Figure 1. Schematic map of the adduction ofCffiE, The Brussels Water Supply

112
The current system is very complex, with many network interconnections, reservoirs and
sources which have varying yields throughtout the year, while demand on the system also varies, but
in a different way.
Cotes (m)
250 t--
~ ~;..J,; \ c.: q
200r-------o~~-;-;1~5-+-++\~+-+-~r,-rTrj-r-r~~~~~~-'~-
1\ .. , . • .,'
-
~ C.Spotin l .~
C. Modave '" 1..... r- ~ \ .. ,
_ r- t.. .. .. Z.S.H.
r-------~~~~~~=*~~~~~-+rrrK~~~~~~*,--
150 t-----_+-+1rt--+-t1-HAr---tHitt-lt1~~\;::±rtt-ln-t--+--t-I\rt-t'6--,.---
v 1\ 1---\ lO'''' 1
Z.H.
II \ ~\ ~, ~.I
\ 1\ IV 1 I r\ W- Z.M.
l00t--------+-~-+-r1r~~r\~~_*~\~+_MI~~'+_~~/+_\-
( .Ta Ire , \ ~ I! I
1--II11t1I-~~n-tlltliriQ..E'&Ec{s;:f\4~t--n-;h,1It\1V1l~1-p-.... Z.B.
, \ , /I ' •
I., !! !! Ii Ii I' " , ""r--{..
~ ., eos
::s
~o o~
50
-
I....
F'
/
I
eSp" .>0:
.M os (.B
• 8
.of
~iD!-I
.::l
.~
,~
•
•
lOOkm 75 50 25 •o
Figure 2. Schematic profile of tile adduction ofCmE
The remote water sources need adequate mains with capacity for current and future needs
and also for the capacity of the source itself The paradox for the water services is that demand is
greatest when resources are scarce, very often during hot dry summers.
113
There are obviously daily, weekly and yearly variation in water demand, and a system of
transfers, regulation works and reservoirs can have great complexity. This is the case for Brussels
(Figures 1. and 2.). The variation in the yield of the different sources of ground or surface water can
also be important and it is necessary to search for increasingly remote resources.
450
400
350
300
250
200
~ I-rl-If , r1 ~ r ~ , i : T 1 : : i
150'
-I ~ i 1
' "
r ' i : i i i COMMUNES
, "
ASSOCIEES I i
,
I:: : :: ~:::IIII PtJt
: :: : :: : ' , : i i : 1':111:"1\1
: :: : :: : T.M.V:W.I! r.Ji.. ...~.~. :
1m:~I~~
100
i J.......
, ...... , r~1 ,..
..L.I.:M:\:"::
~ 1":'" :::: 1
:
:
~
I
I
I
I
I
:- I
I
;--:
I I
1- I 1 I: :.: ~ : :
1
I
:
I
50 : :: COMMUNES CLIENTES ~ :
.,: :!-....
,. """....~-,~-=.:.:-..,.I-rT(-, -,_, , III 1 I..i.., i l l
, "r"l' , ,. ~
! I!
: :
:rri-=-r,
I ! II
o I
I I I
I
I
I
I I I
I
I
I
I
I
I
I
I
I
,
I I
I
I I
1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992
Figure 3. Diagram of the tota\ monthly water conswnption of the ernE from 1972 to 1992
Groundwater sources often have less viable yields than surface resources: the catchment of
Modave, into the Dinantiallimestones, has a medium yield of 6O.000m3 /d, with a minimum of
45.000 m3 /d and a maximum of 85.000 m3/d, while the river Meuse, at its entry in Belgium (near
Givet) has a yield varying between 10 and 1.000 m%.
114
For peak periods of consumption, it is often useful to have a big surface or underground
reservoirs, such as old quanies or mines but the underground reservoirs are the cheapest and most
important when their capacity is big enough. The transfer lines of the eIBE are of following types:
aqueducts, open air channels, feeders and water pipelines.
2.2. THE AQUEDUCTS
The aqueduct is a very ancient system used for the transfer of water. The Romans built a lot
of famous aqueducts all around the Mediterranean. There are also very old quanats in Iran or in
Maghreb.
Figure 4. Cross sections of some aqueducts

115
Their gradient is very swallow to avoid unacceptable velocities. They are always built with
local material: brick or concrete. The valleys were crossed with bridges, such as the 'Pont du Gard'
in France (Figure 5.), they are now crossed with pressure pipelines.
Figure 5. The Mont-Saint-Pont aqueduct, Braine-\' Alleud Figure 6. Internal visit of an aqueduct
Sudden and big variations in yield are to be avoided. Therefore the size of an aqueduct
depends on the importance of the source itself, often spring water. If the spring is big enough it is
better to catch only the minimum available yield, stable all year round. As this is often not the case,
the size of the channel depends on a yield that can be reached, say at least 10010 of the time and,
when the real yield falls below this value, level and velocity need to be regulated in the channel.
116
An aqueduct is a vulnerable system when it crosses some aquifers, but quality control can be
easily achieved; at some valley crossings, for instance, where polluted water can be diverted.
The stability of aqueducts is usually good, especially underground and at depths of>30 m:
the Trajan aqueduct, for instance, is still in service for Rome. The aqueducts of ernE were built in
1855, 1895 and 1920. Their maintenance is easily carried out during visits made every 3 to 6 years
(Figure 6.).
However, some problems exist when these works cross karstic areas.
Except for the largest yield (> 10m3/s) it is good practice to cover the aqueducts with more
than 1 m of ground: evaporation is reduced, temperature is regulated, pollution by air born
pollutants is also avoided.
Figure 7. Concrete gunitiny in the aqueducts Modave-Mazy

117
Two new techniques are useful in the construction of aqueducts:
• microtunelling for new aqueducts at big depths,
• concrete gunite for new linings up to 700m ofaccess shaft, for old aqueducts (Figure 7.).
2.3. THE FEEDERS
From the beginning of this century it has been possible to use pipes of bigger diameter (1 m)
for the new transfers; these pipes are made in cast iron, steel, asbestos-cement or concrete,
reinforced or prestressed. Prestressed concrete is very good for large pipes (> 1.2 m) (Figure 8.).
Figure 8. Trenching and laying of the feeder
Obviously feeders are always to be placed below the piezometric line. In most cases, they
join two big reservoirs, the first situated near the source, the second one near the consumers. That is
118
the option chosen by the CffiE. The pipes generally follow the ground topography; air valves have
to be placed at all high points, discharge valves at the low ones in the valleys. This pipe system has
better flexibility than aqueducts above all when it serves some different reservoirs because it is easy
to connect two feeders through a valve complex.
Feeders are less vulnerable to pollution when they are pressurised unless there is an
accidental pressure fall.
During the layout of a feeder, the biggest problem for the builder is often the provision,
transport and the placement ofits big pipes.
The pipes of each feeder are to be designed to sustain the maximum internal pressure of
water, taking into account eventual surge pressure (water hammer). They have to resist all external
forces, vertical or other (ground pressures, charges of traffic load) and also longitudinal forces
(caused by temperature variations): this is the case for pipes used for the transfer of surface water
the temperature of which may vary between 2° C and 25°C.
Feeders allow rapid and important velocity variations; this is useful in adjusting the yield to
demand. For instance, in less than one hour, the yield of the Hainaut feeder can vary from 1.000 to
4.000 m3/h. This gives great flexibility to every system for water transfer, especially in the case of
accident on an aqueduct or feeder; a good water service has to take into consideration the
probability of such accidents.
Figure 9. Horizontal boring below an express way near Mazy

119
2.4. SPECIAL WORKS
Aqueducts and feeders often have to cross rivers, canals, roads, speedways and railways.
The building of these crossings raise the same problem for both transfer systems. Techniques used
for crossings are quite traditional now except for the horizontal boring ordinarily used in soil, but
also in hard rock (Figure 9.).
For these crossings, especia11y when the soil is loca11y compressible, it is often necessary to
place the water pipe into a deformable channel (an external pipe), almost always in reinforced
concrete. This technique is used for the crossing ofa canal in the surroundings of Mons (Figure 10.)
and also for the crossing of a new expressway south of Brussels (Figure 11.). These protection
structures are highly useful in the case of pipe rupture and in a110wing the repair of the failing pipe
(Figure 12.).
The connections between various transfer lines, aqueducts or feeders, requires the building
of complex valve chambers. At the upper Mazy connection (Figure 13.), a regulation tower was
built, the aim of which was to avoid excess pressure in the downstream part of the feeder (Figure
14.). To avoid surge pressure in the biggest feeders, with great yields and high static pressure it is
useful, as in the Tailfer works (Figure 15.), to reinforce the pipe itself and to connect this pipe to a
special purpose reservoir.
Figure 10. Defonnable channel below the 'Canal du Centre' near Mons
120
Figure 11. Channel below an express way near Waterloo
Figure 12. Replacement of a steel pipe in a channel below an express way, Wauthier-Braine
121
Figure 13. The Mazy valve chambers
Figure 14. The Mazy regulation tower

122
2.5. REGULA nON AND CONTROL
Spatial and temporal transfer of drinking water requires accurate information - quantity and
quality. It is thus essential to place sensors and fittings at the upstream and downstream ends of a
waterline, also often in the middle of them, to measure flow and pressure, turbidity and chlorine
residual (when chlorine is added). The effiE has consequently equipped its waterlines with
flowmeters, pressure gauges, turbidimeters, chlorine monitoring in its catchments, along its
aqueducts and feeders, and in its regulation reservoirs.
In order to manage all this information a dispatching system was built in Brussels in 1973
and recently (1993) renewed. All the information is updated at least every two minutes.
Figure 15. Pwnping station at Tailfer, near Namur

123
I T-C L2 1
amonl
M~;'====j====t~===t3i~ 91 ~
pontin ~
Foret
liteseryes d's\'sl . 175 ()()() m3 I
06110194 15:26:32 BEN-AHIN
Figure 16. Schematic view of the adduction of the ernE as to be seen on the screen of its central dispatching
I pz t96.1m I
TOOL
[Y]ill
§
~ (g£]QID r- Actions-
.~V4k-'eV...4{]JlIJ."";-=-4i"'.L..-A-.....--(
BOIS Df7- L
VILU'RS \ \"
I \
--+ V4"
)...~~......!.CALLOIS
MARBAIS
Supervision
Incrustatton
Inhibition
VIO
I \ V8
LAND~~ L L , ...m--[[D-.....- .....
D2
~ ~O;::iI:::::"
Entree Man .
EM\lI.'ES I \ I Telccornande
\ \ \ VI6 Selection
I I Acqu inement
\ I
\ I VJ1 IMcnul~
\ ~
\ V IS IT§I]I PWDI
MAZY rlliL::JI~
'r-~i-{]J~~®-----1H )--~)--+-l~ AVA!. Zoom & Pan
Imm@81
\
I D6 V!5
\---- -- - -- - ~;.J~~I I
I a£D1SJ1SI
I
I
MAZY -AMONT [.li][]~
00 ~~
I 2
~~ *"-10
SPONT1N
z· "
: MAZY
AVAL
CL:iiiJ~
---..
~
I 06110194 1~ 05 '23 RES. RHODE V6 VANNE 64 DEFAUT CALCULE DF.FAUT
Figure 17. Schematic of the Mazy chamber as to be seen on screen

124
This new computer system is controlleed by six dispatchers who relieve one another
permanently (Beal, 1987). The dispatcher in charge can always see the entire complex of transfer
lines (Figure 16.) and choose to zoom on a specific installation as, for instance, the upper Mazy
connection (Figure 17.) or the Callois Reservoir (Figure 18.). From Brussels the operator is thus
able to check all the waterlines, command valves and pumps (Beal, 1987). In the case of pollution in
a catchment, the operator has the option of opening the discharge valves and closing the line
valves, before the entry of polluted water into aqueducts or feeder. But in all such emergencies, of
course, chemists and engineers in charge of the production service are informed and can override if
necessary.
if]
DIS
MAX
T.P.
~
...
Reservoir 2
C ....... IIJlIIJIIl.1
--
I
~~~: TOOL
Actions
SupelVision
Incrustation
Inhibition
Consignation
Substitution
Entree Man.
Telecomande
o Selection
o Acquittement
SWDE
... --
Lillois
V21
NOEUDDE
PLANCENOn
CALLOIS V Haul 21.0

V Bas 48.1
Oo--mu2J.2Y:$QO V olal
Choix HautlBas
~ ~
07/1 M>4 ()9,23: 11 RES RHODE V3 VANNE DEFAUT
Figure 18. Schematic view of the CalIois complex. as to be seen on screen
2.6. DURATION OF THE TRANSFER
As the mean average velocity of water is about 1 mis, water from each remote catchment
takes less than one day to reach one of the regulation reservoirs built upstream of Brussels, in the
125
southem part of the complex. One or two days later, water can reach the customers tap. That is the
best situation, when all the feeders and aqueducts are nearly at their optimum operation. But, when
a new waterline is put into service, waterflow and speed are low, the duration of the transfer is high
and the quality of the water still has to be checked. To avoid low yields it is often necessary to
balance water between resources.
The temporal transfer of groundwater does not create problems because such sources are
diverted into artificial channels, below the ground level up to closed reservoirs, distribution network
and taps. But, when one has to impound, treat and transfer surface water, and mix those with
groundwater one may have some problems with the growth of bacteria or other micro-organisms:
surface water, even completely sterilised, contains nutrients for micro-organisms always present in
groundwater. One therefore always has to check the chlorine content of those mixed waters and, if
necessary, adjust this by adding more chlorine, and this is usually carried out before the entry in to a
regulation reservoir, and sometimes in the city network.
2.7. REFERENCES
BeaI D.' (1987) The Implementation of Information Technology to the Year 2000 - A Realistic
Approach. Symposium WateriEau 2000, Nice France, pp. 3-1-1 to 3-1-7.
BeaI D. (1988) Integrated Distribution Management Systems. Aqua, Great Britain, No.5, pp.
267-276.
3. Transfer of Water and Data Management in
Wessex Water - UK
Joho Sooxell
WESSEX WATER
WESSEX House, Passage str.
BS2 OJQ Bristol
Duiao ObradoviC
WINS Ltd.
Tower House, Parkstone Road
Poole, Dorset BH15 2DH
UK
3.1. INTRODUCTION
Wessex Water supplies water to some 1.lm customers in the region between the English
Channel and Bristol Channel, see Figure 1. Water is captured at 153 sites of which around two
thirds are in constant use,with the rest at standby or used for stream support. The distribution
system has some 350 pumping stations and over 400 reservoirs, arranged in several hydraulically
independent systems - the linking of these systems is not yet fully completed. The total average
demand reaches 450 MIld.
In the last ten years a number of major trunk mains have been introduced to link areas of
surplus and deficit - these are shown in Figure 1. The major link was between the surface sources in
the west and the groundwater (chalk and limestone) in the east. This was completed in 1990, when
the south of the UK was suffering from drought and was one of the reasons why Wessex did not
have to impose customer restrictions during the period.
With the gradual integration of resources, management has become increasingly centralised
and an efficient data management system for the regional network is essential (Huntington, 1984,
1990). This Section describes the development of such a system and the way in which telemetry
data, GIS, billing, and other data sources have been integrated to provide detailed information for
operational management, leakage contro~ levels of service and so on.

128
BRISTOL WNER
~~v--~-"",..l
Dorset Trunk
Main
Figurel. Wessex regional area
First, the region was organised into three Areas (Dorset, Somerset and AvonIWtlts), which
were in tum subdivided into some 29 Supply zones as shown in Figure 1. Two further layers were
created below supply zone level:
• District Metered Areas (DMAs),
• Control zones.
DMAs are the basic unit of territory now used in the UK for monitoring leakage and are
relatively stable- in Wessex there are now 335 DMAs which are monitored continuously by data
logger or telemetry. Below DMA level there are numerous more Control Zones which are
monitored for changes in minimum night flow etc as a direct aid to leak location.
129
Water source infonnation was collated and checked and detailed schematics were prepared
under a quality management system (BS5750); an example is shown in Figure 2.
I m:~
.,
m~ AJICo6O
.~~---------- ... ~~ :_,..

I
\i, 1Uo\T1!D SAMPU! : TI.WT.D SA).(Pt.£
I
----------r-_
ctl """"""
,"''''''''_~
!-------------;
:
:------------------- ~~~<1J=
,-------t------I MJ'TO.tJi\If'DOW'N No. I l~
st.mDSmLE.PV).OIt.
--.- .... ~~~
W. '--0------,--------,
=
.....
-f\.OWMEmI,
-\,A.LW
~ "NONU!:n.ItN
VAL...
rrrnz--~-'-~-~----~-~-.----_r~
--~-~
-.-'----r_---_+----~ ~
Figure 2. Site schematic produced for BS5750 certification
3.2. REGIONAL TELEMETRY
At the same time that the supply network was being integrated the existing telemetry
system, which dated back to the mid-70's and was a distributed system, was also centralised and
rationalised (Huntington, 1993); this presented the opportunity to collect all data in one central
telemetry database and establish regional systems for the identification and naming of flow meters,
level sensors, pressure transducers and so on, all essential steps in the creation of a centralised
management information system.
The telemetry system is very large, collecting some 1500 analogue signals and expanding
fast towards its design capacity (6000 signals).
130
All operating data are collected in VAX Central Telemetry computers (duty and standby)
and refreshed continuously. Operators in the Regional Control Centre (R.O.C.) can monitor any
part of the system from one central point and can intervene with specific control capabilities where
appropriate.
Other users can have the access through dial-up computers, see Figure 8. (Huntington,
1993). Every 15 min the key data (mostly flows, pressures and water levels) is transferred to a PC
File-server where it is available for all analysis and archiving purposes ( Snoxell at al., 1989, Snoxell
19933, 1993b, 1994).
The telemetry covers practically all sources, transfers between Supply zones and main
distribution trunks, so these data permit computation of:
• Source outputs,
• Total demand (consumption and losses) in the Region, Areas and Supply zones.
Dial-Up Access
o RO.C.
Central Telemetry Computers (VAX)
Duty Standby
Regional Megastream Network
Divisional
PC Fileserver Gathering Points
( etware)
Area Gathering Points
Outstations
Analysis &
Archiving
Measurements & Commands
Figure 3. Regional telemetry
The overall picture is completed with other data, as shown in Figure 4.

131
Billing data is regularly transferred to the same PC File-server as Telemetry data from the
billing centre. There are three type offiles:
• monthly-read customers (roughly 2000), the largest and the most important ones whose
water meters are read every month,
• commercial users (40 000 approximately), controlled twice per year,
• domestic users (over 23 000), also controlled two times every year.
These data are sorted using the UK Post Code System, established and maintained by
Royal Mail. Briefly, a Post Code denotes a small area with up to 50 addresses clustered together,
with known coordinates. These coordinates are linked to the Demand Management Districts
through GIS, as described below, so the total metered consumption can be computed for each
District and hence for Supply Zones, Areas and ultimately the Region by simple summation (note,
however, that this in not the total consumption of water because the system is not fully metered).
II
-
- - - Fresh Dillll
Divisioml
PC Fi1cservCtll
Local Alta Nctworl<
Portable
Computers
t Logger l>ala
Manual Reading<;
t t
Figure 4. Merging all operational data together
132
The data from Telemetry (VAX computers) and Billing system (a mainframe) are easily
processed and stored in the Master Database, on the PC Fileserver I (Figure 9.) because they come
from a single source. The other data come from over 1000 tlowmeters and watermeters located all
over the Region. Some of these instruments are equipped with portable data loggers (350+) and
others have to be read on the spot (manual readings).
Earlier attempts to organise centralised collection & processing of this valuable information
ran into difficulties and it was decided to organise distributed data processing instead.
Operational staff equipped with notebook PCs can see the maps provided by GIS, as well as
operational data for the last 100 days for their respective in great detail. The inspector then can
enter manual readings and/or download data from the data logger, for all the sites which he has to
cover.
GIS(VAX ) Local Of1ice

PC Filescrvcr
,D'' ' ' 'J! .... 0
•
etwork PC
CGM
Conversion Standalone
AquaMap Weekly Updale Procedure

Portable
Computer
Figure s. Regular update of GIS data
The inspector regularly (once per week) reports to a divisional headquarters where a file
server keeps the data base updated every night by the Master database (Figure 9.). Once he
connects his notebook computer to the divisional PC fileserver two tasks will be carried out
automatically:
• fresh data from his notebook will be taken by Divisional PC fileserver,

133
• the database on his notebook will be updated with new telemetry and billing data, and also
data from other sites.
The process takes a few minutes only. The field data are kept in the Divisional PC fileserver
until night when they are transferred to the Master database for finaI processing.
Divisional PC fileservers (now six) provide additional capacity and flexibility for the system;
even in the worst case - when all communication with the centre is lost- the local staff can continue
to work unhindered, nothing will be lost.
The software on notebooks is very user-fiiendly and specifically tailored for field use. It is
very robust and fool-proof No special training is needed. Note also that all the hardware is standard
PC configuration.
The Historic Data for The Region: Wessex Water

Demmd (MIld)
550.0
500.0 I A Mean Value is
N~
428.045 MIld
450.0
)
JJ / P e 8 ks (MIld):
SINGLE Day: 517.595
V THRlili Oays: 509.274
400.0 k\ ~Mt ~ IrJA ~ 1rJ~ l~jI\ / SEVEN Days: 503. 706
350.0
Total Demand
Tune Step: Day

300.0
01A:l2194 Fe!: Mar Apr May Jun Jul 31A:l7194
Figure 6. Wessex Water tota1 production

134
3.3. MAPPING OR GIS (Geographical Infonnation System)
Wessex started to computerise its asset infonnation in the mid-80's and by the time that the
supply network was integrated and the telemetry system centralised and rationalised, this process
was similarly reaching completion.
One major step forward in the last 2 years has been the opening up of this data to personnel
across the company using PC hardware, both on the desktop and with portable equipment for field
use (Glasbrook, 1993, Snoxell, 1993).
The GIS department keeps records for all assets of the company, using highly specialised
equipment for digital processing.
The changes in the water supply system are regularly updated and this updated information
is kept in the GIS database. Once per week the updated files are transferred to the Central File
server from where they are disseminated to several local file servers, see Figure 5.
The finaI users have access to the updated files through their networked personal
computers, standalone computers or portable computers.
So far the paper has described the major management infonnation systems that were put in
place over the last 5 years to improve the control of a supply network which was becoming
increasingly centralised and integrated and involved moving water over increasing distances via an
increasingly complex trunk main system.
The following section descnbes some examples of the information which is now in common
use for the operational management of that network.
3.4. ON-LINE ANALYSIS OF OPERATIONAL DATA
3.4.1. The Production of'Water
The data captured by telemetry permit a rapid calculation of total production of water - see
Figure 6., as well as demand (including losses) in all Supply zones, Areas and Region.
The current production of individual sources could be also seen and compared with historic
data and abstraction license (Figure 7.).
This information is usually available for the previous day, but in a crisis it can be updated
every 15 minutes giving vital data to the manager.
135
NORTWILT - AquaFlow & LCP Data for: NORTWILT
Flow AquaFlow
MIld LCP-Total
20.0 LCP-MNF
IS.0 <B+-I---it"(;
10.0 <Ir---+---+---+----+----+---:-~~>
5.0 G----+----+---+----+-----+--~:I)
0.0 <L.'.:~:;~:.;:::L::=:::::::::·'· . ,": ...... ,,...,.,J-,-~.....- - _...._ _, : End-of-Period: 30f06l94

TIme (days) 30f06f94
0l/OIf94
Figure 7. Production and consumption data for North Wiltshire
3.4.2. Monitoring of Leakage Losses
The demand in all Districts is monitored either by telemetry or by data loggers. The
introduction of dial-up loggers is planned for the end of 1994. The data permit:
• computation of losses in the conveyance system (primary) as the difference between

quantities produced at sources and those delivered to the Districts,
• evaluation of distribution losses within individual Districts taking into consideration

population and billing data as well.
136
The analysis can be performed at each level from the Region down to individual Districts.
Figure 7. shows results for a Supply zone (North WIltshire).
FOXHILL - Demand in the District FOX HILL S44

Thu Fri Sat Sun Mon Tue \lkd Thu Fri Sat Sun
Configuration:
2000
MNTren : -35. m3/d Ii
I AW067 1.00 Logger
Flow Date Qave Qmin

m3/d
10103194 1086.8 662.7
1600 11103194 1073.6 661.0
,
12103194 1064.4 670.5
13103194 1063.3 650.6
14103194 1003.8 684.3
1200
., 15103194 919.3 521.9
16103/94 858.7 539.1
V VI 17103194 803.4 400.0
18103194 790.7 406.9
800
~ ~ 19103194 825.4 391.4
20103194 841.4 387.9
-~ ltf.- __ Total Volume: 10331 m3
:---
l-
I- -I--
400
BathlFoxHiIl
0 10 12 14 16 18 20 End-of-Period: 20103/94
10103194 Time (days) 20103194
Figure 8. Variations of water demand in a district
Note that Production (full line) and Demand (broken line) data are very closely matched,
showing that there are no significant losses in the primary system in that zone. The stronger
oscillations of Production data are due to changes in storage (not taken into account).
Note also the Minimum night consumption line for the whole Region - it is quite steady,
although a bit too high.
The distribution losses in a District could be estimated by analysing the variation of local
demand - see an example in Figure 8. for Fox Hill District (Bath Supply zone, Avon & WIltshire
Area).
137
Note the sudden decrease in demand after 14 and 17 March 1994, after successful repairs,
with obvious savings. However, the leakage losses are still relatively high, as could be proved by
examining the data in a longer time period (Figure 9.).
FOXHILL District: FOX HILL S44 - Zone/Area:. AVONWILT-BIIDI
1250~-----------------------------------,
Flow
(m3/d)
Total Flow
IUOO Min Night Flow
N-r ofPropcrtics: 1335
750
A\erage values:
- Total Flow: 821.77
-Night Flow. 419.21

500
250
o Bath Supp Zone

15/09/93 Days 31107194
Figure 9. Historic perspective for the same district
The present level of night time demand (about 400 m3Jh) is still higher than in September
1993 (below 250 m3h) - therefore the work is not finished! The significant difference between actual
night line and legitimate night use (computed from water bills and average domestic water use) also
points in the same direction.
138
TDWHYCHR District: Whychurch • Zone/Area: AVONWILT·NORTWILT
750 ~----------------------------------~
Flow
(m3/d)
Tolal F10\
_ _ _ Min Night Flow

6UU
N·r orPropcrties: 2656
450
Average \lIIues:
• Tolal Flow: 441.03
. Night Flo\\': 143.37

300
150
Successful Cmpn
Figure 10. Reducing the leakage levels steadily ...
The campaign against leakage losses can never be completed, but continuous efforts can
achieve substantial savings, as shown in Figure 15. for another District (Whychurch). Note the
steady decline of night line during these ten months, even when the total demand was increased
during a dry and warm period (June-July 1994). The sustained effort in pressure management will
also help in keeping the losses within acceptable limits (Lonsdale, 1984, Tustin, 1994).
The analysis is made easier by providing tables for Supply zones, Areas or the region where
Districts are listed in order of high leakage losses (by trend or by average).
139
3.5. DESIGN AND PLANNING
The archived data offer great advantages for all design and long-tenn planning jobs within
the Company, providing a complete and accurate database. A few examples are included here
showing:
• Figure 11.: Variations of water demand in Poole area in the last 14 years - note
the rising trend, decline during the recession period 1990-93 and
recovery in 1994; also the repeatability of seasonal patterns,
The Historic Data for Supply Zone: Poole Zone
Demand (MUd)
!
60.0
1
50.0 Mean \hlue is
J ~~ A~ln 1
_ _ 4O.252MVd
I
<~~\i~m"~~"'-·- -_.. ''t~; I~
40.0 I I. :>
Pea k s (MUd):
SINGLE Day: 66.482
THREE Days: 59.776
SEVEN Days: 59.047

30.0
20.0
Recession Ended
Time Step: Week

10.0
198 82 83 84 85 86 87 88 89 <j() 91 92 93 94
01/01181 31/12194
WARNING - Data averaged with step: 7
Figure 11. Variations of demand in Poole zone

140
• Figure 12.: Total demand at the regional level in August since 1981 set against
rainfall data - note the relative increase in dry years (1983, 1989,
1990) and decrease in wet years (l985, 1986, 1992),
Wessex Water Pic - Monthly Demand

Demand (MIld)
Regression line
100
500. Trend: 6.42 Mlld/Yeor
Yintr: 364.07 MIld
150
i = 22.45 MIld ( 5.5%)
0 200 Year Regres Demand Y

450.
0 - MIld MIld %
1981: 370.50366.96 -1.0
1982: 376.92 359.50 -4.6
0 1983: 383.34 416.13 8.6
1984: 389.77385.73 -1.0
400. 0 1985: 396.19 372.17-6.1
1986: 402.61 389.30 -3.3
1987: 409.04424.47 3.8
Rain 1988: 415.46411.04-1.1
0 mmlm 1989: 421.88446.81 5.9
1990: 428.31 460.64 7.5
350. 1991: 434.73 434.19 -0.1
1992: 441.15402.97 -8.7
81 83 85 87 89 91 93 Month: August
300.
Years
Figure 12. Effects of rainfall on the total demand

141
• Figure 13.: Source output for Ivyfields pumping station and limits set by Daily
& Annual licence,
The Hiltori:: Data for The Source: Ivyfields Source

Demand (MIld)
25.0
Meon Value is
20.0
-- r-' -- - - _ . .- - - I - - - - - - - _.- 11.844 MIld
Licences
f ___ 18.700MlldDaily
tJ
15.0
.. __ .. _ .. 18.685 MIIdAnnly
J\ A ~ f\. 1 VI Peaks (MIld):

~ V \I V
~ I \) \
SINGLE Day: 18.584
v'\
10.0 v
THREE Days: 18.514

SEVEN Days: 18.113
5.0
Source Output
0.0
1994 Tmte Step: Week
01101193 1993 31112/94
WARNING - Data a\eraged with step: 7
Figure 13. Current output ofIvyfieJds source

142
• Figure 14.: Consumption registered by a watenneter belonging to a large

hospital in 1992-1994 period (billing data) -the artificial peak in July
1993 should be halved to cover the gap in June 1993 (human error)
District: GARLAND - C... tomcr: EAST OORSET HEALTI ( AUTI (ORlTY

500.0(b-- - - - - - - - - - -- - -- - - - ,
Reference: 1390123006
Post Code: BH I5 2JB
400.0
Coordinates:
-- Easting: 40190.00
-- Nortbir@: 9130.00
300.
(ncb;lJy Type: 520
Monthly-read Meter
A\erage : 229.624 m3/d
Maximwn: 600.000 mJ/d
Minimt.m: 195.564 m3/d
BiUingpolicy
0.0 (I,.-..........J...;:..:.::;J'--""-..-.L..:;;.J.-"-..L...--L-=--'>-'-"J....:....-......._ ...............- . J
15103193 Days 15103/94
Figure 14. Billing data for a large customer

143
• Figure 15.: A week's data for the same meter captured by a data logger
(temporarily insta11ed) - the data show the real night consumption
which is quite low in this case.
Data for District Meter: POOLE - Name: Poole Hospital ..... .
Thu Fri Sat Sun Mon Tue Wed Thu Fri
750 ~--~~--~---G~--~---@~~~--~--~~--~
Flow
(m3/d)
Date Qave Qmin

600 ~---+--I----+----+--I----+---+---il----~
06101/94 275.8 126.7
07/01194 224.3 63.4
08101/94 151.0 72.0
09101194 148.6 69.1
450 &--+---1-",
10101194 231.7 69.1
11/01194 234.1 80.6
12101194 231.6 66.2
13/01194 208.0 0.0
Total 'kllurne: 1705 m3

Logger Data
06/01194 Time (days) 14/01194

Start on: 06/01194 at 11 :55
Figure 15. Monitoring the consumption of a large user

144
3.6. REFERENCES
Glasbrook D. (1993) The Practical Application of GIS, and its use for AMP2. IWO Conference
1993, CardUl: Great Britain.
Huntington R (1984) Improvement of Cost and Perfonnance of Distribution Systems. Proceedings
of XV Congress of International Water Supply Association, Spec. Subject No.9,
Monastir-Tunisia
Huntington R (1990) WESSEX WATER's Integrated Water Distribution Management System.
Proceedings of the 5th IAWPRC Workshop, Japan.
Huntington R (1993) Updating and Development of A Regional Telemetry Scheme. 6th IAWQ
Workshop on Instrumentation, Control and Automation of Water & Wastewater Treatment
and Transport System, Canada
Lonsdale P.B. (1984) Automatic Control of Pressure in Distnbution Systems. Wessex Water Co.,
Bristo~ Great Britain.
Snoxell J.D., ObradoviC D., Cavor R (1989) The Use of Real-Time Telemetry and Modelling in
Conjunction for the Operational Management Water Supply Systems. Computational
Modelling' and Experimental Methods in Hydraulics, HYDROCOMP '89, Dubrovnik,
Elsevier Applied Science, London 1989, pp. 301-310.
Snoxell J.D. (1991) New Technology in the Water Industry - Making It Stick in the Real World.
Proceedings of the Pipeline Industry Guild, England.
Snoxell J.D. (1993) Network Management in Water Supply. International Conference on Computer
Applications in Water Supply and Distnbution, De Monfort University, Leicester, England,
pp.331-345.
Snoxell J.D. (1993) The Use and Value of GIS in the UK Water Industry; Experience in WESSEX
WATER Services Ltd., IWSA Specialised Conference on GIS, Lyons 1992.
Snoxell J.D. (1994) Management of a Widespread Multi-Plant Water Supply System. IWSA
Regional Conference, Zurich 1994.
Tustin JR (1994) Economic of Leakage in a Water Supply System. Presented to Royal Society of
Civil Engineers, Feb 1994, London, England.
4. Control Valves - Principles of Operation
Marko Ivetic
IRTCUW, Institute of Hydraulic Engineering
Faculty of Civil Engineering
P.O. Box 895, 11 000 Belgrade
YUGOSLAVIA
4.1. INTRODUCTION
The safety and reliability of operation of many water supply systems depend to a
great extent on how well they can be controlled. In the first place, it is a hydraulic
problem, and after that of control theory, system analysis, signal processing, etc.
Many hydraulic systems, which do not respond well to control actions, have
mismatches in selected equipment (pumps, valves etc.) or operating regimes are in
collision with system dynamics. The value of appropriate control equipment should not
be underestimated in cost analysis. However, compared to very high capital costs of
complex distribution systems, reservoirs, pumping stations, etc. and the required
reliability and troubl~-free operation of the system, the cost of control equipment is
minimal. In selecting control valves for a system, the engineer has to consider valves
different from those usually used in distribution systems (gate valves, butterfly valves,
standard globe valves, etc.).
There are several tasks which are performed by control valves (Obradovic,
1991). ·Usually, the same basic valves are used for several different purposes, with one or
more control devices which are responsible for the desired control action. The basic
valve, when combined with appropriate control loops, can cover a wide range of
applications (Shinskey, 1988). Several types of control valves can be identified: pressure
reducing valves, pressure sustaining valves, pressure relief or back pressure valves, pump
control valves, check valves, flow control valves, electric remote control valves, etc.
There are basically two types of action of control valves, permanent and non-
permanent. Control valves which are used for flow modulation, with the aim of providing
desired flow conditions within the system, are of the permanent type. The other group,
NATO ASI Series, Partnership Sub-Series, 2. Environment - Vol. IS

Edited by C. Maksimovie, F. Calomino, and J. Snoxell
146
for example check valves, float, pressure relief valves, and especially air valves, with
non-permanent operation, are sometimes not considered as control valves.
4.2. HYDRAULICS OF CONTROL VALVES
4.2.1. Flow Control
The principle of flow regulation will be explained with an example of flow

between two reservoirs through a relatively long pipeline (Figure 1.). Flow rate can be
determined from
( L) Q2
~n= L~L +A.-+~Z - - 2
.
(1)
D 2gA
Figure 1. Flow control in a long gravity pipeline
where, ~z is variable head loss coefficient of control valve, and, ~<1> = L~L + A. ~ is used
for energy loss coefficients which can not be changed, and which does not depend too
much on flow rate. When the pipeline is longer than 1000 D, this part of the energy loss
is mainly due to friction. This assumption is correct for fully developed turbulent flow in
rough pipes. Equation 1. can be written as
147
(2)
where, q = Q/ Qrnax ' and Qrnax is defined from
(3)
Flow rate of the system depends on the relative head loss coefficient of the
control valve scaled with 1;<1>. This simple, but important relation, is given in diagram
(Figure 2.). For 1;z /1;<1> = 1 flow rate is equal toO.7Qrnax, for 1;z /1;<1> = 3, flow rate is
equal to O.5Qrnax etc., while in order to get O.1Qrnax, head loss at the valve has to be 100
times greater than the friction loss. It should be noted that flow regulation turns out to
be intentional dissipation of the energy by control valve. For maximum flow, energy loss
is uniformly distributed along the pipeline, while for q < 0.7 a greater portion of the
available energy, 8TI = TI1 - TI2 has to be lost at the control valve.
101
,
10' .............. I
...........
Eo
...
,
~
"-VI 10'
r--
-""
2
10·' I .........
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Flow rate q = QIQ_
Figure 2. Flow control requirement in a long gravity pipeline
4.2.2. Head Loss Coefficient
The control valve is an obstacle to the flow. It reduces the area of flow, causing a
significant increase in local velocities, and local pressure drop. Further downstream the
148
flow area is expanded, water fills that space, and in highly turbulent conditions dissipates
most of the locally increased kinetic energy. With the same diameter of pipe upstream
and downstream of the valve, as a result of the valve action, there is a head loss ,
Ml = TIl - TI 2 . The valve is also responsible for a local pressure drop equal to,
Ml max = TIl -TImiD (Figure 3.).
vena contraCla
Figure 3. Variation of piezometric head and total energy in a valve
Head loss of the valve depends on the valve type, opening and also on the valve
size, actually on the Reynolds number (Re =VD/v where V is mean velocity, D pipe
diameter, v- kinematic viscosity of water). It is related to flow characteristics, velocity
head, V 2/ 2g, via head loss coefficient, ~z
V2
Ml=~z- (4)
2g
4.2.3. Discharge Coefficient
The value of ~z varies from a very small value (for open valve) to infinity (closed
valve). That is why it is replaced by a more convenient coefficient of flow, or discharge
coefficient, CD,
149
1
C --- (5)
D - .J1+f,z
104 r--;--,--,--,-----,-----,--,-----,--,---,-----, 1.0

0.9
:J. 103 0.8 cJ
'5
·u-=
II) 0.7
ISII) 102 0.6 ~

0
U
0.5 g
u
:: j
<Il
<Il 101 -
..s!
"i 0.2 .~
100 -
II)
::r: o
0.1
10.1 '""""'''---'----'----'------'-----L--:L---,-L,------:l-:,-------,L,----' 0.0
0.0 0.1 0.2 OJ 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Valve position 'tz
Figure 4. Characteristics of a valve: head loss coefficient and discharge coefficient
A diagram with relationships of these coefficients to the valve opening, called

valve characteristics, is given as an example in Figure 4. The valve opening, 'tz, is given
in relative terms, 0.0, valve closed, 1.0, valve completely open. Some manufacturers
prefer to use the dimensional presentation of valve characteristics (KUBOTA-
NEYRTEC, 1986, etc), like specific flow
(6)
or, in terms of head loss coefficient, f,z
(7)
It is defined as the flow passing through a one meter diameter valve at a head loss
equal to one meter of pressure head.
150
4.2.4. Cavitation
The ratio of MI/ .MImin called the coefficient of recuperation of pressure, is

always smaller than 1. Those valves which have higher coefficients of recuperation are
better in respect of cavitation danger. Cavitation might cause serious problems in the
operation of control valves. The troubles start when local pressures drop to the value of
vapour pressure at ambient temperature (Pv, around 2-3 kPa). Then, small bubbles of
water vapour (cavities) are formed. Further downstream pressure is recovered, causing
cavities to implode. If this is happening in the free stream, nothing but high level noise is
generated, but if it is happening near the pipe wall, small jets of high velocities may
damage the surface.
l
i
Jt
1-
~
I
-4
0 zone 2 1
...·c"
!;::) I
"1
I
~ ~
<>
0.
'" zone 1
coefficient of cavitation a
1
Figure 5. Cavitation diagram for a control valve
zone 1: no cavitation; zone 2: incipient cavitation, risk of damage small;
zone 3: noticeable cavitation, damage of valve after several hours/days of operation;
zone 4: operation is not possible.
Obviously, cavitation starts at the place where minimum pressure occurs (TImin
in Figure 3.). There are two ways to avoid cavitation: to choose valves with relatively
low difference between TImin and TI 2 , and to keep minimum pressure higher than vapour
pressure.
Cavitation characteristics of a valve have to be determined by experimental
investigation. Results of these measurements are usually given via coefficient of
cavitation, as a function of the specific flow, in diagrams like in Figure 5. The coefficient
of cavitation is defined as
151
ll- Pv
(5=--- (8)
ll-P2
Absolute pressure II is measured one pipe diameter upstream of the valve,

absolute pressure P2 is measured 10 pipe diameters downstream of the valve and
corrected for friction losses between sections 1 and 2. The available cavitation number
has to be calculated for all operating regimes of the valve, by Expression (8). It is then
compared with the required cavitation number of the selected control valve (Figure 5.).
Depending on the intensity of cavitation, there are several zones where calculated
cavitation number can fall.
4.3. DESIGN CONSIDERATION
There are several types of valves which are used for control purposes. They are
characterized by considerable energy loss, accurate positioning, durability and high
reliability, ease of manipulation, cavitation free operation etc. Examples of valves with
good control characteristics are given in Figures (7.), (8.) and (9.). The valve in
Figure 7. consists of valve body and two perforated plates, one of which is fixed and the
other is moving. The valve has excellent cavitation characteristics, simple design, full
range sensitivity etc.
10'
butterfly valve
, , - - needle valve
N
103
, _.
>JJ'
t:II> "" ...
·0 10 2
r-... ...
IEII>
0 , ...
V
(I)
<I) 10' ::--......
.£
-g., ,
:I:: 10°
.........
5
2 - -
10~.O 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Valve position 1Z
Figure 6. Head loss coefficients for butterfly and needle valve (ERHARD)
152
•
•
•
•
Figure 7. Scheme of a control valve with favorable control characteristics

(KUBOTA-NEYRTEC)
In Figures 8. and 9. are shown two different valves, based on the same principle.
The needle valve in Figure 8. has a fixed annular streamlined internal body with control
piston which can be moved in the flow direction. Depending on the duty required, there
are two models, with a vaned ring in order to reduce energy loss, and with a slotted
cylinder, in order to increase energy loss and to improve control characteristics in
systems with high friction losses. "Y" pattern control valve (BERMAD) in Figure 9., can
be equipped with V-port throttling plug. In that way the accuracy, stability and reliability
have been improved, and cavitation noise and risk reduced, especially for high pressure
reduction and/or low flow cases.
Figure 8. Two variants of the same valve a) for low head losses, vaned, b) for high head losses,
slotted cylinder (ERHARD)
153
Figure 9. A way to increase head loss in the control valve for high pressure difference services
(BERMAD)
In Figure 6. the characteristics of two valves, butterfly and needle (slotted

cylinder type, Figure 8.b.) are shown. It can be seen that the butterfly valve can not be
used as a control valve in systems with considerable friction losses. In a hypothetical
pipeline 20 Ian long, with diameter equal 0.8 m, with estimated pipe wall roughness 1
mm, at available difference, TIl - TI2 = 70m, the performances of the two valves from
Figure 6. has been examined. Control characteristics, flow rate vs. valve position, are
given in Figure 10. The movement of butterfly valve fromt = 1.0, completely open
position, to t = OJ, has not been felt by the system (change of flow less than 5 %).
Usage with opening less than 10 % is not recommended, so the valve can control flow
between 75 and 100 % of maximum flow, Qmax . On the other hand, the needle valve
equipped with slotted cylinder to promote energy dissipation, can be used for smooth
control over a wide range between 20 and 100 % of maximum flow.
1.0
' J- -1-1
0.9
1-- butterfly valve
"
... 1-'
0.8 -
i 1- - needle valve
OJ 0.7
V
C5l I
.,..
II 0.6
/
~ 0.5
f
~ 0.4
"
'"
0
ii: 0.3 _/
0.2
0.1
0.0 1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Valve position 'l'z
Figure 10. Control characteristics of a butterfly and a needle valve

154
In order to manipulate a valve in steady and transient operating regimes, one has
to overcome significant forces (hydrodynamic, friction etc.) acting on the control and
isolating elements of a valve. It can be done either manually, using a gearbox to increase
the torque (as for the valve in Figure 7.), or by specially designed devices called
actuators. There are several types of actuators: electric, pneumatic and hydraulic.
Sometimes, actuators are equipped with manual devices, which can override the action
of the actuator and which can be used in case of emergency or actuator malfunction.
The principle of operation of a control valve with hydraulic actuator is given in
Figure 11 . With a help of a small three-way pilot control valve, hydrostatic forces on the
diaphragm and washer are used to move the stem and change the valve opening.
Pressure difference on the diaphragm can be supplied from some external system
(pneumatic - for pneumatic actuators, oil), or it can be provided internally from the valve
body and working fluid. Apart from the one-chambered diaphragm unit shown in Figure
11 ., there are two-chambered units, where the diaphragm is either completely separated
from the working fluid, or communication between the inner chamber is only inhibited
(Figure 8.) in order to achieve damping effects during closing.
Figure 11. Operation of a diaphragm actuated control valve equipped with a three-way control
4.4. EXAMPLES OF APPLICATION
The list of possible applications of control valves is very long. Usually the same
basic valve, with small modifications, can be used for different purposes, depending on
the control loop and additional control devices implemented. The most typical
applications are explained in the next subsections, using data and recommendations of
valve manufacturers (KUBOTA-NEYRTEC, BERMAD, MUESCO, ERHARD).
155
4.4.1. Water Level Control
Two simple cases of water level control in a reservoir are given in Figure 12.
They look similar but they are not.
a) b)
Qm
--. --1:*1--....
Figure 12. Control valves used for water level control
Following the simplest rules of control (control action is initiated when there is an
intolerable difference between measured value and set point), control valve in a) case will
have increase-increase action, i.e., increase of flow, Qin will result in increase of water
level and, as a countermeasure, an increase of valve opening will follow. In the second
case, uncertainties come from the downstream side, from Qou/ ' Increase in Qou/ will
cause a decrease of level in the tank and it will require an increase of valve opening and
vice versa (increase-decrease action). There is one more difference between the two
cases. In the second case every discrepancy, no matter how small, between Qin and
Qou/ will cause emptying or over spilling of the tank. This process can not be left
unattended and uncontrolled. The first process, however, possesses a property of auto-
regulation. Small differences between Qin and Qou/ will be compensated by small
changes of water level. (There is a relationship between Qou/ and h, i.e.,
Qou/ oc.Jh +~ .)
The first group of processes is easier for control, but efficient control is an
imperative for the second group of processes.
4.4.2. Control Valves in Pumping Stations
A standard application of control valves in pumping stations is given in Figure 13 .

When used with centrifugal pumps, the check valve (CV) has to be closed during pump
start-up and to open gradually to prevent motor overloading. A relief valve (RV) is used
156
as a protective device against water hammer and to prevent pressure lock-in. It should
open immediately after the system pressure surpasses a set point, and closes slowly in a
programmed manner.
However, these two valves can have complementary roles of a permanent type.
They can be used as flow control valves. Valve (CV), with additional control devices,
can reduce the flow by increasing the energy loss (jIow throttling). Valve (RV) can do
the same job by diverting some amount of water to the upstream tank. Energy
consumption is higher, but the pump is operating closer to the rated point, with higher
efficiency.
RV - Relief Valve
CV - Check Valve
Tank ~ SV - Shut Off Valve
L--_ _- '
~
SV CV SV
Figure 13. Control valves in a standard scheme pumping station
A typical arrangement of valves with a deep well pump is given in Figure 14.
Relief and check valves have the same roles as in standard pumping stations. An
additional valve (pCV) is typical for deep well pumps, because the vertical pump column
is only partially filled with water. The pump is started and stopped with pump control
valve (pCV) opened. The first rush of air and water is discharged to atmosphere,
preventing water from being forced too rapidly into the pipeline.
~CV
Deep Well Pump
-
CV SV
~fP RV
RV - Relief Valve
CV - Check Valve
SV - Shut Off Valve
PCV - Pump Control Valve
Figure 14. Control valves with deep well pumps

157
RV • Relief Valve
CV· Check Valve
SV . Shut Off Valve
SV
SV -
RV
SV
Figure 15. Control valves in a booster pumping station
An arrangement of control valves, very similar to the standard one, is used with
booster pumping stations (Figure IS.).
4.4.3. Pressure Reducing and Sustaining Valves
One of the most frequent services of a control valve is to keep pressure at a

downstream location (called the control node). If the pressure at the control node is
above the set level, the control valve will close, if it is below, the control valve will open.
The control node can be a point immediately downstream of the control valve, or
at a distant place in a network. In the former case it can operate locally (controlled by
local control loop), while in the latter case, the pressure at control node (eN) has to
converted into digital form and transmitted to the pressure reducing valve (PRV).
PRY - Pressure Reducing Valvc

CN - Control Node
M1 - PRY Head loss
PRY
Distribution Network
Figure 16. Pressure reducing control valve

158
A pressure reducing valve can operate as a control valve only if the upstream
pressure is above the set pressure. If it is below, the PRY cannot add the energy to
water. Also, if the PRY is completely closed, it can not influence the pressure at the
control node; the system pressure is controlled by some other elements of the system
(pumps, control devices etc.).
A control valve can be used to maintain a set pressure at a location upstream of
it (Figure 17.). Comparing to the PRY, this is the opposite role. With this service they
are called the pressure sustaining valves. This type of control has similar limitations to a
PRY. It can not raise pressure at the control node after it is completely closed, and it
cannot lower it more after it is completely opened.
PSV - Pressure Sustaining Valve

eN - Control ode
on - PRV Head loss
Figure 17. Pressure sustaining valve
Pressure reducing and pressure sustaining valves can fulfill their tasks only if their
operating range extends over the greater part of valve stem travel. If the control valve
characteristics are like those of the butterfly valve in Figure 10., the overall performance
of the control valve will be poor.
4.5. CONCLUDING REMARKS
Control valves in large water supply systems are very important for safe and
reliable operation. Design criteria for control valves differ from criteria for other valves
used in water supply systems. For example, considerable energy loss is required for flow
control in long gravity pipelines and so on.
159
The most important characteristics of control valves are durability, high

reliability, cavitation free operation, ease of manipulation, accurate positioning, full range
sensitivity, etc.
The implementation of control valves in overall system control is much easier if
their hydraulic characteristics are satisfactory. Otherwise their operation can be an
additional source oftrouble in the system.
4.6. REFERENCES
BERMAD Control Valves (1994) Catalogue. Evron, Israel.

Coulbeck B., Orr C.H. (Ed.) (1988) Computer Applications in Water Supply, Volume
2 - Systems Optimization and Control. Research Studies Press Ltd. and John
Wiley & Sons Inc.
ERHARD Armaturen (1987) Catalogue. Johannes Erhard, H. Waldenmaier Erben,
Germany.
KUBOTA-NEYRTEC Valves (l986) Catalogue. Kubota Ltd.Osaka, Japan.
MUESCO Inc. (1986) Regulations and Control of Liquids. Catalogue. Houston, Texas.
Obradovic D. (1991) Control Valves and Booster Pumping Stations. Lecture Notes, IHE
Delft.
Shinskey F.G. (1988) Process Control Systems, Application, Design and Tuning. Third
Edition, McGraw-Hill Publishing Company.
Chapter III
New Technologies for Water Treatment

Chapter III
New Technologies for Water Treatment
fflTRODUCTORYSTATEMENT
New technologies in drinking water supply are needed basically for three reasons:
1. To obtain excellent drinking water quality for public health,
2. There is a growing concern for nature and conservation ofthe

environment,
3. For a major part of the world population, costs for sufficient and safe
drinking water are a dominating factor.
During the last few decades new insights have emerged between, on the one hand,
quality offood in general and of drinking water in particular, and public health on the other.
ConsenquentIy an increasing number of drinking water quality parameters are covered
by legislation, for example for lead, aluminium, chlorinated hydrocarbons, pesticides, solvents,
etc. In addition the limits are becoming more stringent, and more compounds can be detected
at lower concentrations.
New philosophies are being developed concerning microbiological, toxicological
(mutagenity, carcinogenity, teratogenity), organoleptic and operational quality parameters, and
also for new treatment technologies.
However, there is also a requirement that these new treatment technologies do not
themselves introduce undesirable compounds into the water.
164
In general the pollution of nature and the environment is still continuing. Raw water
sources, nature and the environment are being polluted more and more.
Hence, more and more compounds have to be removed. On the other hand, from the
point of view of the general responsibility for the natural environment, drinking water
companies have to treat and produce drinking water in such a manner that no new polluting
compounds are accumulated in the environment during production, distribution and usage
(liquid and solid waste, lead, copper, soap, detergent, energy, etc.).
Safe drinking water is a first priority for human life, while a major part of the world
population has very limited finances to pay the costs of a safe drinking water supply.
Therefore new technologies have to be developed in such a way that theses
technologies have costs during production and usage of drinking water which are as low as
possible.
The lowest costs during usage mean for example, that no bottled water is consumed
and minimum costs for soap, detergents, energy, maintenance etc. are realised.
These financial purposes must be consistent with the environmental goals.
1. New Technologies for Water Treatment: General
Aspects
J.P. van der Hoek

A. GraveIand
Vogelenzangseweg 21
2114 DA Vogelenzang
THE NETHERLANDS
1.1. INTRODUCTION
Raw water for drinking water production can originate from ground water, surface water,
or river bank infiltrated water. The potable water produced from these raw water sources has to
meet certain standards, described in legislation and in drinking water rules. These standards can be
divided into four groups: (1) microbiological standards, (2) toxicological standards, (3)
aesthetic/organoleptic standards, and (4) operational standards. Pollutants in the raw water can be
divided into six groups. For every group, relevant treatment processes are available.
Nowadays, the quality of ground water and sutface water is deteriorating and new opinions
and views are being developed concerning optimal treatment of water. Ground water is threatened
by acidification and overfertilization, disposal of waste and agricultural practice (pesticides). New
issues in surface water treatment are the awareness of the formation of harmful disinfection by-
products during chlorination, the presence of pesticides and organic micropollutants in the raw
water, and the tendency to tighten the limits for drinking water and environmentaI standards. All
these aspects require the modification of existing treatment plants and the incorporation of new
processes. In sutface water treatment, the focus is on the removal of organic matter, such as
nutrients for biological growth, pesticides and other organic toxic micropollutants, and the
production of hygienically safe and biologically stable drinking water without the use of chlorine.
For the removal of organic matter such as nutrients, pesticides and other organic micropollutants,
three approaches are discussed: (1) the removal of these compounds by activated carbon filtration
or membrane filtration, (2) the chemical oxidation of these compounds by advanced oxidation
processes, and (3) the biodegradation of these compounds by biological activated carbon filtration.
NATO ASI Series. Partnership Sub-Series. 2. Environment - VoL IS
166
To produce hygienically safe and biologically stable drinking water, slow sand filtration or activated
carbon filtration in combination with UV-disinfection are promising techniques. For the removal of
organic matter in general, and pesticides and other organic micropollutants in particular, three
concepts are presented for drinking water treatment in the future: (1) chemical treatment, (2)
physical treatment and (3) biological/chemical treatment.
1.2. SOURCES OF RAW WATER: GROUND WATER, SURFACE WATER, RIVER

BANK INFILTRATED WATER
For production of drinking water, in principle three sources are available: surface water
from rivers or lakes, ground water and water after river bank infiltration. In the latter case, water
can be extracted from wells located near the surface water source. Traditionally many drinking
water companies preferred ground water above surface water as a source for potable water
production. Arguments for this preference are the relatively high quality of the ground water, the
low treatment costs and the ease of operation of the relatively simple ground water treatment plants.
The most important general characteristics of ground water and surface water as were valid until
recently are summarized in Table 1. and in Table 2.
Table 1. Characteristics of ground water
Advantages: Disadvantages:
- hygienically safe - often limited availability
- constant temperature
- constant quality
- not vUlnerable to accidents
Table 2. Characteristics of surfuce water
Advantages: Disadvantages:
-large quantities - disinfection necessary
available most - fluctuating temperature
of the time - biologically and chemically not stable
- fluctuating quality
- vulnerable to accidents
167
By bank infiltration, the advantages of ground water can be achieved. Fluctuating

concentrations and temperatures are balanced, and thus the water becomes less sensitive to
accidents when compared with surface water, although pollution is spread easily through the
catchment. During bank infiltration also the quality of the water is improved: suspended solids are
removed from the water, and many contaminants are biologically degraded. The biological and
chemical stability of the water is increased.
Due to environmental problems like overfertilization, resulting in increasing nitrate and
phosphate concentrations, and the use of pesticides in agriculture, the quality of ground water is
deteriorating and the preference for ground water is becoming less pronounced from the point of
view of quality.
1.3. STANDARDS AND GUIDELINES FOR DRINKING WATER QUALITY
The raw water originating from surface water or ground water mostly contains pollutants in
such a concentration that it is not immediately suitable as drinking water. Therefore, it has to be
treated in order to obtain drinking water. However, in legislation and in drinking water rules, the
concept of "sound drinking water" is not well defined. On the other hand, the World Health
Organization and the European Community describe guidelines, guide levels and maximum
admissible concentrations that should be met by the drinking water. These standards can be divided
into four groups:
1. microbiological standards;
2. toxjcological standards;
3. aesthetic/organoleptic standards;
4. operational standards.
According to the WHO the most important standards concern the microbiological
standards. The first target of drinking water treatment has to be the production of water that meets
the microbiological standards. As soon as these standards are met the other standards have to be
aimed at. Thus the order of standards mentioned above can be considered as a functional sequence.
168
Microbiological standards and toxicological standards are health related. Many

microbiological parameters are used as an indicator for the presence of pathogenic organisms in the
water. Toxicological standards are calculated based on risk-assessment models. For a long-life
exposure, the WHO allows a risk-factor of 10-'. In the calculation of the standards, an ADI
(Acceptable Daily Intake) or TOI (Tolerable Daily Intake) is used. Drinking water only contnbutes
partly to this ADI or TOI, also food and air contribute to the total exposure.
Examples of toxicological standards are amongst others nitrate, nitrite, (heavy) metals and
many organic compounds.
Aesthetic/organoleptic standards are related to the fact that the water should have a high
aesthetic value, and use of drinking water must be attractive and enjoyable. This means that the
water should be bright and free of colour, odour and taste.
Operational standards do not concern health related aspects, but are guidelines to obtain a
good technical operation of the treatment plant and the distribution system, and to guarantee a good
water quality at the consumers' tap. An example of an operational standard is the aggressivity of the
water. Aggressivity may affect the distnbution system. Indirectly these standards are also related to
microbiological, toxicological and aesthetic aspects.
A complete overview of the standards cannot be given here, but an indication of the way in
which the standards affect the normal constituents of the water, and so affect the treatment scheme
necessary to produce drinking water from surface water or ground water, can be sununarized as
follows:
• the presence of phosphorus, nitrogen and carbon containing compounds interfere with the
bacteriological standards and toxicological standards. P, N and C may act as nutrients for
microbial growth, while carbon containing compounds also include pesticides and other
organic micro pollutants which are toxic,
• calcium and hydrogencarbonate detennine the hardness of water. Hardness is related to the
copper and lead uptake of the water in the distribution system, and thus the toxicological
aspects of the water,
• chloride and sulphate result in corrosion of the distribution system, especially when cast iron
pipes are used. This may result in increased turbidity and colour of the water, leading to a
drop in aesthetic quality.
169
1.4. CATEGORIES OF POLLUTANTS AND UNIT OPERATIONS
In the guide lines of the World Health Organization (WHO, 1993) and in the European
Council Directive relating to the quality of water intended for human consumption (European
Community, 1980) a large list of impurities is described that are not allowed to exceed a certain
guideline (GL) or maximum admissible concentration level (MAC). Several processes are available
to treat the raw water in such a way, that the GL-values or MAC-values can be met. In principle,
these contaminants can be divided into six groups, and for each group the most suitable treatment
processes can be chosen. Table 3. summarizes the contaminants and relevant treatment processes.
Table 3. Groups of pollutants with suitable treatment processes
Pollutants Unit operations

I Suspended solids: Coagulation, rapid filtration
clay, algae, bacteria, viruses, microfauna
2 Heavy metals: Coagulation, rapid filtration,
lead, copper, cadmium, chromium, mercury, etc. softening, hyperfiltration
3 Salts:
Cations: Na+, IC, NHt, Ca2+, Mg2+, Fe3+, Al3+, Rapid filtration, softening,
etc. hyperfiltration
Anions: cr, N03-, HCO;, SOl, pol Coagulation, softening,

hyperfiltration
4 Organic compounds: (Biological) activated carbon
-nutrients filtration, hyperfiltration
- toxicological compounds
- smell, colour, taste
- anti-corrosiOOlscaIing
5 (Micro)organisms: Coagulation, rapid filtration,
bacteria, bacteriophages, viruses, cysts, etc. ozonation, hyperfiltration
6 Radio-active compounds Coagulation, rapid filtration,
(biological) activated carbon
filtration, hyperfiltration
The first group concerns suspended solids like clay particles, algae, bacteria, viruses and
microfauna. In practice, suspended solids are removed by a conventional treatment scheme
including coagulation and flocculation using iron or aluminium salts, followed by sedimentation or
170
flotation to allow much of the suspended floes and other matter to settle out of the water, and rapid
filtration to remove most of the remaining suspended matter from the water. Wrth these treatment
processes, suspended solids can be removed from the water to below 5 g/m3 (Boller, 1994).
Removal of bacteria and viruses is limited with these processes. In the Surface Water Treatment
Rules of the United States, a log reduction of2 is given for viruses and a log reduction of2.5 for
Giardia cysts by this conventional treatment process.
The second group of pollutants concerns the heavy metals like lead, copper, cadmium,
chromium, mercury, nickel etc. By the traditional treatment scheme as descnbed above part of these
pollutants can be removed from the water by adsorption on and occlusion in the floes. However,
precipitation of these heavy metals and adsorption/absorption on floes is rather pH-sensitive, and
often rather extreme pH-values are necessary to achieve a good removal efficiency. Another
possibility is hyperfiltration (reverse osmosis) by which in theory all ions can be removed from the
water, and thus also these heavy metals.
The third group of pollutants concerns the dissolved salts. These can be divided into the
possitively charged cations (Na+, K+, NH/, Ca2+, Mg2+, Fe2+lFe3+, Al3+, etc.) and the negatively
charged anions (Cr, N~-, HC~-, sol, pol) Removal of hardness (Ca2+ and Mg2j is possible
by softening processes by means of precipitation or crystallization of CaC03 (Graveland et at,
1983). Removal of ammonia, iron and manganese can be achieved by rapid filtration, as is normally
practised in ground water treatment. For removal of sodium and potassium, hyperfiltration is a good
alternative. The anions may be removed by coagulation (pot), softening processes (HC~), or
membrane processes (cr, sol, N03) For nitrate removal, specific denitrification processes may
be considered.
The fourth group concerns the organic compounds. The total number of different organic
compounds probably amounts to a few thousands. Depending on their nature and amount in the
raw water they can have the following functions:
• as nutrients to ~ulate (micro)biological growth and cause an accumulation of detritus

(DOC, Dissolved Organic Carbon; AOC, Assimilable Organic Carbon) in the distribution
system,
• to cause a toxicity such as mutagenity, carcinogenity or teratogenity: THM

(trihalomethanes), AOX (adsorbable halogenated compounds), pesticides, solvents etc.,
• to cause smell, colour and taste,
• to act as a means of anti-corrosion and/or anti-scaling.

171
Processes to remove organic compounds from water include:
• physical processes such as adsorption on adsorbants (activated carbon),

• chemical processes such as oxidation by means of ozone, hydrogen peroxide or UV
(Advanced Oxidation Processes),
• membrane processes such as hyperfiItration and nanofiltration,
• biological processes like biological activated carbon filtration. In this process a small dose of
ozone is used prior to the activated carbon filtration to increase the AOC-content and
stimulate biodegradation of the organic matter during activated carbon filtration.
The fifth group comprises micro-organisms like bacteria, viruses and cysts (e.g. Giardia).
These parameters are all related to the hygienic quality of the water and health aspects. With the
conventional treatment scheme of coagulation and rapid filtration a log reduction of 2-2.5 can be
achieved. With chemical disinfection (e.g. chlorine, chlorine dioxide, chloramine, ozone, UV
radiation) inactivation depends on the concentration used and the contact time (CT-value), the
temperature and the pH. Recent experiments have shown that disinfection can also be achieved by
membrane processes like hyperfiItration, nanofiltration and ultrafiltration. Log-reductions of at least
5-7 have been reported (Boller, 1994, Anselme et at, 1993).
The last group concerns radio active compounds. These compounds may be removed by
either coagulation and rapid filtration, (biological) activated carbon filtration or hyperfiltration.
1.5. NEW ISSUES IN GROUND WATER TREATMENT AND SURFACE WATER

TREATMENT
For over a century, both ground water and surface water have been used for drinking water
purposes, usually ground water where available, and surface water always had to be treated to
achieve drinking water quality.
Traditionally, key issues in ground water treatment were iron, manganese and ammonia
removal, achieved by aeration and filtration often with pH correction and hardness reduction. An
example of a conventional ground water treatment scheme is shown in Figure 1.
Nowadays, where the quality of ground water is threatened, extension of the treatment
scheme is required. The most important threats are:
172
• acidification and overfertilization: this results in an increase in nitrate, phosphate, sulphate

and hardness in the water, and an increase in the concentration of aluminium, nickel,
cadmium and other heavy metals,
• disposal of waste: this may lead to pollution of ground water by volatile organic
compounds,
• agriculture: the use of pesticides may result in deterioration of the ground water quality.
Figure 1. Conventional ground water treatment based on dual aeration and filtration
Most of these contaminants require additional treatment steps, or modification of existing

treatment schemes. For example, nitrate requires a nitrate removal process, while volatile organic
compounds may require air-stripping. For pesticides, additional processes like (biological) activated
carbon filtration, advanced oxidation or membrane filtration may be needed.
The main issues in surface water treatment so far have been:
• clarity of the water, as most raw river water is high in turbidity,

173
• an acceptable taste and odour, which is not the case for surface water that is highly polluted
with domestic, agricultural and industrial waste,
• hygienic reliability,
• storage of water in open storage reservoirs or in the subsoil, to cover periods of insufficient
quality or quantity and to enable selective intake of water.
An example of a conventional treatment scheme for surface water is shown in Figure 2.
Figure 2. Conventional surface water treatment based on coagulation, rapid filtration and multiple
chlorination (1960-1970)
Turbidity was reduced by coagulation using iron or aluminium salts, followed by

sedimentation or flotation and rapid filtration. Taste and odour problems could be solved through
the addition of powdered activated carbon.
Hygienic reliability was achieved by means of disinfection with chlorine and post
chlorination guaranteed the required bacteriological quality. Selective intake and quality
improvement of the raw water were achieved by storage reservoirs at the beginning of the treatment
process. Dune water supply companies in the Netherlands started infiltration of pre-treated river
water in the subsoil to enlarge the storage and capacity.
174
Also slow sand filtration has been used in this process as a final treatment step to reduce or
to avoid the use of chlorine.
Nowadays this conventional surface water treatment scheme is changing as a result of new
threats and changing opinions concerning robust, safe and reliable treatment schemes:
• hannfuI disinfection by-products are produced during the reaction of chlorine with humic
acids (Rook, 1974). Both chlorinated compounds and brominated compounds might be
produced. Although chloramine is used as an alternative, most attention in Europe is given
to slow sand filtration, artificial recharge, UV-disinfection, ozonation and chlorine dioxide
as disinfection barriers,
• in order to avoid the use of a persistent disinfectant like chlorine, it is necessary to produce a
biologically stable drinking water. This can only be achieved when the content of organic
compounds (DOC), N-compounds and P-compounds which all act as nutrients and result in
regrowth in the distribution system, is very low. Biological stability of the water is related to
the AOC concentration: the concentration of assimilable organic carbon. When the AOC
concentration is below 10 J.lWl, no regrowth in the distribution system will occur (Van der
Kooij et a1., 1990) and no persistent disinfectant has to be dosed. To reach biological
stability of the water, appropriate treatment steps have to be used (dual filtration, slow sand
filtration) and good management of the treatment plant is required. Besides, a reliable
distribution system is required in which no recontamination takes place,
• pesticides and organic micropollutants have been detected in surface water. They originate
from discharges into rivers by industries, and from surface run-off from agricultural land. In
Table 4., the presence of pesticides in the river Rhine and the river Meuse is summarized.
Table 4. Presence of pesticides in river Rhine and river Meuse
Number of pesticides
River Meuse River Rhine
till 1990 1991 till 1990 1991
~
>1 0 0 12 2
>0.1 6 9 43 11
175
Table 5. Current and future standards for drinking water according to the EC and the Netherlands
Waterworks Association
Parameter Current standard New standard
Na 120 mgIl lOOmgll
Total hardness I.S nnnolll 1.0nnnolll
pH 8.3 8.6
Pb SO~ 1O~
Cu 3mg1l I.S mgIl

DOC S mgIl 3mg1l
Colour 20mgPtll 10 mg PtII
AOX 20~
Pesticides O.l~ O.l~
Chloropesticides O.l~ 0.01~
1HM 70~ 30~
AOC 10~
Modified fouling index 3 s/f
Aeromonads SO!lnlleaving plant
Additional treatment. steps such as (biological) activated carbon filtration, advanced

oxidation or membrane filtration is necessary to remove these substances from the raw water:
• in addition to the presence of pesticides and other organic micropollutants as mentioned

above, it must be realised that many more pesticides and micropollutants may be present in
the water. For example, ofa total of320 different pesticides applied in the Netherlands, 225
are significant in drinking water supply, based on use, production, their presence in
surrounding countries, mobility and persistence in the environment. Analytical methods are
only available for 94 of these pesticides (Hopman et al, 1989). Polar pesticides and their
metabolites in particular cause serious analytical problems. This makes it very likely that a
large number of pesticides occur in surface water and bank-infiltrated water that are not
expected to be removed by customary treatment processes, including activated carbon. It is
176
not yet clear whether advanced oxidation processes, ozonation and membrane filtration will
be a sufficiently effective method for these components. As analytical techniques are being
developed and make it possible to detect their presence in water, it must be realised that the
standards that the treatment processes will have to meet are getting tighter,
• there is a tendency to tighten the limits for drinking water and environmental standards. This
will affect parameters that relate to organic matter (DOC, AOX, pesticides, other
micropollutants), and to conditioning (softening) of the water. Moreover, it will also result
in standards for the quality of recharge water, for the use of chemicals and for the
production of sludge.
Table S. summarizes some quality standards derived from current and future EC drinking
water regulation and from the Netherlands Waterworks Association.
1.6. NEW TECHNOLOGIES IN SURFACE WATER TREATMENT IN RELATION TO

THE NEW ISSUES
Looking at the new issues in surface water treatment mentioned in the preceding paragraph,
new technological developments will have to focus on the removal of organic compounds from the
water (DOC, N-compounds, P-compounds, and especially pesticides and organic micropollutants)
and on the production of hygienically safe and biologically stable drinking water without the use of
chlorine.
1.6.1. Removal of Organic Matter as Nutrients, Pesticides and Other Organic

MicropoUutants
With respect to the removal of organic matter (nutrients), pesticides and other organic
micropollutants, in principle three approaches are possible:
• in the first approach the main objective is to remove organic matter, pesticides and other
organic micropollutants from the water phase by activated carbon filtration or by
hyper/nanofiltration. Those compounds that can not be removed from the water by these
processes for economic or technical reasons have to be oxidised (see second approach),
177
• in the second approach the main objective is to oxidise (chemical oxidation) the organic
matter, pesticides and other organic micropollutants with an Advanced Oxidation Process
(AOP). AOC that is produced in the AOP is removed by one or more filtration processes,
• in the third approach organic matter, pesticides and other organic micropollutants are
biodegraded. By means of a low ozone dose the biodegradation capacity is increased and
organic matter, pesticides and other organic micropollutants are partially oxidised to yield
more biodegradable products. In fact AOC is created by ozonation and in the subsequent
activated carbon filtration the biodegradation capacity is increased. Hence, the organic
matter can be biodegraded or adsorbed, depending on the nature of the material. Also the
more persistent organic compounds may be biodegraded, as a result of the high biological
activity in the carbon filter.
The main characteristics ofthe three approaches will be discussed below.
Activated carbon filtration has been applied in the past for removal of smell, taste and
odour, trihalomethanes and also pesticides. It can also be used for the removal of organic matter in
general and micropollutants in particular. However, the extent of removal largely depends on the
polarity of the compounds. Highly nonpolar compounds (e.g. lindane) adsorb strongly on carbon,
but polar compounds like trichioro-acetic acid and dikegulac have a low adsorption capacity,
resulting in short running times of the carbon filters, a high reactivation frequency and high
reactivation costs. Other factors that determine the adsorption capacity for pesticides and other
organic micropollutants are:
• the presence of natural organic matter (humic substances) in the water: due to a relatively
high concentration, they are adsorbed preferentially and reduce the adsorption capacity for
pesticides and organic micropoIIutants,
• kinetic factors like surface loading, empty bed contact time, etc.
As membrane filtration processes, hyperfiltration or nanofiltration can be used. In these

processes, water is driven by pressure through a semi-permeable membrane that (partially) rejects
dissolved compounds. In this way, the feed water is divided into a permeate (product water) and a
concentrate (brine). The principle of this process is shown in Figure 3.
Hyperfi1tration membranes have already been used a long time for desalination of brackish
and sea water and retain almost 100% of monovalent ions. Nanofiltration membranes retain
178
approximately 50010 of these compounds. Both types of membranes retain more than 90010 of nearly
all other components (bivalent ions, micro organisms, DOC). This is also the case for pesticides and
organic micropollutants with a molecular weight of 300 or more. No final conclusion can yet be
drawn about organic substances with a lower molecular weight than 300, as research into this
subject is being carried out at the moment. Besides, removal of pesticides and organic
micropollutants is strongly affected by the size and structure of the compounds, and the
characteristics of the membrane material. In addition it must be mentioned that hyper- and
nanofiltration need an extensive pre-treatment to avoid fouling of the membranes, and post-
treatment to adjust the pH and hardness after membrane filtration. Also the concentrated waste
stream, the brine, has to be disposed of.
.It product
membrane
. .'. brine
;
• •
$" •
• ",<§'. •
~~.
.9.
~.
•
•
•
<$'
<$i-'" •. '0"
-s:."'~
'- ~
<F
Figure 3. The principle ofhyperfiltration and nanofiltration
From the characteristics of activated carbon filtration and membrane filtration it is clear that
these processes are not an absolute barrier against pesticides and organic micropollutants. In
particular, polar compounds (in activated carbon filtration) and low molecular weight compounds
(in membrane filtration) are only partially removed. So, additional processes might be necessary to
form a complete barrier.
Chemical oxidation can be achieved by ozonation or advanced oxidation. Some of these
processes are summarized in Table 6. The effect of AOP is based on the formation of hydroxyl
radicals, an extremely reactive and unstable form of the hydroxide ion lacking one electron and thus
uncharged. Hydroxyl radicals might also be formed as a result of the reaction between 0 3 and H20.
Ozone itself reacts selectively and incompletely with organic material. Reaction rate with
aromatics and unsaturated chains is very high, while ozone does not react with saturated aIiphatics
and chlorinated compounds under practical circumstances. The AOPs are much more reactive.
Reaction rates for all organic pollutants are very high, so in theory it is possible to convert all these
179
compounds to C~ and H:zO, but then a high dose and long contact time are necessary (40-80 mg
<M. 20-400 mg H:z07/l, t = 60-300 min). In practice, the objective is selective oxidation of
individual contaminants at doses and contact times nonnally applied in drinking water production
(2-8 mg <M. 1-3 mg H20il. t = 15 min) (Kruithof et al., 1994). The required dose and contact time
depend on the nature of the compound, the oxidation degree required, the amount and nature of
other organic material present (DOC) and the amount and nature of inorganic material present
(HC03·,C~~.
Table 6. Advanced oxidation opocesses and hydroxyl radical fonnation
3 ~+H:zO => 20H-+4~
203+H2~ => 20H-+3~
~+H20+UV => 20H-+~
H2~+UV => 20H
An important drawback of ozonation with high ozone dose and AOP is the formation of by-
products like brominated compounds and bromate. Only by applying a high H:zOv'~ ratio or a low
pH can this be restricted. Removal of the pre-cursor bromide prior to ozonation or advanced
oxidation is an alternative. Also AOC is produced, which has to be removed by subsequent
biological filtration steps. It can be concluded that ozonation and AOP can be used as barriers
against pesticides and organic micropollutants. However, the applicability is restricted as a result of
the formation of by-products like brominated compounds, especially bromate. AOP and ozonation
have to be followed by filtration for AOC removal.
An alternative for AOP and ozonation is biological activated carbon filtration (BACF)
(Graveland, 1994). In this process a small part of the organic material, pesticides and other organic
micropollutants are converted to AOC by a low dose of ozone (1-2 mgII), causing production of
100-150 ~g AOC/l and a higher biodegradation capacity in the subsequent carbon filtration due to a
higher bacteria density, resulting from the increased AOC concentration. During the activated
carbon filtration both biodegradation and adsorption take place. So, a combination of biological
oxidation and adsorption is achieved, resulting in a prolongation of running times of the carbon
filters with a factor of 10. An additional advantage of ozonation in the BACF process is the
disinfection of bacteria, viruses and cysts. Due to the lower ozone dosage as compared with the
chemical oxidation in the advanced oxidation processes, the risk of bromate formation is much
lower.
CONVENTIONAL CHEMICAL PHYSICAL BIOCHEMICAL
TREATMENT TREATMENT TREATMENT TREATMENT
sto~ge
CI 2
I---Fe intake intake intake
PAC + + +
conventional pre-treatment conventional pre-treatment conventional pre-treatment
+ + +
storage storage storage
~aOHI - HzS°4 . ozonatlOn

Ca(OH)2 anti-scala -"
00
NaOHI a
softening hyper / nano filtration I---ca(OH)2
I
rapid ,Itration activated carbon filtration softening
AOP marble filtration biological activated

carbon filtration
I U.V.
(--C1 2 U.V. / slow sand filtration

I
clear water reservoir l ~r water reservoir I clear water reservoir clear water reservoir
Figure 4. New process schemes in surface water treatment

181
1.6.2. Production of Hygienically Safe and Biologically Stable Water
With respect to the production of hygienically safe and biologically stable drinking water,
the use of chlorine as disinfectant is avoided if possible.
As the main disinfection barrier, the application of slow sand filtration, artificial recharge,
UV-radiation, ozone and chlorine dioxide is considered. Also the disinfection capacity of membrane
filtration is currently being studied (Boller, 1994, Anselme et aI., 1993).
To abandon the use of chlorine as postdisinfectant, the production of a biologically stable water is
the objective. When the AOC concentration of the water can be reduced to 10 J..LgII or lower, no
regrowth will occur in the distribution system and thus no postdisinfection is necessary. As final
treatment steps, slow sand filtration or activated carbon filtration can be used. With slow sand
filtration, it is possible to produce both biologically stable and hygienically safe drinking water. With
activated carbon filtration, it is possible to produce biologically stable drinking water. To achieve
hygienically safe drinking water (colony counts below the admissible concentrations) it might be
necessary to use UV for final disinfection.
1.7. EVALUATION
As a result of pollution of raw water sources and more stringent limits for drinking water,
new processes have to be applied, both in ground water treatment and in surface water treatment.
Focusing on surface water treatment, the traditional process scheme will have to be
modified to meet the new standards and to remove contaminants like organic matter (nutrients),
pesticides and other organic micropollutants. In fact three approaches can be used: the chemical
approach, the physical approach and the biological/chemical approach. Figure 4. shows these
approaches, together with the conventional treatment scheme for surface water.
1ft all three options pre-treatment is similar to the conventional treatment of surface water:
coagulation, rapid sand filtration and storage. Then the new processes are incorporated for
softening, removal of organic matter, pesticides and other organic micropollutants, disinfection and
biological stability.
In chemical treatment, softening is achieved by pellet softening or precipitation, and if
necessary followed by direct filtration for removal of carry-over. The AOP «()y'H2~) results in both
oxidation and disinfection. Activated carbon filtration removes AOC from the water, any remaining
nonpolar pesticides and organic micropollutants. Postdisinfection to produce hygienically safe
drinking water is achieved by UV-radiation or slow sand filtration.
182
The biological/chemical approach resembles the chemical approach, but in contrast to

chemical treatment a relatively low ozone dose is applied to restrict formation of by-products. After
ozonation, the water is first softened by pellet softening, and then treated by biological activated
carbon filtration. Contact times of 30-40 min are needed to stimulate and maximize biodegradation
in this carbon filter, and to reach the lowest costs per cubic meter and to meet the legal standards.
Compounds which are less biodegradable will be removed by adsorption. As with chemical
treatment, biological stability and hygienic safety is achieved by slow sand filtration and/or UV-
disinfection.
In the physical treatment process, hyper- or nanofiltration is used for removal of
pesticides and other organic micropollutants, softening and disinfection (retention of bacteria,
viruses and cysts by the membranes). To avoid precipitation of salts on the membranes, H2S04 and
if necessary an anti-sealant is dosed. After membrane filtration the hardness has to be adjusted, for
example by marble filtration or by mixing with water, conventionally treated. Carbon filtration acts
as an additional barrier for low molecular weight compounds that might pass the membranes. Post-
disinfection is achieved by UV-radiation.
Although these schemes concern direct treatment of surface water, they are also suitable for
indirect treatment. In these cases, the water is stored in the subsoil following treatment. Reasons for
subsoil storage are balancing and maintaining a strategic supply, and improvement in the
bacteriological quality of the water.
It is not possible to give a general preference for one of the treatment schemes. In practice,
existing surface water treatment plants have to be upgraded. So, the choice for one of these
schemes depends strongly on the set-up of the existing plant (can process-units, already present, be
incorporated in the new treatment scheme?), the composition of the raw water to be treated (what
are the most important contaminants, and what is their concentration?), and environmental aspects
such as the possibility to dispose of the brine in cases of membrane filtration all have to be
considered.
Because all thr~ schemes have advantages and disadvantages, in practice also combinations
of the schemes are being investigated and constructed. In the next section examples of such
combinations as applied by the Amsterdam Water Supply will be discussed in detail.
1.S. REFERENCES
Anselme C., Mandra V., Baudin I., MalleviaIle J. (1993) Optimum Use of Membrane Processes in
Drinking Water Treatment. Proceedings of the 19th International Water Supply Congress,
Budapest, 2-8 October 1993, Special Subject 2 ''Membrane Technology": SS2.1-SS2.11.
183
Boller M. (1994) Trends in Water F!ltration Technology. Aqua 40: 65-75.

European Community (1980) Council Directive of 15 July 1980 Relating to the Quality of Water
Intended for Human Consumption. Official Journal of the European Community L229 23:
11-29.
Graveland A (1994) Application of Biological Activated Carbon Filtration at Amsterdam Water
Supply. Paper Presented at the IWSA Specialised Conference on Activated Carbon in
Drinking Water Treatment, Amsterdam.
Graveland A, Dijk J.C. van, Moel P.J. de, Oomen J.H.C.M (1983) Developments in Water
Softening by Means of Pellet Reactors. Journal American Water Works Association
December 1983: 619-625.
Hopman R, Beek C.G.E.M. van, Janssen H.MJ., Puijker L.M. (1989) Pesticides and Drinking
Water Supply in the Netherlands. KIWA Communication 113 (in Dutch).
Kooij D. van der, Hijnen WAM. (199O) Criteria for Defining the Biological Stability of Drinking
Water as Determined With AOC Measurements. Proceedings of the AWWA Water Quality
Technology Conference, San Diego, California, 11-15 November 1990.
Kruithof J.C., Schippers J.C., Dijk J.C. van. (1994) Drinking-Water Production From Surface
Water in the 1990s. Aqua 43: 47-57.
Rook J.J. (1974) Formation of Haloforms During Chlorination of Natural Water. Water Treatment
Examination 23: 234-245.
World Health Organisation (1993) Guidelines for Drinking Water Quality.
185
2. New Technologies for Water Treatment: Case

Studies
J.P. van der Hoek

A. Graveland
Vogelenzangseweg 21
2114 BA Vogelenzang
THE NETHERLANDS
2.1. INTRODUCTION
Amsterdam Water Supply operates two production plants: the River-Dune Waterworks
(RDWW, 70.106 m3/year) and the River-Lake Waterworks (RLWW, 31.1 06 m3/year).
Both are for the main part surface water treatment plants. The basic process scheme for
both plants consists of pre-treatment, storage for quality improvement, levelling, selective intake and
strategic supply (RDWW: dune infiltration; RLWW: self purification basin), and post-treatment.
The historical development of the RDWW and the RLWW is discussed over the last 20
years, and the need for introduction of new processes in the treatment schemes is explained.
New processes concern softening by means of pellet reactors, removal of organic matter
(nutrients for biological growth, pesticides and other organic micropollutants) by biological
activated carbon filtration, and hyperfiltration for desalination, an disinfection as an additional final
safety barrier for the removal of pesticides and other organic micropollutants. Optimal incorporation
of these processes in the existing treatment schemes is discussed.
Finally, the new processes are discussed in more detail, with emphasis on the principles of
the process, the merits and limitations of the process, and some technical and operational aspects of
the process.
As it is said the Amsterdam Water Supply operates two water production plants: the River-
Dune Waterworks (RDWW) and the River-Lake Waterworks (RLWW). The fundamental process
schemes of both waterworks are shown in Figure 1. and are more or less comparable.
NATO ASI Series. Partnership Sub-Series, 2.Environment- Vol.15

Edited by C. Maksimovic, F. Calomino, and J. SnoxelI
186
River-Dune Waterworks River-Lake Waterworks

(RDWW) (RLWW)
Figure 1. Basic process scheme of the River-Dune Waterworks (RDWW) and the River-Lake Waterworks
(RLWW)
In the RDWW water from the Lek Canal is pre-treated in Nieuwegein by coagulation and
rapid sand filtration, and then transported to the dunes west of Amsterdam (Leiduin). There the pre-
treated water is infiltrated in the dunes for quality improvement, to enable selective intake of water
and to cover periods of insufficient intake (strategic supply). After a residence time of
approximately 100 days the water is abstracted from an open collecting basin and after rapid sand
filtration for removal of suspended solids, algae and plant components and nitrification, the water is
further treated to achieve drinking water quality.
In the RLWW raw water from the Bethune Polder and the Amsterdam-Rhine Canal is
treated by coagulation before it enters a self purification basin with a residence time of
approximately 100 days. These treatment facilities are located in Loenen. Before the water is
transported to Weesperkarspel for post-treatment, rapid sand filtration is carried out in order to
remove suspended solids, algae, plant components and the cells of Dreissena polymorpha as well as
to allow ammonia to be converted to nitrate. In this way the rapid sand filtration functions also as a
transport treatment instead of a transport chlorination. In Weesperkarspel post-treatment is carried
out.
The raw water quality of the Lek Canal and the Amsterdam-Rhine Canal may be
characterized as river Rhine water with a high salt and phosphate content, a low organic matter
content and many xenobiotic material (AOX, pesticides). The water from the Bethune Polder
reclamation area may be characterized as having a low salt content and a high content of phosphate,
ammonia and humic acids. Total hardness of both water types is almost equal. Most of the present
time only water from the Bethune Polder is used as raw water source in the RLWW.
187
The average composition of the water, after pre-treatment and storage in the dunes or basin,
which has to be treated in Leiduin (RDWW) or Weesperkarspel (RLWW) is summarized in
Table 1. In this table also the requirements set by Amsterdam Water Supply for the final water
quality are included.
Table 1. Water quality after dune infiltration (RDWW) and after storage and rapid sand filtration
(RLWW), and treatment targets
ItT~SO/(mgII) 1:- I~WW 1--

~lowM J 10
___
I:::
I
12 150{; 1<2.5
AOX(J.igII.)
1 50{;
15 I~~15 1<10
<5
I pesticides (J.igII.) 1 ~0.1 10 1 <0.1
I total hardness (nunol/l) 12.7 12.5 [1.5
I Br"~l 1 200 1 100 1
2.2. mSTORICAL DEVELOPMENT OF THE PROCESS SCHEME OF THE RLWW
Figure 2. shows the historical development of the RLWW over the last 20 years. In 1983,
disiIifection using chlorine has been stopped. The main reason was the formation of harmful
halogenated compounds during chlorination. Depending on the content of organic matter expressed
in mg DOell, bromoform and non-volatile brominated hydrocarbons and/or chloroform and non-
volatile chlorinated hydrocarbons are formed.
The elimination of chlorination was possible as slow sand filtration was already available as
the final purification step. With slow sand filtration it appeared to be possible to produce
hygienically safe and biologically stable water (AOC < 10 ~), thus no postdisinfection was
necessary.
In 1987, softening was introduced in the treatment scheme with pellet softening as the
chosen technique. The process itself will be discussed in one of the next paragraphs. The reasons for
introducing central softening are presented in Table 2. (Van Ammers et al., 1986, Kostense, 1988).
BEFORE 1987 START 1987 START 1994 START 1998/1997
......
rapid sand filtration ozonation CD
CD
I
Chlori~tion Istopped
1983
Figure 3. Historical development of the River-Dune Waterworks

189
Table 2. Arguments for central softening
Public health aspects:

• reduction of the uptake of heavy metals (pb, Cu, Zn) in the distribution system
• exclusion of private softeners, which may lead to corrosion and bacteriological
problems
Economic aspects:
• reduced consumption of detergents
• reduced scaling and wear of kitchen apparatus
• reduced energy consumption of heating apparatus
• reduced wear of clothing
Environmental aspects:
• reduced heavy metals (pb, Cu) in sludge and effiuent from sewage
treatment plants
• reduced detergent use and phosphate content of sewage
• reduced brine discharge from private softeners
Aesthetic aspects:
• prevention of stains
• consumer comfort
In pellet softening the pH is increased by addition of NaOH, Na2C<lJ or Ca(OH):z to

crystallize CaC03• Because a high pH during ozonation results in a fast disintegration of ozone to
oxygen, softening is carried out after ozonation.
In 1992 biological activated carbon filtration (BACF) was introduced to replace
coagulation/rapid sand filtration. Coagulation/rapid sand filtration had been applied for the removal
of organic matter (especially DOC). However, removal of organic matter by this process is limited.
Thus, the load of the subsequent slow sand filters with organic matter acting as nutrient for
microbial growth might be relatively high, which complicates the production of biologically stable
drinking water and stimulates the development of Aeromonads. Besides, the process is costly and
generates large amounts of sludge. Moreover, the presence of pesticides and other organic
micropollutants was becoming more important, while limits for these compounds were becoming
more stringent. With coagulation/rapid sand filtration, removal of these compounds was insufficient.
Therefore, biological activated carbon filtration was introduced, which means light oxidation by
means of ozone, followed by carbon filtration. In the process AOC is increased to 100-150 ~g11 and
190
adsorption as well as biodegradation occurs in the carbon filter (Graveland, 1994). For the
reduction of DOC, an important target in the RLWW, biodegradation plays the dominant role.
Thus, in the BACF process the carbon filter acts like a bioreactor. The process will be descnbed in
more detail in one of the next paragraphs.
Carbon filtration is placed after the softening process for the following reasons:
• it also serves as post-filtration after softening if sludge (carry-over) is escaping from the
pellet reactors. If this sludge should reach the slow sand filters, clogging may occur and
frequent cleaning of the slow sand filters by scraping might be necessary. However, this
approach necessitates the use of activated carbon that resists backwashing, preferably with
water and air,
• due to biodegradation in the biological activated carbon filters, lowering of the pH takes
place, making a small dose oflye necessary of up to SI = + 0.6 (Langelier Saturation Index)
after the carbon filtration. In case of carbon filtration preceding the softening, first more lye
has to be dosed in the softening process and later some hydrochloric acid again would be
necessary. Minimal use of chemicals has been a major objective.
For introduction of the BACF in the existing treatment scheme, the already present pre-
ozonation has been used as the ozonation step in the BACF, while the rapid sand filters, present
after the coagulation process, were converted into carbon filters. Also new carbon filters have been
built.
2.3. mSTORICAL DEVELOPMENT OF THE PROCESS SCHEME OF THE RDWW
Figure 3. shows the development of the process scheme of the RDWW over the last ten
years, and a proposed modification in the near future. As in the RLWW, chlorination was stopped
in 1983, and for the same reasons.
Softening was introduced in the treatment scheme in 1987. In contrast to the RLWW,
softening is carried out directly on the raw water abstracted from the collecting basin, before rapid
sand filtration. To avoid clogging of the nozzles in the pellet reactors the water first passes a sieve
(3 mm) to remove larger particles (aquaguards).
After the discovery that certain pesticides and other organic micropollutants (amongst
others bentazone, metolachlorine, atrazine) passed through the layer of dune sand, in 1989 it was
1976-1987 START 1987 START 1992
Bethunepolder / Bethunepolder / Bethunepolder /

Amsterdam-Rhine Canal Amsterdam-Rhine Canal Amsterdam-Rhine Canal
coagulation coagulation coagulation
self purification basin self purification basin self purification basin
rapid sand filtration rapid sand filtration

,
rapid sand filtration
ozonation ozonation ozonation
softening softening ~
coagulation coagulatlOn
t
biological activated
carbon filtration
rapid sand filtration rapid sand filtration
slow sand filtration slow sand filtration slow sand filtration
chlorination
i
I stopped
1983
,
Figure 2. Historical development of the River-Lake Waterworks
192
decided to extend the treatment scheme with biological activated carbon filtration for the removal of
these compounds. As in the RLWW the main objective of BACF is the removal of organic matter
that acts as a nutrient (DOC) in general and AOX and pesticides in particular, in the RDWW the
main objective ofBACF is the removal of pesticides and organic micropollutants (AOX etc.), albeit
that also in the RDWW the removal of nutrients by BACF is important to increase the biological
stability of the produced water and to extend the running time of the slow sand filters. The sequence
of the process units will be ozonation - softening - activated carbon filtration - slow sand filtration.
Rapid sand filtration will be replaced in the process scheme and will be the first process step in the
post-treatment after dune infiltration.
The main reasons for the new location of the rapid sand filtration are:
• pollution of the pellet reactors with material from the collecting basin will be avoided,
• by two filtration steps (rapid sand filtration and activated carbon filtration) a drastic
reduction of organic material (nutrients, DOC) is reached, which results in prolonged
running times of the slow sand filters and an increased biological stability of the produced
water.
Treatment of the water according to the extended and modified process scheme started in
November-December 1994.
In the near future the capacity of the RDWW will have to be increased from 70 to 83.106
m /year. Extension of the dune infiltration by surface infiltration is not allowed. Therefore,
3
alternatives have been developed and are being studied currently. One of these alternatives concerns
the direct treatment of surface water without dune passage. Pre-treated river Rhine water (taken
from the Lek Canal and pre-treated by coagulation and rapid sand filtration) will be further treated
by ozonation, biological activated carbon filtration, slow sand filtration and finally by hyperfiltration
(reverse osmosis). Hyperfiltration is necessary (1) to lower the relatively high salt content of the
Rhine water, (2) to soften the water, (3) to remove pesticides and other organic micropollutants or
reaction products of ozonation that might pass the biological activated carbon filtration, and (4) to
act as an additional disinfection barrier. Extensive treatment preceding hyperfiItration (BACF and
slow sand filtration) is necessary to avoid mechanical damage and organic, biological and colloidal
fouling of the membranes, and to reduce the cleaning frequency of the membrane units. Besides, as
in this alternative no storage reservoir or dune passage is used which balances fluctuating
concentrations, enables selective intake and serves as an strategic supply, a robust, reliable treatment
scheme is required with several barriers to guarantee continuous drinking water production,
independant of the raw water quality. Hyperfiltration will be discussed in more detail in one of the
next paragraphs.
193
2.4. THE NEW TREATMENT PROCESSES
The new treatment processes that have been introduced in the RLWW and RDWW, or that
will be introduced in the short teon, concern pellet softening, biological activated carbon filtration
and hyperfiltration. In this paragraph, these processes will be discussed in more detail. Special
attention will be paid to the fundamentals of the processes, the merits, advantages and limits of the
processes, and finally some operational aspects.
2.5. WATER SOFTENING BY MEANS OF PELLET REACTORS
2.5.1. Principles ofPeUet Softening
Softening basically involves the dosing of an alkaline reagent (NaOH, caustic soda;
Ca(O~, lime or Na2C03, soda ash) to increase the pH in the case of NaOH and Ca(OHh, and to
increase the C~2- content in the case of Na~~.
By increasing the pH, hydrogencarbonate is converted into carbonate, the solubility product
of calcium carbonate will be exceeded and nucleation takes place:
C03 2- + Ca 2+ => CaC03..1-
A difference can be distinguished between homogeneous nucleation (precipitation) and

heterogeneous nucleation (crystallization). In the presence of a large crystallization area
spontaneous homogeneous nucleation in the water phase is avoided and crystallyzation of CaC~
on seed crystals proceeds at a lower pH compared with precipitation.
Precipitation takes place when the pH, after addition of the alkaline reagent, is about 1-1.5 above
the equilibrium pH. Precipitation:
• takes a relatively long time (1-4 hours),
• produces voluminous sludge floes in the water phase.

194
Crystallization takes place when the pH is about 0.3-0.4 above the equilibrium pH. Crystallization:
• proceeds very fast (10 s),
• produces a solid waste when crystallization takes place on the surface of sand grains,
providing the required crystallization surface. Dry, hard and durable marble grains are
produced.
Figure 4. shows the difference between crystallization and precipitation.
········..,.·:::::::::. ::.:...P.~~~.6
...
....... pr~ipit.ation
...... ",
pH=9.6
.. .. ........
... ...
crystallization ... ..... ......
pH=8.6
o 2 3 4
time (h)
Figure 4. Time required for softening by crystallization and precipitation
The mechanism that determines the rate of the process concerns micro-nucleation. In the
case of precipitation (no crystallization surface) this can be quickened, but then an even higher pH is
required (in Figure 4. pH = 9.6 instead of pH = 8.6).
In a pellet reactor sand grains are fluidized homogeneously by the upflow velocity of the
water and the fluidized bed provides the crystallization surface. The grains are graded over the
height of the reactor and move freely in the upward flow of the water so that cementing of the
grains is prevented. The principle design of the pellet reactor is shown in Figure 5.
The Amsterdam reactor type ("Reductor") (Graveland et al., 1983) is cylindrical over the
entire height of 5 m. The top consists of a cone to prevent wash-out of grains. The bottom of the
reactor is equipped with specially designed nozzles that allow a separate injection of water and
195
caustic soda solution into the reactor. With enough of these nozzles per square meter proper initial
mixing is obtained while the turbulence in the reactor can be kept relatively low in the higher region
to enable a homogeneously fluidized bed.
This, in combination with the cylindrical shape, ensures a stable flow pattern and a complete
classification of the fluidized bed; i.e., the larger grains are found at the bottom and the smaller
grains at the top of the bed, indicating a good plug flow. The Amsterdam reactor can be operated
with caustic soda and soda ash.
overflow weir
softened water
__-+- fluidized bed
~--. - - nozzle
_ supply of seed grains
r--I-- reactor bottom
~ discharge of grown pellets
- supply of soda (25%)
'oill"ort (
'--_ _ water chamber
Figure 5. The Amsterdam reactor: softening ofwater by means of crystallization
Due to the crystallization of CaC~ on the grains, they grow into pellets. The larger and
heavier pellets hydraulically grade towards the base of the reactor where they can be removed
periodically and are replaced by the same amount of smaller diameter seeding grains. As only a part
of the pellets is removed, this procedure can take place during full operation.
196
Because the sedimentation velocity of the grains is much higher than that of CaC~ flocs,
the surface load of pellet softeners is about 20-100 times that of a sludge process (precipitation
process). The high rate of crystallization allows a short detention time.
Some important process conditions are:
• a minimum flow through the reactor is required to prevent cementing of the grains. This is
called the minimum fluidization velocity and is governed by the grain size of the pellets and
the temperature. The minimum fluidization velocity for grains of 1 mm (the size, at which
grains are normally removed from the reactor) is approximately 50 mIh at 20°C,
• the dosage of the alkaline reagent must be constructed in such a way that mixing is fast and
homogeneous over the whole cross-section of the reactor in order not to exceed the pH
value at which precipitation takes place and sludge floes are formed,
• the water and chemical must be mixed in the presence of seed grains with a high specific
surface area to obtain immediate crystallization.
2.5.2. Merits of the Process
The most important advantages of pellet reactors (crystallization) above the sludge process
(precipitation) are:
• the crystallization process produces a relatively small quantity (approximately 10 percent by

volume compared with the sludge process) of dry, hard, and durable marble grains (pellets),
that can be used for other purposes, including treatment of aggressive water, neutralization
of acid wastewater, and utilization for road construction and cement manufacturing,
• the difference between efiluent pH and equilibrium pH for the crystallization process is less
than it is for a precipitation process, indicating a lower required supersaturation. This
involves a higher chemical efficiency and fewer, if any, scaling problems,
• treatment plants that use the crystallization process are reliable and relatively simple in
operation as compared with the precipitation process. In this process, sludge processing
itself requires operational action,
197
• investment costs for the crystallization process are much lower than those for the sludge
process. Due to the high surface load that can be applied (100 mIh and higher) the required
surface area of the reactors is low. Also costs for operation and maintenance are lower
owing to higher reliability and the fonn of CaC~ produced.
With softening, reduction of the Cu- and Pb-dissolving potential is the objective.
Observations in the Netherlands show that by pellet softening the Cu- and Ph-dissolving
potential is reduced by 60-80010 and 40-50010 respectively (Van Annners et aI., 1986).
2.5.3. Limitations
Softening with NaOH and Ca(OH):z affects the concentration of hydrogencarbonate, while
softening with NaOH and Na2C03 affects the concentration of sodium, as can be seen from the
chemical reactions below:
Ca2++HC~- + Na+ + Off ~ CaC~,J, + H20 + Na+

caustic soda
Ca2+ + 2HC03" + Ca2++ 20Ff ~ 2CaC~,J, + 2H2O

lime
Ca2+ + 2Na+ + C~2" ~ CaC03,J, + 2Na+

soda ash
The use of lime results in twice as much reduction in HC~- as compared with caustic soda.
Soda ash does not change the HC~- concentration. Caustic soda and soda ash both result in an
increase in the Na+concentration.
In order to distribute water with enough pH buffering capacity, and to avoid corrosion of
cast iron or copper pipes, it is recommended to maintain the bicarbonate concentration above 2
mmolll (Van der Hoven and Van Eekeren, 1988). In the current standards, the standard for sodium
is 120 mg/l. Hence, the choice of chemicals is determined by the raw water quality.
Where the HC~- concentration of the raw water is already too low, only soda ash can be
used.
198
Where the HC~" concentration is high, caustic soda or lime can be used. With moderate
HC03" concentrations, only caustic soda may be used.
When the Na+ concentration in the raw water is high, the sodium standard may exclude the
use of soda ash and caustic soda.
In practice, post-filtration is necessary to remove carry-over from the pellet reactor.
Especially in the case of softening with lime, carry-over can be high (50-70 mgII, turbidity up to 40
NTU), although with the use of special methods for the preparation of the lime milk (e.g.
application of lime with high chemical purity, or use of mixing water free of carbon dioxide and
hydrogencarbonate) a reduction of 60-80010 in carry-over can be obtained (Van Eekeren et aI.,
1994).
When softening is canied out with caustic soda, the turbidity increase is generally less than
1 NTU. This leads to filter runs of the post-filters of one week or longer (Graveland et a1., 1983).
2.5.4. Technical and Operational Aspects
The dimensions of pellet reactors vary and depend on the construction type (Kostense,
1988). In practice, diameters of 1-3 m are applied and heights of 5-14 m The upflow velocity
ranges from 60-100 mIh in case of lime, and from 50-125 mIh in case of caustic soda. The fixed bed
height is approximately 2 m, while after fluidization the bed height is approximately 1.5-2.5 times
higher.
The grains, used as crystallization area, can be either ordinary sand (silicate) or garnetsand.
Advantages of garnetsand are:
• a high specific weight, so the seed grains can be small at the same upflow velocity without
the rise of wash-out. The small particle diameter results in a high specific surface area on
which crystallization can take place,
• garnet sand has a relatively rough surface structure, promoting nucleation of CaCD3 on the
grains.
199
2.6. REMOVAL OF ORGANIC MATfER BY MEANS OF BIOLOGICAL ACTIVATED

CARBON FILTRATION
2.6.1. Principles of Biological Activated Carbon Filtration
Biological activated carbon filtration (BACF) is based on the combination of ozonation and
activated carbon filtration. As already discussed in the preceding paragraphs, BACF is applied in
order to reduce the content of organic matter for (micro)biological growth (nutrients) in general,
and that of pesticides and other organic micropollutants in particular. In BACF the water is first
ozonated with a low ozone dose « 1-2 mg OYI) to produce AOC (assimilable organic carbon) up
to a concentration of 100-150 Jlgll. As a result, the biodegradation capacity in the subsequent
carbon filters is increased, because the increase in AOC yields a higher bacterial density in these
filters.
Hence, the organic matter can be biodegraded or adsorbed, depending on the nature of the
material. Compared with activated carbon filtration without preceding ozonation, the
biodegradation of the organic matter is strongly stimulated, and also organic matter, already
adsorbed on the carbon, might be biodegraded (bio-regeneration). The biodegradation of natural
organic matter (NOM) favours adsorption of pesticides and other organic micropollutants, because
it lowers NOM-preloading onto the carbon. Also the more persistent organic compounds may be
biodegraded as a result ofthe higher biological activity in the carbon filters.
Ozone preceding activated carbon filtration has two important positive effects:
1. By ozonation the AOC content of the water is increased, stimulating the biological activity
in the subsequent carbon filters. Because both adsorption and biodegradation of organic
matter takes place in the carbon filters, the running time of the carbon filters is prolonged
with a factor of 10. Due to the longer running times between reactivations, costs for
reactivation of the carbon are substantially reduced. Figure 6. shows an example of the
ozonation on the running time of carbon filters. From this figure it can be seen that for a
decrease of DOC, biodegradation plays a dominant role. Thus the carbon filter acts like a
bioreactor.
200
10
• EXPECTED WITH OZONE
9
+ WITHOUT OZONE IN PRACTICE
• WITH OZONE IN PRACTICE
6
~--~
- -.---.....-------+-----+
/
/'
s
/
/
~
/
/
""
I
,!, ~
4 /
//
3
/ /~ ~./ /~
2 / ~~
~/
Y
/~-
o L-
o 2 4 6 8 10 12 14
RUNNING TIME (WEEKS)
Figure 6. The DOC values in the influent (6.5 mgIl) and effluent of the carbon filters as a function of the
running time, with and without preceding ozonation
Figure 7. shows the production of AOC as a function of the applied ozone dose (DOC
content of the raw water 2 mgll).
140
[20
[00
~
~
80
~
...:
60
40
20
0
0 0.5 1.0 1.5 2.0 2.5 3.0
ozone dose (mg/l)
Figure 7. AOC production as function of the ozone dose

201
2. Ozonation has a disinfection capacity (CT = 1-2). Especially in the winter the number of
viruses in the river Rhine water, used in the RDWW, can be high (40.000/m3), and also in
the initial filtration steps. The still present micro-organisms can be sufficiently inactivated by
ozonation with CT = 1-2 mWL'min.
2.6.3. Limitations
Ozonation can have two negative effects:
1. Ozonation may result in the oxidation of halogens, and bromate (BrOi) can be formed as a
result of the presence of bromide (Br) in the raw water. Bromate is known to be
carcinogenic (Kurokawa et al., 1990). The World Health Organization recommends a
provisional guideline value of 25 J1g/l with an excess lifetime cancer risk of 7xl0·5• In the
Netherlands a maximum admissible concentration of 0.5 J1g/l is proposed with an excess
lifetime cancer risk of 1O~.
The amount of bromate that is produced is affected by the DOC content and bromide
content of the raw water, and varies with the ozone dose. For the RDWW and the RLWW
the extent of bromate production has been measured at several ozone doses. Figure 8.
shows that with an ozone dose below 1 mg/l, as is applied in the BACF process, bromate
production is below 1 J1g/l, both for the RDWW and for the RLWW.
140
I
120
/
/
RDWW J RDWW RLWW
E'
co 100
C ! DOC (mgll) 2.0 6.5
Ul / Cl (mgll) ISO 70
~0
80
/ 100
Br (Ilgll) 200
~ 60
/
Ol /
--
40
/
/
/ RLWW
20
/
0
o
.....---:.
2 4
------ 6
OZONE ooSE (mgll)
Figure 8.The amount of bromate, produced by various doses of ozone for the RDWW and the RLWW
202
2. The BACF process may result in too high a content of AOC in the eftluent of the slow sand
filters (AOC > 10 p.gIJ.). However, with adaquate doses of~, AOC production can be
restricted to a maximum of 100-150 J.1g/J., which can be removed from the water by good
operation of the subsequent carbon filters and slow sand filters. Figure 9. shows the AOC
concentration in the water after ozonation, and after activated carbon filtration + slow sand
filtration by three ozone dosages, as measured in a pilot plant treating the raw water of the
RDWW.
140
120
100
80
~ 60
~
~
00(
40
20
0
after ozonation
- after GAC and SSF
t:::::J ozone dose 0.38 mgll _ ozone dose 0.86 mgll [==:::J ozone dose 3 mgIl
Figure 9. Removal of AOC by activated carbon filtration (GAC) and slow sand filtration (SSF) after
ozonation with varying ozone doses
The empty bed contact time of the carbon filters in the BACF process amounts to 40 min.
This contact time is used to stimulate and maximize biodegradation in the filters, while with this
contact time also the lowest total costs per cubic meter are reached. Figure 10. shows the capital
costs, reactivation costs and total costs as a function of the empty bed contact time.
As an alternative for one-stage filtration with a contact time of 40 min, two-stage carbon
filtration can be applied, with in each step a contact time of20 min. When a filter in the first stage is
loaded and shows break-through above a certain value, it will be reactivated. A filter that has
operated in the second stage subsequently can be placed in the first stage. With this approach a filter
can be loaded more heavily before it is taken out of operation for reactivation. Reactivation costs
203
are decreased with this set-up (15%), but investment costs will be higher. In the RLWW one-stage
BACF is used, in the RDWW two-stage BACF is used.
30
28
26
\r\
,
24 ~
~
~ 22
20
\
(,)
\ \
j 18
16
\
...:I
\
~ 14
"'"
--
12 \ --".~
'"
I-<
'"0 10
-...
"'. ~...... -----
-
(,)
8
6
4
. .... ...... r-
2
0
o 20 40 60 80
CONTACT TIME (MINUTES)
• CAPITAL COSTS C REACTIVATION COSTS
6 TOTAL COSTS
Figure 10. Costs per cubic meter of drinking water as a function of the empty bed contact time: capital costs,
reactivation costs and total costs
As a result of biological activity in the carbon filters, biomass accumulates and head loss
increases. For this reason, the filters have to be backwashed periodically (every 1-4 weeks). To
avoid fonnation of mudbaIls that move down into the filter bed during backwash with only water,
backwash has to be carried out with water in combination with air. Due to the biological activity
during carbon filtration, the effluent of the carbon filters might contain a low oxygen content.
Therefore, dosing of oxygen into the effluent of the carbon filters might be necessary.
2.7. REMOVAL OF CONTAMINANTS BY MEANS OF MEMBRANE FILTRATION
2.7.1. Principles of Membrane Filtration
Membrane filtration is the general name for processes that use a semi-permeable membrane
for the separation of water and contaminants. In pressure driven membrane processes, the water is
204
driven through the membrane while the contaminants are retained by the membrane. The membrane
acts as a selective banier, permitting the passage of certain components and retaining other
components, as shown in Figure 11.
The feed water is split in a product stream (the low concentrated permeate) and a waste
stream (the concentrate or brine, a high concentrated waste stream).
/t product
membrane
.. . brine
•
~.
•",.
.}" •
~t>
?$Ir6' •
• •
•
•
\).
~.
rI'.•• ~ §' •
~ •
~<§J. ~
•• • $-o~ ~
Cl
Figure. 11. Principle of membrane filtration
A distinction can be made between microfiltration, ultrafiltration, nanofiltration and

hyperfiItration, the last one also named reverse osmosis. Figure 12. defines the different membrane
technologies as a function of the membrane pore size, related to the size of particles occuring in the
water.
Microfiltration membranes are used for the removal of colloidal and suspended matter,
which are in the pore size of the membrane. Ultrafiltration membranes retain colloids and
macromolecules such as proteins, which have a molecular weight larger than 1000-100,000. The
selectivity of microfiltration and ultrafiltration membranes is primarily based on sieving.
Nanofiltration membranes and hyperfiltration membranes both retain dissolved ions and organic
compounds, the rejection by hyperfiltration being superior to that by nanofiltration: hyperfiItration
retains monovalent and multivalent ions, while the retention of monovalent ions by nanofiltration is
limited. Nanofiltration therefore is a suitable technique for the removal of hardness and higher
molecular weight compounds. Hyperfiltration is a desaIination technique, also removing low
molecular weight compounds, and has been used widely for desalination of brackish water and sea
water. Nanofiltration and hyperfiltration are diffusion controlled processes.
As can be seen from Figure 12., in principle disinfection can be achieved by ultrafiltration,
nanofiltration and hyperfiltration, as these processes retain bacteria and viruses.
205
Organic macromoleculas
Pesticides
Organic
Algae Colloids
'----;:::::::===== compounds
Bacteria I ViruseS I Dissolved
salts
I
Pollens IIYeastsl
I I I I I
100 Jlm 10 O.l 0.01 0.001 0.0001
o o o o
IReverse osmosis I
Red Smallest Polio
hair globule microorganisms virus INanofllttation I
Ultrafiltration
Microfllttation
'ISan-d-fj-lte-r'I
Figure 12. Relation between membrane pore size and the size of the "particles" retained by the membranes
The membrane processes are operated at different pressures. Table 3. summarizes the main
characteristics.
Table 3. Main characteristics of pressure driven membrane processes
Filtration Particle size Pressure MWCO I Specific flux

process (~) (bar) (lIm2/dIbar)
micro > 100 <1 1000-5000
ultra 10-100 5 100-100000 400-1000
nano 1-2 5-10 200-500 100400

hyper 0.2-1.0 15-60 100 50-100(BW1
1-5 (SWl)
I Molecular Weight Cut Off (90"10 of compound is rejected)

2 Brackish Water
3 Sea Water
206
The membrane structure can be either homogeneous, asymetric or composite. In

hyperfiltration and nanofiltration nowadays the Thin Film Composite membranes (TFC membranes:
0.1 ~ separating material, usually polyamide; 50-75 ~ microporous base, usually polysuIphon;
reinforced by a polyester support fabric, 125-150 J.Ull) in particular are gaining application.
Compared with the traditionally used asymetric Cellulose Acetate membranes (CA membranes)
they have superior qualities with respect to stability (no compaction, not biodegradable, no
hydrolysis), retention and flux.
Membrane filtration is a strictly physical, absolute separative technique whose quality

performance depends only on the membrane pore size. Conventional filtration processes used in
water treatment can only provide relative effluent quality, determined by both the raw water quality
and the process operating conditions. In membrane filtration processes the treated water quality is
no longer dependent on the raw water quality, but on the pore size occuring in the implemented
membrane. Raw water quality variations and operational conditions influence only the process
throughput, this might require extensive post-treatment (see "Limitations").
Depending on the type of particles that have to be removed, the most appropriate type of
membrane filtration process can be chosen, and several treatment stages can be combined in one
process unit (e.g. desalination, softening and disinfection by hyperfiltration).
Since the driving force of the process is pressure, automation is relatively simple. The
membrane processes are compact as compared with conventional processes (e.g. microfiltration as
alternative for coagulation), resulting in relatively small treatment plants. Moreover, it is a modular
process and extension oftreatment plants can be quite straight foreward.
2.7.3. Limitations
Although membrane filtration in drinking water treatment is a promising technique that can
be used for many purposes, it has also limitations and disadvantages. The most important are:
1. In the process a concentrated waste is produced, the brine, that has to be disposed of,
2. Energy consumption, especially in the case of nanofiltratiom and hyperfiltration, is

somewhat higher than that of conventional processes (approximately + 0.5 kWhlm\ On
the other hand, with these new processes a higher drinking water quality can be obtained as
207
compared with conventional treatment, and the membrane processes can save costs, e.g.
costs for softening in the case ofhyperfiltration and nanofiltration,
3. The perfonnance of nanofiltration and hyperfiltration for the removal of organic low
molecular weight compounds is not yet fully known. Most of the membranes, available at
the moment, have been designed for desalination of sea water and brackish water
(hyperfiltration), and for softening (nanofiltration). Research into the rejection of these
compounds by the membranes is now being canied out. Development of new membrane
materials might be required to reach the high drinking water standards for these compounds,
4. In principle, disinfection is possible with u1tra-, nano- and hyperfiltration. A log-reduction of

at least 5-7 has been reported (Boller, 1994, Anselme et aI., 1993). Two important
questions however remain:
• can the integrity of the membranes and membrane modules be guaranteed,

what is the chance and risk of leakages (connections of the elements
(O-rings), application of glue, etc.),
• how can the integrity of the membranes be monitored in full-scale plants in a

relatively simple manner. In research pilot-plants, dosing of bacteriophages
has been applied to measure high log-reductions. In full-scale plants this is
not possible. With a relatively simple analytical technique such as
conductivity, only a leakage in the range of log 2 - log 3 can be measured.
So, new analytical techniques have to be developed, or a good strategy
applying, for example, conductivity has to be developed to monitor the
integrity of a full-scale plant.
5. The membranes might release undesirable, toxic substances into the water, especially new
membranes. In many countries, approval for toxicological parameters is required for the use
of membranes in drinking water treatment, stressing the need for migration tests.
In practice, membrane systems consist of a high pressure pump and a number of modules,
with in every module a number of membranes (modules or pressure vessels). The capacity of the
modules is limited, so modules in parallel are required to obtain the desired product flow. In order
208
to obtain proper flow velocities along the membranes (important to prevent fouling) the recovery
(this is the permeate flow as percentage of the feed flow) per module is limited to 40-50010,
otherwise the brine flow would become too low. When a higher total recovery is required, modules
must also be arranged in series. An often applied arrangement is the so called "X-mas tree"
arrangement, as shown in Figure 13. After the first stage, the brine of this stage (50-60% of the
original feed) is sent to the second stage. So, the number of modules in this second stage is about
50010 of that of the first stage. In the same way a third and fourth stage can be used to achieve the
desired overall recovery of the plant.
x - mas,~t:r:ee:ljti=¢=t~~:r--C;:~I-r-r---I)(I-'_
Feed
Figure 13. Membrane staging
In the case of nanofiltration and hyperfiltration, attention must be paid to pre-treatment of

the water. In order to prevent colloidal and biological fouling, which would result in damage of the
membranes and require frequent cleaning of the membranes, the water to be treated should have a
MFI (Modified Filtration Index, a parameter determined by means of filtering a water sample
through a 0.45 J.UIl membrane filter) below 0.5-1 s/l2 and an AOC concentration (Assimilable
Organic Carbon) below 10 11g/l (biologically stable water to prevent bacterial growth). Possibilities
to reach this high quality of the water are the use of microfiltration or ultrafiltration preceding the
hyper/nanofiltration, or the use of ozonation-activated carbon filtration-slow sand filtration
preceding the hyperfiltration, as applied by Amsterdam Water Supply. An advantage of the second
alternative concerns the removal of organics from the water, including pesticides and other organic
micropollutants, before the hyperfiltration. As a consequence, the brine of the hyperfiltration will not
contain organic pollutants and disposal of the brine will be easier.
In the concentrate, high concentrations occur and the solubility of certain inorganic salts can
be exceeded, resulting in scaling of these salts on the membranes (e.g. CaC~, BaSO", Caso",
SrSO", CaF2). In order to prevent the precipitation of CaC~ the pH of the feed is adjusted by acid
dosing to the feed. In order to prevent scaling of the other compounds, an anti-sealant can be dosed
which inhibits precipitation and crystallization of salts. Then the solubility product of the compounds
may be exceeded by a factor up to 50-100 without problems.
209
In the case ofhyperfiltration, the pH and hardness of the product water will be very low and
the C~ content will be high (C~ produced by the acid dose to the feed and passing the
membrane). Therefore, a pH correction is necessary and the hardness has to be corrected, for
example by filtration over CaCO:J grains, which will simultaneously neutraIize C~ and increase
hardness. An attractive alternative set-up is mixing of the product water with hard water in a split-
treatment mode. In this way the desired hardness and pH may be obtained without additional
treatment, depending on the water composition and split-treatment ratio.
As a summary, the main key-design aspects which have to be taken into account in
membrane filtration processes are shown in Figure 14.
Brine disposal
Discharge to sea
Discharge to surface water
Further processing
Raw water Clearwater
Pretreatment Membrane system Posttreatment

Colloids Recovery C02 - removal
Organics Membrane type Hardness correction
Biology Pressure pH control
Scaling
Nanofiltmtion Degasification-marble filtmtion
Flolation-Filtmtion-(SSF) (Hyperfiltmtion for salts) (Mixing in split-treatment mode)
FigUre 14. Key design aspects in membrane filtration
2.S. REFERENCES
Anselme C., Mandra V., Baudin I., MaIIevialle I (1993) Optimum Use of Membrane Processes in
Drinking Water Treatment. Proceedings of the 19th International Water Supply Congres,
Budapest, 2-8 October 1993, Special Subject 2 "Membrane Technology": SS2.1-SS2.11.
Ammers M. van, Dijk IC. van, Graveland A, Nuhn PAN.M. (1986) State of the Art ofPeIlet
Softening in the Netherlands. Water Supply 4: 223-235.
210
BollerM. (1994) Trends in Water Filtration Technology. Aqua 40: 65-75.

Eekeren MW.M. van, Paassen JAM. van, Merks WAM. (1994) Improved MiIk-of-Lime for
Softening of Drinking Water - The Answer to the Carry-Over Problem. Aqua 43: 1-10.
Graveland A (1994) Application of Biologica1 Activated Carbon Filtration at Amsterdam Water
Supply. Paper presented at the IWSA Specialised Conference on Activated Carbon in
Drinking Water Treatment, Amsterdam, 27-28 September 1994.
Graveland A, Dijk IC. van, Moel P.I de, Oomen IHC.M. (1983) Developments in Water
Softening by Means of Pellet Reactors. Journal American Water Works Association
December 1983: 619-624.
Hoven Th.J.J. van den, Eekeren M.w.M. van (1988) Optimal Composition of Drinking Water.
KIWA Communication No 100, Nieuwegein, The Netherlands, October 1988 (m Dutch).
Kostense A (1988) Softening in Pellet Reactors. KIWA Communication No 102, Nieuwegein, The
Netherlands, October 1988 (in Dutch).
Kurokawa Y. et aI. (1990) Toxicity and Carcinogenicity of Potassium Bromate - A New Renal
Carcinogen. Environ. Health Persp. 87: 309-335.
3. Treatment of Drinking Water in the United States
on the Threshold of the 21st Century
Thomas G. Sanders
Tory L. Champlin
USA
3.1. INTRODUCTION
Three basic objectives of water treatment are (ASCE and AWWA, 1990):
Production of water that is safe for human consumption,
Production of water that is appealing to the public, and
Production of water using facilities which can be constructed and operated

at a reasonable cost.
The production of biologically and chemical safe water is the primary goal in the design
and operation of water treatment plants. A properly designed treatment plant, however, does
not guarantee safety. Skilled and properly trained operators that are attentive to sanitary
requirements of water supplies and distribution systems are also equally important.
Today, water treatment professionals are facing multiple concerns including the
elimination of contaminants which cause acute health effects such as pathogenic organisms
along with the removal of trace contaminants which have long-term or chronic health effects.
While water supplies have remained relatively constant, the demand for these resources
continues to increase. Surface waters and ground waters are primarily the sources used for
drinking water. Contamination of surface waters through dilution with treated wastewaters has
continued to degrade surface water quality. Despite dramatic improvements in wastewater
NATO ASI Series, Partnership Sub-Series. 2. Environment - Vol. IS

212
treatment since the 1960s, further degradation of surface waters is still occurring. An
additional challenge is the growing awareness of increased contamination of ground water
sources. Ground water aquifers are threatened by agricultural contamination from nitrates and
other inorganics and infiltration by manufactured synthetic organic chemicals. There is still a
need to improve water quality. With continued economic development and population growth,
demands on water supplies will require innovative approaches to treat and protect water
quality. The demand for water professionals will be extensive in the upcoming century.
3.1.1. Definitions of Public Water Systems in the United States
A public water system in the United States (US) is defined as a system that provides
piped water for human consumption to at least 15 service connections, or serving an average
of at least 25 people for at least 60 days each year (USEPA, 1993). In 1992, 197,060 water
systems in the US including Indian lands and US territories were classified as public water
systems (pontius, 1994a).
Due to the number of different types and sizes of water systems, public water systems
are further subdivided into three types of water systems:
• Community water systems - provide water to the same population year-round,
• Non-transient, non-community water systems - regularly serve at least 25 of the

same people at least six months of the year (e.g., schools, factories, and hospitals with
their own water supply,
• Transient, non-community water systems - serve transitory customers in non-

residential areas (e.g., campgrounds, motels, and gas stations).
Reasons for the different classifications of public water systems relate to length of
exposure to a particular drinking water. For example, persons who drink from a transient, non-
community water system have only a brief exposure to contaminants in the water. In contrast,
persons who drink daily from a community water system are exposed to the contaminants on a
continual basis. Due to these differences in exposure, community, non-transient non-
community, transient non-community water systems are regulated differently and have different
water quality standards.
Community water systems are the main focus of this paper. Non-community systems
are mentioned briefly regarding waterborne disease outbreaks discussed later.
213
3.1.2. Growth of Community Water Systems
Water treatment began in the US in the late 1800s. In 1887, an experimental station
consisting of rapid sand filters was constructed in Lawrence, Massachusetts by the
Massachusetts State Board of Health, after an outbreak of typhoid fever (Culp et a1., 1986).
Soon after slow sand filters were introduced in Albany, New York (Craun, 1986). Other
filtration experiments were also conducted in Louisville, Kentucky, and in New Orleans,
Louisiana in the 1890s. Early experiments, however were not completely successful. Attempts
to filter muddy waters resulted in clogging and short operational cycles. This problem was not
alleviated until aluminum sulfate (alum) was introduced as a pretreatment. Although the use of
alum had been proposed in the 1757, it was not effectively demonstrated in the United States
until 1885 (Vesilind and Peirce, 1982).
In 1908, the introduction of chlorine disinfection was introduced in the United States at
Jersey City, New Jersey. The dramatic effect of chlorination was marked in the years that
followed by dramatic decreases in deaths due to typhoid fever and other waterborne diseases.
Table 1. Growth of community water systems in the United States from 1800 to 1958 (Craun, 1986)
Year Number of Number of Number of

Water Facilities Purification Disinfection
Facilities Facilities
1800 16
1850 83
1870 243
1880 598 1
1890 1,878 NA
1896 3,196 NA
1924 9,850 638 NA
1930 NA 1,531 2,917
1934 10,790 NA NA
1938 12,760 NA 4,054
1940 NA 1,855 4,590
1948 NA 2,054 6,137
1958 17,808 2,517 8,845
As shown in Table l., community water systems were developed for most major cities
in the US during the late 1800s and early 1900s. Disinfection of water to prevent waterborne
disease was used increasingly in the years prior to 1940.
214
As shown in Table 2., by 1904, only 10% of the urban population in the US received
filtered water. By 1958, that number had increases to 58% of the urban population.
The growth of rapid sand filtration has been more dramatic than slow sand filtration.
By 1958, rapid sand filters supplied water 95% of the population receiving filtered water.
Table 2. Growth of sand filtration in the United States (Craun, 1986)
Slow Sand Rapid Sand Percent of Urban

Year Filtration Filtration Population Served by
(population) _(population) Filtered Water
1800
1850
1870 0 0 0
,
"'1\ 1\1\1\ 1\ .... 1
1890 35,000 275,000 2
1900 360000 L500,000 6
1904 560000 2,600,000 10
1910 3,883000 6922,000 28
1914 5,398,000 11,983,000 36
1924 5,054,000 18,610,000 38
1940 3,579200 26751,700 41
1948 2,014,635 39,731,460 47
1958 3,177,380 55,606,245 48
3.1.3. Water Sources
There are approximately 58,000 community water systems in the US serving

approximately 245 million people. The remainder of the population receives its residential
water from private wells or other non-federal regulated systems (i.e., those serving fewer than
15 service connections or 25 people). As Table 3. shows, 10,768 community water systems
(18% of the total number of systems) providing water to 140 million people (57% of the total
population) receive surface water, and 47,898 community water systems (82% of the total
number of systems) providing water to 105 million people (43% of the total population)
receive water from ground water sources. Based on the total number of systems, 62% are
considered very small systems (servicing 25 to 500 people) providing water to only 2% of the
population. In contrast, only 1% of the total systems are considered very large (more than
100,000 people) providing water to 45% of the population. A breakdown of other system
sizes, their number and population served is shown in Table 3.
215
Table 3. Surface and ground water sources used by Community Water Systems (CWSs) in the United
States at the end of 1992 (Pontius, 1994a)
Nwnber of CWSs served by Surface Total Population Served by Surface and

and Ground Water Sources Ground Water Sources
Total Percent Population Percent of
System Size Surface Ground Nwnber of Surface Ground Served by Population
Water -Water ofCWSs CWSs Water -Water CWSs Served by
x10 6 xl0 6 x10 6 CWSs
Very small 3,526 32,985 36,511 62 0.72 4.85 5.57 2
(25-500)
Small 3,678 10,838 14,516 25 5.70 14.35 20.05 8
(501-3,300)
Medium 1,746 2,505 4,251 7 10.35 14.38 24.73 10
(3,301-
10,000)
Large 1,598 1,464 3,062 5 45.99 39.04 85.04 35
(10,001-
100,000)
Very Large 220 106 326 1 76.95 32.85 109.80 45
(> 100,000)
Total 10,768 47,898 58,666 100 139.71 105.47 245.19 100
3.2. mE PROCESS TO IMPROVING WATER QUALITY
As shown in Figure 1., there are three basic steps followed for improving drinking
water quality: (1) the development of criteria based on identified health effects, (2) the creation
oflaws and standards regulating the quality of treated water, and (3) the selection of treatment
goals by community water systems to achieve regulated levels.
The quality of drinking water throughout the US is determined by federally enforced
regulations. These regulations establish water quality standards concerning contaminants which
have a direct effect on public health. The criteria for these standards are developed by various
groups which define contaminants and their associated concentrations which should not be
exceeded in treated drinking waters.
216
Health Effects
•
Developing
Water Quality Criteria
Regulations
•
Promull1llling
Water Quality Standards
Community Water Systems

Establishing
Water Quality Goals
Figure 1. Process of improving water quality in the US
Once detennined, these criteria serve as a basis for establishing water quality standards.
These standards are promulgated into federal laws and regulations. Community water systems
are then required to meet these standards which may involve modifying existing treatment
practices or the design and construction of new facilities. Water quality goals are contaminant
concentrations chosen by community water system to ensure a quality of water that meets
regulated standards.
3.2.1. Health Effects - Identifying Risk
Water quality criteria are developed solely on the basis of data and scientific judgment
without consideration of technical or economical feasibility. These criteria may be developed
for different uses. For a single contaminant, separate criteria can be applied for drinking water
based on either health effects or appearance.
3.2.2. Regulations - Driving Water Quality
There are two major regulations which govern the quality of water in the US: the Clean
Water Act (1948 through 1972) and the Safe Drinking Water Act (1974). The Clean Water
Act, originally enacted in 1948, regulates the discharge of pollutants into US waterways. This
law directly effects the discharge of wastewater from municipal and industrial treatment
facilities into rivers, lakes and surrounding oceans. The Safe Drinking Water Act, enacted in
1974, regulates harmful contaminants in public drinking water systems.
217
Setting water quality standards is important and many times a controversial aspect of
water treatment. Since the passage of the Safe Drinking Water Act in 1974, emphasis has been
placed on linking low levels of contaminants to higher incidence of disease such as cancer. The
ability to measure quantities of contaminants improve and our knowledge of their health effects
increases, water professionals are faced with increasingly complex water quality regulations.
Understanding and forecasting the promulgation of these regulations are important to water
professionals for a number of reasons. Standards affect the selection of raw water sources,
choice of treatment processes and design criteria, ability to modify existing treatment processes
and treatment costs.
Unlike the development of water quality criteria, water quality standards must consider
factors such as economic impact, technical feasibility, and political realities. Establishing
standards is a careful balance between the benefits of protecting public health and the costs of
meeting various water quality levels.
3.2.2.1. Clean Water Act. The objective of the Clean Water Act is the restoration and
maintenance of the "chemical, physical and biological integrity" of the waters in the US. Two
goals were included in the 1972 legislation: zero discharge of pollutants by 1985 and an interim
goal and where possible, water quality that is both "fishable" and "swimable" by 1983
(USEPA, 1993). While these dates have passed and the established goals have not been met,
the goals still remain, and efforts to attain these goals continue. There is still, however, some
question as to the motive of zero discharge. Is it the complete elimination of pollutants in the
discharging effluent, or an effluent having a concentration of pollutants less than or equal to
the concentrations in the receiving stream? Prohibiting the discharge of any pollutant at any
concentration is one possible interpretation.
3.2.2.2. Safe Drinking Water Act. The Safe Drinking Water Act (SDWA) directs the
US Environmental Protection Agency's (EPA's) Administrator to develop national primary
drinking water regulations that incorporated maximum contaminant level goals (MCLGs),
maximum contaminant levels (MCLs) and treatment techniques. Under this act, EPA set
maximum contaminant levels for 10 inorganics, 6 pesticides and one set of disinfection by-
products (Le., trihalomethanes). Since 1974, the SDWA has been amended four times. Each
time, new contaminants have been added. In January 1993, the SDWA was amended again
and now includes a total of 117 regulated contaminants, 14 inorganics, 6 microbiological
contaminants and 6 radiological contaminants (USEPA, 1993). A listing of the regulated
contaminants is provided in Tables A., B., C. and D. shown in the Appendix. Included in these
lists are both regulated and non-regulated contaminants. The status of regulated contaminants
are "final" and are identified in the fourth column of each table. Non-regulated contaminants
218
are cited, because either EPA requires more data to determine a contaminant's significance or
to determine any additional limits that may be required. In cases where water systems are
especially vulnerable to a particular contaminant (regulated or non-regulated), a specific
monitoring program may be requested by EPA
In addition to primary standards, EPA also sets secondary drinking water standards
which specifY secondary maximum contaminant levels (SMCLs). These standards are primarily
to protect public welfare and deal primarily with contaminants affecting drinking water odor
and appearance. A list of secondary standards is provided in Table E. shown in the Appendix.
These standards are intended as guidelines and are not federally enforced (Montgomery, 1985).
3.2.3. Selecting Water Treatment Goals
Water quality goals are set by local governments and community water systems. These
goals are typically more stringent than standards and may include contaminants and other
measurements not identified in the regulations. For example, a local government or community
water system may decide to set a turbidity goal of 0.2 nephelometric turbidity units (NTUs)
even though the standard is 0.5 NTUs. Setting treatment goals below standards ensure
satisfactory performance. Even though many decision makers question the value (and
additional costs) to exceed federal standards. The American Water Works Association
(AWWA) is also an example of a professional organization that sets goals for drinking water
treatment. Alternatively, a community water system may elect to provide a water quality
better than required by standards or for contaminants not regulated. Examples include goals
for trihalomethanes (THMs) that are below standards or based on other surrogate parameters
such as total organic carbon (TOe) or ultraviolet light absorbance at a wavelength of 254
nanometers. The decisions on setting goals involve determination of costs, benefits, and
political philosophy of the local government or community water system.
3.3. CHALLENGES FACING DRINKING WATER TREATMENT
Community water system's concerns are derived from impending regulations. The
impetus of the majority of the research is driven by the following regulations:
219
• Surface Water and Ground Water Treatment Rules which will require disinfection
and, in the case of surface waters, filtration of drinking water for communities serving
more than 25 persons,
• Amendments to the Surface Water Treatment Rule which requires EPA to

promulgate MCLs for and additional 25 contaminants every three years,
• DisinfectantIDisinfection By-Product Rule which will regulate disinfection by-

products that have been identified or are being identified as carcinogens, and
• Lead and Copper Rule which limits the levels of lead and copper at the consumers
tap by corrosion within the distribution system.
Table 4. Utility Research Efforts as identified by AWWARF Survey (AWWA, 1992)
Number of CWSs Percent of Total Response (a>

Research Areas Respondin2
Corrosion Control 447 57.7

Water Conservation 445 57.4
Wellhead Protection 298 38.5
Filtration 297 38.3
Watershed Maruutement 264 34.1
Chlorine 263 33.9
Ozone 197 25.4
Permanganate 172 22.2
Chloramines 159 20.5
Granular Activated Carbon 150 19.4
Chlorine Dioxide 145 18.7
Microbiological Treatment 99 12.8
Air Stripping 91 11.7
Powdered Activated Carbon 84 10.8
Reverse Osmosis 43 5.5
Ion Exchange 31 4.0
Ultrafiltration 29 3.7
(a) 775 community water systems responded to this survey
Table 4. summarizes the results from a recent survey conducted by AWWA concerning
ongoing research activities by 775 community water systems (AWWA, 1992). According to
this survey, nearly 60010 of the respondents are involved in controlling corrosion and water
conservation. Other areas of research include disinfection with chlorine and the use of
220
alternative disinfectants (Le., ozone, chloramines, and chlorine dioxide), filtration including
sand, reverse osmosis and ultrafiltration, and organic removal technologies such as granular
activated carbon (GAC), powdered activated carbon (PAC), and air stripping. Microbiological
treatment is also becoming increasingly important regarding biofilms and biofilm regrowth in
distributions systems. Concerns with biofilms include colonization of pathogenic organisms and
increased levels of disinfection by-products.
3.3.1. Waterborne Disease
3.3.3.1. General Consideration Since the germ theory of disease was developed by
Pasteur and Koch in the 1870s (ASCE & AWWA, 1990), the issue of waterborne disease from
contaminated water has been a concern. During the period between 1920 to 1992, 1,683
waterborne outbreaks were reported in the US resulting in 466,486 cases of illness and more
than a 1,000 deaths. Figures 2. and 3. show the number of outbreaks and cases of illness by
etiology during this period. Of these outbreaks, 43% occurred from community water systems
resulting in 83% of the reported illnesses and 71% of the reported deaths (Centers of Disease
Control, 1985, Craun, 1986, St. Louis, 1988, Levine et al, 1990, Herwaldt et al, 1992, and
Moore et al, 1993).
During 1920 through 1936, waterborne outbreaks were reported most frequently for
community water systems. After 1936, outbreaks were generally more frequent for non-
community water systems (i.e., systems which do not serve residents but serve tourists). In
general, the number of outbreaks reported by community water systems have decreased since
the late 1920s through the mid-1950s (Craun, 1986). Since 1965, however, the number of
outbreaks have been increasing steadily from 56 outbreaks during 1961-1965 to 196 outbreaks
during 1976-1980. The number of outbreaks reported during 1979 through 1980 by
community water systems were the greatest since 1942. The decline in the number of reported
water!>orne outbreaks during the 1940s through the 1960s was not considered associated with
any new developments in water treatment or widespread use of a treatment technology (Craun,
1986). Disinfection and filtration which showed dramatic decreases in waterborne outbreaks in
the early 1900s, were developed and applied in the 1940s (Craun, 1986).
221
300 ~~~~-r~~~-+~J-~~-L~~~-r~~Lf--[~JI-'T)Y~P~h~~~d~F';'V;'e~r----ll
i Gas1roenterilis
! Hepatitis A
i,,· ..· ! " ... .',.' ""',!,' ,,_ ~ Shigellosis

Salmonellosis
......
• I Chemical Poisoning
! , Giardiasis
!,,', '.',
~
.D
:;
200
i i i
. i ""j '· . . · i ' "i''''' '.
,
I
: ~
r:n
Ctyptosporidium
g:hirylobac!erosis
o !
150 t..... _.: ......... !..•...•_!....•...!.......
~ 1 i i
.
.D !
!
E 100
::I
Z
50 ,
.. !!!"'. ...!!! .,...!!!

0
'"N g "'
'"
0
~ '" $
VI ..."' ...Ii!. ~ 2 a VI
~ ......'"
!!! !!! !!! !!!
. !!! !!! !!! !!! !!!
01
.;,
... ...!!!
.;, .;, .,.;, ~ .;, ....~ .;, .;.
~
i e
<'>
~ ~ !!! !!! ; '" !!! ! ~ !!!
!!! !!! ~ lR ~
Figure 2. Total number of waterborne outbreaks reported by etiology from community,

non-community and individual systems
Typhoid Fever
100.000 , Gaslroenleritis
i Hepalilis A
i Shigellosis
80000 ·_·..····'f·,······i ... " .......,..

! !... _ .. ,
Salmonellosis
Chemical Poisoning
..... i
Giardiasis
Cryptosporidium
.,.
i c.ilT'9ylobaclerosis
,g 60000 r ....1 T·i . '!'··. ..,,"·, ..".., !·..•·..

Other
'0
Ii
.D
E
40.000 ~.. "+N;·~l~ ~J.
::I
! ''''10 19!8S
z !
20.000 L~~ ..-
0
., ...
'" ..,!!!0 '"...co ;;. ;;. ...~ :ll... ~ :e'" ...'"
* !.,.
0 0 N
N
Iii
~ ~ .;.~ ~
..
01
...... :e.. ;;!!!

~
.:. .;,
N
!!! !!!
..,.;, .;,
'" !!!'" !!! !!!... !!! '"'"
~ :e 01 i ~
:i: ., ~
01
!!!
Figure 3. Total number of waterborne illnesses reported by etiology from community,

non-community and individual systems
222
3.3.1.2. Identification and Surveillance Techniques Giardia lamblia, which was

considered a benign microorganism for many years, has been related to numerous waterborne
disease outbreaks in the mid-1970s. The identification of this microorganism's pathogenic
nature coupled with recent research at Colorado State University which showed that filtration
without proper coagulation, flocculation, and sedimentation is ineffective in removing Giardia
(Hendricks et al., 1988a, 1988b, 1986a and 1986b) has promulgated the Surface Water
Treatment Rule requiring disinfection and filtration for all surface water sources.
More recently, cryptosporidium has been linked with a number of waterborne

outbreaks in the US (AWWA Satellite Teleconference, 1994). In the well-publicized outbreak
in Milwaukee in April, 1993, 370,000 people were affected by cryptosporidium (Fox and
Lytle, 1993). Inadequate dosing of polyaluminum chloride, used as a coagulant, was
determined as the primary cause. Other treatment deficiencies related to waterborne outbreaks
during the period between 1920 to 1980 are listed in Table S. One major cause for outbreaks
during this period was inadequate filtration and other allied treatments.
Table 5. Treatment deficiencies related to waterborne outbreaks and illnesses in community,

non-community and individual water systems in the United States between 1920 to 1980 (Craun, 1986)
Deficiency Number of Outbreaks Number ofllinesses
Inadequate filtration/allied treatment 37 77 062

Inadequate disinfection when only 104 45,823
treatment
Interrupted disinfection when only 102 16,965
treatment
Inadequate disinfection with other 4 1,785
treatment
Interruption of disinfection with other 8 8,715
treatment
Inadequate control of chemical feed 9 700
Contamination of distribution system 187 72,254
No treatment of surface water 152 10,799
No treatment of well water 522 80,145
Source contamination 20 51283
Contaminated storage facilities 30 8,359
The approach used to prevent waterborne outbreaks is the "system of multiple

barriers." As opposed to relying on only one treatment process for removing pathogenic
microorganisms (i.e., disinfection), a series of processes or barriers are considered (i.e.,
223
coagulation-flocculation-sedimentation, filtration, and disinfection). This system of barriers

begins with the protection of surface water supplies from human and animal contact.
The removal of pathogenic microorganisms is usually expressed using a logarithmic
scale. For example, 90% removal equals I-log, 99010 removal equals 2 logs, 99.9 removal
equals 3 logs, and 99.99010 removal equals 4 logs. Under the SDWA, community water
systems are required to achieve 3-10g removal for Giardia cysts and 4 log removal for viruses.
Table 6. shows the removal credits expressed as logs for different filtration technologies.
Conventional filtration (Le., coagulation, flocculation, sedimentation, and filtration) receives
the highest log-credit of 2.5. New technologies such as cartridge filters, bag filters and
membranes may also be used, but need to be demonstrated under normal operating conditions.
Log credits are also given to other processes. These credits are additive, resulting in a
total credit for a series of treatment processes. The log credit equation for filtration and
disinfection is expressed in the following equation:
Removal Inactivation 3-10g (99.9010)
Credit + Credit for > for Giardia
Filtration Disinfection 4-log (99.99%)

for viruses
In other words, the log removal achieved by filtration, plus the log inactivation
achieved by disinfection, must be greater or equal to 3 logs for Giardia and 4 logs for viruses.
Table 6. Removal credits for various filtration technologies (pontius, 1993)
Turbidity Limit(a) Log

Technology (NTU) Removal Credit
Conventional 0.5 2.5 Giardia

treatment(b) 2 Viruses
Direct filtration(c) 0.5 2 Giardia
Slow sand filtration I 2
Giardia
2
Viruses
Diatomaceous earth I 2
Giardia
filtration I
Viruses
Other technologies I Must be demonstrated
(a) Filtered-water turbidity must be less than or equal to the level indicted
in 95 percent of the measurements taken every month.
(b) Coagulation, flocculation, sedimentation, and filtration.
(c) Coagulation, with or without flocculation, and filtration.
224
Cryptosporidium is not currently regulated, but is on EPA's Drinking Water Priority

List for possible regulation (pontius, 1994b). Cryptosporidium oocysts are resistant to
chlorination. Chlorine and monochloramine at typical concentrations have little effect (pontius,
1994b). Chlorine dioxide has been shown somewhat effective (pontius, 1994b). Ozone is the
most effective and can completely inactivate Cryptosporidium oocyst. Filtration and watershed
protection are the only conventional control practices available for most community water
systems. Greater than 97% removal (1.5 log removal) of Cryptosporidium oocysts can be
achieved with proper coagulation and filtration (pontius, 1994b).
3.3.2. Disinfection By-Products
Based on a study published in the American Journal of Public Health [82, 955
(1992)], people who drink chlorinated surface waters are more likely to develop bladder and
rectal cancer than those drinking un-chlorinated water (Morris et a1., 1992).
According to study findings, the risk of getting bladder cancer was 9% and rectal
cancer was 15% greater for people drinking chlorinated water (Hileman, 1992). It is estimated
that of the 47,000 cases of bladder cancer reported each year in the US, 4,200 cases are
associated with drinking chlorinated water (Hileman, 1992; and Water Quality Association,
1992). Of the 44,000 cases of rectal cancer reported each year in the US, 6,500 cases are
associated with drinking chlorinated water (Hileman, 1992; and Water Quality Association,
1992). Other study results have suggested that the cancer risk from drinking chlorinated water
is 3 times greater than drinking un-chlorinated water (Cantor, 1992).
Termed disinfection by-products (DBPs), these contaminants in treated drinking water
has been the main focus of research in the US in the past two decades. DBPs are the result of
chlorinating surface waters containing levels of natural organic matter (e.g., humic and fulvic
acids). Halides from either chlorine or naturally present bromide are substituted onto the
organic molecules.
The chemical structures of trihalomethane and haloacetic acid species are shown in
Figures 4. and 5., respectively. A complete list of identified disinfection by-products is
provided in Table 7. Important to note is that a significant number ofDBPs have not yet been
identified. Identified DBPs represent only 50% of the total organic halide (TOX)
concentrations measured in most chlorinated surface waters (AWWA Satellite Teleconference,
1993).
Factors which influence DBP formation include contact time, chlorine dose,
temperature, pH, the level of natural organic matter in the water and the bromide concentration
(AWWA Satellite Teleconference, 1993).
225
For examp Ie:
X CI CI
I I I
H- C-X H- C-CI H- C- CI
I I I
X CI Br
X= ClorBr Chbroimn Brorrodi:hbrorrethane

(Tri:hklrorrethane)
Figure 4. Chemical structure oftrihalomethane (THM) species

(AWWA Satellite Teleconference, 1993)
For exatll'le:
X CI CI
I I I
X- C- COOH C l- C- COOH Cl-C-COOH
I I I
X H CI
X=H,qorBr Dichbroaceti: Tri:hloroacetic

AcKl (DCAA) AcKl (TCAA)
Figure 5. Chemical structure ofhaloacetic acid species

A balance between the need to disinfect water to prevent waterborne disease and to
minimize the creation ofDBPs that are linked with causing cancer is essential. Figure 6. shows
graphically the risks of waterborne disease and cancer caused by DBPs with increasing
concentrations of chlorine.
Source water and disinfection
1
technologies affect slopes
Cancer
!Tom DBP's
Optimum
---+ operating range
Increasing chlorine concentration •
Figure 6. Risk trade-offs between waterborne disease and cancer caused by disinfection by-products
(adapted from AWWA Satellite Teleconference, 1993)
226
At low concentrations of chlorine, the risk of waterborne disease is high and the risk of
cancer from formed DBPs is low. As chlorine levels are increased, the risk of waterborne
disease is lowered and the risk of cancer from formed DBPs is increased.
Best available technologies for controlling DBPs include: moving the point of
chlorination to the end of the treatment plant, using alternative disinfection (e.g., chloramines,
ozone), or removing the natural organic matter or precursors to DBPs prior to chlorination
(e.g., enhanced coagulation, granular activated carbon, or nanofiltration). Moving points of
pre-chlorination to later stages in treatment allows for removal of DBP precursors by
conventional processes. Historically, water treatment plants have pre-chlorinated prior to the
treatment plant and prior to filtration as a means of preventing biological growth. Moving
these points of chlorination allows for the coagulation/sedimentation and filtration processes to
reduce DBP precursors prior to chlorination. Research has also shown that the addition of
ammonia to chlorinated water stops DBPs from continuing to form in the distribution system.
Adding ammonia prior to the distribution system stops the formation of DBPs (AWWA
Satellite Teleconference, 1993).
Table 7. Disinfection by-products (DBPs) formed by chlorination

I Disinfection By-Products
I Trihalomethanes (THMs) Haloacetic Acids (HAAs)
-.;MMMI ·fi~·Nb&'S&8~A.cid
Bromodichloromethane Dichloroacetic Acid
Dibromochloromethane Trichloroacetic Acid
Bromoform Monobromoacetic Acid
Dibromoacetic Acid
Cyano2en Halides Bromochloroacetic Acid
Cyanogen Chloride
Cyanogen Bromide Haloacetonitriles (HANs)
Dichloroacetonitrile
Halopicrins Trichloroacetonitrile
Chloropicrin Dibromoacetonitrile
Bromopicrin Tribromoacetonitrile
Bromochloroacetonitrile
Chlorol Hydrate
Haloketones, Haloaldehydes, Halophenols
MX l3-chloro-4-(dichloromethvl)-S-hvdroxy-2(5H)-furanone)
227
While substituting chlorination with ozone eliminates the DBPs listed in Table 7., other
DBPs with unknown health effects are formed. Ozone also does not leave a residual needed to
disinfect the distribution system. A secondary disinfectant which is capable of leaving a residual
in the distribution system is still required. Enhancing or increasing coagulation concentrations
has shown to increase DBP precursor removals effectively reducing the formation of DBPs.
Granular activated carbon and nanofiltration (a membrane technology) have also been shown
to reduce DBP precursors.
3.3.3. Corrosion Control
A total of 11 chemical poisonings have occurred from corrosive waters leaching copper
and lead from plumbing materials during the period between 1971 to 1983 (Craun, 1986). Six
acute copper poisonings with measured copper levels of 4, 12.5 38.5 and 80 mg/1 occurred
during 1971 through 1980. Five copper and lead poisonings occurred during 1981 to 1983.
The concern of copper and lead leaching from piping and their acute and chronic health
effects has promulgated the Lead and Copper Rule. This rule sets what are known as "action
levels" for lead and copper of 15 Ilg/l and 1,300 Ilg/l, respectively, measured in water
discharging from residential taps.
Treatment approaches for reducing copper and lead levels has included: pH adjustment
to achieve a more positive Langlier Index (Le., promote calcium carbonate saturation and
deposition), feeding various phosphate-based corrosion inhibitors (e.g., zinc orthophosphate),
and replacing lead service lines. Results of corrosion control studies have been positive
showing compliance with the action levels (Ramaley, 1993). Another benefit of controlling
corrosion has been the reduction of metal levels in biosolids from associated wastewater
treatment plants.
3.3.4. Biofilms and Biological Regrowth
Biofilms and biological regrowth within the pipes is a recent and new concern in water
treatment. Biofilms refer to microbial cells that have attached to pipe surfaces or on particles
(suspended in the water or stationary within a filter). Biofilms may consist of numerous
species, sometimes in relatively high numbers. Residual chlorine concentrations in distribution
systems have an important impact on the diversity of these biofilms (AWWA, 1988).
Investigations of distribution systems have shown that biofilms occur and may be
responsible for the release of bacteria into treated drinking water. Biofilms have been linked
with coliform outbreaks in New Haven, Connecticut (AWWA, 1988).
228
Conditions conducive to biofilm growth include (AWWA, 1988):
High pH (PH range 6.8 to 7.5) in raw, treated efiluent and distribution water
simultaneously, coupled with low concentrations offree chlorine,
High turbidity in raw, treated eftluent, and distribution water simultaneously,

coupled with low concentrations offree chlorine, and
High coliform counts in the treated eftluent, coupled with low concentrations of
free chlorine.
The problems of biofilms has sparked research into biological treatment at drinking
water plants. Historically, biological treatment was prevented using pre-chlorination in front of
treatment plants and ahead of filters. With the problems of disinfection by-products, pre-
chlorination has been almost eliminated. Now, biological treatment within filters is being
considered as a method for reducing organics in water, resulting in lowering disinfection by-
products and in removing microbial nutrients needed for biofilm growth. Biologically active
tilters have resulted in less biological activity in distribution systems (AWWA, 1988).
Assimilable organic carbon (AOC) is a new analytical measurement used to detect the level of
organics in water that would contribute to biofilm growth. This measurement is currently being
explored in research and by water utilities (Reid, 1991).
3.4. ESTABLISHED AND EMERGING TECHNOLOGIES IN THE UNITED STATES
3.4.1. Overview of Available Technologies
Table 8. provides a listing of the available treatment technologies used in water

treatment. The technologies are organized according to type of treatment needed to remove
key contaminants. Summary data is provided for each technology and alternatives are
compared on the basis of performance, suitability to treatment plant size, degree of acceptance,
conditions required for effective treatment, operational and maintenance requirements, and
compatibility with other treatment processes.
229
Table 8. (PI of2) Experimental, emerging and established water treatment technologies
in the United States (EPA, 1989)
Technologies Stage or Size Comments

Acceptability Suitability
FILTRATION - MICROBIAL
CONTAMINANTS
Conventional Established All Most common: adapted for
adding other processes
Direct Established All Lower cost alternative to
conventional filtration
Slow sand Established Mostly small Operationally simple: low cost,
but requires large land areas
Package plant Established Mostly small Compact, variety of process
combinations available
Diatomaceous earth Established Mostly small Limited applicability; potentially
expensive for small systems
Membrane Emerging Mostly small Experimental, expensive
Cartridge Emerging Small Experimental; expensive
DISINFECTION -
MICROBIAL
CONTAMINANTS
Chlorine Established All Most widely used method;
concerns about health effects of
by-products
Chlorine dioxide Established All Relatively new to the US;
concerns about inorganic by-
products
Monochloramine Established All Secondary disinfectant only;
some by-product concerns
Ozone Established All Very effective; requires a
secondary disinfectant
Ultraviolet radiation Established All Simple, no established harmful
by-products; requires a
secondary disinfectant
Advanced oxidation Emerging All Limited information concerning
(ozone plus hydrogen disinfection aspects of this
peroxide plus ultraviolet process
radiation)
230
Table 8. (P2of2) Experimental, emerging and established water treatment technologies

in the United States (EPA, 1989)
ORGANIC CONTAMINANTS
Granular activated carbon Best available All Highly effective; potential
technology waste disposal problem
Packed column aeration Best available All Highly effective for volatile
technology compounds; potential air
emissions issues
Powdered activated carbon Established Large Requires conventional
treatment processes
Diffused aeration Established All Variable removal effective
Multiple tray aeration Established All Variable removal
effectiveness
Oxidation Experimental All By-product concerns
Reverse osmosis Emerging Medium Variable removal
effectiveness; expensive
Mechanical aeration Experimental All Mostly for wastewater
treatment; high energy
requirements, easy to operate
Catenary grid Experimental All Performance data are scarce;
potential air emissions issues
Higee aeration Experimental Small Compact, high energy
Resins Experimental Small

requirements; potential air
emissions problems
Data scarce
I
Ultrafiltration Emerging Small Primarily for turbidity; data
for organics removal is
limited
INORGANIC CONTAMINANTS
Reverse osmosis Established Medium Highly effective; expensive;
potential waste disposal
problems
Ion exchange Established Medium Highly effective; expensive;
problems
Activated alumina Established Small Highly effective; expensive;
I problem
Granular activated carbon Experimental Small Experimental for

radionuclide removal;
I problem
CORROSION CONTROLS
pH control Established All Potential conflict with other
treatments
Corrosion inhibitors Established All Variable effectiveness
depending on type of
inhibitor
231
3.4.2. Established Technologies
Table 9. lists the most commonly used treatment technologies for treating water
sources (FRDS, 1994). These technologies were identified in the Federal Reporting Data
System (FRDS). Information in the FRDS is supplied by State Agencies in compliance with the
quarterly reporting requirements under the Safe Drinking Water Act. The accuracy of this
information is dependent upon the quality of information provided by State Agencies. The
entire list of treatment technologies used and their significance are listed in Table F. in the
Appendix.
Table 9 . Common used treatment technologies used for treating drinking water (FRDS, 1994)
Water Treatment Practices
Adsorption Disinfection
Ion ExchanJ1;e Gaseous chlorination, Post
Gaseous chlorination, Pre
Aeration Hypochlorination, Post
SlateTrav
Filtration
Conventional processes Rapid Sand
Rapid mix Pressure Sand
Coagulation
Flocculation Fluoridation
Sedimentation
Lime-soda ash softeninE
Corrosion Control
Polyphosphate inhibator
3.4.3. Emerging Technologies
Although many of the so-called emerging technologies listed in Table 8. have been used
either in research or demonstration plants, their use by community water systems has been
limited. For a technology to emerge into practice lies with the effect of new regulatory
requirements. The direction of water treatment research is often driven by regulations. The
more pertinent technologies having a strong potential of emerging in the next century include:
232
Alternative Disinfection Membrane Filtration

Ozone disinfection Ultrafiltration
Ultraviolet disinfection Nanofiltration
Biological Treatment Granular Activated Carbon

Biofilters Computer Control Systems
Granular activated carbon
3.4.3.1. Alternative Disinfection. Alternative disinfection technologies such as ozone

and ultraviolet light will emerge as a result of regulations targeting hannful disinfection by-
products and pathogenic organisms such as cryptosporidium which is more resistant to
chlorination.
3.4.3.2. Biological Treatment. Allowing microbial activity in conventional filters has

shown to reduce microbial nutrients in treated water resulting in lowering the potential for
biofilm growth within distribution systems (AWWA Research Foundation, 1988). With
concerns about increased levels of coliforms due to biofilms, biological treatment in
conventional filters and on granular activated carbon may emerge as promising technologies.
Further research is needed.
3.4.3.3. Membrane Filtration. Membrane filtration is gaining wider spread acceptance

due its ability to remove a variety of contaminants including organic and inorganic
contaminants and pathogenic organisms. Ultrafiltration has shown to be effective in
disinfection and removal of large weight organics. In conjunction with either coagulation or
granular activated carbon, ultrafiltration has been effective in reducing natural organics that are
precursors to disinfection by-products. Nanofiltration, introduced in the early 1980s, is very
effective in reducing disinfection by-product precursors without any other treatment.
3.4.3.4. Granular Activated Carbon. Granular activated carbon will emerge as a

technology used to reduce levels of disinfection by-product precursors as well as regulated
chlorinated hydrocarbons.
3.4.3.5. Computer Control Systems. Computerization continues to be one of the fastest

areas for emerging technologies. Computer control systems allow for real time monitoring of
233
water characteristics and process performance. Sensors on influent waters to treatment plants
can detect quickly abnormal water characteristics which may lead to process upsets or reduced
treatment efficiency. Plant operations such as filter backwashing can be automated by
computer controls.
3.5. TECHNOLOGY COST COMPARISONS
The capital costs, annual operational and maintenance (0 & M) costs, and the total
costs per 10 cubic meters of treated water (capital costs have been amortized over a 20 year
period at a 10% fixed interest) for different treatment technologies are listed in Table 10.
(USEPA, 1990). These costs are based on a small treatment plant size of 400 m3 per day
(100,000 gallons per day).
As indicated by Table 10., each treatment technology has different mix of capital and
operating and maintenance costs. Technologies with high capital costs often have lower
operating and maintenance costs (e.g., slow sand filters); those with lower capital costs often
have higher operational and maintenance costs (e.g., hypochlorite solution - disinfection).
Thus, communities which can not afford the initial capital costs usually become burdened with
higher operational and maintenance costs later.
Based on the costs listed in Table 10., reverse osmosis and covered slow sand filters
are the most expensive technologies with costs of $6.40 and $6.60 per 10 m3 of water treated,
respectively. Reverse osmosis capital and operational and maintenance costs are expensive
relative to other technologies. Reverse osmosis equipment which includes high pressure
pumps, pressure vessels and computer automation is expensive. Operational and maintenance
costs include membrane replacement and also high energy requirements. Costs for covering
slow sand filters is a major capital cost. Land requirements are normally large resulting in large
covers. In comparison, the capital costs for uncovered slow sand filters is nearly half that for
covered filters (Table 10.).
Of the disinfection technologies listed, ozone is the most expensive. Capital costs which
include the cost of an ozone generator are significantly higher than the other disinfection
technologies. Operational and maintenance costs, however, are less for ozone compared with
hypochlorite solution disinfection.
234
Table 10. Estimated costs of drinking water technologies for a 400 m3/day (100,000 gallday) plant
(USEPA, 1990)
Treatment Technologies Capital Cost (a,b) AnnualO&M Total Cost

$1994 (a,b) per
$1994 10m3 (c)
FILTRATION
I
I
Slow sand filters; covered $740,000 $9,800 $6.60
I Slow sand filters· uncovered $427000 $9100 $4.10
I Diatomaceous earth vacuum filters
I I
$131,000 $14,200 $2.00
Diatomaceous earth pressure filters $135,000 $13,500 $2.00
PACKAGE PLANT FILTERS
CoagulationIFiltration with tube $225,000 $14,000 $2.80
settlers
Pressure depth clarifierslPressure $263,000 $13,300 $3.00
filter
Pressure depth clarifierlPressure $314,000 $20,800 $4.00
filter with GAC adsorber
DISINFECTION I
Gas feed chlorination $13,400 $4,500 $0.40 I
Hypochlorite solution $5,200 I $71100 I $0.50
I Pellet feed chlorinators $2,100 $5100 $0.40
I Ultraviolet light $33,200 $2,700 $0.50 I
Ozonation-high pressure $50,100 $6500 $0.80
ORGANIC CONTAMINANTS
Granular activated carbon; pressure $223,000 $18,400 $3.10
vessel (6 mo carbon)
Packed tower aerator $58,000 $3,700 $0.70
INORGANIC CONTAMINANTS
c----lli~ressure rSll:erse osmosi~ I ~J~1 DOLl I Sj,?,J,Q£} ! ViJIl
LaW J2ressyr!il r!ill:ll!l!il osmosl~
Cation exchanae I $193000 I $lOQ~ I C?3Q
Anion exchan~e I ~I~Z QQQ I S,1J, !.DO ! S,?.J./J
Activated alumina I $133,000 I $18700 I $2.40
(a) ConstrUction costs generally include manufactured equipment, concrete, steel, labor, pipes
and valves, electrical equipment and instrumentation, housing, site evacuation, some other
site work, general contractor's overhead and profit, engineering costs, financial and
administrative costs, and interest costs during construction. Construction costs do not
include land costs, legal fees, interface piping, roads, and certain other site work. 0 & M
costs generally include annual energy, labor, and chemical costs. Construction costs can vary
depending on specific data characteristics. 0 & M costs can vary, up to plus or minus 100
percent for some technologies, depending on such variables as feed water characteristics,
flow rate, and chemical dosage requirements.
(b) Capital costs and 0 & M costs have been inflated at an annual percentage rate of 5% from
originally reported 1989 cost figures reported in EPA, 1990 (FIP, 5%, 5 years = 1.276).
(c) Costs include capital costs annualized at 10 percent interest over 20 years (AlP, 10%, 20
years = 0.1175) plus annual 0 & M costs.
235
Granular activated carbon is more expensive than packed tower aeration due to
significant capital costs and the expense of regenerating carbon. Packed tower aeration,
however, is only effective for volatile organic compounds. Non-volatile organics would require
granular activated carbon or some other technology.
For removing inorganic contaminants, ion exchange resins (i.e., cation and anion) are
the least expensive technologies both with respect to capital and operational and maintenance
costs. Reverse osmosis (high and low pressures) have significantly higher capital and
operational and maintenance costs.
3.6. RESEARCH AT COLORADO STATE UNIVERSITY
3.6.1. Pilot-Plant Studies
A Water Treatment Research Facility (WTRF) was constructed at Colorado State

University in November, 1992. This facility includes two pilot-plants: (1) an advance water
treatment pilot-plant and (2) a membrane filtration pilot-plant. The advanced water treatment
pilot-plant was funded by the US Army Corps of Engineers Construction Engineering
Research Laboratory in Champaign, llIinois (Hendricks et a1., 1993).
Processes include low and high G rapid mix, flocculation and coagulation, plate
settling, multi-media filtration, ozone, granular activated carbon, ion exchange and air
stripping. Rated at a flow of75 liters per minute (20 gaVmin), this pilot-plant is at a scale large
enough to simulate full-scale operation. The membrane filtration pilot-plant was funded by the
USEPA Risk Reduction Engineering Laboratory (RREL) in Cincinnati, Ohio (Champlin and
Hendricks, 1994). This pilot-plant consists of six (6) 4"x 40" pressure vessels, each provided
with its own high pressure multi-stage centrifugal pump.
The WTRF occupies an area of about 10 m x 20 m (30 ft X 60 ft) at the south end of
the Hydraulics Laboratory in the Engineering Research Center, located at the Foothills Campus
of Colorado State University in Fort Collins, Colorado. Figure 7. shows the layout of the
different processes and supporting equipment. Tables 11. and 12. list some of the features of
each pilot plant.
The purpose of the pilot plant is to permit testing of alternative treatment trains for a
given set of contaminants and water quality. It has also been used to evaluate individual
processes. Specific waters can be synthesized by adding chemical spikes, pathogenic organisms
or particles. The pilot plant is fitted with sensors and sampling ports to allow determination of
236
various water contaminant profiles as they change through treatment. Thus, the effects of
different variables on processes can be assessed. The selection of an appropriate treatment train
for a given contaminated water results in effective treatment at lower cost. At the same time,
optimization of a unit process design can result in significant cost savings. Being able to
conduct research adds to the knowledge base on the behavior of the respective process which
translates into lower cost designs.
The pilot plant is computer controlled, in that valves can automatically be opened and
closed along with the collection of sensor data. Flow, temperature, pressure, conductivity,
turbidity and particle sensors send data directly to a computer. The computer is programmed
to detect problems in the operation of the pilot plant.
For example, flow sensors on influent and effluent lines are used to detect leaks.
Differences in flows in and out of the pilot plant greater than 10% indicate a leak to the
computer, shutting the pilot plant down. Sensor data also provides a means to trouble shoot
and to explain unexpected events.
... 0006
O~o r~ n
XAD-S Column
NORTH
~ y Operating Platform
~----------------------~
~ ISouth I Computer
Control
Operating ~~
Platform
~~
Membrane
Pilot-Plant
Spike and Floccuant Tanks
Figure 7. Layout of water treatment research facility (WTRF) at Colorado State University
(Hendricks et aI., 1993)
237
Table 11. Available unit processes at the water treatment research facility (Hendricks et aI., 1993)
-------------_ ...
Unit Process Objective
Advanced Water Treatment Pilot-Plant
Low G rapid mix Disperses coagulant
High G rapid mix Disperses coagulant more effectively
Coagulation Charge neutralization of colloidal particles
Flocculation Promote collisions between particles increasing
•particle size
~d«lI;.~·~Aj:1~ll6;.-S",·A· fermit settliM of settlable particles
Dual-media filtration Removal of smaIl charged particles
Ozone (generation, gas transfer, reaction) Oxidize organic molecules and disinfection
Granular activated adsorption Adsorb organic molecules
Clino-thiolite ion exchange Adsorbs cations, particularly metal ions
Air stripping Transfer of volatile organic compounds to the gas
I phase
Membrane Filtration Pilot-Plant
Six Membrane Pressure Vessels Removal of particles, molecules, ions in accordance
with pore size of membranes tested
Table 12. Appurtenant equipment to water treatment research pilot plant (Hendricks et aI., 1993)
Appurtenant Equipment
Components
Head boxes Motors
Pumps Stirrers
Metering pumps Blowers
Chemical storage tanks Computers
Air compressor Storage tanks
Control System
Sensors; flow, water levels, pressure, temperature, conductivity, pH, turbidity, particle counts
Automated on-offvaIves
Animated analogs
Plot generation
Sampling
Tap.s at each node point between
processes
Taps along column towers
Laboratory Support
Gas chromatograph Zeta meter
Total organic carbon Particle counter
UVanalyzer Spectrophotometer
Fluorometer Turbidimeter
Safety Plan
Containment area to capture spills Approvals for discharge to city sewer
Carbon vent traps Secondary containment
Power back-up for computers Personnel health and safety training
Computer shut-down from sensor data Computer alarms
238
3.6.2. Survey of Treatment Technologies
Treatment of contaminated waters is a major issue in the United States and around the
world. Problems vary due to differences between municipal and industrial needs, the varied
types of contaminated sources, the wide range of contaminants, and the increasingly more
stringent regulations. Treatment trains, comprised of an array of unit processes, are selected
typically for the type of contaminants present and the degree of treatment required. With the
wide range of contaminants involved and the need for economical treatment, the selection of
unit processes for treatment trains is uncertain.
As one part of a two year project with the US Army Corps of Engineers Research
Laboratory, Colorado State University was involved in research to delineate performance of
varying treatment technologies for a variety of water contaminants. One delineation has been
the development of a technology/performance matrix, intended to assist engineers in the
preliminary selection of treatment unit processes (Champlin and Hendricks, 1993).
Based upon reviews of data bases and journal articles, a technology/performance
matrix which indicates expected removals for nearly 650 different chemicals and water
contaminants versus 50 different treatment technologies was developed (Champlin and
Hendricks, 1993). Listed contaminants were defined by federal laws and regulations.
The matrix shows clearly the limited availability of treatability information. For the
46,000 cells that comprise the matrix, about 80% show the symbol "NA" (not available) with
respect to treatability information (Champlin and Hendricks, 1993). Thus the matrix shows that
for most of the contaminants listed, there is a dearth of knowledge on treatment efficiency,
design criteria, operational guidelines.
The matrix shows also the reliance that has been placed primarily on the activated
sludge and activated carbon technologies. About 25% of the contaminants listed in the matrix
are effectively treated (50% removal or better) using activated sludge (Champlin and
Hendricks, 1993). By ~he same token, about 10% of the contaminants listed in the matrix are
effectively treated using activated carbon technologies (i.e., GAC and PAC). A smaller
percentage of contaminants are treated effectively by the other listed technologies.
3.7. REFERENCES
ASCE and AWWA (1990) Water Treatment Plant Design. Second Edition, McGraw-HilI
Publishing Co. New York, New York. ISBN 0-07-001542-2.
239
AWWA Research Foundation (1992) WIDB Water Industry Data Base: Utility Profiles.
American Water Works Association, Denver, Colorado. ISBN 0-89867-643-6.
AWWA Research Foundation (1988) Bacterial Regrowth in Distribution Systems, Research
Report. Subject Area: Distribution Systems. American Water Works Association,
Denver, Colorado. ISBN 0-89867-418-2.
AWWA Satellite Teleconference (1994) Preventing Waterborne Disease: Is Your System at
Risk? Participant Guide. Broadcast April 8, 1994. American Water Works
Association, Denver, Colorado. ISBN 0-89867-742-4.
AWWA Satellite Teleconference (1993) Disinfectants and Disinfection By-Products:
Understanding the Proposed DIDBP Rule. Participant Guide. Broadcast
October 15,1993. American Water Works Association, Denver, Colorado. ISBN 0-
89867-708-4.
Cantor K.P. (1992) Cancer Risks Associated with Chlorination By-Products. WaterReview
Technical Report, A Publication of the Water Quality Research Council. Vol. 10, No.
2,1992.
Centers for Disease Control (1985) Water-Related Disease Outbreaks Surveillance. Annual
Summary 1984, US Department of Health and Human Services, Public Health Service.
Issued November, 1985.
Champlin and Hendricks (1994) Nanofiltration For Treatment of Low-Turbidity - Waters
Membrane Fouling and Removals of Disinfection By-Product Precursors. Report
prepared for USEPA Risk Reduction Engineering Laboratory, Cincinnati, Ohio.
Prepared by Colorado State University, Department of Civil Engineering, Fort Collins,
Colorado.
Champlin and Hendricks (1993) Treatment Train Modeling for Aqueous Contaminants.
Volume 2 of 5, Matrix of Contaminants and Treatment Technologies, Report prepared
for the US Army Construction Engineering Research Laboratory, Champaign, llIinois.
Prepared by Colorado State University, Department of Civil Engineering, Fort Collins,
Colorado.
Craun G.F. (1986) Waterborne Diseases in the United States. CRC Press, Inc., Boca Raton,
Florida.
Culp G.L., Wesner G.M., Culp R.L. (1986) Handbook of Public Water Systems. Van
Nostrand Reinhold Co. Inc. New York, New York.
FRDS (1994) Federal Reporting Data System. Information supplied by Primary Agents in
Compliance With Quarterly Reporting Equirments Under the Safe Drinking Water Act.
Search of Reporting Treatment Plant Processes, July 15, 1994.
Fox K.R., Lytle D.A. (1993) The Milwaukee Cryptosporidiosis Outbreak: Investigation and
Recommendations. Proceedings to the AWWA Water Quality Technology Conference,
Part II - Session 3D through ST6. November 7-11, 1993, Miami, Florida.
Hendricks D.W. et al. (1993) Treatment Train Modeling for Aqueous Contaminants. Volume 1
of 5, Summary Report, Report prepared for the US Army Construction Engineering
Research Laboratory, Champaign, llIinois. Prepared by Colorado State University,
Department of Civil Engineering, Fort Collins, Colorado.
Hendricks D.W. et al. (1986a) Removing Giardia Cysts from Low Turbidity Waters by Rapid
Rate Filtration. Journal of AWWA., May, 1986.
Hendricks D.W. et al. (1986b) Rapid Rate Filtration of Low Turbidity Water Using Field-Scale
Pilot Filters. Journal of AWWA., December, 1986.
240
Hendricks D.W. et al. (1988a) Removing Giardia Cysts and Other Particles from Low
Turbidity Waters Using Dual-Stage Filtration. Journal of AWWA Vol. 80, No.2,
February, 1988.
Hendricks D.W. et al. (1988b) Bench-Scale Evaluations of Coagulants for Low Turbidity
Water. Journal of AWWA, April, 1988.
Herwaldt B.L., Craun G.F., Stokes S.L., Juranek D.D. (1992) Waterborne Disease Outbreaks.
1989-1990, Centers for Disease Control Morbidity and Mortality Weekly Report,
1992, Vol. 40, No. SS-3, pl-21.
Hileman B. (1992) Cancer Risk Found From Water Chlorination. Chemical & Engineering
News. July 13, 1992.
Lay T. (1993) SDWA Reauthorization - Good Time to Plan for Future. AWWA Opflow, Vol.
19, No.1, ISSN 0149-8029, January, 1993.
Levine W.C., Craun G.F. (1990) Waterborne Disease Outbreaks. 1986-1988, Centers for
Disease Control Morbidity and Mortality Weekly Report. March, 1990, Vol. 39, No.
SS-I, pl-13.
Montgomery 1.M. (1985) Water Treatment - Principals & Design. 1985. John Wiley & Sons,
Inc. New York, New York.
Moore AC., Herwaldt B.L., Craun G.F., Calderon R.L., Highsmith AK., Juranek D.D. (1993)
Surveillance for Waterborne Disease Outbreaks - United States. 1991-1992, Centers
for Disease Control Morbidity and Mortality Weekly Report. November 19, 1993,
Vol. 42, No. SS-5, pl-22.
Morris R.D., Audet A.M., Angelillo I.F., Chalmers T.C., Mosteller F. (1992) Chlorination,
Chlorination By-products and Cancer: A Meta-Analysis. American Journal of Public
Health. July 1992, Vol. 82, No.7, p. 955-963.
Pontius F.W. (1994a) SDWA Advisor, Regulatory Update Service. American Water Works
Association Publication. Denver, Colorado. ISBN 0-89867-736-X.
Pontius F.W. (1994b) Part II - What You Should Know About Cryptosporidium. AWWA
Opflow, Vol. 21, No.7, ISSN 0149-8029, June, 1994.
Pontius F.W. (1993) Surface Water Treatment Rule Targets Microbials. AWWA Opflow, Vol.
19, No.5, ISSN 0149-8029, May, 1993.
Pontius F.W. (1992) Plan, Prepare Now for Impact of Phase V Regs. AWWA Opflow, Vol.
18, No. 11, ISSN 0149-8029, November, 1992.
Ramaley B.L. (1993) Monitoring and Control Experience Under the Lead and Copper Rule.
Journal of AWWA. Vol. 85, No.2, February, 1993.
Reid 1. (1991) Emerging Technologies - Some Are Years Away, But Others are Coming to a
Plant Near You. AWWA MainStream, July 1991.
St. Louis M.E. (1988) Water-Related Disease Outbreaks, 1985. Centers for Disease Control
Morbidity and Mortality Weekly Report, CDC Surveillance Summaries. June, 1988,
Vol. 37, No. SS-2, pI5-24.
US Environmental Protection Agency (1978) Estimating Costs for Water Treatment as a
Function of Size and Treatment Plant Efficiency. Municipal Environmental Research
Laboratory, Cincinnati, Ohio. EPA-600/2-78-182.
US Environmental Protection Agency (1989) Technologies for Upgrading Existing or
Designing New Drinking Water Treatment Facilities. Office of Drinking Water,
Cincinnati, Ohio. EPAl625/4-89/023.
US Environmental Protection Agency (1990) Environmental Pollution Control Alternatives:
Drinking Water Treatment for Small Communities. Center for Environmental Research
Information, Cincinnati, Ohio. EPAl62515-90/025.
241
US Environmental Protection Agency (1992) Drinking Water Seminar Handbook. US

Environmental Protection Agency, Region vm, Denver Colorado and Department of
Civil Engineering, University of Wyoming, Laramie, Wyoming.
US Environmental Protection Agency (1993) Drinking Water Regulations and Health
Advisors. Office of Water, Washington D. C. EPA 8221F-93-011.
Vesilind P.A., Peirce lJ. (1982) Environmental Engineering. Butterworth Publishers, An Ann
Arbor Science Book. Woburn, Massachusetts.
Water Quality Association (1992) Findings Link Chlorination with Bladder and Rectal Cancer.
WaterReview Technical Report, A Publication of the Water Quality Research Council.
Vol. 7, No.2, 1992.
3.S. APPENDIX OF DATA TABLES
TableA. United States drinking water standards, health effects and best available
technologies (BAT) for treatment of organic contaminants.
TableB. United States drinking water standards, health effects and best available
technologies (BAT) for treatment of inorganic contaminants.
TableC. United States drinking water standards and health effects for treatment of
radionuclides.
TableD. United States drinking water standards, health effects and best available
technologies (BAT) for treatment of microorganisms and surrogate
contaminants.
Table E. United States drinking secondary maximum contaminant level (SMCL) water
standards (USEPA, 1993).
Table F. Surface water treatment practices in the United States (FRDS 1994).
Table G. Ground water treatment practices in the United States (AWWA Research
Foundation, 1992).
242
Table A. (PI of6) United States drinking water standards, health effects and best available
technologies (BAT) for treatment of organic contaminants
Contaminant Effect I 1I I
Health I Advisories
Refer.
Dose (s)
Status
Standards (b)
MCLG MCL
(mglkgld) .. (mgll) (mgll)

ORGANICS
Acenaphthene 0.06
Acifluorfen 0.013 F zero
Acrylamide 0.001 F zero IT PA
Acrylonitrile D zero
Adipate (diethybexyl) Carcinogen, liver and 0.6 0.4 0.4
reproductive system
effects
Alachlor Carcinogen
0.01 F zero 0.002 GAC
Aldicarb 0.001 D 0.007 0.007 GAC
Aldicarb sulfone 0.001 D 0.007 0.007 GAC
Aldicarb sulfoxide 0.001 D 0.007 0.007 GAC
Aldrin 0.00003 D
Am~nm~__________-+______________+-~0~.0~0~9__~~F~~__-__-r___ -__~____~
Ammonium sulfamate 0.28 F
Anthracene (PAR) 0.3
Atrazine Effects offspring 0.035 F 0.003 0.003 GAC
Bayg~o=n____________~____________~~~0.~00~47-~-=F__r-77-~+-__-__-r____-4
Bentazon 0.0025 F 0.02
Bemz(a) anthracene (PAR) zero 0.0001
Benzene Carcinogen F zero 0.005 GAC,
PTA
Benzo(a) I?vrene (PAR) Carcinogen zero 0.0002 GAC
Benzo(b) fluoranthene zero 0.0002
(PAR)
Benzo(g,h,i) perylene
Benzo(k) fiuoranthene zero 0.0002
bis-2-Chloroisop",r..;;.opy~ll...:;et.;;;;;h",e;;..r+-______________t-_o:..:.o.:,.4.:...--+---=FO---+-___
- __+-___
- __+-____~
Bromacil 0.13 F
Bromobenzene - D
Bromochloroacetonitrile - D
Bromochloromethane 0.013 F
Broinodichloromethane 0.02 D zero 0.1
(THM)
I Bromoform(THI'vl) 0.02 D zero 0.1
Bromomethane
~~)benzyl phthalate 1 0. I I 2 zero I
~~~~~..",.;O":.OO:;..:-:I'__t-_F=---+__-_ _t____::~-:___t_----
0.1 I
::i~~:::;~ec-
Butylbenzene tert-
~_~~~_~_~~_·~·_tI~-=-0-=-.0-=-5-=-~I~ ~DF= -~tf- - - -=-~t- - - -=- =-tl- - - - -
..
~----i'f----:--::--t-.-::---1.f-----t.----t._ __
Carbaryl __ --'1_ 0.1 I F I I I
Footnotes are listed at the end of the table
243
Table A. (P2 of 6) United States drinking water standards, health effects and best available technologies
(BAT) for treatment of organic contaminants
Health I Advisories Standards (b)

Contaminant Effect Refer. Status MCLG MCL BAT (e)
Dose (a)
(m2!k£ld) (m2fl) (m2fl)
Carbofuran Nervous and 0.005 F 0.04 0.04 GAC
reproductive system
Carbon tetrachloride Carcinogen 0.0007 F zero 0.005 GAC,
PTA
Carboxin 0.1 F - -
Chloral hydrate 0.0002 D 0.06 -
Chloramben 0.015 F - -
Chlordane Carcinol!;en 0.00006 F zero 0.002 GAC
Chlorodibromomethane 0.02 D zero 0.1
I (TUM)
Chloroethene - D - -
Chloroform (TUM) 0.01 D zero 0.1
Chloromethane 0.004 F - -
Chlorophenol(2~ 0.005 D - -
p-Chlorophenyl methyl
sulfide!
- - - -
sulfone!sulfoxide
Chloropicrin - - - -
Chlorothalonil 0.015 F - -
Chlorotoluene 0- 0.02 F - -
Chlorotoluene D- 0.02 F - -
Chlorovrifos 0.003 F - -
Chrvsene (PAR) - - zero 0.0002
Cyanazine 0.002 D 0.001 -
Cyanol!;en chloride - - - -
C¥mene-p - D - -
2,4-D Nervous system, liver 0.01 F 0.07 0.07 GAC
andkidnev
DCPA (Qacthall 0.5 F - -
Dalapon Kidney and liver 0.026 F 0.2 0.2 GAC
effects
Di[2-ethylhexyl] adipate Carcinogen, liver
and reprod. syst. ef.
0.6 - 0.4 0.4 GAC,
PTA
Diazinon 0.00009 F - -
Dibenz(a,h) anthracene
,(PAR)
- - zero 0.0003
Dibromoacetonitrile 0.02 D - -
Dibromochloropropane
I (DBCP)
Carcinogen - F zero 0.0002 GAC,
PTA
Dibromomethane - - - -
Dibutyl phthalate 0.01 - - -
Dicamba 0.03 F - -
Dichloroacetaldehyde - D - -
Dichloroacetic acid 0.04 D zero -
Dichloroacetonitrile 0.008 D - -
Footnotes are Itsted at the end of the table
244
Table A. (p3 of 6) United States drinking water standards, health effects and best available technologies
(BAT) for treatment of organic contaminants

Dose (8)
(mI!lk2ld) (m2ll) (m2ll)
Dichlorobenzene 0- Nervous system, 0.09 F 0.6 0.6 GAC,
lung, liver and PTA
kidney
Dichorobenzene m- 0.09 F 0.6 0.6
Dichlorobenzene p- Liver and kidney 0.1 F 0.075 0.075 GAC,
PTA
Dichlrodifluoromethane 0.2 F - -
Dichloroethane (1,1-) - D - -
Dichloroethane (1,2-) Carcinogen - F zero 0.005 GAC,
PTA
Dichloroethylene (1,1-) Liver and kidney 0.009 F 0.007 0.007 GAC,
PTA
Dichlroethylene (cis-I,2-) Nervous system, O.oI FF 0.07 0.07 GAC,
liver, circulatory PTA
system and
carcinogen
Dichloroethylene (trans-I,2-) Nervous system, 0.02 F 0.1 0.1 GAC,
liver and circulatory PTA
Dichloromethane Carcinogen 0.06 F zero 0.005 PTA
Dichlorophenol (2 4-) 0.003 D - -
Dichloropropane (1,1-) - D - -
Dichloropropane (1,2-) Carcinogen, liver,
lungs and kidneys
- F zero 0.005 GAC,
PTA
Dichloropropane 0,3-) - D - -
Dichloropropane (2,2-) - D - -
Dichloropropene 0,1-) - D - -
Dichloropropene 0,3-) 0.0003 F zero -
Dieldrin 0.00005 F - -
Diethyl phthalate (pAE) 0.8 D - -
Diethylene glycol dinitrate - - - -
DiethylhexYl phthalate (pAE) 0.02 D zero 0.006
Diisopropyl
methylphosphonate
0.08 F - -
Dimethrin 0.3 F - -
Diiiiethyl meiliYlphosphonate 0.2 F - -
Dimethyl phthalate (pAE) - - - -
13-Dinitrobenzene 0.0001 F - -
Dinitrotoluene (2,4-) 0.002 F - -
Dinitrotoluene (2,6-) 0.001 F - -
tg 2,6 & 2,4 dinitrotoluene - - - -
Dinoseb Tbyroidand 0.001 F 0.007 0.007 GAC
reproductive organ
effects
Dioxanep: - F - -
Diphenarnid 0.03 F - -
Diphenylamine 0.03 F - -
Footnotes are lIsted at the end oCthe table
245
Table A. (P4 of 6) United States drinking water standards, health effects and best available
technologies (BAT) for treatment of organic contaminants

Contaminant Effect Refer. Status MCLG MCL BATCc)
Dose Ca)
(mglkgld) (mgIJ) (mgIJ)
Diquat Liver, kidney, GI
tract effects;
0.0022 - 0.02 0.02 GAC
cataract formation
Disulfoton 0.00004 F - -
Dithiane (l 4-) O.oI F - -
Diuron 0.002 F - -
Endothall Liver, kidney, GI 0.02 F 0.1 0.1 GAC
tract, reprod.
system effects
Endrin Liver, kidney and 0.0003 F 0.002 0.002 GAC
heart effects
Epichlorohydrin 0.002 F zero IT
Ethylbenzene Kidney, liver, 0.1 F 0.7 0.7 GAC,
nervous system PTA
Ethylene dibromide (EDB) Carcinogen - F zero 5E-5 GAC,
PTA
Ethylene gl~col 2 F - -
ETU 0.00008 F - -
Fenamiphos 0.00025 F - -
Fluometron 0.013 F - -
Fluorene (PAH) 0.04 - - -
Fluorotrichloromethane 0.3 F - -
Fog Oil - D - -
Fonofos 0.002 F - -
Formaldehyde 0.15 D - -
Gasoline, unleaded (benzene) - D - -
Glyphosate Liver and kidney 0.1 F 0.7 0.7 OX
effects
Heptachlor Carcinogen 0.0005 F zero 0.0004 GAC
Heptachlor epoxide Carcinogen IE-05 F zero 0.0002 GAC
Hexachlorobenzene Carcinogen 0.0008 F zero 0.001 GAC
Hexachlorobutadiene 0.002 F 0.001 -
Hexacb!orocyclopentadiene Carcinogen 0.007 - 0.05 0.05 GAC,
PTA
Hexachloroethane 0.001 F - -
Hexane (n-) - F - -
Hexazinone 0.033 F - -
HMX 0.05 F - -
Indeno(l 23,-c,d) JlYIlle (PAH) - D zero 0.0004
- -
--
Isophorone 0.2 F
IsoPfOllVl methvlDhoSDhonate 0.1 D - -
Isopropylbenzene - D - -
Lindane 0.0003 F 0.0002 0.0002 GAC
Malathion 0.02
Footnotes are listed at the end of the table.
246
Table A. (p5 of 6) United States drinking water standards, health effects and best available
tecbnologies (BAT) for treatment of organic contaminants

Contaminant Effed Refer. Status MCLG MCL BAT (c)
Dose (a)
(ma/k£/d) (mRII) (mRII)
Maleic hydrazide 0.5 F - -
MCPA 0.0015 F - -
Methomyl 0.025 F - -
Methoxychlor Liver, kidney, 0.005 F 0.04 0.04 GAC
nervous and reprod.
systems
Methyl ethyl ketone - F - -
Methyl J)3f3thon 0.00025 F - -
Methyl tert butyl ether 0.005 D - -
Metolachlor 0.15 F - -
Metribuzin 0.025 F - -
Monochloroacetic acid - D - -
Monochlobenzene Kidney, liver, 0.02 F 0.1 0.1 GAC,
nervous system PTA
Naphthalene 0.004 F - -
Nitrocellulose (non-toxic) - F - -
Nitroguanidine 0.1 F - -
Nitrophenol p- 0.008 F - -
Oxamyl (Vydate) Kidney 0.025 F 0.2 0.2 GAC
Paraquat 0.0045 F - -
Pentachloroethane - D - -
Pentachlorphenol 0.03 F zero 0.001 GAC
Phenanthrene (pAH) - - - -
Phenol 0.6 D - -
Picloram Kidney and liver 0.07 F 0.5 0.5 GAC
Polychlorinated biphenyls
. (PCBs)
Carcinogen - P zero 0.0005 GAC
Prometon 0.015 F - -
Pronamide 0.075 F - -
Pr0J)3chlor 0.013 F - -
Prooazine 0.02 F - -
Propham 0.02 F - -
Propyibenzene n- - D - -
I Pyrene (pAH) 0.03 - - -
RDX
Simazine Carcinogen,
0.003
0.005
F
F
-
0.004
-
0.004 GAC
circulatory system
Styrene Liver and nervous 0.2 F 0.1 0.1 GAC,
system PTA
24 5,-TP (Silvex) 0.01 F - - GAC
2 3,7,8-TCDD (Dioxin) Carcino.e;en lE'()9 F zero 3E-8 GAC
Tebuthiuron 0.07 F - -
Terbacil 0.013 F - -
Terbufos 0.00013 F - -
Tetrachloroethane (1,1,1,2-) 0.03 F - -
Footnotes are hsted at the end of the table.
247
Table A. (p6 of6) United States drinking water standards, health effects and best available technologies
(BAn for treatment of organic contaminants

Contaminant Effect Refer. Status MCLG MCL BATCc)
Dose Ca)
(m2fk2ld) (m2lI) (m2lI)
Tetrachloroethane (1 1,22-) - D - -
Tetrachloroethylene Carcinogen 0.01 F zero 0.005 GAC,
PTA
Tetranitromethane - - - -
Toluene Kidney, nervous 0.2 F 1 1 GAC
system and lung
Toxaphene Carcinogen 0.1 F zero 0.003 GAC,
PTA
2,4,5-TP Nervous system, 0.0075 F 0.05 0.05
liver and kidney
1, 1, 2-Trichloro-l, 2, 2-
trifluoroethane
- - - -
Trichloroacetic acid 0.1 D 0.1 -
Trichloroacetonitrile - D - -
Trichlorobenzene (1,2,4-) Kidney and liver 0.01 F 0.07 0.07 GAC,
effects PTA
Trichlorobenzene (1,35-) 0.006 F -
Trichloroethane (1,1,1-) Liver, nervous 0.035 F 0.2 0.2 GAC,
system, circ. svstem PTA
Trichloroethane (1,1,2-) Carcinogen, kidney 0.004 F 0.003 0.005 GAC,
and liver effects PTA
Trichloroethanol (2,2 2-) - - - -
Trichloroethylene Carcinogen - F zero 0.005 GAC,
PTA
Trichlorophenol (2,4 6-) - D - -
Trichloropropane (1,1 1-) - D - -
Trichloropropane (1,2,3-) 0.006 F - -
Trifluralin 0.0075 F - -
Trimethylbenzene (1,2,4-) - D - -
Trimethylbenzene (1,3,5-) - D - -
TrinitroJ!lycerol - F - -
Trinitrotoluene 0.0005 F - -
Vinvl cItloride CarcinoJ!en - F zero 0.002 PTA
Xylenes Liver, kidney, 2 F 10 10 GAC,
nervous system PTA
(a) Reference dose IS an estimate of a daily exposure level to the hwnan population that IS likely to be Without
appreciable risk of deleterious effects over a lifetime (EPA, 1993). Units are mg contaminant per kg mass
ofhwnan per day.
(b) Status refers to the cummt state of regulations pertaining to the particular contaminant (F refers to final, D
refers to draft and P refers to proposed). MCLG refers to maximwn contaminant level goal. A non-
enforceable concentration that is protective of adverse hwnan health effects (EPA, 1993). MCL refers to
the maximwn contaminant level. This concentration is the maximwn pennissible level which is delivered
to any public water system. IT refers to the treatment technique required, no MCL are stated (EPA,1992).
(c) BAT refers to the best available technology for treatment (EPA, 1990; EPA, 1992; and Opflow, Nov.
1992). Abbreviations: M-Activated A1wnina; AER-Aeration; CC-Corrosion Control; CIF
Coagulation/flocculation; DDF Direct and Diatomite Filtration; DIS - Disinfection; ED Electrodialysis; FIL
- Filtration; GAC - Granular Activated Carbon; IE - Ion Exchange; LS - Lime Softening; PA - Polymer
Addition; PAC - Powdered Activated Carbon; PTA - Packed Tower Aeration; and RO - Reverse Osmosis.
248
Table B. (PI of2) United States drinking water standards, health effects and best available
technologies (BAT) for treatment of inorganic contaminants

Dose (a)
(m2lkeld) (melI) (mefl)
INORGANICS
Aluminum - D - -
Ammonia - D - -
Antimony Decreases 0.0004 F 0.006 0.006 CIF,RO,
growth and GAC
lon2evity
Arsenic - D - 0.05
Asbestos (fibersIL
> 10JUIIlengili)
- - 7MFL 7MFL CIF,DD,CC
Barium 0.07 F 2 2 IE,LS, RO,EI

Beryllium Carcinogen, 0.005 D 0.004 0.004 AA,IE,RO,
damage to LS, CIF
bones and
lungs
Boron 0.09 D - -
Cadmium 0.0005 F 0.005 0.005 CIF, IE,LS,
RO
Chloramine 0.1 D 4 -
Chlorate - D - -
Chlorine 0.08 D 4 -
Chlorine dioxide 0.003 D 0.08 -
Chlorite - D - -
Chromium (Total) 0.005 F 0.1 0.1 CIF, IE, LS,
RO
Copper - - 1.3 17
Cyanide Spleen, brain, 0.022 F 0.2 0.2 IE,RO,CH
and liver
effects
Fluoride 0.12 - 4 4
Hypochlorite - - 4 -
Hypochlorus acid - - 4 -
Lead (at tap) - - zero 17
Manganese 0.005 D - -
~ry (inorganic) 0.0003 F 0.002 0.002 CIF,GAC,
LS RO,PAC
Molvbedenum 0.005 D - -
Nickel Heart and liver 0.02 F 0.1 0.1 IE,RO,LS
effects
Nitrate (as N) 1.6 F 10 10 IE RO,ED
Nitrite (as N) 0.16 F 1 1 IE,RO
Nitrate + Nitrite (both
asN)
- F 10 10
Selenium 0.005 - 0.05 0.5 AA, CIF, LS,

RO,ED
Footnotes are hsted at the end of the table.
249
Table B. (P2 of2) United States drinking water standards, health effects and best available technologies
(BAT) for treatment of inorganic contaminants

Contaminant Effect Refer. Status MCLG MCL BAT (c)
Dose (a)
(m2lk2ld) (m2fl) (m2ll)
Silver 0.005 0 - -
Sodium - 0 - -
Strontium 0.6 0 - -
Sulfate
Thallium
Gastronteritis
Kidney, liver
-
0.00007 F
- -
0.0005
-
0.002 AA,m
and brain
effects
Vanadium - 0 - -
White phosphorus 0.00002 F - -
Zinc 0.3 F - -
Zinc chloride (as Zn) 0.3 F - -
(a) Reference dose IS an estimate of a daily exposure level to the human population that IS likely to be Without
of human per day.
(b) Status refers to the current state of regulations pertaining to the particular contaminant (F refers to fmal, D
refers to draft and P refers to proposed). MCLG refers to maximmn contaminant level goal. A non-
enforceable concentration that is protective of adverse human health effects (EPA, 1993). MCL refers to
the maximmn contaminant level. This concentration is the maximmn pennissible level which is delivered
1992). Abbreviations: AA - Activated Almnina; AER - Aeration; CC - Corrosion Control; CIF
Coagulation/flocculation; ODF Direct and Diatomite Filtration; DIS - Disinfection; ED Electrodialysis; Fll,
Table C. United States drinking water standards and health effects for treatment of radionuclides

Contaminant Effect Refer. Status MCLG MCL
Dose (a)
(mglkwd) (mgll) (mgll)
RADIONUCLIDES
Beta particle and photon acitivity - - zero 4mrem
Gross alpha particle activity - - zero 15 p(iIL
Radium 226 - - zero 20pCi/L
Radium 228 - - zero 20pCi/L
Radon - - zero 300pCi/L
Uranium 0.003 - zero 2OIlg/L
(a) Reference dose IS an estimate of a daily exposure level to the human population that IS likely to be Without
of human per day.
the maximmn contaminant level. This concentration is the maximmn pennissible level which is delivered
to any public water system. IT refers to the treatment technique required, no MCL are stated (EPA, 1992).
250
Table D. United States drinking water standards, health effects and best available technologies (BAT)
for treatment of microorganisms and surrogate contaminants

Contaminant Effect Refer. Dose (a) Status MCLG MCL BAT (c)
(m2fkeld) (meIL) (meIL)
Crptosporidium L - -
Giardia larnblia F zero IT FIL, DIS
Legionella F zero IT FIL, DIS
Standard Plate Count F NA TT FIL, DIS
Total Coliforrns F zero - FIL, DIS
Turbidity F NA PS FIL
Viruses F zero TT FIL, DIS
(a) Reference dose IS an estimate of a dally exposure level to the hmnan populatIOn that IS likely to be Without
of human per day.
(b) Status refers to the current state of regulations pertaining to the particular contaminant (F refers to fma1, D
enforceable concentration that is protective of adverse hmnan health effects (EPA, 1993). MCL refers to
the maximmn contaminant level. This concentration is the maximmn permissible level which is delivered
1992). Abbreviations: AA - Activated Alwnina; AER - Aeration; CC - Corrosion Control; CIF
CoagulatiOn/flocculation; DDF Direct and Diatomite Filtration; DIS - Disinfection; ED Electrodialysis; FIL
Table E. United States drinking secondary maximum contaminant level (SMCL) water standards
(USEPA, 1993)
Contaminant Status (b) Standard(a) (SMCL)

Aluminum F 0.05 to 0.2
Chloride F 250
Color F 15 C.u.
Copper F 1.0
Corrosivity F non-corrosive
Fluoride F 2.0
Foaming Agents F 0.5
Iron F 0.3
Manganese F 0.05
Odor F 3 TON
IpH F 6.5 to 8.5
Silver F 0.1
Sulfate F 250
Total Dissolved Solids F 500
Zinc F 5
(a) Secondary standards provide guidance relating to aesthetic quality. These standards are only guidelines
and are not enforced.
the maximmn contaminant level.
251
Table F. (PI of2) Surface water treatment practices in the United States (FRDS 1994)
Water Treatment Practices Number of Plants (a) Percent of Total Plants

Treatment Facilities
Public water systems 52,567 100.00
Treatment applied by seller 1,101 2.09
Treatment applied by plant 47,990 91.29
Treatment applied at entry point 55 0.10
No treatment/not applicable 9,004 17.13
Treatment not reported 413 0.79
Other treatment 84 0.16
Adsorption
Activated alumina 19 0.04
Granular activated carbon 687 1.31
Powdered activated carbon 412 0.78
Ion Exchange 3,008 5.72
Aeration
Cascade 445 0.85
Diffilsed 125 0.24
Packed Tower 236 0.45
Slate Tray 1,666 3.17
Spray 134 0.25
Algae Control 219 0.42
Conventional Processes
Rapid mix 817 1.55
Coagulation 2,615 4.97
Flocculation 1,075 2.05
Sedimentation 1,774 3.37
Corrosion Control
Biometallic phosphate inhibitor 120 0.23
Hexametaphosphate inhibitor 130 0.25
Orthophosphate inhibitor 161 0.31
Polyphosphate inhibitor 838 1.59
Silicate inhibitor 20 0.04
Disinfection
Ammoniation 106 0.20
Chloramines 271 0.52
Chlorine Dioxide 83 0.16
Gaseous chlorination, Post 6,647 12.64
Gaseous chlorination, Pre 2,145 4.08
Hypochlorination, Post 11,758 22.37
Ozonation, Post 18 0.03
Ozonation, Pre 47 0.09
Permanganate 780 1.48
Peroxide 7 0.01
Ultraviolet Radiation 659 1.25
Footnotes at end of table
252
Table F. (P2 of2) Surface water treatment practices in the United States (FRDS 1994)
Water Treatment Practices Number of Plants (8) Percent of Total Plants

Distillation 5 0.01
Filtration
Unfiltered, avoiding filtration 45 0.09
Unfiltered, must install filtration 661 1.26
Not subject to SWTR 5,668 10.78
Cartridge 316 0.60
Diatomaceous Earth 67 0.13
Greensand 484 0.92
Microscreening 49 0.09
Pressure Sand 1,073 2.04
Rapid Sand 2,385 4.54
Slow Sand 263 0.50
Ultrafiltration 67 0.13
Reverse osmosis 256 0.49
Electrodialysis 2 0.00
Fluoridation 2,955 5.62
Iron Removal 203 0.39
pH adjustment 2,141 4.07
pH adjustment, post 597 1.14
pH adjustment, pre 638 1.21
Reducing Agents
Sodium bisulfate 3 O.oI
Sodium sulfite 2 0.00
Sulfur dioxide 10 0.02
Sequestration 915 1.74
Softening 186 0.35
Lime - soda ash 1,669 3.17
Taste and od~r 136 0.26
<a) Infmmation provided by data search of Federal Reporting Data System (FRDS). FDRS
contains information supplied by Primary Agents in compliance with the quarterly
reporting requirments Wider the Safe Drinking Water Act. The accuarcy of all the
infmmation in this database is dependant on the quality of the infmmation provided by
the Primary Agents.
253
Table G. Ground water treatment practices in the United States

(AWWA Research Foundation, 1992)
Ground Water Number of Percent of

Treatment Pradiees Plants <a> Plants Surveyed
Aeration 155 26.3
Predisinfection/Oxidation
Chlorine 437 74.1
Chlorine dioxide 3 0.5
Chloramines 31 5.3
Ozone 1 0.2
Permanganate 31 5.3
Reverse osmosis 8 1.4
Ion exchange 21 3.6
Lime/Soda Ash Softening 92 15.6
Recarbonation with CO2 33 5.6
Iron & Manganese Removal 137 23.2
Filtration 176 29.8
Organics Removal
PAC addition 2 0.3
GAC adsorption 22 3.7
Air Stripping 35 5.9
Off gas treatment 2 0.3
Resin Adsorption 1 0.2
Radionuclide Removal
Radon 2 0.3
RadiumlUranium 0 0.0
Fluoridation 256 43.4
Det1uoridation 0 0.0
Corrosion Control
pH adjustment 146 24.7
Alkalinity adjustment 33 5.6
Corrosion inhibators 124 21.0
Post-Disinfection
Chlorine 182 30.8
Chlorine dioxide 1 0.2
Chloramines 26 4.4
Any Disinfection (pre and/or Post) 513 86.9
No Treatment 53 9.0
...
<a> 590 utilities responded to this survey
Chapter IV
Emerging Problems and Technologies

in Water Treatment
Chapter IV
Emerging Problems and Technologies in Water

Treatment
INTRODUCTORY STATEMENT
Almost from the introduction of public water supply, the emphasis on water treatment
shifted from point of use to centralised treatment. Due to the development of more stringent
water - quality standards, potable water is now being treated in centralised facilities to very
high standards at substantially increased costs. In approaching minimum risk from waterborne
disease, environmental engineers have begun to focus more on subtle risks from cancer. As we
exhaust more resources in the pursuit of a longer list of chemical contaminants regulated at
lower concentrations, we must not loose sight of fundamental requirements for public health. A
long standing philosophy of producing water that "tastes good" has resulted in design practices
that frequently emphasise the use of granular activated carbon and ozone. At the same time,
water disinfection practices have also evolved. First, a concern for halogenated by-products of
disinfection resulted in reduction of chlorine doses, substitution by alternative disinfectants,
and better application of chlorinating technology. More recently, it has been found that
treatment - resistant pathogens (viruses and pathogenic protozoa, especially Giardia and
Cryptosporidum) may limit the degree to which reduction in chlorine dosage may be used to
control disinfection by product formation without posing unacceptably high risk of infectious
disease.
258
Changes in drinking water regulations are causing a stronger coupling between the
level of drinking water treatment and the quality of the water supply. Lower quality supplies
may require such capital intensive processes as filtration, high - level disinfection (such as
ozonation) and perhaps membrane treatment. In some cases, this provides strong incentive for
starting and / or strengthening watershed protection programs. In other cases, it may provide
an incentive to locate and reduce major sources of contaminant loading that contribute to a
water supply used for drinking water. This may be particularly important in the case of
microbial loading.
Development of new technologies and approaches to water treatment can not entirely
eliminate the need to further improve currently available treatment options. As a matter of fact,
enhanced coagulation is designated as a possible alternative to meet more stringent disinfection
by-product regulations. In addition optimisation of conventional settling and filtration
processes would certainly have major benefits for water utilities.
Intermediate transfer of risks in water treatment must also be avoided, since we are
merely sweeping risk from potable water to the air we breathe, and to our landfills. Risk and
benefit - cost analysis must be extended to consider life - cycle impacts on the water treatment
industry. These actions are needed not only to improve drinking water quality but to provide
information needed by decision makers in balancing environmental quality with other valuable
objectives of our society (schools, hospitals, etc.). We must also carefully examine our human
health ethic to be better prepared for water recycling that is already being implemented in many
arid areas.
1. Control of Disinfection By-Products in Drinking
Water: Regulations and Costs
Radisav D. Vidic
Department of Civil and Environmental Engineering
University of Pittsburgh
949 Benedum Hall, Pittsburgh PA 15261
USA
1.1. HISTORY OF DBP FORMATION AND REGULATIONS
The regulation of disinfection by-products (DBPs) in the United States was based on
the discovery of the presence of chloroform as a by-product of chlorination in drinking water
in the early 1970s. This suspected human carcinogen was indirectly regulated as one of the
four trihalomethanes (THMs) under the interim total THM (TTHM) regulation established by
the US Environmental Protection Agency (USEPA) in 1979. An interim maximum
contaminant level (MCL) of 0.10 mg/I was established for water utility systems serving more
than 10,000 people, based on the running annual average offour quarterly samples.
It is significant that the standard was established based on then-available health effects
data on chloroform (CHCI3)' The National Academy of Sciences and the USEP A's Carcinogen
Assessment Group estimated the incremental risks associate with exposure to chloroform in
drinking water. Assuming that the population consumes 2 I of water (containing 0.10 mgll of
chloroform) daily, for 70 years, they estimated a risk of 3-4 excess cases of cancer per 10,000
people. Adoption of 0.10 mgll MCL for TTHMs was understood to be a compromise position;
public health considerations had to be balanced with technological and economic
considerations. In the development of regulations in general, two categories of risk targets are
established. The first is the Maximum Contaminant Level Goal (MCLG). The second is the
Maximum Contaminant Level (MCL). The MCLG is a health-based target, usually established
at a 10-6 risk level. The MCL is an enforceable standard and might be as high as 10-4 because
of limitations in technology for removing or measuring the contaminant. The intent in
establishing the MCL for TTHMs, however, was to revise the MCL in future regulations
"based upon an updated assessment of technological and economic feasibility, implementation
of experience and additional toxicological information". Few data existed at the time on the
NATO ASI Series, Partnership Sub-Series, 2.Environment - Vol. 15

Edited by C. Maksimovic. F. Calomino. and J. Snoxell
260
other three THMs: bromodichloromethane (CHCI2Br), dibromochloromethane (CHCffir2),

and bromoform (CHBr3), but their structural similarities to chloroform led the USEPA to
conclude to be inappropriate to exclude them from regulation.
Besides THMs, di- and trichloroacetic acids were identified as a second major class of
DBPs in chlorinated water (Quimby et a1., 1980, Miller and Uden, 1983). Other frequently
identified halogenated DBPs include haloacetonitriles, haloketones, chloropicrin, cyanogen
chloride, and chloral hydrate. A comprehensive study by the National Academy of Sciences in
1987 (National Academy of Sciences, 1987) found that a number of these other halogenated
DBPs also have adverse effects. It is believed that dichloroacetic acid is a more potent
carcinogen than any of the THMs (Bull and Kopfler, 1991) while a halogenated furanone (3-
chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone, also known as MX) has been found to
be extremely mutagenic, even at concentrations as low as 0.05 J.1g/1.
In the mid-1980s, Congress became disenchanted with the pace at which the USEPA
was setting drinking water regulations. Spurred by the perception of a substantial public health
risk being posed by the US drinking water, Congress passed a Safe Drinking Water Act
(SDWA) amendments in 1986. Congress prescribed, with limited flexibility, the first 83
compounds to be regulated and the fu~re rate at which the USEPA would establish standards
(25 new standards every three years).
The most common treatment modification implemented by the water supply industry to
comply with the new MCL for TTHMs include moving the point of chlorination downstream
in the treatment process, decreasing chlorine doses, and utilizing chloramines as alternative
primary or secondary disinfectant instead of free chlorine (McGuire and Meadow, 1988).
These modifications generally lead to the compliance with the new rule but the questions were
raised about the microbiological quality of the finished water. To ensure that the microbial
quality of finished drinking water was not compromised, the USEPA promulgated the Surface
Water Treatment Rule (SWTR) and the Total Coliform Rule (TCR) in 1989 (USEPA,
1989a.,b.). The SWTR states that all public water systems using any surface water or ground
water under direct influence of surface water must disinfect and may be required to implement
filtration as well. It further requires that the treatment must reliably achieve at least 99.9
percent (3-log) removal and/or inactivation of Giardia Lamblia cyst and 99.99 percent (4-log)
removal and/or inactivation of viruses. The key concept in the SWTR is the requirement that
the water must be in contact with a disinfectant concentration for a sufficient period of time in
order to provide specified degrees of inactivation (C·t concept). Furthermore, a measurable
disinfectant residual must be achieved in at least 95% of distribution system samples analyzed.
The regulation of DBPs has been complicated by the discoveries that alternative
disinfectants to free chlorine also produce by-products that have adverse health effects.
Ozonation of bromide-containing waters will produce bromate which has been identified as a
261
probable human carcinogen while the addition of chlorine dioxide to water results in the
formation of chlorite which causes hepatotoxicity in animals and possible hemolytic anemia
(Haag and Hoigne, 1983, Couri et aI., 1982).
Balancing between the risk associated with DBPs and the risk associated with microbial
disease is complicated by the following factors: lack of occurrence data for various DBPs;
insufficient information concerning the nature of unidentified (50%) TOX; lack of information
regarding the by-products formed by using the alternative disinfectants to free chlorine
(chloramines, ozone, chlorine dioxide, etc.); uncertainty regarding the health effects of various
DBPs; and the uncertainty regarding the C·t values required for the inactivation of Giardia and
Cryptosporidium cysts (Singer, 1994).
Because of the lack of sufficient technical data on DBPs, the USEPA could not, within
the prescribed time frame, establish defensible standards for all of the DBPs included in the
initial set of contaminants. They did issue a "strawman" regulation in 1989 for public comment
(pontius, 1990). This rule was to serve as basis for discussion and to seek input from the
affected parties early in the rule-making process. Because of the technical complexity and
uncertainty that exists in various aspects of this rule, the agency instead chose to develop the
proposed rule using the negotiated rule-making process - referred to as regulatory negotiation,
or "Reg-Neg". In this approach, all interested parties that have a stake in the regulation,
including the USEPA, meet as a group and, with the assistance of a third party that acts as a
facilitator, collectively develop the regulation. This Reg-Neg process was initiated in
September 1992 and was completed in mid-1993. The current status of negotiated D-DBP is
presented at the end of this manuscript.
1.2. OCCURRENCE OF DISINFECTION BY-PRODUCTS
THMs and other halogenated DBPs result from the interaction of the disinfectants with
organic precursor (in most cases natural organic material (NOM» as shown by:
rPrecursor1 [DiSinfection 1
lMaten·al J+ [Disinfectant] ~ By - Pro dncts J
262
The major halogenated DBPs identified in finished waters are:
- Trihalomethanes (THMs) - Haloacetonitriles (HANs)

Chloroform Dichloroacetonitrile
Bromodichloromethane Trichloroacetonitrile
Dibromochloromethane Dibromoacetonitrile
Bromoform Tribromoacetonitrile
Bromochloroacetonitrile
- Haloacetic Acids (HAAs)

Monochloroacetic acid - Cyanogen Halides
Dichloroacetic acid Cyanogen chloride
Trichloroacetic acid Cyanogen bromide
Monobromoacetic acid
Dibromoacetic acid - Haloketones
Tribromoacetic acid 1,l-dichloropropanone
Bromochloroacetic acid 1, 1, I-trichloropropanone
Bromodichloroacetic acid 1,I-dichloro-2-butanone
Dibromochloroacetic acid 3,3-dichloro-2-butanone
1, 1, I-trichloro-butanone
- Aldehydes - Halopicrins
formaledehyde Chloropicrin
acetaldehyde Bromopicrin
dichloroacetaldehyde
trichloroacetaldehyde -MX
A recent survey of DBPs in finished drinking water in the United States, which
included 35 water-treatment facilities utilizing a broad range of source water qualities and
treatment processes, revealed that THMs were the predominant class of DBPs present, with
the HAAs being the second most dominant class (Krasner et al., 1989). The median TTHM
concentration was 39 J.l.g/l, which compared very well with the values reported in previous
surveys (AWWARF survey of 727 utilities in 1988). The median HAA concentration was 19
mg/l, approximately half of the TTHM concentration. Formaldehyde and acetaldehyde, by-
products of ozonation, can also be produced by chlorination. Cyanogen chloride was produced
preferentially in chloraminated waters.
263
1.3. FACTORS INFLUENCING THE FORMATION OF nBPs
The fonnation of DBPs occurs primarily through the substitution reactions of chlorine
with the potential reactive sites in the organic molecule. The factors influencing these reactions
include: pH, contact time, temperature and season, nature and concentration of organic
precursors, disinfectant type and dose, and bromide concentration.
pH. Historical data continned that the increase in pH leads to higher concentration of
THMs in finished water (Oliver, 1978, Stevens et aI., 1989). On the other hand, higher pH
leads to overall reduction in TOX formation since many of the halogenated DBPs other than
THMs tend to hydrolyze at alkaline pH-values (Johnson and Jensen, 1986, Stevens et al.,
1989). The most obvious implication for water treatment is the direct trade-off between
TTHM control (lower pH) and the control of most of the other DBPs (higher pH).
Contact time. Concentrations of THMs and HAAs in chlorinated generally increase

with increasing contact times (Stevens et al., 1989) while some of the halogenated DBPs
(haloacetonitriles and haloketones) fonn rapidly upon chlorination but then decay in the
distribution system due to hydrolysis and continuing reaction with residual chlorine. This
behavior has major implications on the exposure of consumers to DBPs depending upon their
proximity to the treatment plant or the point of rechlorination.
Temperature. Fonnation ofDBPs is highly influenced by the seasonal changes in raw

water quality. Higher temperatures in summer months lead to faster kinetics and therefore,
increased chlorine demand which causes more chlorine to be applied and higher DBP
concentrations are observed. Such behavior is very well documented for THM (White, 1986).
It is reasonable to assume .that most of the other DBPs will exhibit similar behavior. On the
other hand, hydrolysis rates for chloral hydrate, dihaloacetonitriles, and possibly chloropicrin
are also likely to increase, possibly reducing the presence of these species. Another aspect of
seasonal changes are the changes in the nature of organic precursors and the concentrations of
bromide.
Chlorine dose. As the chlorine dose increases, TOX increase in proportion to the
available organic precursor. Higher chlorine doses favor the fonnation of HAAs over THMs
and the fonnation of trihalogenated HAAs over di- and monohalogenated HAAs
(Pounnoghaddas et al., 1993).
264
Bromide concentration. In the absence of bromide ion, only chlorinated by-products

are formed. In the presence of bromide, free chlorine (HOCl) rapidly oxidizes bromide to
hypobromous acid (HOBr), which then also reacts with organic precursors to produce mixed
chi oro-bromo substitution products. The extent of bromine incorporation depends on the ratio
of bromide ion concentration to the applied free chlorine. As that ratio increases, the extent of
the formation of brominated DBPs increases (pourmoghaddas et al., 1993). This is true for
both THMs and HAAs.
A number of these factors have been incorporated into DBP predictive equations that
are being used by the USEPA and by the Reg-Neg group to asses impact of various candidate
MCLs on the water supply industry and the extent to which utilities treating waters of different
quality would be able to comply with these MCLs (Malcolm Pirnie, 1992).
1.4. CONTROL OF DISINFECTION BY-PRODUCTS
There are several options for DBP control. As discussed earlier, DBPs tend to increase
with time so that moving point-of-disinfection to the end of the treatment process will
minimize the time for by-product formation. This approach however, does not have any impact
on the formation ofDBPs in the distribution system which is of increasing concern in the US.
Other options for the control ofDBPs include: source water quality control; removal of
precursor material or NOM before it interacts with the disinfectant; removal of the DBPs once
they are found; and use of a disinfectant that minimizes the formation of by-products. The
following section provides some insight into the treatment options, effects of process variables,
and the treatment technology costs associated with some of these options for controlling
DBPs. Since the impact of control of THMs is similar to the control of other halogenated by-
products, the costs developed will be those associated with THM control.
1.4.1. Source Water Quality Control
Many studies have documented that the concentrations of DBPs tend to directly
correlate with the increases in algal metabolites present in the source water. Consequently, one
of the approaches in preserving the quality of source waters with respect to algal blooms is to
implement an adequate nutrient management program that may include: storm-water detention
265
basins to trap nutrients, land-use controls (limit developments on watersheds used for water
supply), installation of aeration systems and implementation of harvesting programs to control
nutrient cycling in reservoirs.
Since the presence of bromide ions was shown to have a pronounced impact on the
formation DBPs (drives the formation of more brominated species that are believed to be more
harmful to human health), it is also advisable to implement programs that would limit the
presence of bromide ions in the source water. Such programs could include the control of salt
water or brine intrusion into the water source by installing structural or hydrodynamic barriers.
Recent studies showed that the water management concept known as aquifer storage
and recovery (ASR) has a positive impact on halogenated DBPs in finished water (Singer et
aI., 1993). This concept involves underground storage of treated drinking water by injection
into a suitable storage zone during those periods of the year when available capacity of
treatment facilities exceed system demand. The stored water is later recovered to meet system
demands that exceed treatment capacity, usually without the need for additional treatment
other than disinfection. The same well is used for both recharge and recovery. Such practice
was shown to result in the elimination of THMs and HAAs produced during treatment and in
decrease in the concentration of residual DBP precursors.
1.4.2. Precursor Removal
There are three ways to effectively remove DBP precursors: enhanced coagulation,
activated carbon adsorption, and membrane processes.
Enhanced coagUlation. Both aluminum and ferric salts have effectively been used to
coagulate NOM (Kavanaugh, 1978). The optimal pH range for alum (Al2S04·xH20)
coagulation is between 5.5 and 6.0. This pH range is easily achieved in water oflow alkalinity
(less than 50 mg/l as CaC03) by the addition of alum. However, in waters of very low or very
high (above 100 mg/l as CaC03) alkalinity, it might be necessary to add an appropriate amount
of a base (lime or caustic soda) or an acid (sulfuric acid), respectively, to maintain the pH in a
desired range. Alum and ferric chloride dosages depend quite heavily on the TOC content and
alkalinity of raw water, the hydrophobic/hydrophilic distribution of the TOC (hydrophobic
fraction of TOC is believed to constitute 30-70% of TOC and is more susceptible to
coagulation), and the pH of coagulation an have to be optimized for each raw water source
thereby, precluding general conclusions with respect to design and cost parameters.
266
GAC Adsorption. Most applications of GAC for the control of precursor (NOM)
would require a separate postfiltration adsorber bed to accomplish appropriate removals.
Empty-bed contact times (EBCT) in excess of 15 minutes are required with regeneration
frequency on the order of three to six months.
USEPA Drinking Water Research Division developed cost estimates for the utilization of GAC
adsorption for the removal ofDBP precursor based on the field studies conducted in Cincinnati
(OR), Jefferson Parish (LA), and Manchester (NH) (Clark et al., 1994). The cost estimates
given in Table 1. were established based on the assumption of 20 min EBCT and for two
different scenarios of MCLs for TTHMs (Ce < 100 Ilg/l and Ce < 50 Ilg/l). For systems of
380,3,800, and 38,000 m3/day, pressure contactors were assumed and for systems of 95,000,
190,000, and 380,000 m3/day, concrete gravity contactors were assumed. For systems of380
and 3,800 m3/day replacement of spent carbon with virgin carbon was used in calculations,
while the on-site reactivation was assumed for the other four systems.
Membrane Filtration. To achieve TOC removal in excess of 75%, membrane

filtration generally requires the use of nanofilters, with membranes having a molecular weight
cutoffs (MWC) of200-500 daltons. For most applications, pretreatment is required to prevent
fouling ofthe membrane. For the cost analysis, it is assumed that a 20.3xI00.6 cm element
removed organics having the molecular weights greater than 200 daitons (Clark et al., 1994).
Based on field experience, the removal of DOC, TOX and DBP can be assumed to be in the
excess of 95%. Figure 1. shows the impact ofMWC on TOX formation potential (TOXFP)
Table 1. GAC cost summary for THM precursor removal in ¢/m3
Plant Design Flow Plant Average Flow Ce < 100 Ilg/l Ce < 50 Ilg/l
(m3/day) (m3/day)
380 190 64.3 -76.0 66.5 - 85.5
3,800 1,900 24.4 - 35.2 26.5 - 43.8
38,000 26,600 14.1 - 18.2 15.0 - 21.1
95,000 64,600 9.4 - 12.8 10.1 - 15.3
180,000 126,600 7.6 - 10.6 8.2 - 12.8
380,000 266,600 6.4 - 9.8 7.0 - 12.1
267
100
'"
~
'Xl
!l)
~N
"t:l
Q>
e~ 70
~ 6J ~
'"
~
~ 50
0 t'-.
E-<
""'
0 40
.g 30 t--..
! 20
~
10 " ~
o
l.()e+2 1.0e+3 1.0e-I4 l.()et5
Figure 1. Impact ofMWC on TOXFP removal in membrane processes
Table 2. Nanofiltration cost summary (¢/m3)
Ground-Water Svstem Surface-Water Svstem

System permeate Surface Deep Well Pretreatment with alum,
capacity Concentrate Injection settling, filtration, deep
(m3/dav) Disposal well iniection
380 53.9 63.1 120.9
3800 35.1 40.7 79.7
19,000 27.7 31.9 63.1
38,000 25.6 29.6 58.3
95,000 24.6 28.2 55.4
190,000 22.2 25.6 50.7
380,000 20.3 23.2 45.9
Two types of systems were considered in cost analysis: (1) groundwater with an
average feed pressure of 1 MPa, an average flux of 0.61 mlday, and 10 membrane skids with
268
nine on line and one on standby; and (2) surface water with an average pressure of O. 7 MPa, an
average flux of 0.41 mlday, and 13 membrane skids with 12 on line and one on standby. Both
systems are operated using a three-stage membrane configuration at 85% recovery.
Assumptions for the cost analysis of groundwater systems included both brine disposal to and
existing surface pond or stream as well as deep well injection. Surface water treatment by
nanofiltration included pretreatment using alum coagulation, solids contact clarification and
rapid sand filtration. Table 2. summarizes the costs for all three options.
As can be seen from Figure 2., which compares the costs of the three alternatives,
nanofiltration is more expensive for the removal of DBP precursors from surface water than
GAC adsorption. On the other hand, small systems utilizing ground water (below 38,000
m3/day), could find nanofiltration an attractive option to GAC adsorption.
140
120
• GAC
100 I • Nanofi1tration - Groundwater
1 80
\
A Nanofiltration - Surface Water
ll..
~ 60 ~--~
-----~-------------.
40
20 --'-:--:"-.-"-"-"-"-"-"-"-"-"-"-.
.. = =
OL-~~~~~L-~~~~~--~~~~~--~~~~~
OetO 1.e+5 2.e+5 3.e+5 4.e+5
System Size (m3/day)
Figure 2. Cost for removing precursors (ISO I18Il to 50 118Il)
1.4.3. Removal ofDBPs
Once the DBPs are formed, it is quite a task to remove them from the finished drinking
water prior to distribution into the system. There are basically two options for the removal of
low levels ofTHMs: (1) GAC adsorption and (2) air stripping. To illustrate the cost impact of
removing DBPs (in this case THMs), costs were calculated for these treatment options
269
assuming the initial THM concentration of 150 J..1g/1 (Clark et al, 1994). It should be noted
that the composition of THMs in the finished water may vary according to source water
characteristics. Cost calculations were performed using the design and operational parameters
given in Table 3.
Chloroform was selected to represent THMs that are difficult to remove by GAC
adsorption (low adsorptive capacity) but are easily stripped using a packed column aeration
(Henry's Law constant = 0.087). On the other hand, dibromochloromethane is much less
susceptible to removal by air stripping (Henry's Law constant = 0.030). Two design options
were considered for cost analyses: reducing the influent concentration of TTHMs of 150 J..1g/1
(75 J..1g/1 of chloroform and 75 J..1g/1 DBCM) to levels below 100 J..1g/1 (Option 1.) and below 50
J..1g/1 (Option 2). The assumptions in calculating the costs for GAC adsorption are the same as
those used in the analysis of GAC adsorbers for precursor removal. Table 4. summarizes the
costs for GAC adsorption, air stripping and air stripping with off-gas control. As is apparent
from this table, removal of THMs from finished water is quite expensive with GAC adsorption
being the most expensive alternative.
Table 3. Design options for packed-bed air stripping tower
Parameter Option 1. Option 2.

AirlWater ratio 40 40
Depth (m) 3 4.5
Diameter (m) 2.1 2.1
Influent (J..1g/1)
Chloroform 75 75
DBCM 75 75
Effluent (J..1g/1)
Chloroform 15 6
DBCM 23 12
1.4.4. Use of Alternative Disinfectants
Performance and costs associated with the use of alternative disinfectants to free
chlorine is described in the following Section.
270
Table 4. Cost for removal ofTHMs from an influent level of 150 11811 (¢/m3)
Ce< 100 J..lg/l Ce< 50 ug/l

Plant Size Air Air Stripping Air Air Stripping
(m3/day) GAC Stripping with offgas GAC Stripping with offgas
control control
380 72.3 14.6 36.5 72.8 16.l 40.0
3,800 31.9 4.0 12.5 32.2 5.1 15.1
38,000 16.9 1.9 5.6 17.2 2.4 6.9
95,000 11.7 1.7 5.2 11.9 2.2 6.5
180,000 9.6 1.6 5.0 9.8 2.1 6.3
380,000 8.9 1.6 5.0 9.1 2.0 6.2
1.5. CURRENT STATUS OF DBP REGULATIONS
The negotlatmg process among the parties involved (Reg-Neg) resulted in the
development of essentially three rules (pontius, 1993):
• An Information Collection Rule (ICR),
• An Enhanced Surface Water Treatment Rule (ESWTR),
• The Disinfectant-Disinfection By Product (D-DBP) Rule itself
The ICR is intended to provide information concerning raw-water quality, treatment

practices, and DBP concentrations in finished drinking waters across the United States for the
surface water systems serving more than 10,000 people. These systems will also be required to
collect information concerning the occurrence of pathogenic microorganisms (Le., Giardia,
Cryptosporidium, and viruses) in raw-water supplies.
The ESWTR is aimed at potentially increasing the requirements for removal and/or
inactivation of Giardia and viruses above those specified in the existing SWTR. Such prospect
came as a result of several waterborne outbreaks of cryptosporiodiosis and the fact that the
current SWTR does not address Cryptosporidium and the degree of protection against this
organism. Subsequently, there is a need to include the required C·t values for Cryptosporidium
271
in the newly developed regulations. One response to this issue could come through the
requirement for a 3- to 6-10g removal of Giardia based on the source water cyst concentration.
This would however, require greater doses of disinfectants which would increase the level of
DBPs in the finished water. The ESWTR provisions will be based on the findings of the ICR
monitoring.
Table 5. Disinfection/disinfection by-product rule: proposed MCLs and MRDLs
Compound MClJMRDL
Total trihalomethanes SO J.1g/l
Haloacetic acids (total offive) 60J.1g!l
Bromate lOug/l
Chlorite 1.0 mg/l
Chlorine 4.0 mg/l
Chloramines 4.0 mg/l
Chlorine Dioxide O.S mg/l
The D-DBP Rule will be developed in two stages. Stage I contains compromise MCLs
for TTHMs, total HAAs, bromate, and chlorite along with maximum residual disinfectant
levels (MRDLs) for chlorine, chloramines and chlorine dioxide based on the running annual
averages computed quarterly (Table 5.). The MCL for TTHMs and HAAs in the Stage II are
now proposed at 40 and 30 J.1g/l, respectively but will be renegotiated following collection of
additional information under the ICR.
Best available technology (BAT) for meeting the Stage I MCLs for THMs and HAAs
is enhanced coagulation (or enhanced softening) or GAC adsorption with a 10-min EBCT and
ISO-day reactivation cycle. If the TOC concentration (used as a surrogate for DBP
precursors) exceeds 2 mg/l at the point of first continuous disinfection, enhanced coagulation
must be implemented. Water systems that serve more than 10,000 customers and whose water
contains TOC above 4 mg/llevel at the point of first continuous disinfection will be required to
conduct pilot-scale studies to evaluate GAC or membrane processes for precursor removal.
Ground water systems that serve more than 50,00 people will have to conduct similar studies if
the treated water TOC exceeds 2 mg/l.
The bottom line for DBP regulation is that promulgation of a long-term DBP rule must
await additional scientific findings and that the ICR will provide the key data to support this
rule-making process.
272
1.6. REFERENCES
Bull R.J., Kopfler F.C. (1991) Health Effects of Disinfectants and Disinfection By-Products.
American Water Works Association Research Foundation, Denver, CO.
Clark RM., Adams lQ., Lykins B.w. (1994) DBP Control in Drinking Water: Cost and
Perfonnance. 1 Envir. Engrg. Div., ASCE, 120(4),759.
Couri D., Abdel-Rahman M.S., Bull RJ. (1982) Toxicological Effects of Chlorine Dioxide,
Chlorite, and Chlorate. Envir. Health Perspectives, 46, 13.
Haag W.R., Hoigne 1 (1983) Ozonation of Bromide-Containing Waters: Kinetics of
Fonnation of Hypobromous Acid and Bromate. Environ. Sci. Techno!., 17(5),261.
Johnson lD., Jensen J.N. (1986) THM and TOX Fonnation; Routes, Rates and Precursors. 1
AWWA, 78(4), 156.
Kavanaugh M.C. (1978) Modified Coagulation for Improved Removal of Trihalomethane
Precursors. J. AWWA, 70(11), 613.
Krasner S.W., McGuire M.l, Jacangelo lG., Patania M.L., Reagan K.M., Aieta E.M.
(1989) The occurrence of disinfection by-products in U.S. drinking water. J. AWWA,
81(8),41.
Malcolm Pirnie Inc. (1992) Water Treatment Plant Simulation Program. Version 1.21. User's
Manual. USEPA, Washington, D.C.
McGuire M.l, Meadow RG. (1988) AWWARF Trihalomethane Survey. 1 AWWA, 80(1),
61.
McGuire M.J., Marshall D.K., Tate C.H., Aieta E.M., Howe E.W., Crittenden lC. (1991)
Evaluating GAC for Trihalomethane Contro!' 1 AWWA, 83(1), 38.
Miller lW., Uden P.C. (1983) Characterization of Non-Volatile Aqueous Chlorination
Products of Humic Substances. Environ. Sci. Techno!., 17(3), 150.
National Academy of Sciences (1987) Drinking Water and Health: Disinfectants and
Disinfection By-Products, Vo!' 7, National Academy Press, Washington, D.C.
Pontius F.W. (1990) Complying With the New Drinking Water Quality Regulations. 1
AWWA, 82(2), 32.
Pontius F.W. (1993) Reg-Neg Process Draws to a Close,. 1 AWWA, 85(9), 18.
Pounnoghaddas H., Stevens AA., Kinman R.K., Dressman RC., Moore L.A., Ireland J.C.
(1993) Effect of Bromide Ion on Fonnation of HAAs During Chlorination. 1 AWWA,
85(1),82.
Oliver B.G. (1978) Chlorinated No-Volatile Organics Produced by the Reaction of Chlorine
With Humic Materials. Canadian Research, 11(6),21.
Quimby B.D., Delaney M.F., Uden P.C., Barnes RM. (1980) Detennination of the Aqueous
Chlorination Products of Humic Substances by Gas Chromatography With Microwave
Emission Detection. Anal. Chern., 52, 259.
Singer P.C., Pyne RD.G., Avs M., Miller C.T., Mojonier C. (1993) Examining the Impact of
Aquifer Storage and Recovery on DBPs. 1 AWWA, 85(11), 85.
Singer P.C. (1994) Control of Disinfection By-Products in Drinking Water. 1 Envir. Engrg.
Div., ASCE, 120(4), 727.
Stevens A.A, Moore L.A, Miltner R.l (1989) Fonnation and Control ofNon-Trihalomethane
Disinfection By-Products. 1 AWWA, 81(8), 54.
273
US Environmental Protection Agency (1979) National Interim Primary Drinking Water

Regulations; Control of Trihalomethanes in Drinking Water. Federal Register, 44(231;
Nov.29), 68624.
US Environmental Protection Agency (1989a) National Primary Drinking Water Regulations;
Filtration and Disinfection; Turbidity, Giardia Lamblia, Viruses, LegioneIIa, and
Heterotrophic Bacteria. Federal Register, 54(124; Jun.29), 27468.
US Environmental Protection Agency (l989b) Drinking Water; National Primary Drinking
Water Regulations; Total Coliforms. Federal Register, 54(212; Jun.29), 27544.
White G.C. (1986) Handbook of Chlorination. 2nd edition, Van Nostrand Reinhold Co., Inc.,
New York.
2. Control of Disinfection By-Products in Drinking
Water: Case Studies of Alternative Disinfection
Technologies
Radisav D. Vidic
Department of Civil and Environmental Engineering
University of Pittsburgh
949 Benedum Hall, Pittsburgh PA 15261
USA
2.1. CHLORAMINES
Chloramines have been used in water treatment since the early 1900s (Wolfe et aI., 1984)
and renewed interest in their use occurred in recent years since they represent a viable
alternative for the control oftrihalomethanes (THMs). Other advantages of using chloramines
include increased residual stability, reduction in odor problems caused by chlorination, and the
lowest cost among main alternatives (Wolfe et a1., 1984).
Chloramines are formed when hypochlorous acid (HOCI) reacts with ammonia or
organic nitrogen in the water according to following reactions:
The species of chloramines present in water are dependent on pH, amount of chlorine relative
to ammonia nitrogen (Table l.), and temperature (formation of chloramines is significantly
retarded below 10 oC).
NATO AS! Series, Partnership Sub-Series, 2. Environment - Vol. IS

Edited by C. Maksirnovic, F. Calomino, and J. Snoxell
276
Table 1. Existence of chloramine species in water
Predominant species pH range Chlorine: Ammonia-N

(mg CI2Il:mg NID-NIl)
monochloramine >6.5 < 5:1
dichloramine 4.3 - 6.5 5:1 to 7.6:1
trichloramine <4.3 after breakpoint chlorination
Normally, product water from treatment plants has a pH ranging from 6 to 9 with
monochloramine as the predominant species. Furthermore, a large majority of utilities operate
at low chlorine to ammonia nitrogen ratio (3:1 to 4:1) which also favors the formation of
monochloramines. Some utilities however, choose to operate at higher ratios of 6:1 to 7:1,
where dichloramine is the dominant chloramine form, thereby increasing disinfection efficiency
and reducing ammonia usage (Trussell and Kreft, 1984). Dichloramine and trichloramine in
drinking water, however, may cause taste and odor problems at concentrations of 0.8 and 0.02
mg/l, respectively (White, 1986).
Controversy exists regarding the effectiveness of chloramines as a microbiocidal agent.
In 1987 EPA labeled chloramines as a secondary disinfectant (White, 1986). However, the
effectiveness of chloramines under laboratory experimental conditions is apparently lower than
in drinking water treatment plants where results show that coliform bacterial counts have been
maintained within the MCL of 1 colonylIOO ml using the chloramines as disinfectants (Wolfe et
al., 1984, Thompson and Ameno, 1987).
Water utilities using chloramines have employed the following patterns for the addition
of ammonia and chlorine in the treatment process: (a) prechlorination upstream a substantial
distance from ammonia application, (b) preammoniation and mixing of ammonia before
application of chlorine, and (c) concurrent addition of ammonia and chlorine. Option (a) allows
for primary disinfection by chlorine prior to the addition of ammonia to form chloramines that
would provide secondary disinfection in distribution systems. Additional advantages of this
option include reduced chlorine demand of the finished water and longer lasting chloramine
residuals. Some utilities were capable of maintaining THM levels below 0.10 mg/l even with 2
hours of free chlorine contact time. Several utilities have added ammonia and chlorine at the
same location and have maintained residual throughout the entire treatment process and
distribution system (Trussel and Kreft, 1984, White, 1986). Most utilities using
monochloramine as a secondary disinfectant switch back to free chlorine periodically in order
to eliminate biological growths that may have colonized various equipment in the treatment
plant as well as the distribution system.
277
Care should be exercised to prevent the appearance of excess ammonia in the finished
water since that might cause nitrification (biological conversion of ammonia to nitrite (N02-)
and nitrate (N03-» to occur yielding a decrease in the total chlorine and ammonia
concentrations in the storage reservoirs and distribution systems. This is especially important
in warm waters. One possible method to prevent this problem is to adjust the CI2:NH3-N ratio
(4:1 to 4.5:1 is recommended) and maintain adequate combined residual (0.5 to 1.0 mg/l)
(White, 1986, Lieu et aI., 1993).
Another advantage of mono chloramine over free chlorine is that monochloramine
produces appreciably lower amounts of disinfection by-products (DBPs). Some dichloroacetic
acid can be formed, however, and cyanogen chloride formation is greater than with free
chlorine (Jacangelo et aI., 1989, Krasner et aI., 1989).
2.1.1. Ann Arbor Water Treatment Plant (George et aI., 1990)
System Description. The Ann Arbor Water Treatment Plant is a split treatment Iime-
softening plant which uses water from two sources: 85% was surface water from the Huron
River, and 15% was from groundwater extraction wells. The overall capacity of the plant is
131 m 3/min divided between two plants. The combined raw water contained high hardness
(279 mg/l as CaC03) and low turbidity (5 NTU). Well water was used for split treatment and
for pH control together with recarbonation. The treatment process incorporated the addition of
polymer in the secondary rapid mix chamber while phosphate and sodium hypochlorite were
added immediately prior to filtration and sodium silicofluoride was added after filtration along
with postchlorination. The final total chlorine residual at the plant was 4.0 mg/l.
After 1978, chlorine was added at the raw water pumping station for disinfection and 1.5
to 3.5 mg/l potassium permanganate was used for taste and odor control. TTHM levels in the
finished water averaged 105 J.l.g/1 while the maximum trihalomethane potential (MTP) varied
from 250 to 600 J.l.g/I. These high THM levels may have been enhanced by the addition of
chlorine and KMn04 at the same location while high pH of the finished water (9 to 10.8)
certainly promoted THM formation.
In September 1981, ammonium sulfate was introduced to a primary rapid mix chamber
to react with free chlorine and thereby reduce THM concentrations. However, being that
ammonium sulfate is highly hygroscopic and therefore, extremely difficult to feed, this practice
was quickly replaced by the addition of liquid anhydrous ammonia through the existing
chlorinator at the raw water pumping station. Chlorine to ammonia nitrogen ratio was
maintained at 3:1 which provided approximately 0.25 mg/l residual free ammonia.
278
System Performance. With the plant expansion and continuous use of coagulant the
plant was producing good quality water with respect to turbidity (average below 0.2 NTU).
Since the conversion to chloramines, TTHM concentrations in the finished water were reduced
to approximately 15 ~g/l while still providing product water with <1 total coliform
bacteria/100 mI at the final combined chlorine concentration in the plant effluent of 4.0 mg/l.
Heterotrophic plate counts (HPCs) within the distribution system were generally less than 10
count simI (when the residual concentration of 3.5 mg/l was maintained during the initial
period, HPC averaged 20 counts/mI). A slight increase (from 2.4 to 4.1 pcu) in the color of
the finished water was observed since chloramines are ineffective in oxidizing coloring agents.
Problem Areas. The introduction of gaseous ammonia into the raw water stream softens
the water, resulting in the precipitation of the salts that caused a scale to form around the
injector. Fish loses were reported by students attending the local university since they were
not aware of chloramine levels in water and did not take appropriate precautions to remove it
prior to use in their aquariums like most of the commercial fish industry in the area.
Economic Analysis. For the implementation of anhydrous ammonia, a 908 kg/day

chlorinator was converted to ammonia applicator together with a purchase of a pump and a
daily storage tank. Table 2. at the end of this section presents capital costs together with an
increase in operation and maintenance (O&M) costs incurred as a result of conversion to
chloramines. The labor requirements increased by 2.3 hours per week and the time required
for laboratory analyses increased by 1.5 hours per week. In addition, an ammonia dosage of
1 mg/I combined with a chlorine dosage increase of 1.1 mg/l comprised all increases in
chemical costs. The total annual capital and O&M costs for the conversion were $20,712
(amortization of capital expenses is based on 20 year life span and six percent compound
interest) which translated to an increase in treatment costs of $0.99/1,000 m3.
2.1.2. Contra Costa Water District, Ralph D. Bollman Water Treatment Plant (George et
al., 1990)
System Description. This is a conventional surface water treatment plant. The raw
water source for this plant is the Sacramento-San Joaquin River delta. Water from the river
delta is pumped into Mallard Reservoir (total volume of 3.8.106 m3) where it is held until
treatment. This water is moderately hard (94 mg CaC03/l) with an average pH and turbidity
of 8.1 and 29 NTU, respectively. In dry years, chloride ion concentration can be as high as
300 mg/l resulting in high THM levels, particularly the bromide-containing compounds.
279
During this period, the raw water contained nearly 1 mg bromidell and the treated water had
THM levels of275 ~g/l, 75% of which was bromofrom.
To address this problem, a chloramination system was added to the plant in September
1981. A small amount of chlorine was added to the raw water as a coagulant aid (0.75 mg/l),
which produced no residual after mixing. Chlorine was also added prior to the filters in the
amount sufficient to produce a 0.75 mg/l residual in the filter etlluent. Ammonia was then
applied at a rate of approximately 0.2 mg/l or at a 3.75:1 chlorine to ammonia weight ratio.
The location of ammonia addition was selected to reduce a free chlorine contact time to only
10 min. Since the conversion to chloramination, it has also been a standard practice to add
enough chlorine, for about one hour each night, to the influent water to produce a 0.5-0.7 mg/l
residual in the filters. This practice was used to reduce the amount of algal growth in the
sedimentation basin and filters.
System Performance. THM concentrations were reduced to an average of 47 ~g/l by

the conversion to chloramination. During a six-months period in 1985, TTHM levels in
drinking water increased when state of California required the plant to stop using the
chloramination because of some problems experienced by dialysis patients in southern
California. Bacteriological quality of chloraminated water showed some improvement over
chlorinated water. Total coliform levels remained nearly undetectable while HPCs were
lowered from an average of 39 counts/ml to an average of 21 counts/mI. 95 percent of
distribution system samples contained a combined chlorine residual. Chlorine-like odor
complaints were eliminated and while some other utilities using the same raw water source
experienced taste and odor problems caused by algae, this was not a problem at Ralph D.
Bollman treatment plant due to significant free chlorine contact time at the plant.
Problem Areas. Hemodialysis patients are very susceptible to health problems that
might be caused by chloramines. On one occasion, a local hospital removed a water
purification process from a patient treatment step and chloramines reached the dialysis
equipment. Several patients suffered anemia for a short time until the hospital personnel
corrected the problem. Change to chloramination has increased the amount of chlorine
required to clean storage reservoirs in the distribution system from 4 to 12 m3 of 12% chlorine
solution per 106 m3 of reservoir capacity.
Economic Analysis. The process changes required the purchase and installation of an
ammoniator, diffusers, and piping while the 2.37 m3 anhydrous ammonia storage tank was
provided at no cost. O&M costs included the chemical costs for the addition of ammonia
(0.26 mg/l), ammoniation system maintenance, and savings realized through the reduction in
280
the chlorine dosage of 1.1 mg/l. The conversion to chloramination at this plant resulted in the
overall savings of$0.1411,000 m3 (Table 2.).
Table 2. Increase in annual (1988) costs for conversion to chloramines ($)
Item Ann Arbor Contra Costa

Capital Costs
Chlorinator conversion kit 847 --
Ammoniators -- 5,240
Storage tank 678 0
Pump 460 --
Hose, valves, fittings 862 10,480
Engineering -- --
Total 2,847 15,720
Annual amortized capital cost 248 1,370
Operation & Maintenance Cost Increase
Labor 1,555 480
Laboratory time 438 --
Chlorine 12,710 -17,932
Ammonia 5,761 10,000
Potassium permanganate -- --
Annual O&M Costs 20,464 -7,452
Total Annual Costs 20,712 -6,082
TOTAL UNIT COSTS ($/1,000 m3) 0.99 -0.14
2.2. CHLORINE DIOXIDE
The application of chlorine dioxide (CI02) has been implemented as an alternative to

meet THM regulation. In addition, CI02 is effective for taste and odor control, decolorization,
iron and manganese oxidation, organic oxidation, enhanced coagulation-filtration, in-plant
control ofbiofouling, and disinfection (Lykins and Griese, 1986, Aieta and Berg, 1986). The
use of CI02 as the primary disinfectant followed by free chlorine or chloramines for
281
distribution residual is a common and effective alternative treatment practice for THM
reduction.
Chlorine dioxide is an unstable gas that, like ozone, must be generated on site. It does
not react with water to the same extent as chlorine; however, it is volatile and can easily be
removed by aeration. Therefore, facilities for the application of CI02 should be designed to
reduce volatilization of the gas out of the solution. Chlorine dioxide is considered to be a
powerful oxidant, but at the pH typical in drinking water treatment, the full oxidation strength
of CI02 is not utilized. Chlorine dioxide will not hydrolyze in water like chlorine and it exists
as dissolved gas in the 2 to 10 pH range. The following reactions define reduction of CI02 in
aqueous solution:
(I)
(2)
Equation (1) is the complete reaction used to describe reduction of CI02, but the reaction that
often occurs in drinking water application is described by Equation (2) since the oxidizing
power of this agent is pH-dependent and the Equation (2) is predominant reaction at pH above
5. Chlorine dioxide can also dissociate in alkaline solutions to chlorite (CI02-) and chlorate
(CI03-) ions:
CI02- is also an effective oxidizing agent and will be consumed in oxidation-reduction

reactions, although at a much slower rate than chlorine dioxide. This reduction half-reaction is
given by:
During water treatment, approximately 50-70 percent of chlorine dioxide reacted will
immediately appear as chlorite and reminder as chlorine. The residual chlorite continues to
degrade in the water distribution system via the above reaction, presumably in reactions with
oxidizable material in the finished water.
282
There is some concern about the by-products of chlorine dioxide decomposition. Cl02,
Cl02- and Cl03- have been implicated as a cause for hemolytic anemia, methemoglobinemia
and other hematological effects, with Cl02- identified as the most potent. However, if the
recommended total residual oxidant level of less than 1.0 mgll is maintained, no adverse health
effects are anticipated (Lykins and Griese, 1986).
Reactions of Cl02 with NOM do not produce THMs (Aieta and Berg, 1986) and the
total production of total organic halides are from 1 to 25% of those produced by chlorine
under similar conditions. Recent research has shown that some of these by-product are most
likely similar to those produced by ozone (Richardson et aI., 1994; see discussion about ozone
by-products). However, Cl02 reacts with THM precursors to render them unreactive to THM
formation by as much as 50%, thereby reducing the amount of THMs that could be formed by
post chlorination. Postchlorination dosages required to maintain a residual however, are
generally higher than before the change to Cl02 (Aieta and Berg, 1986).
Chlorine dioxide is superior to free chlorine as a biocide. However, during the period of
high turbidity (> 100 NTU), an increase in taste and odor and a decrease in disinfection
efficiency may occur. This behavior can be attributed to an increase in organic matter rather
than suspended matter. In general, most studies found chlorine dioxide to be an effective and
rapid bactericide (unaffected by pH in the 6.5 to 8 range) and equal or superior to chlorine on a
mass-dosage basis (Aieta and Berg, 1986).
Generation of chlorine dioxide in water treatment is most commonly performed through
the reaction of sodium chlorite and chlorine (gas or liquid) (Miller et aI., 1978):
2NaCl0 2 + CI 2(g) ~ 2NaCl + 2Cl0 2
5NaCl0 2 + 4HCl ~ 4Cl0 2 + 5NaCl + 2H 2 0
The objective is to produce a solution with highest purity and the greatest yield (purity is the
ratio ofCl02 to the total of all chlorine oxidative compounds). If a gas chlorinator is used, the
chlorinator effiuent must be at pH 3.5 or lower and contain at least 500 mg/l chlorine to
accomplish the above objective. This condition is typically achieved by setting the chlorine-to-
sodium chlorite ratio at 1: 1 in order to provide excess chlorine. It is also possible to add acid
to achieve the desired pH. The most application nowadays use the reaction of gaseous
chlorine with sodium chlorite under vacuum (Aieta and Berg, 1986).
283
2.2.1. Bristol County Water Authority's Child Street Treatment Plant (George et aI.,
1990)
System Description. This treatment plant obtained its raw water from a 1.76.106 m3
reservoir that was fed from three other reservoirs. Raw water was treated in a conventional
alum coagulation, rapid sand filtration system rated at 11 m3fmin with an average water
production ofS.9 m3fmin. The reservoir was subjected to salt water intrusion for a few days
each year when high winds, high tide, and low water levels in the reservoir occurred
simultaneously. Raw water had high color (S7 pcu) with low alkalinity (17 mg CaC03/l), low
turbidity (2 NTU), and significant amounts of iron (0.6 mg/l) and manganese (0.19 mg/l). The
maximum trihalomethane potential of raw water occasionally exceeded 3000 ~g/l which forced
this utility to seek alternative treatment method. One of the criteria for selecting the treatment
scheme was the requirement for the oxidation of iron and manganese. Typically, chlorine,
alum, coagulant aid and lime were added at the raw water intake. Adding potassium
permanganate at the low-lift pump station and moving the point of chlorination to just before
the rapid sand filters reduced the TTHM levels only from 227 to 125 ~g/l. Subsequently, a
new treatment approach had to be adopted.
Based on the results of trial runs, the modifications included the addition of chlorine
dioxide before the pretreatment clarifier (detention time of 3.75 hours at average flow) and
moving the point of chlorine addition to clear well. Chlorine dioxide was applied at 1 to 2 mg/l
to provide approximately 0.1 mg CI02/l residual after the pretreatment clarifier. Chlorine
dioxide was produced from a 25% sodium chlorite solution and chlorine gas in stoichiometric
quantities under vacuum. PAC was added into the settling basin for taste and odor control.
Lime was added to adjust pH and alkalinity just before filtration. Chlorine and fluoride were
added at the clear well which provided 25 min of detention time.
System Performance. The THM levels in finished water were reduced significantly from
an average of 112 to 53 ~g/l. Chlorine dioxide also oxidized iron and manganese to average
finished water levels of 0.043 and 0.037 mg/l, respectively. With the use of chlorine dioxide,
the average chlorine dosage requirements were reduced from 4.1 to 3.2 mg/l. Correspondingly,
the chlorine residual was reduced from 1.1 to O.S mg/l. Even though the raw water color
increased from an average of 74 pcu during the period of chlorine usage to an average of 101
pcu after the change in the treatment process, the new system was capable of producing a
product water with a slightly less color (4.7 pcu) than achieved before chlorine dioxide
treatment (5.1 pcu). In addition, the plant continued to produce water with an average
turbidity of 0.57 NT and odor of I.S TON.
284
Problem Areas. During the summer of 1986, the system experienced taste and odor
problem which was not solved even with the addition of PAC and the cause was never
identified. The only other concern was with the analytical procedure for residual CI02
measurements (low confidence levels in the results obtained by the procedure).
Economic Analyses. Capital expenditures included a chlorine dioxide generator, sodium

chloride handling equipment (mixing tank, barrelluft, mixer) and a block building. The use of
0.97 mgll potassium permanganate was discontinued and chlorine dosage was reduced by
0.9 mgll. Sodium chlorite was added to the water at a concentration of 1.72 mgll. Other
O&M costs included 2 hours/week labor for the operation of generation facilities and 16 hours
per year for equipment maintenance together with some replacement parts (needle valves,
o-rings, etc.) The combined annual capital costs and O&M costs equaled $13,556 resulting in
the overall increase in treatment costs of$2.8/1,000 m3 (Table 3.).
2.2.2. Kentucky American Water Company (George et al., 1990)
System Description. Raw water for the plant was withdrawn from the Kentucky River.
The plant had a design capacity of 105 m3/min with an average production of 74 m3/min.
Prior to conversion to chlorine dioxide, a dosage of 5-7 mgll of chlorine was added to the
influent water to maintain a residual throughout the treatment process. A cationic polymer was
added as the primary coagulant and alum was used as a coagulant aid when needed. Optimal
pH for coagulation was achieved by the addition of lime. Water treatment was accomplished
in 10 Aldrich units (Hydrotreators). THM analyses revealed average concentrations of
125 Ilg/l with only 37% of the total being chlorinated species.
Once the utility recognized that bromide in the Kentucky River was aggravating the
THM levels in the finished water, an alternative treatment method had to be implemented. A
chlorine dioxide generating system capable of generating 304 kg of chlorine dioxide per day
was installed at the plant and 25% sodium chlorite solution was used to feed the generator.
Choice of solution over a dry powder was made because the dry powder is a strong oxidizing
agent that poses a significant hazard if spilled and the mixing of a dry powder would require an
intensive labor. A 2-month trial run was conducted by feeding chlorine dioxide at a dosage
from 0.6 to 1.25 mgll to provide a residual prior to filtration of 0.1 mgll. During this period,
THM levels were reduced by about 60% but a significant number of complaints from
customers about chlorinous- or petroleum-like odors were reported. There was no success in
identifYing source of the problem and it was decided to use chlorine dioxide only during
periods when excessive THM production occurred. After the addition of potassium
285
pennanganate at the river intake to control CI02 odors, a number of complaints significantly
decreased. The point of chlorination was then moved from the rapid mix to the top of the
filters and the addition of chlorine dioxide was reduced only to two quarters throughout the
year when THM levels became excessive.
System Performance. A major concern during the trial run was consistently higher
bacteriological counts throughout the treatment system even though they were always below
state regulations. After portions of distribution system were flushed to control odors,
distribution system bacteriological quality improved tremendously indicating a requirement for
postchlorination. THM levels decreased from 140 to 90 Ilg/l while the average amount of
chlorine added to the system remained practically unchanged (516 vs. 477 kg/day) yielding an
appreciable decrease in chlorine dosage (from 5.1 to 3.9 mg/l) due to the increase in average
flow.
Problem Areas. Due to an observed decrease in the efficiency of chlorine dioxide

generator from 95 to 90010, it was necessary to clean the media once a week with a mild acid
and replace it twice in four years. The cause of odor problems instigated with the use of
chlorine dioxide were never identified and the problems occurred even though no chlorine
dioxide could be identified in the effluent (pretreatment with potassium pennanganate seemed
to eliminate many of the problems). In conjunction with poor biological control capabilities of
CI02 observed at this plant, this problem forced the utility to investigate other disinfection
alternatives. In addition, analytical methods for chlorine dioxide, chlorate, and chlorite that
were capable of providing accurate results were too time consuming and complex for everyday
use by a water treatment plant operators.
Economic Analysis. The capital cost consisted of the purchase of a chlorine dioxide
generator, two mixing tanks and mixers, metering pump, jet pump, two filament-wound tanks,
and two stainless-steel centrifugal pumps totaling $20,340. The average sodium chlorate
dosage was 1.55 mg/l while the chlorine dosage was reduced by 1.2 mg/l. An estimated
additional labor hours associated with laboratory analyses, maintenance, and operation of the
generator was 31 hours per week for a 6-month period. The combined annual amortized
capital cost and O&M costs totaled $55,299 which translated in the increase in treatment costs
of $1.4311,000 m3 (Table 3.).
286
Table 3. Increase in annual (1988) costs for conversion to chlorine dioxide ($)
Item Bristol KY-American

County Water Co.
Capital Costs
Cl02 generator 12,180
Handling Equipment 3,606
Block Building 23,189
Pump --
Hose, valves, fittings --
Engineering --
Total 39,975 20,340
Annual amortized capital cost 3,398 1,773
Operation & Maintenance Cost Increase
NaClO2 22,907 69,124
Chlorine -2,397 -26,758
Potassium permanganate -12,204 --
Labor 220 11,160
Laboratory time 1,432
Replacement parts 200
Annual O&M Costs 10,158 53,526
Total Annual Costs 13,556 55,299
TOTAL UNIT COSTS ($/1,000 m 3) 2.80 1.43
2.3. OZONE
Ozone is one of the most powerful oxidizing agents available for practical applications in
water treatment. It does not impart tastes and odors to water and has beneficial effects on
microflocculation, enhanced suspended solids removal, and pretreatment for biological
activated carbon. Ozone, however, does not provide a disinfection residual and must, in most
cases, be followed with secondary disinfection to maintain residual in the distribution system.
Ozone is a highly reactive gas that follows two basic modes of reaction:
287
0 _ _ _ _ _---=D_ir:.. .:e"'ct-'-::;oXl'-·-=d.c:.at:.c.io";n.:. .o.:. .f;-cs'-'u"'b.:. .st=--ra"'tc.:.e_ _ _ _ _

" Products
3
Slow, selective
/ ---------~--------------------------, ,
I/--------E------------ . . . ~ \R·
Radical formation Radical oxidation
OH·--------------~
Fast, nonselective
C02- Products
3
HCO; Radical consumption CO·-
'----------"'----~ 3
HCO·
3
The rate of ozone decomposition is a complex function of temperature, pH, ozone

concentration, type and concentration of organic compounds, and type and concentration of
inorganic constituents. Conditions oflow pH favor direct oxidation reactions while high pH or
high concentrations of organic matter favor the auto-decomposition of ozone via the hydroxyl
radical (Hoigne and Bader, 1976, Gurol, 1985).
Ozone has received considerable attention for THM control since it is the most effective
oxidant and disinfectant used in water treatment practice and a number of studies showed that
in combination with mono chloramine provides effective treatment while minimizing
halogenated DBP formation. Ozone, however, reacts with NOM to produce a variety of
oxidation by-products like aldehydes (e.g., formaldehyde, acetaldehyde, glyoxal, methyl
glyoxal), aldo- and ketoacids (e.g., pyruvic acid), acids (e.g., oxalic, succinic, formic, acetic),
and hydrogen peroxide (Glaze et al., 1989, Weinberg et al., 1993). Of these, the aldehydes
appear to be of greatest concern from a public-health standpoint at this time (Bull and Kopfler,
1991), but insufficient information is currently available to ascertain the health risks due to
exposure to aldehydes. All of the oxidation products listed are for the most part biodegradable
and contribute to the assiniilable organic carbon (AOC) content of the water which can lead to
nuisance biological growths in flocculation basins, sedimentation basins, filters, and the
distribution system if not properly controlled. One effective way of controlling these growths
is through the use of biologically active filtration (Reckow et al., 1992, Weinberg et al., 1993).
Although ozone itself does not produce halogenated DBPs, it can result in the
production of brominated DBPs (bromate, bromoform, brominated acetic acids, bromopicrin,
brominated acetonitriles) when bromide-containing waters are ozonated. Ozone will oxidize
bromide (Br-) to hypobromous acid (HOBr), which in turn can react with NOM to produce the
fully brominated analogs of the chlorination by-products (Siddiqui and Amy, 1993). In
addition, hypobromite (OBr-) in equilibrium with HOBr can be oxidized further by ozone to
produce bromate (Br03-) which has been classified by EPA as a B2 carcinogen (probable
288
human carcinogen) and is likely to be regulated with an MCL equal to its practical
quantification level of 10 Ilg/L Bromate formation can be controlled by ozonation at acidic
pH-values, at which hypobromous acid dominates over hypobromous ion (Siddiqui and Amy,
1993).
Virtually all-large scale ozone generators incorporate the corona discharge method for
the production of ozone with varying cell configuration, operation voltage, operation
frequency, cooling method, and air preparation. Ozone production depends mainly on the
geometry of the generator, the dielectric properties, the current frequency, the dryness and
content of the feed gas, the pressure, and process temperature. Since the ozone yield is
decreased with increasing temperature and dew point, it is necessary to provide process
cooling and feed-gas (air or oxygen) preparation system. The use of oxygen as feed gas allows
higher ozone production (by a factor of 2), larger ozone concentrations, smaller generators,
and lower energy consumption. Nonetheless, ozone is produced from air in the majority of
cases.
2.3.1. BeD Glade Water Treatment Plant (George et al., 1990)
System Description. Raw water for the Bell Glade plant was taken from Lake
Okeechobee (large, shallow lake) and had average color between 59 and 74 pcu (peak value as
high as 500 pcu), high TOC content of 20 to 30 mg/l (up to 75 mg/l during summer
agricultural drainage periods), and MTP average of 808 Ilg/l. In addition, the underlying
limestone combined with natural acids in the water caused high hardness (average monthly
concentration of 242.4 mg CaC03/l with the average maximum concentration of 512 mg
CaC03/l) and high mineral content of the raw water. Treatment process utilized softening,
coagulation, flocculation, filtration and disinfection. Prior to treatment change, 30 to 60 mg/l
alum and 15 to 50 mg/l of chlorine (breakpoint chlorination) were added to water as it entered
the plant. Alum was added for color removal but did not work well because of the high pH
(10.5) utilized in the excess lime softening process. Treatment process modification
incorporated replacing the chlorine addition to the raw water with ozone to control the color,
biological growth in the plant, musty and obnoxious taste and odor compounds, and to oxidize
organic THM precursors. The average monthly dosage of ozone was 9.5 mg/l (range from 4.9
to 15 mg/l) while postchlorination was used to maintain residual in the distribution system.
System Performance. One of the most obvious improvements of finished water was in
the aesthetic quality with a significant decrease in color resulting in the elimination of customer
complaints concerning color, taste and odor. Ozone addition resulted in significant decrease in
289
the instantaneous TTHM concentration average from 643 to 60 IJ.gIl (prior to change, the
actual measured TTHM concentration occasionally exceeded 1,200 IJ.gIl) while no significant
changes in TOC concentrations were observed. Water softening operation improved slightly
after the change to ozonation. The reduction in the finished water pH from 9.7 to 7.8 after the
conversion to ozonation was not to minimize ozone decomposition, but to control the
precipitation of calcium in the distribution system. A decrease in water turbidity from 0.4 to
0.2 NTU may have been due several factors, including rnicroflocculation and an increase in
polymer dosage from 0.04 to 0.08 mgll.
Problem Areas. No serious problems associate with the change to ozonation were
reported by the principal operator. Minor maintenance (e.g., installation of filters in the seal
water system of the air compressor, replacement of the steel sleeves in the dryers with stainless
steel, and the development ofa few pin holes in the dielectric sleeves) were easily resolved and
the system has operated reliably since installation.
Economic Analysis. The capital and operation and maintenance costs for Belle Glade
water treatment plant are given in Table 4. The estimated total maintenance costs averaged
$4,OOO/year while power requirements for the ozone generator were estimated at 89,913 kw·
hr/month for an average monthly cost of $5,700. Additional labor requirements included 2
hr/day for operation and 16 hr/year for maintenance. After converting to ozone, the chlorine
dosage was reduced from 19.3 to 9.4 mg/l resulting in savings of $3,523/month. The
estimated total yearly increase in O&M costs was $17,367 which, together with annual
amortized capital costs, resulted in total costs of$78,506/year or $14.5711,000m3 .
2.3.2. Monroe Water Treatment Plant (George et al., 1990)
System Description. This conventional water treatment plant used lake water with
relatively,low turbidity (average of 10 NTU) and moderate hardness and alkalinity of 121 and
87 mg CaC031l, respectively. However, because of the shallow lake depth and uneven bottom,
frequent strong winds caused violent wave action that disturbed the bottom sediments and
produced high turbidities (1,200 NTU maximum). These conditions also created objectionable
odors. Spring run-offs and odorous compounds produced by algae and actinomycetes during
Summer also affected the raw water quality. Odor control efforts included application of PAC,
clarification with alum and intermediate chlorination with contact time prior to filtration. These
efforts were successful until the increase in the flow rate decreased the chlorine contact time
and the system was altered to incorporate ozonation for the control of taste- and odor-causing
290
compounds. The two ozone generators, each rated at 102 kg ozone/day at concentrations of 1
wt % using air as the feed gas, were installed at that time. Air treatment prior to ozone
generator included passage through an inlet filter and compression which increased its
temperature to 149 0C. The dew point was lowered to 20 oC by a refrigerant dryer and finally
to -50 to -100 0C by transporting the air through molecular sieves and silica gel. Carryover of
desiccant was removed using a 10 Ilm filter prior to the corona cells. The contact chamber
consisted of a countercurrent gas-diffused (353 porous aluminum oxide diffusers), positive-
pressure (12.4 kPa), recirculating contactor that consistently yielded inlet-to-outlet efficiency
from 93 to greater than 99 percent. The contactor had two passes with three stages in each
pass (60-70% distributed in the first stage of pass 1 and the remaining 30-40% distributed in
the second stage) with an average contact time of approximately 10 min at a flow rate of 47
m3/min.
System Performance. The raw water was treated with an average ozone dosage of 1.2
mg/l and a maximum of 3.1 mg/l during the period of March through May of each year when
taste and odor problems were traditionally most pronounced. As a result, taste and odor
complaints were decreased from a high of 100 in one day to zero. In addition, the chlorine
dosage was lowered from 4.1 to 2.9 mg/l to produce the residual of 0.7 mg/l. It was also
discovered that the chlorine demand of ozonated water increased presumably due to partial
oxidation of complexed organic compounds by ozone which produced by-products that were
more susceptible to chlorine oxidation. TTHM concentration in finished water prior to
ozonation averaged 60-80 Ilg/l while the average of 70 Ilg/l was observed during the periods
of ozone addition. Chloroform was the predominant halomethane in the finished water,
followed by dichlorobromomethane and traces of dibromochloromethane « 7 Ilg/l) while no
bromoform was ever detected. During the periods when ozone was not used, and average
TTHM concentration of 44 Ilg/l was observed. However, due to the limited number of data,
this difference in avera&e concentrations was not found statistically significant. Ozonation did
not significantly improve the product water turbidity (0.2 NT prior and after ozonation) but
enabled a reduction in alum dosages from 25 to 19 mg/l and longer filter runs (average of 88
hours compared to 55 hours). Occasional presence of cyanide originated from steel mils was
successfully treated by ozonation through oxidation to cyanate and subsequent hydrolysis.
Problem Areas. No major problems with the use of ozone was experienced at Monroe
Water Treatment Plant. Several interruptions in ozonation occurred due to malfunction in the
dew point hygrometer and a complete renewal of desiccant had to be performed once. During
the start-up, all rubber material that came in contact with the ozonated water had to be
replaced with silicone rubber.
291
Economic Analysis. The total capital costs amounted to $754,500. Amortization and
insurance comprised more than half of the annual operating costs for this system with the
electrical power being the second most important item. Reduction in chlorine and alum
dosages by 1.2 and 6 mg!1, respectively, resulted in chemical costs savings of 6,575 and 3,945
$/month, respectively. The overall annual cost for this system is $84,650 which resulted in the
increase in treatment costs of$8.5111,000 m3 .
Table 4. Increase in annual (1988) costs for conversion to ozone ($)
Item Belle Glade Monroe

Capital Costs
Ozone generator 620,100 312,700
Installation 58,300 192,700
Piping and valves -- 249,100
Training, prestart-up 4,770 --
Lab test 5,512 --
Start-up 9,434 --
Assistance, poststart-up 3,180 --
Total 701,296 754,500
Annual amortized capital cost 61,139 65,578
Annual O&M Costs 17,367 8,552
Total Annual Costs 78,506 84,650
TOTAL UNIT COSTS ($/1,000 m 3) 14.57 8.51
2.4. REFERENCES
Aieta E.M., Berg J.D. (1986) A Review of Chlorine Dioxide in Drinking Water Treatment.
Journal AWWA, 78(6), 62.
Bun R.J., Kopfler F.C. (1991) Health Effects of Disinfectants and Disinfection By-Products.
American Water Works Association Research Foundation, Denver, CO.
George D.B., Adams Y.D., Huddleston SA, Roberts K.L., Borup M.B. (1990) Case Studies
of Modified Disinfection Practice for Trihalomethane Control. AWW A Research
Foundation, Denver, CO.
Glaze W.H. (1986) Chemistry of Ozone, By-Products and Their Health Effects. Proc. AWWA
Annual Conf, AWWA, 1-15, Denver, CO.
292
Glaze W.H., Koga M., Cancilla D. (1989) Ozonation By-Products. 2: Improvement of an

Aqueous-Phase Derivatization Method for the Detection of Formaldehyde and Other
Carbonyl Compounds Formed by the Ozonation of Drinking Water. Environ. Sci.
Technol. 23(7), 838.
Gurol M.D. (1985) Factors Controlling the Removal of Organic Pollutants in Ozone Reactors.
Hoigne J., Bader H. (1976) Role of Hydroxyl Radical Reactions in Ozonation Processes in
Aqueous Solutions. Water Research, 10,377.
Jacangelo J.G., Patania N.L., Reagan K.M., Aieta E.M., Krasner S.W., McGuire MJ. (1989)
Ozonation: Assessing its Role in the Formation and Control of Disinfection By-Products.
Krasner S.W., McGuire M.l, Jacangelo lG., Patania N.L., Reagan K.M., Aieta E.M. (1989)
The Occurrence of Disinfection By-Products in US Drinking Water. Journal AWWA,
81(8),4l.
Lieu N.I., Wolfe RL., Means E.G. (1993) Optimizing Chloramine Disinfection for the Control
of Nitrification. Journal AWWA, 85(2), 84.
Lykins B.W., Griese M.H. (1986) Using Chlorine Dioxide for Trihalomethane Control. Journal
AWWA, 78(6), 88.
Miller G.W., Rice RG., Robson C.M., Scullin RL., Kuhn W., WolfH. (1978) An Assessment
of Ozone and Chlorine Dioxide Treatment Technologies for Municipal Water Supplies -
Executive Summary. EPA report EPA-600/8-78-018, USEPA, Cincinnati, OH.
Reckow D.A., Tobiason J.T., Switzenbaum M.S., et al. (1992) Control of Disinfection By-
Products and AOC by Ozonation and Biologically Active In-Line Direct Filtration. Proc.
AWWAAnnual Com., AWWA, Denver, CO.
Richardson S.D., Thruston A.D., Colette T.W., et al. (1994) Multispectral Identification of
Chlorine Dioxide Disinfection Byproducts in Drinking Water. Env. Sci. Technol., 28(4),
592.
Siddiqui M.S., Amy G.L. (1993) Factors Affecting DBP Formation During Ozone-Bromide
Reactions. JournalAWWA, 85(1), 63.
Thompson lC., Ameno J.l (1987) Preventing the Formation of Trihalomethanes in Florida
Groundwater. Journal AWWA, 79(1), 38.
Trussel RR., Kreft P. (1984) Engineering Considerations of Chloramine Applications.
Proceedings AWWA Seminar on Chloramination for THM Control: Principles and
Practices, AWWA Annual Conference, Dallas, TX, p. 47.
White G.C. (1986) Handbook of Chlorination. 2nd edition, Van Nostrand Reinhold Co., Inc.,
New York.
Weinberg H.S., Glaze W.H., Krasner S.K., Sclimenti M.J. (1993) Formation and Removal of
Aldehydes in Plants That Use Ozone. Journal AWWA, 85(5), 72.
Wolfe R.L., Ward N.L., Olson B.H. (1984) Inorganic Chloramines as Drinking Water
Disinfectants: A Review. Journal AWWA, 76(5), 74.
3. Technological Aspects of Advanced Processes of
Coagulation and Flocculation in Water Treatment
Peter Kalinkov
Jean Louis Monzo
Higher Institute of Architecture
and Civil Engineering
1 Christo Smirnenski, 1421 Sofia
BULGARIA
3.1. INTRODUCTION
Creating the optimal conditions for an effective coagulation-flocculation process in raw

water treatment is always a question of current interest for specialists working in that field.
Our previous study (Kalinkov and Monzo, 1992) was devoted to the investigation of
the influence of initial rapid mixing duration and the major objective of this paper is to
investigate the influence of geometrical characteristics of mixers in these processes.
It is well known that the effectiveness of the coagulation-flocculation process depends
on the physico-chemical characteristics of water, the type and dose of the coagulant, as well as
the duration and intensity of initial rapid mixing.
Contemporary theoretical models for this process combine the physical aspects, such as
the hydrodynamics of mi~ing, with the mechanism of destabilisation of colloidal particles.
Initial rapid mixing at non - effective hydrodynamic conditions leads to a higher coagulant
consumption as well as to a smaller degree of particle aggregation at a given coagulant dosage.
Different approaches can be used to estimate the effectiveness of mechanical mixers.
Camp and Conklin (1970) use the so called turbulent gross grad coefficient Ct.
They consider that at a condition of full turbulence of the stream the necessary relative
power to start the impeller in water with regards to the unit volume may be expressed with:

Edited by C. Maksimovic, F. Calomino, and J. SnoxeJl
294
(1)
where p water density (kgm-3);

f projected area of the impeller blades normal to the direction of the
tangential velocity (m2);
C, turbulent gross drag coefficient (-);
N speed of impeller (S-I).
Vrale and Jorden (1971) use apparent aggregation rate Kapp, which according to them,
can be defined by the equation:
dn
- = Kappn 2 (2)
dt
dn
where is the rate of changing water turbidity during aggregation;
dt
Kapp aggregation rate constant
n water turbidity
Equation (2) can be integrated in the interval for n (Mo, Mt) and t (0, T).
After integration Equation (2) in the upper interval:
1 1
-=-+KappT (3)
M/ Mo
where Mo water turbidity after initial mixing;

Mt water turbidity after T min flocculation (slow mixing);
T flocculation or slow mixing time.
It is possible, using the theory of mixing presented by Vrale and Jorden, to obtain
experimentally the apparent aggregation rate at different conditions of initial rapid mixing and
the same (constant) flocculation conditions, and to draw a conclusion for the intensity of
destabilisation, e.g. the influence of hydrodynamic conditions of mixing on the effectiveness of
the coagulation - flocculation process.
295
3.2. EXPERIMENTAL SET UP AND PROCEDURE
Laboratory investigation has been carried out in 5.0 I plexiglass vessels and different
geometrical configurations (Figure 1.) and impellers (Figure 2.) the dimensions of which are
valid under the condition of ratio D/C = 0.56; at a constant area of the cross section f,
perpendicular to the plane of rotation.
fID
a) without baflles b) without baflles c) with simple baflles d) with cogged
and C/H=O.5 and C/H= 1 and C/H= 1 perforated baflles
and C/H= 1
Figure 1. Geometrical configuration of mixers used in this study
,lOmm"
~:~+~I
5,~1~
~ 20mm ~
f= 6fo = 15.60 cm2
Figure 2. Impeller - disc turbine type

296
4.0
Tr= 60s, Tslow = 25 min

3.5
oM = 110 mg/l, Alum dose = 21 mg/l
~ 11 M = 340 mg/l, Alum dose = 39 mg/l
.E. 3.0
~ 0 o
~
1'"
o
2.5
o
~ ---+--'0
2.0
I.S
100 200 300 400 SOO 600 700 800
Rapid mix velocity gradient G [lis)
a) Classical coagulation and optimal Alum dose
4.0
3.5
~
.E. 3.0
~
~ ~
~~
r---------------------------~
2.S Tr = 360 s, Tslow = 15 min
oM = 110 mg/l, Alum dose = 11 mg/l
2.0 11 M = 340 mg/l, Alum dose = 19 mg/l
I:
I:
II
I.S
100 200 300 400 SOO 600 700 800
Rapid mix velocity gradient G [lis)
b) Prolonged coagulation and reduced Alum dose
Figure 3. Optimal rapid mix velocity gradient

297
Four mixer configurations have been investigated with square bottom without baffles,
at ratio CIH = 1 and the same vessel with CIH = 0.5; with square bottom with simple baffles
built in and ratio CIH = 1; with square bottom with perforated cogged baffles and CIH = 1. For
all configurations G = f{N), where G is the velocity gradient (S-I) and N is the impeller rotation
speed (min-I).
An artificially obtained turbid water is used, obtained by adding bentonite clay to
distilled water. A 5% solution of Ah(S04)318H20 is used as a coagulant.
Water temperature, pH and alkalinity were held constant and only the physical
parameters of rapid mix were varied. pH and alkalinity were adjusted with a buffer made of
sodium hydroxide, sodium bicarbonate, sodium chloride and acetic acid as suggested by
Packham (1962).
Two series of experiments have been carried out with different water turbidity, as
follows MI = 26.2 EM, corresponding to 110 mgll according to Si02 scale and M2 = 80.9 EM,
corresponding to 340 mgll according to Si02 scale. Preliminary optimal coagulant doses for
these turbidities have been defined by classical method with conventional jar - test equipment.
They are as follows: DCI = 21 mgll and DC2 = 39 mgll Alum.
An experiment has been implemented to determine the optimal values of velocity
gradient G for classical water treatment (60 s rapid mixing and 25 min slow mixing) and
prolonged initial rapid mixing (360 s rapid mixing and 15 min slow mixing) with 50% reduced
dose of coagulant for a vessel without baffles at CIH = 1.
The results obtained are shown on Figure 3a. and Figure 3b. It is obvious that in the
first case the optimal values of the velocity gradient are 550 S-I and 500 S-I for water turbidity
of 110 mgll and 340 mgll respectively according to Si02 scale. And in the second case these
values are 250 S-I and 230 S-I.
The investigation of all configurations is carried out by two procedures: classical
coagulation with optimal coagulant dose and prolonged coagulation (prolonged initial rapid
mixing) and coagulant dose reduced with 50%.
The rapid mixing is carried out in the range of optimal velocity gradient values. In the
course of every experiment the rapid mixing is followed by a slow one with different duration -
from 5 to 25 min and constant impeller speed of about 45 min-I which corresponds to a
velocity gradient of 40 S-I.
Seventy two series of experiments have been carried out which give the opportunity for
the apparent aggregation rate for the different cases to be defined. The results of the
experiments with vessels without baffles and ratio CIH = 1, implemented with the classical
procedure are given in Table 1. and in Table 2. the results for prolonged initial rapid mixing are
shown. In all experiments after the period of slow mixing follows a 15 min period of settlement
and after that is measured the residual water turbidity.
298
Table 1. Residual turbidity values as a function of slow mix time under classic coagulation
Raw water turbidity: 26.2 EM = 110 mg/I according to Si02 scale

Alum dose: 21 mg/l
Initial Rapid mix time:!R. = 60s· C/H = 1
Initial Rapid mix Velocity Gradient G (S·l)
slow mix time 150 250 400 550 650 750
to Mt IlMt Mt IlMt Mt IlMt Mt IlMt Mt IlMt Mt IlMt
min mg/l l/mg mg/l l/mg mg/l l/mg mg/l l/mg mg/l l/mg mg/l l/mg
5 10.05 0.099 8.69 0.115 8.40 0.119 7.94 0.126 8.77 0.114 9.98 0.100
10 6.36 0.157 6.25 0.160 4.75 0.210 4.63 0.216 4.69 0.213 6.78 0.148
15 4.25 0.235 4.01 0.249 3.78 0.264 3.34 0.299 3.58 0.279 4.06 0.246
20 3.43 0.291 2.90 0.345 2.86 0.350 2.85 0.351 2.90 0.345 3.40 0.294
25 2.64 0.379 2.62 0.382 2.34 0.428 2.32 0.431 2.32 0.431 2.88 0.347
Table 2. Residual turbidity values as a function of slow mix time under prolonged coagulation
Raw water turbidity: 26.2 EM = 110 mg/l according to Si02 scale

Alum dose: 21 mg/l
Initial Rapid mix time: tR = 360s; C/H = 1
Initial Rapid mix Velocity Gradient G (S·l)
slow mix time 150 250 400 550
to Mt IlMt Mt IlMt Mt IlMt Mt IlMt
min mg/l l/mg mg/l l/mg mg/I l/mg mg/l l/mg
5 6.25 0.160 6.37 0.157 5.81 0.172 7.52 0.133
10 3.94 0.254 3.69 0.271 3.71 0.269 3.92 0.255
15 2.87 0.348 2.69 0.371 2.75 0.364 2.87 0.348
20 2.35 0.425 2.04 0.489 2.23 0.448 2.34 0.427
25 1.77 0.564 1.76 0.567 1.73 0.577 1.99 0.502
Based on the results of the experiments presented in Table 1., curves for the relation
IIMt = f (t slow mixmg) are shown in Figure 4. and respectively the results from Table 2. are
plotted in Figure 5. Similar values like those presented in Table 1. and Table 2. and plotted
curves in Figure 4. and Figure 5. are obtained for all investigated configurations at conditions
of initial rapid mixing and initial water turbidity. Only pan of these results are given as an
illustration, for brevity.
299
0.7
TF60s
M=IIOmgll
Alwn dose=2lmgll
0.6
~ o G=150 lis
~ 0.5 6 G=250 lis
~ o G-400 lis
0.4
\1 G=550 lis
.§
i <>
"*
G=650 lIs
~ 0.3
co G=750 lis
<IS
'a
0.2
~
CI>
~
0.1
0.0
0 5 10 15 20 25 30
Slow mix time - Tslow [min]
Figure 4. Residual turbidity as a function of slow mix time for various Grapid under classical
coagulation
0.7
T.-360s
M=l1Omg11
0.6 Alwn dose=llmgll
~
g 0.5 0 G=150 lis
2. f::, G=250 lis
~ 0.4 0 G=350 lis

:5
.:: \l G-450 lis
gb
~ 0.3
co
<IS
'a
0.2
~
~
0.1
0.0
0 5 10 15 20 25 30
Slow mix time - Tslow [min]
Figure 5. Residual turbidity as a function of slow mix time for various Grapid under prolonged
coagulation
300
2.1
2.0
1.9
<5
~ 1.8
gg
1.7
~
.S!
~
=
1.6
1.5
E3
~
D
TF60s
1.4
tb Alum
M~IIOmg/l
""'"
dose~19mg!l
1.3
-----fr-
i
With cogged perforated baffies CIH~1
== 1.2
---e--
-.-
With simple baffies - C/H~1
« 1.1
--.-
Without baffles - CIH~1
1.0
Without baffies - CIH~,5
0.9
100 200 300 400 500 600 700 800
Rapid mix velocity gradient G [lis]
2.1
2.0
...... 1.9
0
o·
~ 1.8
""
~ 1.7
~ 1.6
e
.g=
g!,
1.5
1.4 TF60s
E3
tb M~34Omgll
""'" 1.3 Alwn dose='39mg1l

D
I
== 1.2 -----fr- With cogged peIfonnated baffies - CIH~1
---e--
-.-
1.1 With simple baffles - CIH~l
--.-
Without baffles - CIH~l
1.0
Without baffles - CIH~,5
0.9
100 200 300 400 500 600 700 800
Figure 6. Apparent aggregation rate as a function of Grapid under classical coagulation and optimal
Alum dose
301
2.4
2.3
8 2.2
0
~ 2.1
EJ
I)()
I)()
~ 2.0
1.9
~
6
'J:>
gb
1.8
1.7
Tr=360s
M~llOmgll
Alum dose~l)mg/l
D
~
I)()
1.6
'"
=
[ 1.5 ----b-
--e-
With cogged perforated baffies - CIH~)
With simple baftles - CIH~)

~ 1.4
-8- Without baftles - CIH~)
1.3
1.2
100 200 300 400 500
Rapid mix velocity gradient G (s-l)
2.3
2.2
.....
0 2.1
0
~
I)() 2.0
E3
I)()
~ 1.9
2
ec: 1.8
.g
'"~
I)()
1.7
Alum
Tr=360s
M~340mgll
dose~)9mgll
D
I)()
I)()
1.6
'"
=
[
1.5 ----b- With cogged perforated baftles - C/H=)
~
1.4 --e- With simple baftles - C/H=)
1.3 -8- Without baftles - CIH=)
1.2
100 200 300 400 500
Figure 7. Apparent aggregation rate as a function of Grapid under prolonged coagulation and reduced
Alum dose
302
3.3 RESULTS AND DISCUSSION
Analysing the results of the apparent aggregation rate presented in Figure 6. and Figure
7. definite conclusions can be made. On the first place the curves ofKapp = f (G) increase at
the beginning, reach the maximum and after that decrease for all studied configurations. This
shows that optimal hydrodynamic conditions exist in all configurations and if they are reached
the process of destabilisation and flocculation proceeds in the best way. The fact that the value
of this maximum is different means that the different configurations have different abilities
concerning insuring the optimal hydrodynamic conditions for the coagulation-flocculation
process. In classical treatment the hydrodynamic conditions for the process of destabilisation
(coagulation) of colloid particles are improving upwards from a vessel with simple square
bottom without baffles to a vessel with simple baffles to a vessel with cogged perforated
baffles. While the difference between the results of the vessel with simple and perforated
cogged baffles is not very big, the difference between the results of the vessel without baffles
and that one with built in baffles is significant and is 27% for the vessel with baffles. The
difference between vessels with ratio CIH = 0.5 and CIH = 1 is almost insignificant. A slight
priority at CIH = 1 exist, but because of the limited vessel dimensions it is not clearly
distinguished.
The above mentioned tendencies can be observed in the experiment with prolonged
initial rapid mixing and with 50% reduced coagulant dosages. Here again better results for
vessels with baffles are obtained and the difference is 12 - 14%. The most evident feature here
is that considerable high values of the apparent aggregation rate are obtained at prolonged
initial rapid mixing and reduced coagulant dosages than at the classical coagulation -
flocculation. The difference reaches 22% for 110 mgll and 11 % for 340 mgll according to SiO z
scale water turbidity.
A possible explanation of the results obtained is that the neutralisation of colloid
particles depends on the number of collisions between them and the coagulant. The
achievement of efficient collisions depends on one hand on the number of particles
participating in the contact and on the other hand on created turbulence vortices and their
quality. Thus in a vessel with baffles, because of the microturbulences, the possibilities of
collision in the whole volume of the vessel are probably better. In other words, with all other
conditions equal, the effectiveness of the process does not depend only on the energy
parameter G defined by Camp, but also on the character and dimensions of caused turbulences
and in that connection on the shape and dimensions ofthe vessel where mixing is taking place.
The analysis of the obtained results can be done also using the turbulent gross grad
coefficient Ct. From equation (1) it can be written that:
303
The apparent aggregation rate Kapp for every particular case of study can be obtained
from the plotted curves IIMt = Mo + Kapp.T, using the Equation (3). In the present research
the apparent aggregation rate value is computed using the least square method. The full study
Kapp values as a function of velocity gradient are given in Table 3. and plotted in Figure 6. and
Figure 7.
Table 3. Kapp values for various vessel geometric configurations
Kinetic conditions Geometric configuration

Water turbidity square without baffles square with baffles
and Alum dose CIH=1
Rapid mix Gs- I CIH= 0.5 CIH= 1 simple Perf. cogged
timetR
150 0.0132 0.0139 0.0129 0.0138
M= 110 250 0.0140 0.0144 0.0152 0.0156
mg/l 60 s 400 0.0150 0.0152 0.0177 0.0179
550 0.0156 0.0159 0.0190 0.0195
classic DCI =21 650 0.0152 0.0153 0.0184 0.0186
mg/l 750 0.0131 0.0128 0.0173 0.0167
150 0.0109 0.0115 0.0139 0.0153
coagu- M=340 250 0.0131 0.0133 0.0158 0.0164
lation mg/l 60 s 400 0.0146 0.0148 0.0186 0.0191
550 0.0149 0.0154 0.0199 0.0201
DC2 = 39 650 0.0140 0.0142 0.0188 0.0188
mg/I 750 0.0122 0.0117 0.0170 0.0161
M= 110 ~ 0.0192 0.0214 0.0226
mg/l 360 s 250 0.Q208 0.0232 0.0235
-
prolonged Dc= 11
- 350 0.0198 0.0218 0.0218
mg1 450 0.0182 0.0214 0.0192
coagu- M=340
- 150 0.0192 0.0219 0.0220
lation mg/l 360 s 250 0.0200 0.0221 0.0221
c--
Dc= 19 0.0189 0.0214 0.0201
~
mg/l 450 0.0174 0.0189 0.0188
304
Wo
C t = -----:-,-:- (4)
124.p.f.N
where the relative power Wo, which is necessary to operate the impeller, in equation (4) is
substituted with the expression Il.G2 = Wo, which is obtained from the well known Camp's
equation G = (WO / 11)112.
Substituting 11 = 1.01375.10.3 N.s.m-2 for water temperature at tests 19° C, p = 1000

kg.m-3 and f= 15.6.10-4 m2, we receive:
Ct = 5.24064.10-{) ~ (5)
Table 4. Values of turbulent gross drag coefficient Ct for different vessel geometric configurations
Geometric configuration of vessel

Rapid mix velocity square with simple square with
gradient square without baftles baftles perforated cogged
baffles
G (S-I) N C/H= 0.5 N C/H= I N C/H= 1 N C/H= 1
230 196 0.00795 177 0.01080 143 0.02047 149 0.01810
250 208 0.00786 187 0.01082 151 0.02057 158 0.01794
500 331 0.00780 296 0.01080 240 0.02047 251 0.01790
550 352 0.00785 316 0.01085 256 0.02040 266 0.01819
In Table 4. the values of the coefficient Ct are presented for the four investigated
configurations and optimal G values. In this computation we have used the function G = f (N)
which we have obtained in our previous studies, the results of which are already published.
It is obvious that at one and the same velocity gradient, the turbulent drag coefficient
Ct increases in the following sequence: vessel without baffles with CIH = 0.5; vessel without
baftles with CIH = 1.0; vessel with perforated cogged baftles and CIH = 1.0; vessel with simple
baffles and CIH = 1.0.
Certain contradictions can be noticed at once: according to the Camp and Conklin
method the coefficient Ct has its maximum value at a vessel with simple baftles, while using the
305
method ofVrale and Jorden, the apparent aggregation rate has its maximum for the vessel with
perforated cogged baftles. A contradiction exists also with the results obtained by Mhaisalkar
et al. (1991), where the coefficient Kapp is proportional to Ct. In our opinion this is probably
correct only for a certain correlation between the baftles' width and bottom dimensions, e.g.
for Bw/C < 0.10. We consider that the proportion Bw/C must not be increased without
restriction and it would be better the baftles to be perforated and cogged at Bw/C 2! 0.10. In
this way dead zone generation behind the baftles is avoided and micro-turbulences are created
which favours the process of destabilisation and agglomeration of the particles.
3.4. CONCLUSIONS
The experimental results of this study lead to the following conclusions. The
parameters of rapid mixing, velocity gradient, initial mixing duration and geometrical
configuration of vessels, influence the effectiveness of coagulation process. Their optimal
combination depends on the coagulant dosage at a definite water turbidity. The experimental
results confirm once again the conclusion made by us (Kalinkov and Monzo, 1992) that it is
possible at prolonged initial mixing to reduce significantly the necessary coagulant dose
without decreasing the water clarifYing effect.
The most successful parameter for functional assessment of the vessel effectiveness
regarding the creation conditions for particle destabilisation turns out to be that proposed by
the Vrale and Jorden apparent aggregation rate. Assessing the mixers by the universal turbulent
drag coefficient to a certain extent is a wrong step, because the contribution of effective
collisions between particles can't be recognised. All results show that vessels with square cross
section and perforated cogged baftles have better results than the other configurations which
were investigated.
3.5. REFERENCES
Camp T.R., Conklin G.F. (1970) Towards a Rational Jar Test for Coagulation. lNew
EngI.Wat. Wks Ass., 84, 325.
306
Kalinkov P., Monzo IL. (1992) Decreasing Coagulant Dose by Prolonging Initial Rapid
Mixing Time. Annual report of University of Architecture, Civil Engineering and
Geodesy, vol. XXXVI, scroll X.
Mhaisalkar VA, Paramasivam R, Bhole A.G. (1991) Optimising Physical Parameters of
Rapid Mix Design for Coagulation - Flocculation of Turbid Waters. Wat.Res., 25, 1,43 -
52.
Packham RF. (1962) The Coagulation Process: Effect of pH and the Nature of the Turidity.
J.appI.Chem., 12, 556 - 564.
Vrale L., Jorden RM. (1971) Rapid Mixing in Water Treatment. J.Am.Wat.Wks Ass., 63, 52
- 58.
4. Study of Floc-Blanket Clarifier, Combined with
Lamella Settlement under the Condition of
Prolonged Coagulation
Peter Kalinkov
Jean Louis Monzo
Higher Institute of Architecture
and Civil Engineering
1 Christo Smirnenski, 1421 Sofia
BULGARIA
4.1. INTRODUCTION
Water clarification working on the principles of contact flocculation is well known from
60 - 70 years ago. Since then different kinds of clarifiers have been constructed and used,
where the separation of pre-destabilised colloidal particles from water is effected by the floc-
blanket of previously formed flocs. Setting up and the quality of the floc blanket in these
facilities working on the principle of hydraulic fluidization, depend on different factors as well
as on coagulant dose, quality of formed flocs, upwards stream velocity in the floc - blanket
zone, etc. (A3epBhep, 1952, MHHM and KaCTaJIhcKHA, 1962; Tesarik, 1967, Miller and
West, 1968). Recently, lamellas are set up above the floc-blanket. It is considered that the floc-
blanket is stabilised and high treatment effect is achieved at a considerable increased upward
flow velocity, about 1.6 up to 1.8 mm/s, while in common conditions it is 0.5 up to 0.9 mm/s.
According to the data in Marin Galvin (1992), fixing up lamellas gives the opportunity of
decreasing the coagulant dose by 14% in comparison with facilities without them.
Using the results of our previous works (Kalinkov and Monzo, 1992, Monzo and
Kalinkov, 1989) concerning the possible decrease of the coagulant dose by up to 50010 by
means of mechanical mixers with prolonged up to 6 min initial intensive mixing, we
investigated the behaviour of floc - blanket clarifier and lamella settling, working at the above
conditions.
NATO AS! Series, Partnership Sub·Series, 2. Environment - Vol. 15

Edited by C. Maksimovi~, F. Calomino, and 1. Snoxell
© Springer·Veriag Berlin Heidelberg 1996
308
4.1.1. Experimental Set Up and Procedure
For this purpose we used the equipment shown in Figure I.
I. Reservoir for raw water

2. Centrifugal pump
3. Raw water batcher
4. Coagulant batcher
5. Mechanical mixer
Sa. Impeller
5b. Electric motor for
continuous current
6. Flocculant batcher
7. Clarifier
7a. Floc - blanket
Q)
7b. Laminar settlement zone
® 7c. Sludge compacting zone
L - -- - -....!==jr -' 8. Funnel for clarified water
Figure 1. Pilot plant
The investigations are carried out with water, which is turbid with freshly taken sludge
from the Pancharevo lake (near Sofia, Bulgaria). Some of the water quality parameters are
shown in Table I., and in Figure 2. settlement curves at natural conditions without coagulant
are shown.
Table 1. Physico - chemical parameters ofraw water
Turbidity Colour Oxidiza-

mg/l acc. to degrees at Tempe- pH Alkalinity Hardness bility mg/l
Si02 scale Rubluov's rature °C mgeq/l HO O2
scale
50 10 13 ± I 6.8 ± 0.1 1.38 2.35 1.4 - 1.5
100 10 13 ± 1 6.8± 0.1 1.40 2.40 1.4 - 1.5
200 15 13 ± 1 6.8 ± 0.1 1.40 2.40 1.4-1.5
300 15 13 ± I 6.8 ± 0.1 1.40 2.40 1.4 - 1.5
309
E [%) 50
40
30
20 --Jif- M=300mg/l according to Si02 scale

-a- M=200mgll
10 -6:--- M=lOOmgll
--e- M=50mg/l
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Time settlement - t [h]
Figure 2. Curves of water settlement at natural conditions without coagulant
At the time of the tests, the water quantity passing through the mechanical mixer was
regulated by a proportioning device for raw water 3, and the residence time by the rate of
filling up vessel 5. For coagulant, a 0.25% solution of Ah(S04)3 is used, dosed by the
proportioning device 4.
A period of about seven hours is necessary for the normal work of the installation after
its start. The effect of water clarification and discolouration is defined by the Expression (1):
M-M
E == 0 .100% (1)
Mo
c -c
E=~.100%
o
where: MoandCo Turbidity and colour of raw water respectively in mg/l

according to SiOz scale and in degrees ofRubluov's scale,
MandC Turbidity and colour of clarified water respectively in mg/l
according to SiOz scale and in degrees ofRubluov's scale.
310
The optimum values of the necessary coagulant quantity defined by the classical
method with jar - test for raw water turbidity shown in Table 1 are respectively 20; 22.5 and
39 mg/l Ah(S04)3.
All tests are carried out at a prolonged coagulation with residence time in the mixer
of 6 min, using 50% of the optimum coagulant dose, and velocity gradient in the range
230-300 S·I.
4.2. EXPERIMENTAL DATA PROCEEDING AND RESULT ANALYSIS
Investigations are carried out at different initial raw water turbidity for examining the
work of the equipment and defining the behaviour of the floc-blanket in a condition of
prolonged initial rapid mixing and reduced coagulant dose. Turbidity and colour of the clarifier
water are controlled at different velocity of water flow through lamellas. The residence time in
the mixer is 6 min in all tests and velocity gradient in the range of 230 - 300 S·I. The results
from the tests are presented in Table 2. and plotted in Figure 3.
100
<;;7-
'0' 90 V v v
~
~
I
80
G--..o 0
=----e------~------------
0 ____D
·1
~
C)
70
~ ,..,
~-------fr----------------6
.!;j ~ e e
'0
Hf.b=l OOcm, without flocculant
'S 60 M=300mgll, Alum dose=20mgll
!
~ 50
M=200mgll, Alum dose=ISmg/1
M=IOOmgll, Alum dose=llmgll
M= sOmgll, Alum dose=lOmgll
40
2 3 4 5 6 7 8 9 10
Water velocity through lamellas - V [mm/s]
Figure 3. Clarifying effect as a function of water velocity through lamellas

311
Table 2. Effect of clarification at different water velocity through lamellas and floc-blanket
height 1.00 m.
Water velocity Water Turbidity, mg/l according to Si02 scale, to = 13 ± 1°C

M0=50mgll MO=100mg/l M0=200mgll M0=300mgll
through lamellas Alum dose = Alum dose = Alum dose = Alum dose =
IOmgll lIm gil 15m!¥! 20mg/l
mmls M-mgll E% M-mgll E% M-mg/l E% M-mgll E%
3.05 15.15 69.70 24.95 75.05 30.06 84.97 28.50 90.50
3.66 15.83 68.34 25.02 74.98 32.08 83.96 30.33 89.82
4.57 16.50 67.00 27.05 72.95 34.16 82.98 32.85 89.05
I 6.10 17.22 65.56 28.05 71.95 34.86 82.57 35.31 88.23
I 9.15 18.77 62.46 29.15 70.85 38.08 80.96 39.99 86.67
It can be seen that the clarifYing effect increases with the increased turbidity of the
influent and decreases with velocity increase of water through the lamellas, which is expected.
The fact which draws attention is that the effect of clarification is high, the same as at normal
conditions in spite of the restricted (50%) coagulant dose. These results can be explained in the
following way:
a) Neutralising and following adsorption of destabilised colloidal particles depend on the

number of achieved useful collisions between them and the coagulant. On the other
hand, useful collisions depend on the number of particles participating in the contact
and on the character and quality of occurred turbulences. So in the conditions of
decreased coagulant doses it is necessary to prolong the time of initial rapid mixing for
reaching the same quality of useful collisions (Kalinkov and Monzo, 1992),
b) The hydrodynamic conditions in the mixer and its influence on the process of
coagulation and flocculation has been thoroughly investigated by the authors in a
previous study (Monzo and Kalinkov, 1989). The regime of mixing in the present study
provides the obtained floc particles to have grain structure, high density and low water
consistency,
c) The lamellas provide continuous downward flow of partially compacted large heavy
sludge particles, which directly take part in creating and work of the floc - blanket,
d) The contact coagulation, taking part in the floc - blanket is realised on one hand by
small floc particles with characteristics described in b) and on the other hand by larger
ones, like these in c.).
312
Due to all this a floc - blanket of thick, compact, heavy, large flocs is formed, which
provide effective flocculation regardless of decreased coagulant doses and low influent
turbidity.
The test results of the effect of decolouration at a velocity through the lamellas of3.05
mmls are shown in Table 3.
Table 3. The effect of decolouration at velocity through lamellas on.05 nun/s
RAW WATER CLARIFIED WATER

Colour
Turbidity, Colour Degrees at
mgII degrees at Alkalinity Hardness Rublouv's Alkalinity Hardness
according Rublouv's pH mgeqll If' scale E% pH mgeq/l If'
to Si02 scale
scale
50 10 6.8 ± 0.1 1.38 2.35 <5 >50 6.65 0.88 - 1.05 2.35
100 10 6.8 ± 0.1 1.40 2.40 <5 >50 6.64 0.88 - 1.05 2.35
200 15 6.8 ± 0.1 1.40 2.40 <5 > 60 6.61 0.88 - 1.05 2.35
300 15 6.8 ± 0.1 1.40 2.40 <5 >60 6.58 0.88 - 1.05 2.35
4.3. STUDY THE INFLUENCE OF FLOCCULANT ADDITION ON THE

CLARIFICATION EFFECT
It is well known that a flocculant is added to improve the coagulation-flocculation

process at low temperatures and low turbidity.
A series of tests were carried out to answer the question - what will be the influence of
flocculant addition in the specific working conditions of the equipment and at prolonged
mixing and decreased coagulant doses.
The studies were carried out at two values of water turbidity, respectively 50 and 100
mgll according to Si02 scale and temperature 14 DC. As a flocculant a 0.05% solution of active
silicon acid is used.
It should be noticed without describing the tests in detail, that the floc - blanket could
hardly be held in fluidized condition, using doses higher than 0.5 mgll.
313
That is why regardless of the optimum flocculant dose of 1.2 mgll for Mo = 50 mgll
and 0.8 mgll for Mo = 100 mgll according to Si02 scale at jar-test, the tests were conducted
with a dose of 0.25 mgll. Two working conditions of the equipment are studied:
a) At existence of 1.0 m floc - blanket height,
b) Without floc - blanket.
100 Without flocculant With flocculant
- 6. -
--.-
M=IOOrng/l,Alumdose=l1mgil - . . - M=IOOmgil,Alumdose=llmgil
90
-A-
-G - M=50mg1l, Alum dose=IOmgil -. - M=50mg1l, Alum dose=l Omgil
80
70
-e-
----
: : : : : : = :: :: - - ~
40
30
20
10
2 3 4 5 6 7 8 9 10
Water velocity through lamellas - V [mm/s]
Figure 4. Clarifying effect vs. water velocity through lamellas

314
The results are shown in Figure 4. It can be seen that the influence of the flocculant is
significant when the equipment works without a floc - blanket. It is obvious that the quantities
of the floc-blanket formed at the conditions described above don't deserve flocculant addition.
There is no effect when adding flocculant with a floc - blanket in existence. It is possible of
course that flocculant addition will have significant effect at lower influent water temperatures
than these at which the tests were conducted.
4.4. STUDY OF INFLUENCE OF FLOC - BLANKET HEIGHT OVER

CLARIFICATION EFFECT
A series of tests were conducted to investigate the influence of floc - blanket height
over the clarification effect in concrete conditions and prolonged initial rapid mixing and
decreased coagulant doses.
100
90
7:- : : :
v
80 0
~ 70
w, 0
t)
~
Q) 60
""
.5
.Q
.l;j 50
0 v='3.0Smmls
40 V M~300mgll, Alum dose=20mgll
0 M~20Omgll, Alum dos~lSmgll
30 D M~lOOmgll,Alum dose~llmgll
0 M~SOmgll, Alum dose~lOmgll
20
0 10 20 30 40 50 60 70 80 90 100
Floc-blanket height - Hfb [em)
Figure 5. ClarifYing effect vs. height of the floc -blanket

315
The study was carried out at different initial turbidities of raw water, constant velocity
in lamellas, equal to 3.05 mm1s and different heights offloc - blanket.
The results are shown in Figure 5. It can be seen that at all values of turbidity the effect
of clarification increases with the increase in the floc - blanket height.
This increase is considerable at a height from 0.0 to 0.60 m, and after that the process
diminishes. In practice, the increase of the floc - blanket height stops its influence over the
clarification effect at 1.0 m for lower turbidity.
This considerable decrease of the necessary floc - blanket height from 1.5 to 2.5 m for
ordinary clarifiers to 1.0 m in the concrete case can be explained again by reference to the floc
- blancket characteristics described previously.
4.5. COHESION COEFFICIENT ASSESSMENT OF THE FLOC PARTICIPATING

IN FLOC - BLANKET SET UP
In order to explain the excellent results obtained so far, the tests were carried on by
determining the cohesion coefficient of the floc, which participate in floc - blanket setup, as
well as assessment of the concentration of compacted sludge in sthe ludge settler.
In water treatment the cohesion coefficient is defined as the maximum ascending
velocity in a zone of floc - blanket, which causes its tearing and dragging out. In our studies it
is defined by that described in the Degremont (1989) method and Equation (2):
(2)
where: Ka cohesion coefficient,

V sludge volume at flogging,
Va initial sludge volume defined from the curve.
316
The scheme of conducting the test is presented in Figure 6.
250_ 1 - Measuring cylinder

200
2 - Funnel
100 3 - Measuring cylinder for adding water

50 :.:.:
.',
" ,
Figure 6. Scheme of conducting the test
Into a measuring cylinder with a volume of 250 ml about 50 ml fresh sludge from the
upper third of the floc - blanket is added. As a result of the ascending water velocity, which is
increased in vessell, the sludge is flogged. This flogging is relatively proportional to the
velocity of the upward flow. It is defined by the equation:
A
V.=- (3)
I t
where: A = 100 mm and it is a height corresponding to a volume of 100 ml in the

measuring cylinder,
the time for regular water inflow through the funnel in vessell.
By changing the time t different values of flogged volume Vi are assessed. Using the
graph, Equation (2) is a straight line, where Va is a segment from the abscissa axis and the
cohesion coefficient is an ordinate of a point from it with abscissa 2Vo.
317
v [mmls) 8
M = 50 rngII according to Si02 scale
7
o _ Without flocculant
V
v = 3.289 (57.56 - 1) ..
/
•
0
6
Vo = 57.56 rnl
• - With flocculant
5 V
v' = 3.619 (-5- - 1)
8.87
4 Vo' =58.87 rnl
/ I
3 /
o
Kso = 3.289 nun/s
2
K'so = 3.619 nun/s
0
0 50 100 150 200
V[ml)
v [mmls] 8
M = 100 rngll according to Si02 scale
o _ Without flocculant /
7 e
V /
v= 3.416 (54.10 - 1) /e 0
e/ 0
6 Vo = 54.10 rnl
• - With flocculant /0
5 V eO
= 3.746 (53.47
/
v' - 1)
4 Vo' = 53.47 rnl 0
3
KlOo = 3.416 nun/s
2
K' \00 = 3.746 nun/s
0
0 50 100 150 200
V[ml)
Figure 7. Determining the cohesion coefficient of sludge

318
v [rnmIs] 7
6 M = 200 mg/l according to Si02 scale

V
5
v=3.596 (64.21 -1)
Vo=64.2l ml
4
3
I
o
IK200 = 3.596 nun/s
2 I
o ~-----------,--~-------,------~----,-----------,
o 50 100 150 200
V [ml]
v [mm/s] 7
6 M = 300 mg/l according to Si0 2 scale

o
V
v = 3.830 (65.25 - 1)
5
Vo= 65.25 ml
4 o
o
2 o
K300 = 3.830 nun/s
o ~----------,---~------,-------~--,-----------,
o 50 100 150 200
V [ml]
Figure 7. Determining the cohesion coefficient of sludge (continued)

319
The studies were conducted for water with initial turbidity 50 and 100 mgll according
to Si0 2 scale with and without using flocculant; and 200 mgll and 300 mgll without flocculant.
On the basis of received data a curve v = f (Vi) is plotted, and using regression analysis the
equations of the curves are defined, in respect to the cohesion coefficient. The results are
presented in Figure 7.
The cohesion coefficient varies from 3.29 for 50 mgll turbidity without using flocculant
up to 3.83 for influent water turbidity 300 mgll. The values are considerably higher than the
cohesion coefficient for a clarifier with a floc - blanket, working at normal conditions.
According to Degremont (1989), for equipment operating efficiently, it is in the range of 0.8 to
1.2 mm/s. In practice this means that clarifiers with a floc - blanket working in the conditions
of the test could be dimensioned for considerably higher velocities in floc blanket zone, like 3.0
up to 3.5 mmls.
The investigations confirm once again our conviction, that the flocs obtained after a
prolonged phase of intensive initial mixing and contact coagulation in condition of floc-blanket
with sludge sliding down from lamellas are extremely compact and heavy, ready for quick and
effective separation. The results prove once again, that in these conditions, flocculant addition
is not necessary for increasing floc density.
With these tests the minimum velocities of ascending flow have also been investigated,
and these still manage to sustain the sludge layer in suspended condition.
4.6. DEFINING THE CONCENTRATION OF COMPACTED SLUDGE
Defining the concentration of the compacted sludge was performed as samples offloc-
blanket were left at rest after period t, their concentration of suspended particles was defined,
using well known methods (KOMHTeT no TepIITOpHaJIHO H CeJIHWHO YCTPOl1CTBO npH MC
KXHIIHBKC "BO~OKaHaJInpOeKT", 1987). The conducted test results are presented in
Figure 8. It can be seen that a residence time of 5 hours is enough for relatively full sludge
compaction and in the same time the concentration is several times higher than sludge in
traditional clarifiers.
In Table 4. the comparative data of sludge concentration, max. and min. ascending
velocities in the floc - blanket zone are presented. It was obtained in this and other studies for
equipment working at normal conditions.
320
~
('<)
100
8
~
OIl
8 90
0
0
:::; 80
~
Q
0
OIl 70
""~
""g
~
60
p.
80 50
....0
0
40 M~300mgll according to Si02 scale
..
c:
..:::
0
30
M~200mg!1
l:l
c:
Q
M~IOOmgll
0
c: 20
M~50mgll
8
Q
10
...~
<
Q
0
0 2 3 4 5 6
Compact time - t [h)
Figure 8. Sludge concentration vs. compact time
Table 4. Comparison of same dimensioning parameters for a floc-blanket clarifier
Water Floc - blanket clarifier working at normal Investigate equipment (test data)
conditions (according to literature data)
turbidity Dimensioning velocity Average Dimensioning velocity Average
mg/l rrun/s concentrat. of rrun/s concentrat. of
compacted compacted
V min Vrnax sludge after 24 Vrnin Vrnax sludge after
hours 5 hours
50 to \00 0.40 to 0.60 1.40 to 1.60 15000 to 1.46 to 1.72 3.24 to 3.42 75000 to
40000 80000
100 to 400 0.60 to 0.80 1.60 to 1.80 40000 to 1.72 to 1.86 3.60 to 3.83 80000 to
60000 \00000
321
4.7. CONCLUSIONS
The investigations confirm the possibility that equipment for water clarification works
very well even when the coagulant dose is reduced by 50% of the optimum and the initial
intensive mixing is prolonged by about 6 minutes. Especially, clarifiers with a floc-blanket,
combined with lamellas settlement and working in process of the above conditions result in the
formation of thick compact, large, heavy flocs which enable a quick and effective coagulation-
flocculation. This guarantees on one hand a high clarification effect and on the other results in
equipment volume, and lower height of floc blanket, with increased water velocity; in addition
sludge is more concentrated and the sludge compaction zone consequently smaller.
4.8. REFERENCES
A3epBbep C.x. (1952) Bop;oCHa6:lKeHHe H :lKeJIe30p;OpO:lKHOM TpaHCnOpTe.

TpaHC:lKeJIp;0pH3p;aT. MocKBa.
Degremont (1989) Metmento Technique de l'eau. Neuvieme edition, Tome 1.
Kalinkov P., Monzo lL. (1992) Decreasing Coagulant Dose by Prolonging Initial Rapid
Mixing Time. Annual report of University of Architecture, Civil Engineering and
Geodesy, vol.XXXVI, scroll X.
KOMHTeT no TepHTOpHaJIHO H ceJIHllIHO yCTpOACTBO npH MC KXlfl1MBKC
"Bop;oKaHaJInpoeKT" (1987) C60PHHK OT MeTop;HKH 3a H3CJIep;BaHe Ha
OTnap;'b'lHH BO)l;H H yraAKH. C0<pHSI.
Marin Galvin R. (1992) Lamella Clarification in Floc Blanket Decanters - A Case Study.
lWat. SRT. - Aqua, 41, I, pp. 28 - 32.
Miller D., West J. (1968) Pilot Plant Studies of Floc Blanket Clarification.
lAmWat.Wks.Assoc., 60, 2, pp. 154 - 164.
MHHp;IJ; M., KaCTaJIbCKHA A.A. (1962) IIop;rOToBCKa BO)l;bI P;JISI nHTHbeBOrO H
npOMHIIlJIeHOrO Bop;OCHa6:lKeHHSI. MocKBa.
Monzo lL., Kalinkov P. (1995) Influence of Geometrical Configuration of Mixer on the
Effectiveness ofInitial Rapid Mixing in Water Treatment With Alum. Annual report of
University of Architecture, Civil Engineering and Geodesy.
Tesarik I. (1967) Flow in Sludge-Blanket Clarifiers. lSanit.Engng Div., Proc,
Am.Soc.Civ.Eng., 93, 6, pp. 105 - 120.
Chapter V
Diagnosis and Rehabilitation

of Existing Systems
Chapter V
Diagnosis and Rehabilitation of Existing Systems
ffiTRODUCTORYSTATEMENT
Diagnosis, measurement, data acquIsitIOn and processing as well as operational

management of Water Supply Systems (WSSs) have undergone significant changes during the
last two decades under direct influence of information technologies.
The classical sensors and instruments for measurements of flow, levels, velocities and
pressure have gained in usefulness when converted from stand alone devices to an element in
an integrated system for operational management. In this way a multipurpose use of the data
obtained by measurements becomes more feasible for both on-line action and tuning current
control devices.
This Chapter consists of two Sections. Some basic principles of flow measurement and
their application in Water Supply Systems for monitoring water balance and for quantification
of losses are shown in the first Section. General principles of integrating these measurements
into system for data acquisition, processing and operational management are also discussed.
The second Section presents these principles in a more detailed way using examples
from Wessex Water in the UK. Since the ifldividual consumption on the UK is not metered, the
problem of assessment of the key element of water balance is discussed for this advanced
system.
1. Measurement Methods for Diagnosis and
Rehabilitation of Existing Systems
Cedo Maksimovic
IRTCUW, Insitute of Hydraulic Engineering
YUGOSLAVIA
1.1. INTRODUCTION
Existing water supply systems are under .continuous demand to supply an adequate
amount of water of a required quality (health and technical aspect) at the required pressure.
These systems are in continuous need of improvment in their operational management in order
to meet new regulations, reduce costs and increase levels of service. Modem technology
provides the means for such improvements either by the introduction of high level of informatic
support after rehabilitation of the existing, aged systems, or by introduCing appropriate new
technologies into the newly designed and built systems. The computerized operational
management systems have to be based on a number of components such as measurement,
telemetric transmission, data acquisition, relational data base systems, data processing and
hydrodynamic and water quality modeling. It is of upmost importance that the characteristics
of the existing systems are known. This can be achieved by a proper combination of field
measurements and use of simulation models. Thus, the reliable measurements of certain
hydrodynamics quantities such as discharge, velocity, water level, pressure, etc. in addition to
the data on status of flow regulation and similar elements in the systems are the major source
of information on both the characteristics and conditions in existing systems.
Some of the major tasks of operational management of water supply systems are:-
determination of the parameters to be measured, number and placing of sensors and regulating
devices, sampling frequency and duration of measurements as well as the level of local and
central data processing and/or operational control.

Edited by t. Maksimovii:, F. Calomino, and J. Snoxell
328
1.2 SCOPE
The scope of the present chapter is to present the tools and methods by which a reliable
set of information concerning the existing system can be obtained to serve either for its
redesign and rehabilitation or for the operational management. Since the characteristics of
elements of a water supply system change over time (the roughness of pipes increases, the
capacity of pumps decreases, leakage from pipes increases, etc.) the diagnostic procedure
should be applied to the existing systems regularly even if they will not undergo any
rehabilitation. Depending on the scope, measurements can be continuous (sensors remain built-
-in the system) or occasional, ie. sensors are occasionally brought to a particular location for
measurements of the "snapshot" type. The types of measurements expressed in terms of data
sampling are shown in Figure 1.
_
_ ...................
continuous
1't?r::1.
't2b11~ i+l j+m
Figure 1. Three major types of data sampling - recording determining three types of measurements
The first type of data sampling is applicable in permanent measuring stations where a
sensor (flow meter, level meter, pressure transducer, etc.) is placed permanently in the pipes,
reservoir, pumping station, etc. and readings are taken continuously at regular or irregular time
intervals.
329
The second type is usually applied when the single measurement is performed at the
particular location. In this case it is usually necessary to prepare the site before that particular
measurement (drilling, mounting of a sensor, calibration, etc.).
The third type of measurement (occasional data sampling) is applied at the locations
which are accessible and where a portable instrument can be brought and easily mounted for ad
hoc measurement and data sampling.
Table 1. Groups of measurements forWSS's depending on the phase in the life span
of a particular part of the system
Phase Group of measurements

Initial field survey Hydrological, meteorological,
hydrodinamic, geologic, etc.
(water balance)
Design of a new system Concrete location, levels, capacity source,
etc.
Construction Pipe roughness (capacity), sealing
(tightness)
Commisioning and trial work Sensor calibration, tightness
Normal operation Pressure, levels, discharge;
for operation management,
water losses, water quality·
Rehabilitation and expansion Reduced capacity of mains, leakage,
pump characteristics
I Shut-of Final capacity of sources and mains
Regardless to the type of measurement shown in Table 1., one can again distinguish
between two groups of measurements serving two different goals ie:
• indication ie. the data obtained serve only as an indication of the situation (normal or
emergency) in the system, for example main bursts, high night consumption, excessive
night pressure etc.; several water companies in the UK follow this concept (described in
the next Section),
• quantification in which the reliable values of the measured parameter are requested.
330
In what follows, the preference will be given to the second group of measurements,
although they can be used for the first group as well.
In Table 1. the groups of measurements are shown in the function of the place in the
life span of water supply system.
Water supply systems have a finite life span. Depending on its age, various types of
measurements are performed in the system: before design, in the design phase, during
construction and commissioning, during their trial and full application periods, before
rehabilitation, and possibly before shut down. Most of the measurements presented here are
applicable in all phases of the life span ofa WSS.
1.3. MEASUREMENTS FOR WATER BALANCE IN WSS
1.3.1. General Water Balance
Water flows in water supply systems are mostly under pressure. However there are
some exceptions such as reservoirs, wells for pumping of ground water, in treatment plants etc.
where water flows with a free surface. This fact determines the type of sensors and measuring
principles since they normally differ for systems under pressure and those with free surface.
Before drawing conclusions about the system's performance, one has to have the access to the
reliable results of measurements. The reliability is affected not only by the accuracy of
measuring systems (sensors, amplifiers, data transition and recording equipment) but also by
the way in which the measurement in performed i.e. where sensors are placed, how to interpret
the value of an intermediate quantity, how to average and finally process the signal etc. It also
depends on the level of knowledge of the physical characteristics of the system (where pipes
are, their characteristi.cs, existence of connections, valves etc.) and upon the status of flow
regulation devices (open, closed) etc. One of the major tasks of measurements in the systems
is to provide information on water balance (input, own consumption, leakage, consumption by
metered consumers, non-accounted for consumption etc.). For this purpose reliable flow
measurements are essential (Maksimovic, 1993).
The problems of measurement and data acquisition are different in the simple systems
which just start to introduce informatic technology and in the complex systems which have a
long experience in measurements and data processing such as described by Snoxell in the next
Section.
331
For the beginning, let us consider a very simple water supply system consisting of one
source (intake), one treatment plant, one pumping station, network and one reservoir
connected by one main conduit with the network (Radojkovic et a1., 1986)..
Reservoir
Treatment
Pumping station ®
@/~
Pumping main Distribution
Transport of
raw water Q4
Figure 2. Scheme of a simple water supply system
.~
i'l
I+
~ 11)
~~
t:~
o~
~
l
~~
~
Q8!:
.
0
Q2 Q3 Q4 e Qs i:
I =
·S·o
::s
rn
rn
.~ e
11)
rn
...... o.j e "0

B
el1) o~ B.~
c..,
Q2-Q3 alll
~~ ~.g
Pumping
Q4-Q5 leakage
Be ~ ~
.a.§ efficiency ~ e
",CD 1;:1 ~ .1;:1 11 Q5-Q6
0'E
1;:1 ::s
Ze .sf! ,st;:l ti:.§ .so
Figure 3. Positioning of flow meters for water balance
In order to keep track of water balance, at least six or seven various positions of
discharge measuring equipment has to be considered. Taking, as an example, the system
shown in Figure 2. the overall water balance can be made if the following quantities are
monitored (Figure 3.).
332
• natural flow of raw water QI or its integral over a given time periods (day, month, year),
through the water course, or inlfow into reservoir, through an underground aquifer etc.,
• withdrawal (abstraction) of raw water Q2 from a natural resource,
• discharge at the inlet to treatment plant Q3 which can differ from Q2 and QI because of
raw water losses during transportation, pumping, storage etc.,
• discharge of the treated water at the beginning of pumping - trunk main Qs and at the end
of trunk main - Q6 (Q6<QS<Q4),
• discharge in various pipes in the network or its closed parts (districts),
• discharge to or from the distribution reservoir.
In this case it is relatively easy to measure the amount of water delivered into the
system by installing a flow meter (reliable, regularly calibrated and maintained) downstream
from the pump - at the beginning of the water main (ie. outlet of the treatment plant). Placing
the same meter at the downstream end of the main introduces a higher uncertainty because the
information on water losses is lost, that is the amount of water that leaks from the main or is
withdrawn by illegal connections (if any), between the treatment plant and the network.
The uncertainty is reduced if flow meters are placed on both ends of the pipe. For
example, one can measure both Q4 and Qs provided that both meters are correct. In that case
the difference in discharge (or in volume obtained by integration of discharge) represents
leakage Qosses) and illegal consumption. A similar problem occurs if one tries to determine
leakage from a pipe ~ city networks.
Usually it is very difficult to find out any of these data with a satisfactory reliability.
One often finds that certain figures are available by WSS staff, but, if one tries to analyse the
source of data (methods of measurement, data recording and analysis) and their accuracy, one
is usually faced with the fact that reliable data do not exist. One of the frequently available
information is the record on the duration of pumping, presuming that the pump (or pumping
station) pumps water with a constant discharge. Simple measurements of the pumping head
and power reveals that pumps often operate out of the range of favorable efficiency and that
determination of water balance suffers from considerable errors (Maksimovic, 1993).
333
1.4. METHODS OF FLOW MEASUREMENTS IN WSS
There are many books and standards dealing with flow measurements (for example).
International (ISO) standard No. ISO/TR 3313 and others also deal with this issue. The
discharge measurement methods can be classified in four groups depending on the principle
applied as follows:
a) Integral flow measurements by the intermediate measurement of energy loss or

drop of piezometric head caused by construction or other obstacle in the flow
Ventury meter is an example of such a principle offlow measurement (Figure 4.).
VJ
Perfect fluid
----- - - £2 - --_-_-_-_-_- - - - - -
""--th,--- -- _- .: _- - - - - J ---
~
Total head forreal f1~d- - - - - -
E3
2g
p vj
a)
3 19
v3
2g
P3
a) pg
b) P2 3
P2
2 pg
Vi
_ +-------\,h--------~___td
d3
V3
Z3
Z=o
b)
Figure 4. Venturi constriction as a flow measuring device
One of the disadvantages of this type of flow/measurement is that the obstacle in flow
causes permanent energy loss in the pipe. The energy loss can be expressed in terms of
piezometric head difference ~. The flow rate - discharge can be obtained if ~ is measured.
334
100
90 _ :rI Orifice with square edge
80 1":::::-
70
t~·lOO
-lW
...... ::r-,.- ........
.......... ................
60 Nozzle / .......
.......... ,"""-.
"
50
"-
"
40
30
~ kExit cone 15
20
Vi
entun. '-.,
~
10
Exit cone>----- f--
'--1-- '7
9 UVT (Unive~ Venturi Tube)
8
7
'\.- =---1
A
~---~~
...............
'"
6 ,/
'\..
'" ~~
5 .......
Low Loss
4 "'- "-
3
c-J~k, II ~ -D~ISOf
2 r-- d1
l' 7° T
----.--- -------.
i d2 l' c:::>
anUVT
_._._.-
--dii''i.--- --tilT ------------Diffu~;--------------
d
~
1
o
I
0.1
I
0.2
I
0.3 0.4
I I
0.5
I
0.6
I0.7 0.8
Penn anent energy loss for various Venturi type flow meter
Figure 5. Permanent energy loss for various ventury orifice and nozzle types flow meters
The normalized energy loss (Lilli I1pJpg) is shown in Figure 5. for several types of
Venturi flow meters. For comparison, the same data is shown for orifices with square edge and
for a nozzle flow meter. This function is to be taken into account for selection of a new flow
335
meter or upgrade of the exsisting one. In the system in which a big difference in water
consumption in various seasons is pronounced it may even be necessary to use different
obstacles for different seasons.
b) Integral flow measurement by active elements such as electromagnetic flow

meters
The principle of flow measurement by this method is shown in Figure 6. There are
several types of this flow meter: with homogeneous and inhomogeneous magnetic fields etc. A
disadvantage of this flow meters in the past was that they had to be dismounted and carried to
laboratory for calibration purposes. This problem can be solved by using insertion (portable)
flow meters for on-line calibration.
e=e(Q)
a) The conductor is a wire of the length b) i=D pipe diameter. The conductor
e = e(1. vav) is a fluid element of the volume A . dx
A = cross section area
The principle of electromagnetic induction applied to electromagnetic
flowmeter
Figure 6. The principle of electromagnetic induction applied to electromagnetic flow meter

336
c) Intermediate determination of discharge by correlation with average velocity

measurement along a single or multiple path of ultrasound
The principle on which this method is based along with the scheme of an ultrasonic
flow meter are shown in Figure 7a. and 7b.
a) Fluid at rest b) Movable fluid with constant velocity
c»vavcos@
a) Definition of velocities for application of an ultrasonic sensor
Signal
emitedby
~ sensor2
tu
emitedby
~ sensor I
b) Scheme of an flow meter
Figure 7. The principle of an ultrasonic flow meter

337
The following definitions and abbreviations are used in the basic equations:
• travel time of sound in the upstream direction: tlb
• travel time of sound in the downstream direction: trJ,
• average velocity along the path of sound v.",.
These tenns are explained in the Figure 7.a. and 7.b. and are calculated as follows:
upstream time of travel
L d
t =----
U C-Vav cose sin e(c + VSll cose)
downstream time of travel
L d
td -
- C+VSII cose sine(c+vSII cos e)
average velocity
v=u t -td _ _d__

av tutd 2sinecose
d) Indirect determination of discharge by point velocity measurement

(ie. measurement of velocity profile) in a cross section ofa pipe
Figure 8. shows the principle based on point velocity measurement by a miniature two
velocity component electromagnetic flow meter.
In addition to the above mentioned best known types of principles there are also
several other instruments based on various principles of fluid mechanics such as: Corriolis
force, vortex, drag force, Anubar, etc. (Bernard, 1988). Detailed presentation of all these
principles is beyond the scope of the present text. Although, all of the previously mentioned
instruments are well known and well documented, it is surprising how often they are misused
in practice.
338
The major shortcomings and sources errors in practice are caused by the ignorance of
basic principles of fluid mechanics and of flow measurements, or by lack of proper calibration
or by poor maintenance and by an inadequate or incomplete data recording and processing. For
example when a Ventury tube is used for flow measurements, the rating curves are normally
provided by the manufacturer.
Figure 8. The principle of a two velocity components electromagnetic sensor used for flow
measurements
However, application of various types of geometries such as UVT - Universal Ventury

Tube, low loss causing permanent energy losses !:ill (Figure 5.), or differences in piezometric
heads (~) are obtained. Thus it is necessary to perform calibration both after introduction and
in operation.
Traditional techniques for measurement of differences in piezometric head (PI - P2
(Figure 5.» based on mechanical principles can be easily improved by adding a differential
pressure transducer (also shown in Figure 4.). After the meter has been installed, occasional
(ad hoc) recalibration is rarely done in practice, after the meter has been installed. When the
local velocity method is used for flow measurement it is important to determine the distance of
the sensitive element of the sensor (electrodes of a miniature electromagnetic flow meter for
example) from the wall. The velocity distribution in a circular pipe is known to be dependant
on flow regime and pipe roughness as schematized in Figure 9.
339
0.9
tulmarL
ISmooth now regimel
\ l!KoU m tlow re~ ~mel

-
0.8 - ~ ~ -
~Pr :,......- r--
~
~
----- ~f
0.7
~
U max
~ -
0.6 ~
'\.
/
./
0.5
,/ ~ I~
7
Laminar flow Turbulent flow
0.4
10 2 2 4 103 10 4 2 4 105 2 4 106
2000 4000
Figure 9. Ratio of average and maximum flow velocity in a circular pipe as a function of flow regime
and pipe wall roughness
Similarly, the ratio between average and maximum flow velocity changes accordingly.
Positioning of a sensor is normally prescribed or done by the manufacturer.
However, in practice, conditions change during the period of use of a sensor. For
example roughness can change and so the rating curve (ie. function of discharge Q v.s. the
electrical signal e generated by the point velocity) can be shifted or changed. The problem can
be solved by calibration measurements of discharge using an alternative method. Measurement
of velocity field in one or more radial positions can be applied to compute the discharge. To
perform this kind of measurement it is necessary to have an appropriate access to the pipe. The
access to the pipe and a place prepared the calibration is rarely done during the construction.
340
Thus the drilling and mounting of sensors and auxiliary devices has to be done on the
spot before calibration. The modern term "insertion flow meter" is used to denote a portable
device that can be inserted in the pipe where needed. In most cases it is a local velocity sensor,
used for discharge monitoring, required for water balance or leakage detection. An
arrangement for determination of pipe's carrying capacity (ie. hydraulic pipe roughness) is
shown in Figure 10. In addition to flow measurement, flow regulation might be needed in
order to achieve the desired range of flows. Piezometric heads (or the appropriate differences)
have to be measured in two cross sections. Temperature measurements are to be used for
determination of viscosity, thus the following can be determined by the sensor used for flow
measurements:
• flow (discharge) as a function of flow regime and pipe wall rougness,
• roughness and carrying capacity by measurements of piezometric heads and discharge.
vi /:lE!W!
g
vl,
PII 2i
.J----------r=""'-*""'~-....I.. Thcnnomctcr
Z =o
Figure 10. Setup for detennination of the pipe's hydraulic roughness and carrying capacity by
measurements of piezometric heads and discharge
341
1.4.1. Measurements for Determination of Water Losses
In detennination of water losses there are two major tasks (i) finding--spotting leaks
and (ii) quantifYing the amount of water that leaks. In a WSS which has a sufficient number of
reliable flow meters to keep the control over principal elements of water balance (intake,
production, delivery to the network etc.) and in which individual consumers are metered, the
information on water losses should be deducible from the data collected regularly. However,
many factors make this task difficult:
a) in many systems (for example in the UK) individual consumers (households) are not
metered regularly or not at all,
b) where metered, many meters are unreliable, out of order or not functioning properly,
c) consumption for fire fighting, street washing, pipe flushing, is not metered in general,
d) public consumption (fountains for instance) are also not metered.
Thus the determination ofleakage is in many cases arbitrary unless special attention is
paid to improve the accuracy of data.
Losses are monitored by several methods as follows:
a) passive methods in which the leak is spotted only where water appears on the surface,
b) active method in which the location ofleak is detennined by tracking the noise
generated (leak spotters, correlators, etc.),
c) direct flow measurements in a selected main (or in its reach) by two fixed flow meters
(when existing) Figure II.a. or by two insertion flow meters - (Figure lI.b).
d) Detennination of leakage in a part of network by special pumping, monitoring and

control equipment mounted in a vehicle. In this case the vehicle with powerful pumps,
flow and pressure meters, regulators and other required equipment is driven to the part
of the network to be investigated. By closing valves, a defined area (part of network) is
isolated from the rest and water is pumped from a small reservoir on the vehicle (outlet
discharge) to circulate through the isolated sub network and received back to the
vehicle (inlet discharge). All consumptions within the sub network are either
342
disconnected or metered. The difference between the two meters (inlet and outlet)
minus the metered consumptions is accounted as leakage or water loss recorded by the
change of water level in the reservoir. There are also options without the reservoir. In
this case the information on losses are deduced from the changes in pressure or in the
pump load.
e) other methods.
Water losses (leakage Qloss = 7)

Qin a QOII' + Qlo..
6H
QOUI - QQOf1JumptiOl'1 + An:I ' -
t.T
a) Two fi;o(ed (c;o(isting) clectromagnetic flow melcrs
a) Two immersible
flow mctcrs Notebook
vclocity
distribution
b) Two fixcd built-in

flow melcrs
Figure 11. Determination ofleakage from a main and quantification oflosses by two calibrated flow
meters a.) two fixed (existing) flow meters, b) two insertion flow meters - inspection of velocity fields
for determination of discharge
343
In the systems well equipped for leakage detection and reduction, it is argued that it is
economically feasible to perfonn regular leakage spotting annually or even twice a year. The
savings in water losses exceeds the costs involved. In the other systems which do not have a
regular program of a leakage control the losses of the order between 30 and 45% are
considered regular whereas in extreme cases they reach as high as 700/0. Irregular delivery of
water, rationing, frequent epidemics, and inconvenience caused to the consumers are very
frequent in such cases.
Thus, the leakage detection and water losses management should have a high priority
in setting up rules for rehabilitation and modernization of the existing systems.
1.5. MEASUREMENTS FOR PRESSURE CONTROL AND OPERATIONAL

MEASUREMENT
1.5.1. Pressure Control
One of the goals of operational management in WSSs is to secure continuous water

supply with sufficient pressure and of the satisfactory quality to all consumers. Distribution
networks are usually devided in several supply zones depending on ground (terrain)
elevations. However, the dynamic pressure in the network (in each of the zones) varies in time
and in space and it often happens that certain parts of the network are exposed to excessive
pressure either continuously or during some periods of a season or of a day. This excessive
pressure has adverse effect on home appliances, valves, faucets, etc.; it increases danger of
leakage and increases water losses wherever there is a crack. Modem approach in operational
management aims at an efficient pressure control within the network.
Pressure can be controlled within network as a whole or an automatic pressure
reduction can be applied locally. In any case, for pressure control it is necessary to have an on-
line indication of the local or global pressure (Figure 12). Moderately priced pressure reducing
valves with actuators are available on the market.
By application of pressure control measures the following improvements can be

achieved within WSSs:
• reduction of water losses,

344
• reduction of frequency of main bursts,
• reduction of energy consumption,
• more stable pressures in the network,
• reduction of damages on the services.
a) Remote indication for pressure reduction within a WS networks
- ~_-!!ght without pressure reductio on ~

----..,.".____
--~ti -------
.
...~.l!ie~
~Omct·
__ '!I~t'p!e~l!1~tric head .......... !!.c..lJ~d
WIth pressure redueuon --------
varriable
speed pump or
pressure reduction
valve on the upstream transducer for indication of
side of the pipe pressure at the far and of
the pipe
-,
b) Local pressure measurements and reductions
: piezometric head (regulated)

------ ------,
L _____ _
pressure
reduction
valve
actua:or·"~"
:t;. ,,,P.. D
eedb '"'~
~c~ b-lli}$o~
~cC!J'
Figure 12. Global and local pressure control

345
1.5.2. Operational Management
In addition to measurements related to pressure reduction, operational management of

WSS (excluding management of water treatment processes) requires other parameters to be
monitored such as:
• ground water level in wells and in control holes (piezometers),
• water level and in reservoirs and other facilities with free surfaces,
• status of opening of flow modulating valves,
• states and operating range of pumps (pressure differences--pumping heads),
• flow in water mains,
• pressure at control points,
• parameters of water quality (turbidity, chlorine residual etc.).
Depending on the level of automatic control of a WSS, measurement can be

accompanied by a degree of mechanical, automatic or software controlled flow regulation, that
can be at local or at the whole network level (Koelle, 1993).
Recent tendencies in development of operation supervision and operational control rely
on central data acquisition systems integrated with local instruments and data loggers.
Communication system is based on either cables or radio links or public telephone lines.
One of the possible arrangement in handling operational data and commands is shown
in Figure 13.
In developing countries the application of the advanced computer-aided operational
management system should be planned realistically and implemented in steps. A special
attention should be paid to the reliability and accuracy of data being transmitted and stored.
The data obtained in real time or statistically processed time series of particular
measuring spots are stored in data files used later for simulation models and operation models.
These quantities are of a dynamic nature. For application of the above mentioned models the
additional data on network topology are needed. The latter data are almost fixed or change
much slower as new pipes are laid or old ones have been rehabilitated.
346
Sufficient amount of measuring points, positioned in the system, the (reliable) data
collected and transferred to the operator's control rooms is one of the prerequisites for
introduction of the additional functions to the system's operational management.
If-JV~H",,_=--j '=r
IMeasnrements I
Data source
~
Data
I Conversions I ICommnnicationl IProcessorsI
Comman(is
~
~~
Radio or
dedicated UHF
cable
transmission
s Ground water Legend:

~1iI"i!::~--+- level e - An electric
quantity
Q - Discharge
H -Level
e(Q) p - Pressure
1-'1=1 W
Data
c::=::>
s-Drawdown
Q e(p)rJ
Pipeline
(discharge, pressure)
e(Q)
e(p)
Commands
~ Hy drodinamic
e(H) "
.. quantities
e(N) ~ C:::::> Information .. Status (On, Off,
-~....,,.......,~-;.~~ Flow regulation Open,Closed, etc,)
valve ¢=:J Commands
Figure 13. Scheme of a centralized system for operational management of a WSS
Some of the possible functions of integrated operational management are:
• energy consumption optimization,
• leakage and water losses management,
• rehabilitation scheduling,
347
• resources optimization etc.,
• application of GIS technologies.
reliable
calibrated
measurements
SUPPLY
hydrodynamic & water quality models

integrated with real world data
C operatioual management models
energy
operational consumption demand
management management
changes
Figure 14. The data sources, techniques and functions of an integrated

operational management in a WSS
The scheme of an interaction of measurements, telemetric data transmission, CAD &

GIS operational and hydrodynamic modelling with the operational management functions is
shown in Figure 14.
One should distinguish between hydrodynamic model and the operational management
model. The hydrodynamic models, capable of performing detailed simulations are used for
reproduction of measurements and for predictions of future condition on both short term and
long term basis, such as presented in the Chapter VI of this book.
The information obtained are essential for planning and redesign of rehabilitations and
reconstructions, for development and calibration of operational management models.
Operational management models are based on a simplified geometry of network and other
parts of the system and are based on the results of modelling of various scenarios simulated by
348
hydrodynamic models. It has to be robust and rugged, reliable and user friendly. It has to have
been played with in advance in order to elaborate the and facilitate handling emergency
situation as well as regular ones. An example of such a system is given in the next Section by
Snoxell.
1.6. CONCLUSIONS
Introduction and widespread of both simulation and operational management models

require reliable sets of data obtained by measurements. In planning the upgrade of an existing
WSS for introduction of advanced operational management it is necessary to set up priorities.
In the systems with limited funding, such as in developing countries, the recommended
priorities are as follows:
• placing a minimum number offlow meters in order to monitor a general water balance and
for securing their long term observability, detection and quantification of major sources of
losses,
• introduction of advanced simulation models, preceded by reliable documenting of network,
• development of a robust operational management model, based on an achievable sets of

data telemetered or available through a number of data loggers.
1.7. REFERENCES
Bernard C.l (1988) Fluid Flowmetering. The Trade & Technical Press Ltd., England.
Koelle E. (1993) Automation Levels in Water Supply Systems. Measurement and Control
Variables. Water Supply Systesm, (Ed. Cabrera E., Martinez F.), Computational
Mechanics Publications, Southampton Boston, pp. 297-317.
Maksimovic C. (1993) Merenja u hidrotehnici. Measurements in Hydrotechnics (in Serbian).
Faculty of Civil Engineering, Belgrade.
Radojkovic M., Obradovic D., Maksimovic C. (1986) RaCunari u komunalnoj hidrotehnici
Computers in Urban Waters Engineering (in Serbian). Gradjevinska knjiga, Belgrade.
2. Diagnosis and Rehabilitation of Existing Systems
John SnoxeU
WESSEX WATER
BS2 OJQ Bristol
UK
2.t.INTRODUCTION
Over a long period of time all water supply networks deteriorate, and losses increase;
although this is an obvious and fairly simple fact it presents very complex operational and economic
problems for water company management. This Section looks at these questions, how they are
dealt with in the UK and how technology is being applied to solve these problems. The Section
firstly considers the measurement of inputs to the water supply system by looking at the types of
flow metering commonly used and the way in which this data is handled. Calculating a basic water
balance on the network is a fundamental element of management and the paper looks at the ways
that this is achieved in the UK. Leak detection and quantification is currently undergoing a major
review in the UK and the results of that review are outlined in the Section along with the methods
that can be employed to reduce leakage.
The following sections go on to look at ways in which existing system can be upgraded
using a number of techniques, and suggests that a "total integrated" approach offers the best overall
solution where a number' of complementary techniques are used to provide the most cost-effective
package of measures.
2.2. MEASUREMENT OF INPUT TO WATER SUPPLY SYSTEMS
Firstly, we should look at the basic measuring devices used on typical networks for
monitoring inputs to the system at various points.

350
In the UK all water sources have a daily and annual abstraction rate set down in the licence
issued by the UK regulatory body, the National Rivers Authority (NRA). These abstraction rates
are measured by the water company, which reports regularly to the NRA and is also audited by the
Director General of Water on such matters as meter accuracy, calibration etc. Magnetic flowmeters
are now commonly used and these are often telemetered to a central point for data capture,
archiving, monitoring and the production of routine output statistics. This basic measurement of
water produced forms the basis of all other measurements of consumption, leakage and so on. It is
common for modem supply networks to be divided on a geographical basis into demand zones
which can be monitored individually or in groups. Each zone will typically have one or more
sources, together with imports and exports to or from the zone, and thus has a certain
configuration offlow meters which must be carefully controlled and monitored.
Generally, major transfers of water are also measured by magnetic flow meters and
telemetered back to a central point for continuous monitoring.
Wessex Water has 1million customers, supplied from over 120 sources and the territory is
divided into 29 demand zones, monitored by around 450 telemetered flow meters. This monitoring
system is used both for source outputs and zonal demands and can be used to check for major
losses on the trunk main and reservoir system where, in the UK, it is estimated that around 4% of
leakage occurs.
Figure 1. shows part of the Wessex Water territory with zonal boundaries.
Figure 1. Wessex Water - zonal boundaries

351
The distribution network within each zone is further subdivided into districts, called District
Metered Areas (DMA's see Figure 2.), each with between 1000-3000 properties and each metered
by one or more waste meters which are generally of a mechanical type (e.g. Kent Helix). Below
this level are smaller zones, sometimes called control zones, which are genera1ly used to monitor
night flows during step-testing activity (Report No.26,Natioanl Leakage Initiative, Managing
Leakage, 1994). The vast majority of domestic customers in the UK are not metered and this
means that calculations of "unaccounted" for water are, at best, estimates which include a number of
assumptions.
Distriet meter measures

flow Into dlstriets eg 1000-
3000 properties (DMA)
measures total
output ,,
'" ,
-- ____________ 1
,,,
" "
···...
\-r- __
'
·,·'...
··.
,, ' '
, '
~ '\ ....
, "
Figure 2. Typical layout of district metered area (DMA) and control zone
Until fairly recently meter readings from DMA's were collected and analysed manually but it
is now common that these meters are logged by smal~ battery-powered data loggers which are
collected and downloaded to a central point at regular intervals, thus providing a continuous record
of daily flow profiles for each DMA.
352
Wessex has over 350 DMA's monitored by over 600 data loggers and this presents a major
logistical exercise in collecting, collating, analysing and monitoring the data, a task that has to be
largely automated ifit is to be canied out efficiently. There are a number of trends emerging in the
UK to streamline this operation, including "dial-up" loggers which can be interrogated by public
telephone line (including cellular phone networks) or logger data can be downloaded into
convenient points on the company PC network which can now reach remote points inexpensively
via new communication services available (e.g. Integrated Services Digital Network - ISDN).
Thus in the UK we monitor the network at 2 or 3 levels, but do not have the benefit of
customer metering as the final element in the water balance.
2.3. CALCULATING THE WATER BALANCE
The mass of data coming in from different levels in the system presents major problems in
data handling, archiving, analysis and so on, to produce a complete water balance for the network at
regional level. The data is derived from different instruments, is collected in different ways and is
transmitted to different points in the organisation - collating, integrating and analysing the data into
meaningful information is something which Wessex has solved through the development of its
Aquaware software (AquaWare, 1996). This PC-based package reads data from different sources
through standard interfaces from telemetry and data loggers and enables the User to configure flow
meters in any combination to balance inputs, outputs, transfers etc. The package also reads metered
consumption data from corporate billing files and Post Code information on numbers of properties
in particular geographical areas - this is useful in the UK, which is predominantly unmetered.
The system allows the User to reconcile production data from a zone with consumption
data as measured at DMA level; this gives a good indication of losses on the primary system:
pumping mains, reservoir overflows, trunk mains and so on. These losses are generally held to be a
small proportion of the total leakage but can be vel)' significant and, most important, can be very
cost-effective to repair. The Aquaware system also allows the user to monitor complex
configurations ofDMA meters and to monitor night flows from DMA's continuously and to build
up these zonal night flows into divisional and regional data. This is probably the most reliable form
ofleakage monitoring in the UK and can be used to monitor the effectiveness ofleak detection and
repair expenditure and so on. Changes in DMA boundaries and configurations of meters can be
handled easily through the software with changes in metered demand allocation automatically
adjusted.
353
In this way the complete water balance for the company can be computed and monitored
regularly.
Clearly, to assess leakage rates, some estimate of customer consumption has to be made
and a number of studies have been carried out in the UK to this end; while some companies still
report relatively high or low consumptions, most now fall in a band between 130-160 Up/day as the
annual average. This can vary considerably with the weather, however, since it is common practice
to water the garden with the domestic water supply pushing consumption well above 200-300
Up/day.
Customer leakage is a further problem - it is currently included in the specific consumption
figure and is generally held to be around 8% of total production on average in the UK (Water
Metering Trials, 1993).
2.4. LEAK DETECTION AND QUANTIFICATION
Since privatisation of the UK industry in 1989, leakage has become a major concern and the
industry has recently reviewed all of its methodologies and procedures in order to produce realistic
estimates on which to base major investment decisions; demand management is now probably the
most important subject in the UK industry and most companies have embarked on major initiatives
to suppress demand. Traditionally, the UK industry has been "resource-led"; it is now definitely
"demand driven" and it is quite possible that the UK will not see major new resource developments
for the next 5-10 years at least. Consequently, capital expenditure has reduced considerably while
the effort and resources put into demand management has increased. Current leakage rates in the
UK range between 15-40010 expressed as a simple percentage of water produced, which can be very
misleading; it is more meaningful to express leakage in terms ofJoss per Km of water main.
These levels of leakage are high by comparison with some reported continental leakage
levels, where 10010 losses are often considered to be high. However, leakage levels in the UK are
related to the cost-benefits of suppressing leakage by the means now available, and it is the case that
on purely economic grounds relatively high leakage levels can be justified, though this may not be
politically acceptable; in practice all companies repair small leaks reported by customers purely on
PR grounds even though this may be totally uneconomic.
Leakage can be reduced by a variety of means:
• Leak detection, location and repair,
• Rehabilitation (replacement and relining),

354
• Metering,
• Pressure Reduction.
2.5. LEAK DETECTION, LOCATION AND REPAIR
The business of controlling leakage on a water supply network is inherently complex,

especially in the UK where the majority of domestic customers are unmetered. The UK industry
has for many years followed the principles and methodologies set out in Report No 26, published by
the National Water Council in 1980; this report collated the results of a national survey and
recommended "best practices" as perceived at the time. Over the last 10 years this report has been
the guiding light for the industry but more recent experience and research has led to the need for a
review and revision of some of the guidelines. Thus a national Water Industry body, the National
Leakage Initiative, has been set up to carry out this review and the first reports are now starting to
appear in draft form.
One major change from previous thinking is related to target levels for leakage. Report 26
set guideline values for urban and rural areas based on best practice in the UK industry. These were
expressed in Vproperty/hour during the night flow period (0200-0600) after due allowance had been
made for legitimate domestic night use and legitimate night use by major metered customers. It is
now recognised that meaningful targets at zonal level must relate to the basic cost-benefit equation
of expenditure on detection and repair and revenue savings on water production or capital savings
on new facilities.(National Leakage Initiative, 1994)
Progress has been made in this area in the UK by the use of software which has been
developed to understand and quantifY annual losses in a water supply area and, specifically, to
analyse the components of leakage. The 'Bursts and Background Leakage Estimate' (Babe)
software suit has emerged from the UK National Leakage Initiative and can be used to provide
an analysis of the components of night flows (IWEM, 1994).
When DMA's with high leakage have been identified the next step is the detection and
location ofleaks. Leak detection is carried out via a step-test; this is a systematic investigation of a
DMA which has been isolated by valve closure and is usually carried out at night when legitimate
usage is low (Figure 3.). Valves within the district are closed sequentially and at the input meter the
"step" drop in flow is recorded. Traditionally this has been carried out with data loggers which are
analysed the next day, but it is now possible to transmit results from loggers on the system back to a
central point (using devices such as the Spectrascan Tele-link Step Tester) so that the results are
immediately available and visible to the operator - in this way unnecessary valve operations can be
avoided. Figure 3. shows a step test and illustrates advances in step-test technology.
355
Graph produced from Tele-Link Step Tester showing

valve closure
Flow t
Time -+
Figure 3. Principle of step testing
Leak detection (ie establishing whether leakage is present and which part of the network it
is occurring in) leads to the problem ofleak location; here again there have been major advances in
technology in recent years.
The most obvious example is in the use of acoustic methods for leak location- so-called leak
noise correlation. Figure 4.. illustrates the principles involved and shows the basic equipment which
is required (Biwater IndjJstries, 1993).
Microphone sensors are placed on each side of the suspected leak at convenient points (eg
valve chambers); generally one microphone is connected directly to the correiator while the other
transmits its signal to the correiator via radio link. Modern correlators now have sophisticated
capabilities for filtering extraneous noise and for accurate operation with relatively small leaks.
These units can also be used with 2 radio units and can be used with non-metallic pipe over
relatively long distances.
However, this technology does require proper operation and training is essential - the
effectiveness of this equipment can be increased greatly with correct training.
356
1---. -iIIillbJii- •
1------ __
l3 Zoom BIaDk Peak Options Help Active
~ ~;\~---
E:lJ U::::=:::J"'IC:=:::::::::===========1E]
Total Leogth 164.618
Figure 4. Principle ofleak-noise correlation and typical acoustic pattern
2.6. REHABILITATION (REPLACEMENTIRELINING)
Wessex has around 10,000 km of water mains and replaces less than 1% per annum. This is
not uncommon in the UK and contrasts with many continental countries where 2% replacement is
widespread. Often, higher replacement rates such as these correlate with lower reported leakage
rates but there is no clear economic argument for increasing replacement rates to reduce leakage
where the marginal cost of water only is taken account of In addition to replacement, UK
companies now have extensive relining programmes generally driven by water quality directives.
Wessex now relines more water mains than it replaces, using either cement mortar or epoxy resin,
though the effect of relining on leakage is not clear - in some cases an increase in leakage after
relining has been identified.
Mains renewal is generally much more expensive than relining, owing to the obvious
factors of excavation and so on. However, technology is advancing in ail of these areas,
making the options more difficult to evaluate.(IWSA, 1994).
357
The cost of digging a trench, laying a main and reinstating the surface makes open cut
renewal an expensive option but this can be overcome by using narrow trenching techniques or
guided boring methods which will drill a hole in the ground into which the new pipe is
subsequently pulled.
Relining techniques are either structural or non-structural: structural techniques
include pipe bursting, slip lining and other insertion techniques; as the name implies they
physically reinforce the existing pipeline as well as improving it's hydraulic and water quality
characteristics. Non-structural techniques include cement or epoxy lining; these improve water
quality characteristics but do not physically reinforce the pipeline. Cement relining obviously
reduces the internal diameter significantly which in tum reduces the hydraulic capacity of the
pipe. Thus each rehabilitation technology has a variety of functional and cost characteristics
which have to be assessed carefully before a choice is finally made.
2.7. METERING
Universal domestic metering is still a contentious subject in the UK and although a new
means of charging customers has to be found by the year 2000, it is not certain that metering will be
the method adopted. Extensive metering trials have been conducted in the UK, however, and this
revealed more fully the extent of domestic leakage i.e. leakage which is primarily the responsibility
of the customer. This is now put at around 8% on average and the metering trials showed that
demand was suppressed in metered areas both via charging tariffs and through reductions in
leakage. It is now considered more cost-effective to install domestic meters than to build new
resources in many cases and this argument is now being actively pursued by the UK regulators
when capital expenditure pr,ogrammes are promoted by the companies. Several UK companies are
now committed wholly qr in part to domestic metering, generally those in the south and east of the
country where resource problems are greatest.{Water Metering Trials, 1993)
2.S. PRESSURE CONTROL
Pressure contro~ when applied in the right situation, is a very effective means of reducing
background leakage in distnbution networks. It is particularly popular in the UK, which is
358
untnetered and where pressure and flow variations do not affect income directly. Pressure control
also reduces the frequency of bursts and consequential damage and generally results in a better
service to customers, though pressure has to be managed carefully in the UK in order to meet the
conditions of the operating licence issued by the regulator, which is 10 m head at the boundary, at a
flow of 9 Vrnin for a single property. In practice this is generally interpreted to mean that a service
pressure of 12-15 m at the stoptap must be maintained at all times (Hegarty 1994).
100
/
90
80
70
V
.8.S /
V
60
~
1
50
...l
~ /
31
V
/
/
/
20
10
0
0
V 20 40
I
60 lIJ 100
Average zone night pressure head (m)
Figure 5. Relationship between pressure head and leakage
The relationship between leakage rates and pressure head is show in Figure 5., taken from
the latest work by the NLI; it is derived from work carried out on selected areas where large leaks
were repaired prior to the tests. Average Zone Night Pressure (AZNP) is defined as the mean of
the lowest and highest pressures measured in the zone, generally weighted to take account of
topography and layout of domestic properties. Leakage in each zone was measured in the approved
way using net night flows before and after pressure control was applied. The relationship is shown
in Figure 5. and indicates, for example, that if AZNP can be reduced from 60m to 30m, then the
Leakage Index reduces from 45 to 20 or more than half; and this will be the approximate effect on
the actual leakage rate in the zone. Thus, provided that the zone is suitable, quite large and cost-
359
effective savings can be made; unfortunately, in real world conditions these savings often do not
materialise because, for example, a customer on high ground within a zone or with a long service
pipe requires that pressure across the whole zone has to be kept much higher than necessary for
reduced leakage. Flow modulation is a relatively new feature which has been added to the more
typical pressure reducing valves. It adjusts the action of the valve to produce a more stable pressure
profile at a target point in the distribution system- where system headloss is significant, flow
modulation can improve leakage savings by more than 20010 over the non-modulated type of valve
in such types of system.
The technology in pressure control has improved enormously in recent years and a wide
variety of valves is now available for pressure reduction, pressure sustaining, with or without flow
modulation, profiling and so on. In one example where flow modulation was employed average
daily metered throughput to the district decreased by 20010 and pressures at the target point were
improved at peak demand - the results of this type of installation are reduced leakage, reduced
incidence of burst mains and reduced customer complaints. Installations such as this also represent
rapid pay-back and there have been improvements in the design ofPRV's to improve flexibility and
reliability and to reduce maintenance. A typical graph showing the relationship between flow and
pressure head in a water supply network is shown in Figure 6.
More sophisticated control can be applied using electronic devices such as the Autowatt
logger/controller now used extensively in Wessex Water. These may utilise remote pressure
readings telemetered from a distant control point or may follow a local control setting stored in the
controller which may be battery powered. Time switching and/or flow modulation can be used with
this device, giving a high degree of flexibility and consequently higher savings.
25
30 41 ~ 6) ~ lI) ~ 100 110 1:xl Jl) 140 1~ 16)
Time
Figure 6. Relationship between flow and pressure

360
2.9. UPGRADE OF EXISTING SYSTEM
The previous sections of this paper have described some of the technologies now used in
the diagnostics and rehabilitation of existing systems. We have seen how data available from
telemetry and data logging systems is now collected routinely for monitoring and analysing
pressures, flows and levels throughout the network. Wessex' Aquaware suite of programs has
been developed specifically to automate the handling of this data as far as is possible. Some of the
results of this work are described in Chapter II.
In addition, a number of techniques are used to assist in planning the upgrade of an existing
water supply network, including modelling and data integration - if used properly these can lead to
an optimised network where the capital investment yields the maximum return.
System Analysis
Network Analysis I
System Improvement
I I I l
Rehabilitation Applied System Operational Demand
Technology Redesign Changes Management
Reline TelemetIy Mains Pump and Leakage control

Replacement Data logging Strengthening valve controls Metering
Software Reservoirs Pressure control
I I I I
O~EDSOLU110N
Figure 7. A methodology for upgrading existing water supply systems

361
For example, it may be that pressure control is more cost-effective than mains replacement
for parts of a network, or that relining is better than replacement. Network modelling often reveals
major bottlenecks or "black holes" in a network, where relatively small expenditure on system
redesign is much more cost-effective than major replacement.
In these situations technologies for data capture, analysis, modelling, leak detection and
location, logging, pressure reduction, metering and so on often means massive savings in capital
expenditure on major schemes. Similarly, simple changes to controls on the network may lead to
economies in capital expenditure. This can only be achieved by approaching the problem in an
integrated manner.
The major elements in such an integrated approach are shown in Figure 7., and the major
elements in the process are described below.
2.10. SYSTEM ANALYSIS
The first step is to establish the basic facts about the network and produce a clear simplified
schematic showing all of the main elements and their interrelationships; the schematic should include
all main pumping units, reservoirs, water mains, major metered customers and so on. Boundary
valves and main metering points, PRV's, non-return valves etc. should also be shown clearly. This
process is a vital first step in the process of analysing basic water balances in the network and
involves using every source of basic data which is available. This includes the "local knowledge" of
the supply staff (much of which usually turns out to be wrong on investigation), schematics used for
valve operation, GIS records on paper or computer file and so on. Once this basic information is
cross-checked, verified and put onto a schematic, the basic system operation can be analysed by
using as much dynamic data as is available concerning flows, pressures, levels, pump and valve
controls etc. These can be manual records, data collected via portable data loggers, or ideally, data
from a telemetry system. It is useful if all of the data collected during these stages can be stored,
collated and integrated in one software package. This has been one of the objectives in the
development of WESNET, Wessex' network management software package, which allows
schematics, graphics, GIS, telemetry and logged data to be integrated and analysed to establish the
basic operation of the network. This data also establishes a database for the construction of a
network mode~ if this is deemed appropriate, at a later stage in the process.
Collating this basic information is crucial to the overall success of the project and it is
essential that, once collected and verified, the resulting documentation is put under proper
document control procedures so that it can be maintained over a long period of time - this can be
362
achieved by 2 simple means - firstly the basic schematic should be put under central control via, say,
the corporate GIS, and secondly authority for changes must reside with one individual only - i.e.
responsibility for the basic information must lie with one individual in the organisation.
The schematic will reveal the overall structure of the network and suggest ways in which
the overall water balance can be established via the major inputs and outputs.
Other data also gives an insight into the system; burst main records overlaid onto the GIS
background are a good indicator of possible pressure problems, while customer records on dirty
water complaints give an indication of internal condition, for example, tuberculation due to iron
deposits.
This stage of the exercise - System Analysis - should be properly documented and naturally
leads to the next phase ofthe exercise - Network Analysis.
2.11. NElWORK ANALYSIS
This phase involves the construction of a computer model of the network and is vital with
some types of network where the system is hydraulically sensitive, where we wish to experiment
with different types of control or where options for mains replacement or system redesign need
testing. These decisions often involve major expenditure and it is vital that a number of options can
be evaluated hydraulically before such expenditure is made. Modelling is not always necessary,
however, nor desirable in some cases. Often much can be deduced simply by analysing the basic
data available and concentrating on, for example, isolating areas of major losses which can be
rectified most cost effectively.
Network modelling is a specialised task requiring judgement, experience and knowledge,
there is no widely accepted methodology which covers all situations in all networks though attempts
have been made to pr~duce codes of practice for, say, calibration.
WESNET, the network modelling and management package used in Wessex, has been
developed specifically to aid the modeller in the construction, and specifically the calibration, of the
model by integrating GIS, telemetry, data logging, and other data sources into one package in a way
which is unique in the water industry. Some systems have considerable telemetry coverage and this
invaluable data can be used to establish good basic calibration of the model - often, a field test will
be needed where key data is logged on the system, usually over a period of several days, to establish
the calibration of the model.
Once confidence in the model has been established it can be used to test a wide variety of
operational changes and system redesign options, as well as confirming or otherwise areas of high
363
leakage, bottlenecks or other anomalies including meter errors and other faulty instrumentation. A
well-calibrated model is, in fact, a way of verifYing the current state of knowledge of the network,
much of which is often found to be wrong.
This phase, network modelling, leads to a set of proposals for improving the system. These
will generally comprise:
• Rehabilitation - It may be apparent that certain areas would benefit from rehabilitation -
replacement or re1ining. Replacement may be the only viable option in high pressure zones with
old mains and high leakage rates, for example, associated with particular pipe materials and/or
jointing compounds. Re1ining, a much cheaper option, may be appropriate where the structural
condition is sound but flows are restricted by internal encrustation and pressures are low.
• Applied Technology - It may be apparent that certain instrumentation is faulty or missing, that
certain points should be telemetered or logged, or that software systems could be used to great
effect for data handling and analysis. Wessex' Aquaware package has been developed to provide
the complete integrated software system for water supply network management. Low-level PC-
based telemetry systems are now available in addition to the more traditional SCADA systems
and these offer cost-effective data capture for the key data from the system.
• System Redesign - The investigation may reveal that the network can be redesigned at lower
overall cost than any of the alternative options. This often applies to old complex systems which
have evolved in piecemeal fashion over a long period.
• Operational Changes - A wide variety of pump and valve controls are now available and major
savings can be made by implementing these changes. Obvious areas of saving include pump
control and tariff selection, valve controls for pressure sustaining, reduction and so on. It is
often the case that the effective life of the system can be extended by relatively simple operational
changes rather than major capital expenditure on pipeline replacement.
• Demand Management - As stated earlier in the paper, in the UK at least, the major emphasis in
system management is now in Demand Management, which includes leak detection and location,
pressure control, metering and mains rehabilitation. Most of these topics have been described
earlier but they should be considered as part of the overall integrated approach.
The result of this process is an overall plan for the effective upgrade of the network at
lowest cost and inevitably this means changes and improvements in the management of the network;
in Wessex the following key areas have been identified:
364
• Infonnation - Key information must be made available to the staff running the system at regular
intervals and basic disciplines established in the use of that data. .
• Training - Network management is becoming increasingly technical, through the widespread

use of computers and also through the kinds of technology described in this paper - data loggers,
telemetry, valve and pump controls, system analysis, modelling, relining techniques and so on.
Most companies in the UK are now investing in GIS' (Geographical Information Systems),
which are invaluable in more detailed analysis of geographically-based data, in addition to
providing an automated means of record-maintenance. All of these techniques require fewer
people but higher levels of technical ability and training.
• Operating Manuals - Tighter control of the network can only be achieved through better
management disciplines and in Wessex this has been approached by introducing operating
manuals for each network - in this way basic disciplines are introduced for rezoning, valve
operation, pump controls and so on. Controlling demand effectively can only be achieved in this
way though it can be a major culture change and difficult to implement.
2.12. CONCLUSIONS
Over recent years the technology for diagnosing problems with existing water supply
systems has improved enormously; in the UK, where domestic customers are still predominantly
unmetered, there has been major investment in instumentation, data capture equipment and
information systems for producing accurate water balance data at various levels in the network.
The technology for rehabilitating these networks has also advanced, providing the
management of such !!}'stems with a number of options for improving performance and stretching
the operational life of the network. The optimum solution generally now involves a combination of
techniques and the paper suggests that a methodology is needed which enables the management to
assess this best combination.
365
2.13. REFERENCES
Technical Working Group (1985) Report on Leakage Control Policy and Practice, No. 26., Water
Authorities Association, WRc ISBN 0 904561 95 X 1985.
Water Metering Trials (1993) The National Metering Trials Working Group.final report. ISBN 0
947886338.
IWSA (1994) Rehabilitation. Regional Conference. Zurich.
National Leakage Initiative (1994) Managing Leakage. Final Report, WSA
lWEM (1994) Accounting for Losses - The Bursts and Background Concept. lWEM Journal,
Vol.8 No.2 April.
Biwater Industries (1993) Water Industries Manual.
AquaWare Systems (1996) WESNET User Guide, AquaWare Systems. UK.
The Water Mains Rehabilitation Manual (1989) WAAlWRc. second edition.
Hegarty M. (1994) The Strategic Role of Infonnation Technology in Leakage Control. BICS
Conference.
Chapter VI
The Application of Modelling in the

Management of Water Supply Systems
ChapterVI
The Application of Modelling in the Management of

Water Suppply Systems
ffiTRODUCTORYSTATEMENT
Computer modelling in water supply is an old technique, often referred to as network

analysis. In recent years a number of advances have been made in the application of this technology
to the management of water supply systems and this chapter looks at some of these new concepts
and applications. The basic principles of modelling are reviewed, including the traditional
applications in Planning and Design.
The Chapter then goes on to look at the way modelling is being introduced into operational
management, using examples from Wessex Water (UK), and the WESNET software package, and
the advances that are now being implemented in integrating complementary sources of data,
including telemetry, GIS, data loggers and post code data. The paper also looks at the introduction
of a simple expert system module into the WESNET network modelling software package to
improve the accuracy, speed and consistency in the calibration process. The conclusion is that the
future for network modelling lies in making models relevant to everyday operational management,
and in having links to live data for ongoing calibration checks - this data can be derived from
traditional SCADA systems, data loggers or from the increasingly common telephone dial-up type
loggers which are now commonly used in the UK for monitoring leakage at DMA level.
The second Section in the Chapter looks at new techniques in the computer software used
in some models and the impact of object oriented programming techniques, using the HYPRESS
package; the paper also looks at the way in which GIS can be used for presentation of results and as
a complementary system to hydraulic modelling.
1. Water Supply Network Modelling and its
Application in Design, Planning and Operational
Management
Dulian Obradovic
WINS Ltd.
Tower House, Parkstone Road
Poole, Dorset BHl5 2DH
John Snoxell
WESSEX WATER
BS2 OJQ Bristol
UK
1.1. INTRODUCTION
Computer modelling in water supply is an old technique, often referred to as network

analysis. The applications and objectives of such techniques are well documented and will be found
in the references; this Section looks at new ways in which these techniques are being used for
operational management and the way that complementary data sources, from telemetry, GIS, data
loggers and Post Code data are being used to facilitate this process by reference to work in Wessex
Water in the UK.
Water supply has become a limiting factor in development, along with employment, food
and transportation problems (UNESCO Workshop, Zandvoort, 1977). The community requires
better levels of service, expansion of facilities to meet rising demand, but also tighter control over
waste water treatment and disposal.
In recent years there has been rapid progress in the implementation of new technology into
water supply management. It is generally accepted that technology is the means by which increasing
pressures for better service and higher efficiency can be achieved (Brammer and Schulte, 1993,
Cosgriff, 1985, Cubillo and Mercier, 1993, Cullen, 1987, Haddon, 1994, Halpern and Pascal, 1987,
Huntington, 1984, Nguyen, 1994, Snoxell, 1994, Takagi, 1983, also other references).
NATO ASI Series, Partnership Suh-Series, 2.Environment- Vol.lS

372
The modernization of a water company, however, should not be a revolution. The service
must continue without serious interruptions, while important changes within the company take
place. All actions must be carefully planned, having in view both the desired ends and available
means, then executed as efficiently and economically as possible. Errors can be very expensive with,
in some cases, very damaging consequences. It is to be expected that they could be avoided by
studying the methods of the most advanced companies and learning from their experience. This text
will try to summarize the achievements in information technology and mathematical modelling as
applied in such water companies today.
1.2. FUNDAMENTALS OF MATHEMATICAL MODELLING
A mathematical model is an artificial entity which describes the behaviour of a real system in
various situations. In the case of a water supply system this means that the model can simulate
variations of flow and pressure in a distribution network, the operation of pumping stations, the
functioning of all devices such as control, check and float valves, changes in water quality at service
connections, compute costs, forecast demand patterns, optimize the use of pumps, etc. - and all this
in normal, peak demand or emergency situations.
1.2.1. Oassification of Models
Obviously, the building and testing of a model will take time and effort, besides significant
costs. No one will do it without good reasons, least of all a water engineer pressed with other
urgent tasks. But such reasons do exist; to name just a few:
• the model will collate all known facts and data for the real system, with all known
relationships between its elements,
• the model will permit the analysis of all operational data in much greater detail, helping to
understand the complex process within,
• all interested parties: operators, designers, engineers and managers can participate in the
building of the model, make a contribution and later use it for various tasks,
373
• the current control policy can be tested on the mode~ verified or improved, without risky
experiments with the real system,
• the model should be a vehicle for design and planning, greatly reducing the time and costs
for such activities,
• the response of the system in various norma~ emergency and catastrophic situations may be
analyzed beforehand and adequate strategies developed and perfected,
• the model is a wonderful tool for education and training of new hands: from operators to
managers.
These arguments, as well the availability of computer hardware and software, have caused a
proliferation of models and modelling techniques in the water industry in the last few decades. One
can argue that modelling was always an integral part of the design process; the novelty was its
application on real water supply systems.
This complexity has forced most of investigators to concentrate just on one facet of the
problem, say, hydraulics - the other option being an unacceptable oversimplification of the system.
Models therefore can be sorted in three main groups according to their main purpose (ObradoviC,
1993):
I. Hydraulics: models which simulate changes of flow and pressure in water supply and
distribution systems;
ll. Water QUality: models which simulate changes in water quality in ambient waters and in
the water supply system;
Ill. Economics: models which optimize functioning of water supply systems.
Hydraulic models may analyze either steady state (one "snapshot" or a series of steady states
interlinked with an integration scheme) or transient regimes (rigid water column models, water
hammer, hydraulic vibrations, two-phase transient regimes) - see for example Bhave (1991),
Hoogsten and Zwan (1985), Rao and Don (1977), Walski (1984), Zwan (1988).
Water Quality models may analyze the situation either in ambient waters (surface or ground water
or both) or in the water supply system itself (transport of raw water, water treatment plants or
374
deterioration of WQ in the distribution network) - see Clark (1994), Cohen (1990), Filip and
ObradoviC (1986), Males et aI. (1985), Nguyen and Montiel (1994), among others.
Economic models analyze either Operational Efficiency (optimum pump scheduling, water sales,
operation and maintenance costs, specific consumption of various resources) or Planning and
Investments (water demand in future, capital investments, development of alternative solutions for
new systems, optimization of various plans and designs) in order to find optimum solutions. A good
example are models for optimum control of water distribution systems, see Chase and Ormsbee
(1993), Chase and Jones (1993), Coulbeck and Orr (1988, 1989), Fallside (1977), Jowitt et aI.
(1989), Schulte and Maim (1993).
Beside these models, specially developed for the water industry, many other software tools
and general purpose programmes are used in water companies: CAD, GIS, demographic studies
and others.
The models can be classified alternatively by their Scale, Time Step, Depth & Sophistication
or numerical method applied. The user might not be aware of these technicalities; a more important
distinction is between:
• Simulation Models, which describe the behaviour of a real system as closely as possible,
and
• Optimization Models, which prescribe what should be done in order to achieve a certain
objective.
Hydraulics and Water Quality models belong to the first group.

All these models can not be covered in one book. After all, they do not have the same
importance and frequency of use in practice. This text will concentrate only on hydraulic simulation
models, for steady-state analysis.
1.2.2. Structure of a Hydraulic Model
The current view is that any water supply system may be quite adequately represented by a
set of nod e s and lin k s, arranged in the proper order.
A nod e is a location within the system where either head or inflow/outflow is known.
The first group includes:
375
• faxed-head nodes, where water level remains unchanged during the simulation because the
storage is infinitely larger than all withdrawals and/or inputs (examples: aquifers, lakes, large
pools, springs, wells, etc.),
• variable-head nodes, where water level rises if the inflow exceeds the outflow and vice
versa (examples: reservoirs, water towers, contact and balancing tanks).
Y--
Demand
a) Ordinary Node
b) Transfer Node
~~
/ In
Gravity flow
PumPingt ~
~ c) Fixed-Head Node
Inflow
d) Variable-Head Node
Figure 1. Nodes in a model
In both cases water level is a known quantity while outflows and inflows need to be
detennined through the computation.
The other group is called ordinary nodes. There the total sum of all inflows and outflows is
known and the head is to be computed. Examples are:
376
• a transfer point, where a known quantity of water is either delivered to (or taken from) the
system,
• a junction (branching point) of two or more pipes, with or without some local consumption,
• a service connection to an important user, where local consumption of water is known in

advance,
• any other point within the system which is important due to any other reason.
Examples of nodes used in modelling are shown in Figure 1.

A lin k connects two nodes and can convey a certain amount of water. There are three
types of links:
• a pipe, which simply transports water from one node to another,
• a pumping station, which adds energy to water,
• a control valve (any type), which regulates the flow of water by reducing its hydraulic
energy ("throttling").
These names are mnemonic: they describe the behaviour of one element of the model by
using the name of the similar physical device.
This does not mean that each valve in a water supply system will be represented by a
"valve" in the model - this is seldom practicable! Quite the opposite, only a few important physical
valves (control valves, for instance) will be shown as such in the model- and the bulk (zone valves)
will be simply included as minor losses in their pipes.
Next, a "pipe" in, the model represents any conduit, a tunnel, a water main, a large or a smaIl
pipe section - anything that is used to convey water from one place to another under pressure.
As an opposite example, a real pumping station which houses several pumps under the same
roof may be represented in the model as several "pumping stations" (in modelling sense) if the
suction or the delivery node is not the same one for all units.
Examples oflinks are shown in Figure 2. Note that hydraulic energy faIls along the flow in
the case of "pipe" and "valve", while it is boosted in a "pumping station".
It can be easily shown that all elements of any water supply system could be described by
these items - either directly or by a suitable combination.
377
Flow
• ~
• a) Pipe
•
Valve
u~ • b) Valve
Pressure Node
c) Pumping
station
Suction Node
~~ // P \
Pump Characteristics Q
Figure 2. Links of a model
1.2.3. Modelling of Steady State or Gradually Varying Flow
Thus far it has not been defined whether the model will serve for analysis of steady or
transient regimes. Further development will be limited to relatively slow changes - where inertial
forces and the compressibility of water can be ignored. This assumption is valid for all normal and
for most of the emergency situations in water supply systems, especially iflonger time periods (up
to several days) are being analyzed. Of course, when analysing transient states - which last just a
few minutes at most - inertial forces, the elasticity of pipe walls and compressibility of water must
be taken into account, using more sophisticated analytical tools.
A further simplification of the real system is about water withdrawals. It is clearly impossible
to represent each tap in the model so whole buildings and blocks of houses are represented as
"sinks" - nodes with aggregated demand (total consumption plus some losses). At the very best, this
can be only a crude image of the real process - but still useful! Even with all these assumptions the
task can not be solved without a computer.
378
The resulting system of non-linear equations can be solved by various numerical methods,
see, for instance, Bhave (1991), Obradovic (1993) or Walski (1984).
The first task is reading all necessary data: from model files or from telemetry files. The data
are then cross-checked to discover all errors and inconsistencies - before the simulation. The user is
warned and the procedure halted to pennit the corrective actions from user's side.
After this stage has been successfully completed, the program formulates the initial state of
the system, makes necessary preparations (rearranges the data, closes input files, opens runtime
files, sets pointers, etc). Then the "clock" is set "on" and the simulation cycle starts.
The program solves steady-states in this system for given boundary conditions, using one of
a few available methods described in the next section. The results are then analyzed in order to
determine eventual changes in boundary conditions:
• compute the current water demand (option: take into account service pressure as well),
• increase/decrease of water level in all reservoirs, water towers, contact and balancing tanks,
• change of status of any pump - due to changed water level in control reservoir or pressure
in the control node,
• change of valve opening - for the same reasons, etc.
Relevant messages are displayed - so the user may intervene through the interface - for
instance, to review current results. Then the procedure is resumed: the clock increases time for a
certain increment (from 1 to 30 minutes, rarely more) and determines the water demand at this
moment before repeating the same cycle again.
The simulation lasts until the set time is reached - or until the user has stopped it, for any
reason. The main results. are stored in data files for analysis. The user may get information on
extreme values of flows, pressures or water levels in selected nodes, the operating range and time
for pumping stations and individual pumps, maximin openings and flows through control valves,
various diagrams and snapshots at certain moments, etc. The data could be shown on maps
(provided by GIS) and schematics; or set against the current or historic operational data.
1.2.4. Model Application
The abundance of options available by modern software is such that the need for additional
on-line help in the form of an EXPERT system is recognized by many users.
379
The main applications of modelling are in:
• Operational Management of existing water supply systems,
• Design & Analysis of existing and future water supply systems,
• Education, Study and Research in hydraulics and water supply.
It follows that the principal users are:
• engineers and operators in Control & Information Rooms of water companies,
• engineers and technicians in Design Bureaus,
• engineers and students in Universities, High Schools, and in Research Departments and
Institutes.
Typical applications for models are:
• analysis of how the system really operates,
• verification of current control policy,
• evaluation of storage reserves in emergencies,
• optimum pump scheduling,
• discovering 'bottlenecks' within the :;ystem,
• selection of post-chlorination sites,
• decisions concerning which tariff should be used,
• completion of an existing telemetry system,
• design of new facilities,

380
• design of enlargements and reconstruction of the system,
• design of a new controVtelemetry system,
• 'war games' for the staff,
• training of personnel at various levels, etc.
Modern software products are labelled "user-friendly" because they do not ask for any
special knowledge or training prior to use (this is only partly true). The characteristic features of
such a program are:
• is extremely easy to learn and use, because it is tailored to suit user's demands,
• is completely menu-driven, so the user does not often have to refer to the manual,
• has an comprehensive system for capturing errors, with an explanation in plain language,
so the user is protected against common blunders and mistakes,
• on-line help is available at any stage,
• can model a wide range of possible control policies for pump stations and control valves,
but keeps all technicalities invisible to the user,
• an ability to keep the user informed about the state of the system during a simulation run,
• always provides an answer to the user regardless the size or complexity of the case - and
never loses control!
• has good computer graphics available to the user both for input and output jobs,
• can compare operational data with simulation results on the same graph,
• permits the user to create the output according to his (her) particular wishes in a very
versatile way,
• is a very efficient tool!

381
As for the user himselt: it is assumed that:
• the user is well informed about the water supply system helshe is modelling, but does not
have to be an expert in mathematical modelling or computers,
• the user should describe the real system as it really is, without "inventions" - and the
programme must be versatile enough to cope with real-life complexities.
1.3. SIMULATION MODELS AS TOOLS FOR DESIGN
The design process has its creative phase, when alternative solutions for water supply
systems are being formulated and its routine phase, when accepted solutions are elaborated in
greater detail. These two phases cannot be completely separated, because every proposal has to be
evaluated from various points of view: functionality, controllability, capital expenditure, operational
& maintenance costs, feasibility in given circumstances, etc. This evaluation is based partly upon the
Designer's experience and judgement, and partly on computation. Here the models come
into play, permitting the Designer to examine quickly and efficiently different alternatives with
sufficient precision and detail. The idea is as old as designing itself, but Computer Science has
opened new possibilities which did not exist before, as amply demonstrated in references - see, for
example, Brammer (1993), Coulbeck and Orr (1988), Hoogsten and Zwan (1985), ObradoviC
(1986, 1993), Rao and Don (1977), Walski (1984) or Zwan (1988).
1.3.1 A Simple Case
To illustrate this point a simple case is included here, see Figure 3. The new system has to
supply four different communities (denoted as "Demand A" to "Demand D") from a single source.
One possible solution is to build a main pumping station, four large reservoirs (I to IV), two booster
stations (ll and III) and control the pressure at two places (A and D), as shown below.
Obviously, this is not the only solution. Maybe Reservoirs n and III could be merged into
one and then ''Demand Cot could be fed by gravity flow from either Reservoir n or III - or both.
Clearly many similar alternatives can be offered, each one with its advantages and
shortcomings. Without modelling, one could only make arbitrary decisions (not necessarily the
wrong ones!) which cannot be demonstrated to other interested parties, nor could the consensus be
382
reached on the basis of facts. With models the situation is far better: every proposal can be
examined on the model of the water supply system in relatively short time. All concerned can see for
themselves why one solution is better than another. Expensive mistakes could be then avoided and a
sound consensus reached at the earliest possible stage of the project development. This consensus
cannot be taken for granted; the owner of the future system must be consulted almost from the
beginning of the design process. Another simple case is included to illustrate the importance of good
cooperation.
Reservoir IV
Rese!Wir I
Conveyance System
Reservoir II Booster PST
o
Reservoir III
Source PST
B Distribution System
Figure 3. A water supply system
Figure 4. shows the simplest possible problem: satisty a given Demand from a Source.
Suppose that water must be pumped into the demand area; but there are still (at least) three possible
solutions:
1: pump all water through demand area to a balancing reservoir at the opposite side;
II: pump all water first in a high reservoir from which it can flow to the demand area - the
pressure can be controlled by a PRY;
ill: pump all water directly to demand area by variable-speed pumps, without any storage
space.
383
1 1-----1. . . ~
Source The Problem
Reservoir
Solution I
Reservoir
Solution II
PST
Demand
Solution III
Figure 4. Alternative solutions
There are pros IUld cons for each solution (without even mentioning widely different capital
costs); for instance, Solution I is simple, can be controlled easily, offers some security in
emergencies - but the network will be submitted to relatively high and variable pressures. Solution
ill can be the cheapest to build, but operating costs could be high, and the maintenance of variable-
speed pumps might be difficult. Moreover, tI'1ere would be no reserve capacity if the power system
fails (is this probable? How often?). Next, pressure management would be almost impossible in
Scheme I and quite simple in the other two, etc. Obviously, the water company staff should be
consulted and their preferences taken into account from the very beginning of the design process.
The model offers excellent pOSSIbilities for quick and systematic evaluation of various alternatives so
many costly mistakes could be avoided in good time.
384
1.3.2. The Design Procedure
The Designer is well advised to start the job by carefully examining the existing water
supply system - the cases where there is none are extremely rare. The benefits are many: one can
gather valuable infonnation about local circumstances(example: specific consumption and seasonal
effects), then make contacts with the local staff and mobilize their support for the project. Local
experience is usuaIly irreplaceable, so the additional effort is more than justified.
This exercise could be performed best on a mathematical model; in some cases this model
will already be available in the water company, while in others it has to be made by the Designer. In
any case modelling of the existing system provides plenty of opportunities for the development of
cooperation and for collecting of information vital for the design. The list of tasks include:
• Analysis of demand in time & space: current level, trends, specific consumption for various
uses, patterns of use, local peculiarities;
• Unaccounted-for-water levels: true losses, pressure management, billing system, non-billed

use;
• Existing Control system: equipment, instrumentation, level of technology and expertise

within the water company;
• Data Management: organisation, sensors, communications, available hardware and

software;
• Control policy in normal and emergency situations: remote and local facilities, reaction time,
spare capacity.
The good and bad sides of the existing system will surface in this process, enabling the
Designer to make sound solutions with the maximum benefit/cost ratio. Even more important, the
stages between the existing system and the future one will be clear to all concerned, not only to the
Designer.
The steps in design procedure therefore should be:
I Make or update the mathematical model of the existing water supply system;
II. Calibrate the model in cooperation with the local staff using fresh operational data;
385
ill. Analyze the existing system in great detail and find its good and bad features;
IV. Define several feasible alternative solutions for the future system;
V. Optimise acceptable alternatives by using multiple criteria like Internal Rate of Return,
average and initial price of water, etc.;
VI. Elaborate stages of development from the present system to each alternative solution;
VII. Analyze various Normal and Emergency states on the model and discuss the relative
merits with the staff,
VIIT. Decision - select the most promising alternative, possibly with one or two major
subvariants.
The model for design must include all sources, reservoirs, pumping stations and control
valves, while the distribution network can be reduced to major pipes only.
The model must be verified before the beginning of any serious analysis. This is done best
by simulating the performance of the real system during one or several days, and then comparing the
results of simulation with observed values.
A computerised database with complete operational data (flows, water levels, pressures)
will greatly facilitate this task; if such facilities are not readily available the Designer should organise
data capture effort in cooperation with the local staff.
1.3.3. Calibration
One area where. there has been major progress in recent years has been in the use of
operational data for calibration.The results of simulation will not agree well with observation,
certainly not at the first attempt.
The reasons are many, from imperfections of the model to false data. A careful analysis of
differences can be very rewarding, showing which instruments are faulty, what data are wrong or
which hypothesis was bad.
The model is then adjusted accordingly. This process of "fine tuning" is called calibration
and must be repeated at regular intervals in order to keep the model well trimmed. Data from the
real system can come from many sources and traditionally this may have been from manual records-
over the last decade a wide variety of portable, inexpensive data logging instruments have become
386
available which provide an easily accessible source of real data for calibrating models. These data
loggers can now be interrogated remotely via telephone lines and this approach is now common in
the UK for continuously monitoring leakage at DMA level (a DMA, or District Metered Area, is
the basic zonal unit of around 5000 properties at which flows are measured continuously for
analysis of Minimum Night Flows).
Most water supply networks are now monitored by telemetry and if this data source can be
opened up for modelling then the process of calibration can be made even more efficient and cost
effective. The natural consequence of this is also that modelling can realistically become part of
operational management because all staff can see that the model is producing similar results to the
real system.
Examples of results taken from a typical model using telemetry data for calibration in the
Wessex region are shown in Figures. 5 to 7.
Calibration 7th Hau 91 'X'Link te5t 38 12 91

Reservoir :
188T-____-+______~----~------~----~~----,
• RGouldsH
+ RGouldsH- Te I
or: 1.13 StDe"

oS: 1.3;.: Dire
-
E ................; ................ ~.- ....... .
·····-··········~·············· .. f················~···· ............. ~ ................ '~ ............... .

. -
.: .:
................:................ :-.............. -_ .... -.
887/11/91 : 88/11/91
~----+-----~----~----~-----+-----4
8 .88 1.88 8 .88 12 .88 16.88 2B . 88 B.BB
Th... (h)
Esc- E ~ l t F2- Plot FJ- Prlnt
Figure 5. Variations of water level in "Gouldshill" reservoir

387
Celll'NOtion 7t.h Mou 9 1 ' X'L1 nk t.e. t 38 12 91

PIP E
,"143
r= 1 . 18 StDeu
'" 7.5 l< Dirt
BlV lV'J1
8'.88
~----~~---8~. 88
~----
12~.~8-----
8 1~6+.88
~---2B
~.~ee
-----8-+.88
TI_ (h)
Figure 6. Outflow from Chabury Reservoir, Weymouth
Cellbratlon 7th Mou 91 ' X'Llnk t.ast 38 lZ 91
s . T-----~----~----~----~-----+____-, •" Top/lill

0 D E
+ TOp/llll - Tel
4.
",. B.18 StDeu
'" 3 .7 l< Dirt
3, ,+ ................ ~............... ~
l
j Z.
1. ... ............ I. . . . . . . . T" ..•....•...••.. .....•.......•...:..........•.....••................

~
87/1l/91 ~ i 88/11/91
4 .88 8 . 88 1Z . BB 16 . 88 21.88 8 . 88
T1.e (h)
Ewe-Exit rz-Plot FJ-Prlnt
Figure 7. Water demand ofIsle ofPortIand, Weymouth

388
Figure 5. compares computed variations of water level in a main reservoir with the data
collected by the telemetry system for the same day. Figure 6. shows the computed and observed
values of outflow from another reservoir. Figure 7. contains similar data for water demand in a zone
within the model. Note good agreement between computed and observed values in all three
examples - they are close to each other, but are not identical; the "noise" is peculiar to the observed
data, quite understandably. By studying these examples one can get the idea about the range of
differences between the real system and its model. Further calibration will not improve substantially
our knowledge about the real system.
1.3.3. Using GIS Data in Modelling
Over the last decade many modern water companies have invested heavily in computerising
their asset record data with mapping systems, now generally referred to as "GIS"- Geographical
Information Systems. The link between such systems and network modelling has long been
recognised but rarely achieved in any significant way (excluding, that is, research or demo projects).
Wessex has made much progress in this area in the last 3 years in converting all of its major models
into geographically correct format by using actual GIS data from its corporate system as a
background to the model. The model can be constructed by tracing over the background and results
can be displayed against the background, greatly enhancing overall understanding.
Iji&iU":"M'
U l • '" Jool. Print
i"rtl,
' v'
I
:r.-
~
Figure 8. Using GIS in modelling

389
In addition, a wide range of other data can be integrated into the modelling environment to
complement the results of modelling work, including, for example, burst main locations, customer
complaint data, post code information and so on.
The example (Figure 8.) shows an example of a network model superimposed upon the GIS
background (water main features and houses only have been included in this example).
The natural evolution in this process is to use the actual network data held in the GIS as the
network geometry for the model and various demo projects are to be found in the literature-
however, few, if any can be sustained in the real world and there are serious problems in doing this
in a manner which is cost effective and manageable in real-world conditions. These problems are
beyond the scope of this paper but it is likely that in the next 5-10 years this will indeed become
properly understood, technically feasible and manageable.
1.4. NElWORK MODELLING IN OPERATIONAL MANAGEMENT
In reality Operations managers rely these days on telemetry and on judgement, experience
and knowledge. Because of this the methodology for creating, maintaining and managing computer
models for Operational management does not exist and many questions remain about the basic
viability of such a management philosophy. It is worth considering this in 4 main areas:
• What are the benefits to Operational management of creating and maintaining computer
models of water supply networks?
• What are the problems (mcluding cost) in implementing this philosophy; why has it not been
achieved?
• How can such a philosophy be implemented and what are the managerial and technical
problems and solutions.
• What have been the actual experiences in Wessex Water over the period of implementation?
This Section describes how, through the development of WESNET, Wessex Water has
introduced computer modelling into Operations management and the results and experiences of the
exercise.
390
1.4.1. The Potential Benefits
These are generally recognised as follows:
Accounting for Water: Any major production operation in any branch of industry must
know how much product is being produced, and where it is being consumed. In the case of
water the model provides a factual base for this accounting process, including an assessment
of how much water is being lost in each of the demand areas. To be able to do this
regularly over a long period of time in a consistent manner is a very powerful tool in basic
management.
Cost Control: A major cost in running the network is the pumping costs and these can be
reduced significantly through a combination of source management (eg "cheapest first"),
pump control (eg leve~ profile, etc) and tariff selection (eg "Demand", "Day/Night", STOD
- Season Time of Day). The permutations are very complex, however, and a modelling aid
is now essential to answer these questions.
Standards of Service: The Director General of Water has recognised formally that
network modelling is a valid method of demonstrating compliance with pressure standards.
An operational model would be an ideal vehicle for this.
Resources (Bottlenecks): It is often the case that resources cannot be made available at the
point of consumption at peak: demands because of bottlenecks in the system. IdentifYing and
removing these is essential to the operational manager.
Emergencies - Burst/Contaminants: The ability to simulate bursts (ie exceptional

demands), or pollutants introduced to the network has long been recognised as of major
potential value to the operational manager in dealing with emergencies.
Capital Expenditure: An operational model could potentially identifY situations where

minor changes to the network could avoid major capital expenditure.
Duplication of ModeUing Effort: One common feature of current practice is the wasted
effort in duplicating models over a period of years, mainly because of a lack of ongoing
maintenance. Operational models, kept live through telemetry, would provide a central
library for others to use as a basis for other work, thus removing much of this duplication.
391
1.4.2. Problems in Implementing Computer Modelling in Operations Management
In the previous section we outlined the major areas where, potentially, computer modelling
could have a major benefit in Operations management. The major question which arises from this
is: Why has so little been achieved?
The experiences in Wessex Water through the development ofWESNET have focused on
some of the difficulties and also the solutions to these problems.
1.4.3. Software
Firstly, the software for creating and maintaining such models has been very limited. The
available packages are not suited to modelling the actual dynamic behaviour of the network,
including the wide variety of pump and valve controls now commonly in use on real networks. In
practice it is common to find network modellers (who are usually engineers) using a wide variety of
tricks to simulate the behaviour of the network; indeed it is often a source of pride and discussion
how to "fiddle" the network geometry to reproduce a comparatively simple feature. Often people
have become used to "hacking" data files to achieve their ends rather than use the software through
the User Interface, partly because it is not very User friendly or because this "expertise" has become
second nature to the individual.
This approach is simply not viable in the Operations environment where day-to-day changes
have to be made quickly and reliably to keep up with the dynamics of the actual system. The
software must reflect this environment and WESNET has been specifically developed to answer this
need. If the User has to "fiddle" or "hack" then something is wrong - the software must satisfY the
User's requirements, not require the User to be a computer expert.
Current software packages are very weak in simulating pumping equipment accurately, yet
it is this which is often the most expensive single cost centre in the running of the network, and
therefore of prime concer:n to the Operations manager. This is a reflection of the fact that hitherto
modelling has been the exclusive province of the Planner or Designer who is concerned primarily
with the hydraulic limitations of the distribution network, for example what are maximum and
minimum flows and pressures under certain conditions of demand. It is quite common for such
models to exclude most or all of the significant pumping equipment altogether; often models start at
the service reservoir and go into immense detail in the distribution network, much of which adds
nothing to the overall understanding of the network but which adds a considerable maintenance
overhead.
It is a common experience to find that such models are, in practice, useless as a basis for an
operational model and it is a mistake to try to convert them for operations' use.
392
1.4.4. Calibration
A major obstacle to the ongoing use of network models in opemtional management is the
problem of calibmtion; there are two aspects to this: firstly the costs and feasibility of maintaining
the model over a period of time, and secondly the actuaI process of calibmtion, which to date has
been developed solely on a "snapshot" basis, a method which is inadequate for operational use.
The development of WESNET has been largely driven by the realisation that the first
problem can largely be solved by the integmtion of telemetry and data-logger data into the
modelling package, thus providing a constant, free source of data for the fine-tuning of the model
over a long period.
The latest version of the software (V 6.0) now provides extensive capabilities for
integmting logger data from DMA monitoring, thus providing an additional source of on-line
calibmtion data from the low level telemetry which is now employed on many networks for leakage
monitoring. The experience has also proved that the existing calibration techniques, as expounded in
the WAA Code of Practice are not appropriate, or adequate, for opemtional modelling.
As a result Wessex has included in WESNET a new set of criteria for judging the adequacy
(or otherwise) of the dynamic calibmtion ofa model. This provides a much better guide to how
well the model is actually simulating the behaviour of the real network, and speeds up the ongoing
calibmtion process. WESNET produces a "verdict" in terms of "good", "fair" or "poor" for each
specified calibmtion point on the network, based on statistical criteria over the whole period of
simulation.
This approach formaIises the mther subjective judgements which would be made about the
agreement between the model and reality over a period of time, and also points up how important is
the dynamic calibration, as opposed to the "snapshot" approach. It is clear that the industry should
now start to encompass the dynamic approach in a national Code of Pmctice as a complement to
the existing Code, perhaps based on the ideas in WESNET - the refinement of these techniques will
only come through experience and actual implementation.
1.4.s. Mythology
A further problem in implementing modelling in opemtions management is that the subject

has been surrounded by considerable mystique and mythology, partly through a lack of real data and
experience. One common fallacy concerns the inherent complexity of demand modelling; it is
argued that demand is made up of many components, each of which has to be modelled sepamtely
for the model to balance properly. If true, this would make the ongoing maintenance of the model
quite difficult. Experience in opemtional modelling over the last three years has shown quite clearly
393
that this is not the case, however, and demand patterns have not proved to be a major problem in
the ongoing maintenance of the Wessex models. In practice, demand patterns have proved to be
stable, repeatable, and easily manageable.
Another common view is that the system changes so frequently due to rezoning, mains
reinforcement, pump changes, new sources etc, that it is impossible to keep the models up-to-date.
In practice, with the link to telemetry this has not proved to be the case; the operations models act
as a focus for different departments once they understand what the objectives are and
communications can be set up to account for all such operational changes.
1.4.6. Operational Reliability
One sometimes hears telemetry data being derided, especially flow data, the implication
being that telemetry per se is inherently unreliable and unusable for modelling purposes. This stems
partly from the early days of water industry instrumentation, when certain devices did have design
faults which led to "drifting" or other problems. Nowadays, however, most of these problems have
been resolved and telemetry devices generally are reliable, provided they are installed and
maintained correctly. Experience in Wessex is that telemetry is an invaluable and reliable source of
raw data for the calibration of operations models; in practice the models sometimes show up
anomalies resulting from transducer problems in a way that would be difficult to detect by other
means and in this way act as a perfect complement to a telemetry system.
These are some of the problems that have prevented the perceived benefits of modelling
being implemented in Operations management; it is a daunting list and it is no ~rise that little real
progress has been made. A corollary of this is that there is virtually no methodology for people to
follow to implement this in reality. One assumption is that "network modelling" is just one
technique - in practice there are now at least two classes of network models: "classical" as
developed traditionally by Planners and Designers, and "operational" as developed and implemented
through WESNET. They. are quite distinct in character, implementation and use and the following
section describes in detail the methodology developed in Wessex Water over the last three years to
implement and realise this philosophy for operational modelling.
394
1.5. EXPERIENCE AND BENEFITS OF OPERATIONAL MODELLING IN WESSEX

WATER
1.5.1. Model Construction
The initial phase when the model is constructed is a period when all current knowledge
about the network is collated and verified. This invaluable information typically becomes history
very quickly with conventional modelling, but with operational modelling, where the ~odel will
continue to be "live" for an indefinite period, this presents an opportunity to create a database which
can be used as a source of basic information on a day-to-day basis. WESNET has been developed
to exploit these possibilities in the following way:
• GIS data is available as background and is user selectable such that any key features can be
included.
• Photographic data of key features is available.
• Schematics can be overlaid and organised in one place.
• Basic pump curve information can be organised and catalogued.
• User defined or imported polygon data can be used for a variety of spatial analysis tasks
Overlays of point data (eg customer data, burst or leak locations etc) can be incorporated
into the model to provide greater insights into modelling results.
The major benefits are, therefore, in verifYing current knowledge about the network and in
organising and managing key information about the network.
This is a hitherto unexploited feature of network modelling which can only be used with
confidence if the model is constantly updated, which, of course, is the key feature of the operational
model.
1.5.2. Initial Calibration
This phase usually throws up numerous anomalies concerning current "knowledge";

typically these might be methods of pump control, zoning, closed valves, et cetera. The
395
achievement of the overall balance will usually indicate areas of possible leakage - the Wessex
experience is that the overall balance nearly always indicates such areas and that quite significant
leakage can be found on pumping mains and trunk mains upstream of main service reservoirs, areas
which are sometimes ignored with the traditional UK method ofleakage control. The key factor in
model construction is to ensure that the complete network is modelled, from source pumping to
demand area.
1.5.3. Pumping Costs
A key cost centre in any operations budget is power; energy costs for water production in
Wessex are £4-5 millions per annum. Containing and reducing those costs is one of the key
objectives in maintaining an operational model and WESNET has been developed to reflect that
need.
Pump controls and tariffs are User selectable so that any combinations can be tested and
evaluated. Wessex estimates are that 2% reductions in electricity costs should be made annually
(roughly equivalent to £lOOK per annum) through a combination of pump control, tariff selection
anti waste control; the savings may be far greater than this - Wessex has concentrated on energy
cost reductions for some ten years and thus the annual savings still possible are around 2%.
1.5.4. Telemetry
Anomalies between telemetry and model results may be caused by transducer error, and
WESNET provides a useful first-line check on flow-meter calibration. Flow meters may be
configured in groups to balance against one another, thus highlighting possible errors. The Wessex
experience is that telemetry data has proved generally reliable but that there will be occasional
errors. Since 9()01o+ of flow data is now collected through telemetry (at a major saving in manual
data collection) it is essential to have low-cost routine flow data verification - WESNET provides
this facility.
In addition, WESNET provides an excellent means of analysing telemetry data off-line in its
own right; that is without any reference to the model. Most telemetry systems are very limited in
their ability to analyse historic data in a flexIble way; the operational model opens up this
information to users across a wide area at very low cost, for example using existing PC networks,
and in Wessex is used almost exclusively to analyse historic data.
396
1.5.5. Standards of Service
A primary concern to the Operations manager is to achieve the set Standards of Service at
least cost; but how does he demonstrate that this is being achieved? In the UK, network analysis
has been accepted as a primary means of demonstrating this to the Director General of Water, and
WESNET has been developed such that an operational model can do this on a regular basis. Each
demand area has one critical node modelled in it specifically to represent the "worst case" - typically
this may be a block of flats in a remote part of a zone, or properties on high ground. WESNET
includes a module which enables the duration of a pressure below a user-selectable standard (usually
15m) to be recorded and viewed. This feature is invaluable in highlighting potential areas of low
pressure since it can be used as a filter.
1.5.6. Emergency Management
The Operations manager has to live constantly with the risk of an emergency on the
network. The most common are mains burst, power disruptions, and contamination of the network
(pollution). It has long been recognised that a reliable operational model of the network should be a
useful decision support tool in the management of the emergency. Experiences in Wessex have
confirmed this view though the reality is that, in practice, real emergencies are not managed by
running simulations of various scenarios at the time of the emergency. The model does aid the
manager by enabling the nuYor scenarios to be tested in advance rather than flying blind at the time;
in particular the model confirms the limitations or bottlenecks in the system, enabling difficult
strategies to be clearly worked through in advance.
1.5.7. Duplication of ModeUing Effort
Finally, an unexpected benefit of operational modelling has been a reduction in duplication

of modelling effort throughout the Company, because it is now possible for all departments to make
use of an up-to-date, fine-tuned model of the major urban supply systems, rather than repeat the
effort at regular intervals because confidence in previous versions has diminished. PC-networking
now means that Users across the Company have access to a central library ofup-to-day models on
line, together with all of the relevant management information, including graphical, telemetry, data
loggers etc. This is invaluable and greatly improves the efficiency of other departments.
In summary, therefore, the experiences in Wessex have demonstrated that the concept of
using operational models in water supply management can be translated into reality; that the benefits
397
are many and fully jusillY the small cost incurred in construction and ongoing maintenance - the
benefits increase with time. Many other benefits have been realised which were not foreseen at the
outset, and the narrow concept of operational modelling has now evolved to a much broader one of
NETWORK MANAGEMENT; ie the model becomes an essential element in a much wider
concept, that the supply network is actually effectively managed through the system, which
becomes the managers "toolkit" for this task. This concept is now overlapping with that of Quality
Assurance on the management of the network; the system confirms that the supply network is
operating to targets of cost and standards and demonstrates that this is the case.
Perhaps the most telling evidence that these techniques can have a major impact on
operational management is that Wessex has consistently avoided any restrictions on water supply to
its customers during a period of consistently low rainfall from 1989-1996 and has been consistently
regarded most highly among the UK plcs for its standards of service, in contrast to many of its
neighbours.
1.6. REFERENCES
Bea1 D. (1988) Integrated Distribution Management Systems. Aqua, Great Britain, No.5, pp.
267-276.
Bessey S. G. (1985) Progress in Pressure Control. Aqua, Great Britain, No.6, pp. 325-330.
Bhave PR (1991) Analysis of Flow in Water Distribution Networks. Technomic Publishing
Company Inc., Lancaster, USA
Bland P., Townend IC. (1987) An Integrated Approach to Telemetry - North West Water
Authority - A Regional Telemetry Scheme. Symposium WaterlEau 2000, Nice, France, pp.
9-2-1 to 9-2-14.
Brammer L.F., Schulte AM. (1993) A Joint US and UK Approach to Water Supply and
Distribution Modelling. Journal of the Institution of Water and Environmental
Management", Gr~tBritain, Vol. 7, No.5, pp. 471-480.
Chase D.Y., Ormsbee L.E. (1993) Computer-Generated Pumping Schedules for Satis1)ring
Operational Objectives. Jour. of American WaterWorks Assoc., Vol. 85, No.7, pp. 54-61.
Chase D.Y., Jones G.L. (1993) Lessons Learned from Application of Optimal Control Technology.
Integrated Applications in Water Supply, ed. B. Coulbeck, Vol. 2, JOHN Wll.EY &
SONS, pp. 377-394.
Clark RM. (1994) Modelling Water Quality Changes and Contaminant Propagation in Drinking
Water Distribution Systems: A US Perspective. Aqua, Vol. 43, Nr. 3, pp. 133-143.
Cohen I (1990) The Development of A Dynamic Calculation Model for Drinking-Water Networks.
Aqua, Vo1.39, No.3, pp. 172-187.
Cohen I, de Visser P.M. (1992) The Development and Application of Monitoring Systems for the
Distribution of Drinking Water. Aqua, Vol. 41, No.6, pp. 352-359.
398
CosgriffG.O. et al. (1985) Interactive Computer Modelling:Monitoring and Control of Melbourne's

Water Supply System. Water Resources Research, Vol. 21, No.2, pp. 123-129.
Coulbeck B., Orr C.H. (1988) Computer Applications in Water Supply", Volume 2: "Systems
Optimization and Control. Reseach Studies Press, Ltd., Herts, Great Britain.
Coulbeck B., Orr C.H. (1989) Computer Programs for Real-Time Optimized Control of Water
Distribution Systems. HYDROCOMP '89: Computational Modelling and Experimental
Methods in Hydraulics, Elsevier Applied Science, London 1989, pp. 289-300.
Cubillo F., Mercier M. (1993) An Approach to Water Quality Modelling in the Paris and Madrid
Distnbution Networks. Water Supply, Vol. II, Nos. 3/4, pp. 377-385.
Cullen N. (1987) Computing - Not Yet Half-Way There. Symposium WaterlEau 2000, Nice,
France, pp. 3-3-1 to 3-3-12.
Fallside F. (1977) On-line Control of a Water Supply System. International Workshop on
Instrumentation and Control for Water and Wastewater Treatment and Transport Systems,
PaperNo. 40, London-Stockholm, pp. 6-40-1110.
Filip A, Obradovic D. (1986) An Advanced Method for Prognosis and Control of Water Quality in
Complex Water Distribution Systems. Proceedings of XVI Congress of the International
Water Supply Association, Spec. Subject No.l3, Rome, Italy.
Fontaine 1. (1983) Towards Integral Automation in Water Distribution (in French). Aqua, Great
Britain, No.6, pp. 302-312.
Glasbrook D. (1993) The Practical Application of GIS, and its use for AMP2. IWO Conference
1993, Carrlifi: Great Britain.
Grombach P. (1986) Centralized Monitoring and Control of Water Supply System of the City of
Zurich: An Example of the Systematic Optimization of Operation by Computer. Colloque
Eau et Informatique, Paris, France, pp. 216 - 225.
Haddon M. (1994) Ringing the changes for Meter Reading. Water Bulletin, Great Britain, No. 60S,
pp.9-1l.
Halpern 0., Pascal O. (1987) Water Distribution System for the 3rd Millenium" (in French).
Symposium WaterlEau 2000, Nice, France, pp. 1-1-1114.
Hoogsten KJ., van der Zwan 1.T. (1985) Distribution System Modelling. Aqua, Great Britain, No.
5, pp. 247-25l.
Huntington R (1984) Improvement of Cost and Performance of Distribution Systems. Proceedings
of XV Congress of International Water Supply Association, Spec. Subject No.9,
Monastir-Tunisia.
Huntington R (1990) WESSEX WATER's Integrated Water Distribution Management System.
Proceedings of the 5th IAWPRC Workshop, Japan.
Huntington R(1993) Updating and Development of A Regional Telemetry Scheme. 6th IAWQ
Workshop on Instrumentation, Control and Automation of Water & Wastewater Treatment
and Transport System, Canada.
Jeffery 1., Taylor CR (1993) Development in Leakage Control at North Surrey Water Ltd - A
Case Study. Aqua, Vol. 42, No.4, pp. 223-232.
Jowitt p.w., Garrett RT., Cook S.C., Germanopoulos G. (1989) Real-Time Forecasting and
Control for Water Distribution. Coulbeck B. and C. H. Orr: "Computer Applications in
Water Supply", Vol. 2: "Systems Optimization and Control",Research Studies Press, Ltd.,
Herts, Great Britain.
Kado M., Itoh H. (1987)A Satisfactory Water Service Expert. Symposium WaterlEau 2000, Nice,
France, pp. 2-1-1/20.
Lonsdale P.B. (1984) Automatic Control of Pressure in Distribution Systems. Wessex Water Co.,
Bristol, Great Britain.
399
Males RM., Clark RM., Wehrman PJ., Gates W.E. (1985) Algorithm for Mixing Problems in
Water Systems. Transactions ofthe ASCE, Jour. of the Hyd. Div., Vol. Ill, No.2, USA
Morita S., Arakawa T. (1989) Water Distribution Control System. WATER NAGOYA '89,
ASPAC IWSA(1S-22-B22), pp. 626-635.
Nguyen B. (1994) Automatic Operation of the Water Distribution of the City of Paris. IWSA
Regional Conference, Zurich, Switzerland, pp. 311-317.
Nguyen B., Montiel A (1994) On-line Quality Control in Distribution Networks. IWSA Regional
Conference, Zurich, Switzerland, pp. 299-309.
Obradovic D., KordiC M. (1986) Studying a Disastrous Situation Before It Actually Happens.
Proceedings of XVI Congress of the International Water Supply Association, Spec. Subject
No.4, Rome, Italy.
ObradoviC D., Filip A (1986) The Value of Data Banks: The Case of Belgrade Water Supply
System. Colloque Eau et Informatique, Paris, France, pp. 111-126.
Obradovic D. (1989) Modernisation of Urban Water Supply Systems. Proceedings of the NATO
Urban Water Resources Advanced Research Workshop, Isle of Man, UK, pp. 150-165.
Obradovic D. (1993) Mathematical Modelling of Water Supply Systems. IRE - International
Institute for Infrastructural, Hydraulic and Environmental Engineering, Delft, The
Netherlands.
Rao HS., Don W.B. Jr. (1977) ExtendedPeriod Simu-lation of Water Systems. Transactions ofthe
ASCE, Journal of the Hydraulics Div., - part A, Vol. 103, HY-2, pp. 97-108, - part B,
Vol. 103, HY-3, pp. 281-294, USA
Robertson J. (1993) Optimum Network Management through Integration of Operations Processes.
Integrated Applications in Water Supply, ed. B. Coulbeck, Vol. 2, JOHN WILEY &
SONS, pp. 317-329.
Santoni A et al. (1987) EXPERT Systems - New Tools for Water Treatment. Symposium
WaterlEau 2000, Nice, France, pp. 6-1-1/8.
Schulte AM., Malm AP. (1993) Integrating Hydraulic Modelling and SCADA Systems for
System Planning and Control. Journal of American Water Works Association, Vol. 85, No.
7, pp. 62-66.
Snoxell J.D., Obradovic D., Cavor R (1989) The Use of Real-Time Telemetry and Modelling in
Conjunction for the Operational Management Water Supply Systems. Computational
Modelling and Experimental Methods in Hydraulics - HYDROCOMP '89, Dubrovnik,
Elsevier Applied Science, London 1989, pp. 301-310.
Snoxell J.D. (1991) New TeChnology in the Water Industry - Making It Stick in the Real World.
Proceedings of the Pipeline Industry Guild, England.
Snoxell J.D. (1993) Network Management in Water Supply. International Conference "Computer
Applications in Water Supply and Distribution", De Monfort University, Leicester, England,
pp.331-345.
Snoxell J.D. (1993) The Use and Value of GIS in the UK later Industry. Experience in WESSEX
WATER Services Ltd.
Snoxell J.D. (1994) Management ofa Widespread Multi- Plant Water Supply System.
Takagi T. et al. (1983) Distribution Network Control for Water Supply System. Hitachi Review,
Japan, Vol. 32, No.5, pp. 259-264.
Walski T.M. (1984) Analysis of Water Distribution Systems. Van Nostrand-Reinhold, New York,
USA
Zwan J.T. van der (1988) Computer Utilization in Distribution. Aqua, Great Britain, No.1, pp.
9-13.
2. Object-Oriented Hydroinformatics Tools for
Water Distribution Analysis
Petr Ingeduld
Vladimir Havlik
Stanislav Vanecek
Pavel Jirousek
Charles University
Prague
CZECH REPUBLIC
2.1. INTRODUCTION
A notion proposed by Abbott (1991) concerning primarily the software and hardware
development of simulation tools seems to have come true. New achievements in hardware and
software, such as Dual-processors, high density media opens a space for new operation
systems, Windows NT for example. The gap between compilers that can be used to design
different modules of software, such as model management, computer graphics, database
connections and the numerical algorithm itself, is rapidly deepening. The outlook of present
word processors, spreadsheets, high standard level of pre-and-post processing techniques
makes new demands on developers of mathematical models to change their programming
paradigm. Knowledge encapsulation and integration of several tools into a numerical model
may easily navigate a user through a modular part of a software. The authors would like to use
the HYPRESS package as an example of possible way satisfying some of the required features.
NATO AS! Series. Partnership Sub-Series, 2. Environment - VoL 15

402
2.2. WATER DISTRIBUTION
The problem of the analysis of water distribution networks consists in determining both
flows in every link, such as pipe, pump, valve etc. and the piezometric head at every node in
the system. Regardless of the fact that most of the recent published studies solving the water
distribution problems are based on standard methods, such as Gradient Methods, Newton-
Raphson method, The Linear Theory or The Optimization techniques, there exists an unique
approach published by Verwey (1994) based on a Finite Difference Method. Since the idea
seems to be applicable the authors expand on this technique in this Section.
2.2.1. Algorithm Based on Global Gradient Method
The steady flow in water distribution networks is based on the simultaneous fulfilment
of a mass conservation and an energy conservation law. The flows in the links (branches) and
piezometric heads at the nodes are interrelated via the head loss formula and the water
distribution problem is restricted to determining either link flows or nodal piezometric heads.
An algorithm is based on one of the most advanced and robust technique, the Global
Gradient Method with an incomplete Choleski conjugate gradient algorithm, as the base matrix
is always symmetric, positive definite, highly sparse, large and Stieltjes type. This algorithm has
been tested under several modes including Windows 3.1 enhanced mode, DOS 32-bit
protected mode, in order to provide as short computation times as possible.
Pipe network consists of branches (pipes), nodes with unknown heads and nodes with
fixed heads. It may contain hydraulic components located along pipes such as pumps and
valves, but their influence is included in equations describing head loss in pipes. Nodal
hydraulic components such as simple tanks, surge chambers, feeder tanks and air vessels are
schematised by means of nodes with fixed or unknown heads. Steady flow in the network is
achieved when both the heads in nodes and the flows in branches don't change in time.
2.2.2. Branches
There are two parameters characterising each branch, namely a and 13. For a given flow
rate Q be the flow in a particular branch, relationship between flow and head loss (or head
difference) may be written as:
(1)
403
where a contains the summary infonnation about roughness of the pipe, its length and local
losses such as losses in fittings and valves, while 13 is an additional parameter which is non zero
only if branch contains a pump.
If a branch connects two nodes with heads HI and H2 and is characterised with
parameters a and 13, the flow Q in this branch (if the "positive" direction is from node with
head HI) must meet this condition:
(2)
where hL is the head loss. This equation can be then expressed as:
(3)
2.2.3. Nodes with Unknown Heads
Each node is characterised by its consumption. Ground level does not have to be
specified. Consumption can be either a positive or a negative number. Positive value means
real consumption, negative one represent of a source with a fixed flow.
If there is n branches connected to a node and the the flows QI, ... Qn in these branches
are positive or negative depending whether they are coming in or out, the mass conservation
law for the nodal demand q must be satisfied:
q=QI + ... +Qn (4)
2.2.4. Nodes with Fixed Heads
In order to solve the system of linear equations, at least one node with fixed head is
required. These reference nodes (usually represented by reservoirs) can provide arbitrary flow
(even keeping their water level constant). Consequently they do not have to satisfy the
Equation (4) in the present fonn.
2.2.5. Algorithm
The recursive algorithm is used to generate an initial flow solution satisfying the nodal
balance in the first iteration. The initial solution is generated by taking the identity matrix I
404
(n=I). The linear system of simultaneous equations in the flows is coupled to the nodal balance
linear equations and solved iteratively to produce consecutive flows. The explicit inverse of the
matrix on the right-hand side of the equations is not needed, an iterative method based on a
preconditioned conjugate gradient method can be effectively used for the solution of linear
system of equations.
2.2.6. The Incomplete Choleski Conjugate Gradient Algorithm
As the coefficient matrix A is always symmetric, positive definite, highly sparse, large
and it is of Stieltjes type, it seems to be the best way to solve the set of simultaneous equations
using the incomplete Choleski conjugate gradient algorithm. Working with sparse matrix, only
non-zero elements are stored in the computer memory to reduce the storage and computation
time requirements.
2.2.7. Finite Difference Method
The domain of the finite difference method consists of branches connected one to
another by means of nodes. Minimal number of grid points is generated along branches and
they represent in fact the place, where governing equations are computed. Various hydraulic
structures (such as simple tanks, surge tanks, feeder tanks, air vessels, valves, pumps, etc.) can
be included any time at selected places of the domain (Figure I).
Figure 1. Discretization of network
Branches represent pipes of constant properties and may include various controls e.g.
valves, pumps, etc. Nodes represent free branch ends, branch connections or a specific storage.
Boundary conditions are defined at the nodes by setting up values of hydraulic gradient line or
405
discharge as a constant value or as a function of time. For a branching junction the continuity
equation is used and a common head is assumed since the minor losses are neglected. Simple
tanks, surge chambers, feeder tanks and air vessels are located in the nodes. The system
requires a computational grid to define small sub-sections on a branch over which the
governing equations are numerically solved.
The governing equations describing one-dimensional unsteady pressurised flow in a
pipe are the continuity and momentum equations:
oQ gA oH
-+--=0 (5)
ox a 2 of
and the momentum equation has been modified as:
oQ oH f
a -+gA-+-QIQ\=o (6)
m of ox 2DA
where: Q discharge [m3/s], H- piezometric head [m],
f Darcy-Weisbach friction factor [-],
D diameter of the pipe [m],
A area of the pipe [m2],
x, t space and time co-ordinates [m], [s],
momentum diffusion coefficient.
Governing Equations (5, 6) are solved by Preissmann implicit finite difference scheme.
The space centred scheme is stabilised by the weight parameter e equal to unity and high value
of a. m - momentum diffusion coefficient. The value of a. m has its signifficant effect especially
during the first several steps of fiding the equilibrium. The initial conditions are usually set as
follows: zero discharges and constant pressure head corresponding to the highest pressure
head in the system (including pump characteristics).
During the first ten time steps the boundary conditions achieve the desired value of
discharges and pressure heads at the system boundaries.
The loop algorithm is based on the fact, that a loop network contains elements known
as nodes which represent the confluence of several flow paths, some of which originate from
406
other nodes, some from boundary points. A system of simultaneous linear equations is
developed where the piezometric head H changes at each node serve as the only unknowns.
The solution of this system by any matrix elimination technique yields the H values at each
node. The H and Q values at all intermediate grid points between the nodes are expressed in
terms ofH variation at the nodes.
Figure 2. Scheme of computational grid
Supposing, that there are three links, 21, 23 and 24, along branch 23, b grid points and
c grid points along a link 24 (Figure 2.). For any computational grid point the Equations (5, 6)
may be written as:
(7)
(8)
where L, M, N are functions of coefficients (A, B, C, D, E)I,2 found through a double sweep
elimination.
These Equations express the partial dependence of the unknown variables Q and H at
any grid point on a branch on the value ofH in the two adjacent nodes.
At internal nodes a compatibility condition must be satisfied. The most simple one is
that of continuity and common hydraulic piezometric heads:
(9)
407
h !1+1
II = h!1+
12 =... = h!1+
1 ik =... = h!1+1
1 un (10)
where n+ 1 indicates the (n+ 1).1.t time level in the solution, k is index of the links which
emanate from the node i, and m is number of links. These relations can be written for each
node ofM nodes and leads to a system ofM linear equations having as unknowns piezometric
heads at each node.
[S] {h} = {T(L)} (11)
where [S] is a coefficients matrix, M x M elements, {h} is a vector of unknowns, M elements;

{T(L)} is a vector of the free terms, M elements.
This system of linear equations may be solved by any matrix inversion techniques.
Once the increments of piezometric heads R are known at the nodes, it is possible to
recompute Q(i) and R(i) values for all intermediate grid points by means of Equations (7, 8).
The loop algorithm may be expressed by the following steps:
• the coefficients of the implicit Preissmann scheme discretize the governing equations
between two successive grid points on a branch;
• local elimination method is used to express Q and R grid point values on each branch in
terms ofR at the branch ends (nodes);
• one equation for each node leads to the system of linear equations that is solved by a
matrix elimination method;
• substitutions inside the branches yield the values of Q(i) and R(i) values for all
intermediate grid points from the known values ofR at the branch ends.
The solution of this system by fast topology-based searching technique (Verwey, 1994)
yields the R values at each node. The required Rand Q values at all intermediate grid points
between the nodes are expressed in terms ofH variation at the nodes.
408
2.3. DESIGN OF A NUMERICAL SIMULATOR
The modem software package architecture should consists of several independent

modules that are coupled together by means of database and internal data files. The data pre
and post-processing is based on a storage data in a standard database that gives QBE (query by
example) and/or SQL (structured query language) options to classify groups of pipe diameters,
roughness, wave celerities. The same facility may be used for the visualisation of
computational results, such as pipe discharges, velocities etc. The graphical user interface may
be designed on the basis of multi document interface. This implies, for example, that during
editing of data the changes are visualised immediately in graphical windows. All stored data in
the data base can be imported or exported in ASCn format. This can always be neatly
organised for standard data, such as values associated with the description of pipes and nodes.
The output processing part covers the most important features as time series of longitudinal
profiles of pressures, discharges and pipe velocities with optional envelope drawing. Colour
animation; time series plots of the same variables with interactive graphical selection of point
location; colour animation of time series in the horizontal plan; zooming of any graphical
screen and co-ordinate reading on any function or network plot by means of the cursor.
Data-base files Database
I Jnputpart
II Output part J
I Data Export I '1\ Data Importl
I tt.. ASCII files, binary filesI
Itt..
\VI \VI
FINITE DIFFERENCE MEmOD GLOBAL GRADIENT MEmOD
IWater Distribution I I Water Distribution I

IInitial State for Water Hammer I IQuasi-steady Flow I
ISlow Transients I
IWaterhammer I
Platfonn : Platfonn :
Windows 3.1 Windows 3.1
DOS
DOS 32-bit mode, DBOS extender
UNIX
Figure 3. Software architecture

409
Computation models can perfonn water distribution calculations, slow transients and
water hammer simulations. The models are independent programs that read data from internal
data files prepared by data preprocessing part. The results are stored on hard disk in Ascn
and binary files. These files are transferred into the output module. Software architecture with
respect to computational models is shown in the Figure 3. describing an example the
HYPRESS architeture.
This implies, for example, that during editing of data the changes are visualised
immediately in graphical windows. The data post-processing part covers the most important
features, such as:
• time series of longitudinal profiles of pressures, excessive-pressures, discharges and pipe

velocities with envelope drawing and their colour animation;
• time series plots of the same variables with interactive graphical selection of point location
• colour animation of time series in the horizontal plan;
• zooming of any graphical screen and co-ordinate reading on any function or network plot
by means of the cursor.
2.4. THE HYPRESS EXAMPLES
2.4.1. Loop Network
The following example shows the layout of a proposed pipe loop network for a fire
fighting system (Figure 4.). Under conditions with zero hydrant discharge, a pressure of
approximately lObar is· maintained in the system by small pumps. In the event of a fire (one
fire at a time is assumed) an electric duty pump supplies the demand flow to any of the hydrant
points.
The steady-state head-discharge characteristics of a centrifugal pump has been
approximated as H= 101.9-875 Q2. The rated speed is 1000 r.p.m. The pump sump is
connected to one of the nodes. The results of computation for a steady state discharge
Q = 0,035 m3ts with flow distribution within the pipe network is indicated by arrows in
Figure 4. Pump operating point is Q=0.034 m3ts with the corresponding head H=108 m. The
steady state calculation gives the initial conditions for a transient flow analysis which has to
410
solve the pump trip, opening and closing of the hydrants, proper behaviour of a check valve
and the air vessel performance.
150.00
500.00
250.00
0.00
0.00 250.00 500.00 150.00 1000
Figure 4. The layout of a proposed loop network
2.4.2. Large Networks
Algorithm was successfully tested on large pipe networks with several thousands of
nodes and pipes (Figilre 5.). The computational requirements were significantly lower using the
Global Gradient Method than in case of numerical process based on the unsteady flow
algorithm. Since both these algorithms have been implemented under different programming
languages it is very difficult to get comparable results.
The Global gradient method with the Incomplete Choleski Conjugate Gradient
Algorithm seems to be more robust and has a faster convergence. Basic tests using the Water
Distribution Models have been caried out on the existing network topology of 2304 branches
and 2255 nodes.
Several test covering different boundary and initial condition have been computed to
demonstrate HYPRESS ability to simulate the staedy flow. The total length of the pipe
411
branches was 74938 meters, the shortest pipe was 1.00 m long and the longest one was 236 m
long; diameter varied between O.lOm to 1.00m. There were 574 boundary nodes with a given
pressure head and/or discharge type of the boundary condition.
536000.00
535000.00
534000.00
533000.00
251000.00 258000.00 253000.00 260000.00 261000.00
Figure 5. Large network with several thousands of nodes and pipes
2.5. COUPLING GIS WITH A NUMERICAL SIMULATOR
Numerical models for solving water distribution analysis are based on a certain
topological schematization. The solution domain consists of branches connected one to
another by means of nodes and of various hydraulic structures (such as simple tanks, surge
tanks, feeder tanks, air vessels, valves, pumps, etc.) that can be included at selected branches
or nodes. The models usually have their internal database in which the data are stored together
with results of hydraulic calculations. Most of the models support import and/or export of
ASCII data. Hence the data can be presented in specialized graphical packages or can be
412
imported into geographic infonnation systems. GIS model can be efficiently used for data
preparation and verification and for data post-processing and visualization. Input part for the
model may cover creating of digital model of terrain, map plans, digitizing, scanning and
vectoring of data, creating cross-references for the numerical model etc. Output part may use
GIS as a graphical data processor for displaying results from numerical simulator in the fonn of
graphs, maps, thematic maps, 2D, pseudo 3D and 3D images, browsing results, using of grid-
analyst tools to work on the computed data and for geocoding of data.
GIS commands for selecting objects with the help of querying, selecting and SQL
Select enable the preparation of input data for the model and the analysis of the results. Using
SQL Select it is possible to create query tables containing infonnation which was only implicit
in the data base tables. SQL utilisation in combination with other basic units greatly improves
the efficiency of the application of simulation packages.
Data from Numerical Models

CAD oriented data
NODE BRANCH
NODE
o ,,-y.(z)
,,-y. (z) ,,-y.(z)
BRANCH NODE BRANCH

NODE
node number branch number (x,y) coordinates (x,y) coordinates
(x,y) coordinates branch characteristics (node number) (branch number)
ground elevation cross.reference data (olber data)
(oU,er data)
cross-reference data flrst and last mode
number of attached branches results of hydraulic modeling
list of attached branches
results of hydraulic modeling
GEOCODING ----......::0>..
-E-- CREATING OF CROSS·REFERENCES
Figure 6. Two different types of the input spatial data describing a model topology
There are two different types of the input spatial data describing a model topology, we
can distinguished between CAD oriented data models and data from numerical models
(Figure 6.). CAD models usually work with entities such as points, lines, polygons and
surfaces. Each object has its own spatial co-ordinates and the objects are stored and visualised
independently. Descriptive data attached to these objects may contain attributes such as colour,
shape of the objects or a data base attributes such as number of en element, values from
catalogues etc. On the other hand, data from numerical models do not necessarily have spatial
413
co-ordinates (x,y co-ordinates are usually used for pipe network visualisation but not really for
the simulation itself), but ground elevation data may be included. There are several types of
cross-reference data describing the pipe network topology (start and end node attached to a
branch). Data exchange between these two models requires creating of additional operations
such as geocoding, creating cross-references and data pre-processing. As a result the data can
be either used in CAD systems and visualised with respect to their spatial co-ordinates or the
same data can be used as an input into a numerical model.
Raster data are mostly used for creating an intelligent background of the map layers.
The map layers contain additional information that neither is stored in a database nor available
in a vector form. Raster data can describe descriptive data into more detailed level, image of a
certain place of the pipe-network, schematic pictures, etc.
The choice of GIS modules should respect the specific requirements of the end-user,
the type of data that will be maintained and the size of the projects. It is possible to
distinguished between 'small' and 'large' GIS because the hardware and software requirements
(the choice of GIS modules, database, hardware platform) limit the project.
The coupling of GIS with numerical models can be one of the following three types.
• GIS does not exists: In case that there is no GIS in the beginning of the project, it is
possible to develop own GIS and use it both for data preparation, verification, analysis and
for presenting of the hydraulic modelling results.
• GIS exists, but does not include layers with data we need for hydraulic modelling. In this
case the existing GIS is a source of general information. We need to develop additional
layers and data attributes and this layer is later on included into GIS.
• GIS exists and contains a layer with data we need. Then GIS is a source of data, there is a
structure already created to contain results from the hydraulic model and the structure may
be only slightly modi-fied if necessary.
The Figure 7. shows a map consisting of different layers; pipe networks (drawn by lines
between two points with known x,y co-ordinates), manholes (a symbol on x,y position),
several layers with map information such as block of flats, green areas, river (drawn by the
means oflines, polylines and areas). Despite of their common cross-references, the individual
objects on the map are drawn independently. Since the descriptive data attached to graphical
objects, the database information can be displayed interactively on the screen or stored in a
browse table. Each graphical objects can be shaded by a value corresponding to the value or
function of its descriptive or graphical attribute, for example the object length as shown on the
414
picture. Descriptive data without any graphical representation can be geocoded, i.e using a
common attribute they can be assigned to an existing geographical data and can be visualised
on that data. It is also possible to create join tables or to aggregate data.
A powerful SQL language gives the possibility of selecting objects satisfYing certain
conditions. It is possible to display pipes within a given length range or with a diameter higher
than a given value etc. Moreover, geographical operators "within", "intersects", "contains",
"contain entire" can query data accordingly to their geographical location. Hence it is possible
to display intersecting pipes, manholes that are located within certain street, etc. The results of
SQL queries are stored in separate tables and they can be later used to perform desired
operations, or for creating a new mapper. Graphs and simple statistical functions are also
usually a part of GIS modules. Data is mostly stored in an internal or external database and can
be imported or exported in several data types. ASCII and xBASE files describing both
geographical and descriptive data are almost always supported, most systems usually recognise
DXF files for graphical data.
Figure 7. Map with different layers
The Figure 8. shows a connection between Digital Elevation Model and results of
water distribution analysis. A part of pipe network consists of several pipe branches, nodes and
a control valve. The results of water distribution model are pressure heads in nodes and flow
rates in links. Digital Elevation Model (OEM) can create a terrain contour lines or can be used
as well to create pressure isolines from the known pressure heads in the nodes. This additional
415
information is drawn together with pipe network on a map and graphical objects-polylines may
be linked to descriptive data of corresponding pressure head. It is possible to use SQL tools to
query pipes and nodes according to the pressure isoline using standard GIS SQL tools.
Coupling of DMT with a numerical simulator can bring a different view on computed results
and can help user to carry out water distribution analysis. Different GIS modules as statistical
tools, grid-analysts, databases and data visualization packages have very powerful general
tools that can extend the use of numerical models and support analysis of the results.
_ .00
800.00
100.00
'''-~'.
600.00
/
'r;
Figure 8. Connection between DEM and result of water distribution analysis
2.6. CONCLUSIONS
A modem software development is nowadays based on the technology of object linking

and embedding, dynamic exchange of data between different applications, unifying database
platform and user-friendly Application Programming Interface. The present technology of
416
geographic infonnation systems based on the visualisation of infonnation gives new

possibilities to handle an increasing amount of data. GIS as a part of hydroinformatics systems
have the potential to perfonn a complex data analysis. Data visualization techniques and
analytical tools can improve the everyday work of engineers, scientists and researchers. The
link between these models and numerical simulators supports integration of modelling systems,
knowledge-based tools and infonnation systems.
A complex package for water distribution analysis has to cover hydraulic design and
operation of the pipe networks. The software has been designed on the basis of interactive
graphics support to data editing through a multi document interface. During editing, all logical
and topological data types are processed immediately and, wherever appropriate, visualised.
The information system on a pipe network is based on the storage of data in a standard
database.
The numerical models can solve on a pipe network both water distribution and
hydraulic transients including real time control emulation on a variety of hydraulic components
such as pipes, pumps, valves, reservoirs, surge chambers, feeder tanks and air vessels. Two
different numerical algorithms have been used for the steady-state simulation, a numerical
solution based on a finite difference scheme with minimal number of grid points and
an alternative algorithm based on The Global gradient method with an Incomplete Choleski
conjugate gradient method, with examples demonstrated on the HYPRESS package.
2.7. REFERENCES
Abbott M.B. (1991) Hydroinfonnatics: Infonnation technology and the aquatic environment.
The Avebury technical, A1dershot, UK and Brookfield, USA.
Ajiz M.A., Jennigs A. (1984) A robust incomplete Choleski conjugate gradient algorithm.
International Jourrlal for numerical methods in Engineering, Vol. 20
Deakin R. (1994) Data to infonnation : GIS and decision support for coastal engineering.
Hydroinfonnatics '94 Delft, 1994, Balkema, Rotterdam.
Fabri A.G. (1992) Spatial Data Analysis in Raster-based GIS: An Introduction to geometric
Characterization. In Introduction to GIS edited by C.R. Valenzuela, lTC, 1992.
Kershaw D. (1987) "The incomplete Choleski-conjugate gradient method for the iterative
solution of systems of linear Equations", Journal ofComputationaJ Physics, 1987
Maksimovic C., Prodanovic D., Eigy 1., Fuch L. (1994) GIS (or GIM) in water projects -
Tools or Toys. Hydroinformatics '94 Delft, 1994, Balkema, Rotterdam.
Salgado R., Todini E., O'Connell P.E. (1987) Comparison of the Gradient Method with some
Traditional methods for the Analysis of Water Supply Distribution Networks.
Leicester.
Tullis 1.P (1989) Hydraulics of Pipelines. John Wiley & Sons, 1989.
417
Valenzuela C.R. (1992) Basic Principles of Geaographic Information Systems. ITC.

Verwey A., Yu I.H. (1993) A Space-Compact High-Order Implicit Scheme for Water Hammer
Simulations. Proceedings ofXXVth IAHR Congress, Tokyo, 1993
Verwey A. (1994) Linkage of Physical and Numerical Apects of Models Applied in
Environmental Studies. Proceedings of the Conference on Hydraulics in Civil
Engineering, Brisbane.
Tomplin C.D. (1990) Geographic Information Systems and Cartographic Modeling. Prentice
Hall.
Chapter VII
Transient Flows
Chapter VII
Transient Flows
mTRODUCTORYSTATEMENT
The transport of water in pipes nonnally involves some sort of control and, the
imposition of certain constraints over the water flow. As the size of the water supply system
grows, the close control of flow becomes more necessary. The system is designed to meet
required demand profile, while on the other hand, water treatment and pumping could be
carried out according to some other (different) schedule. Adjustment of system parameters by
valve control and pump switching is required all the time.
In standard design procedures it is nonnally assumed that the system operates under
steady state conditions and that transient flow (unsteady flow) occurs when the system is in
transition between two steady state flows. Every change of flow rate in a pipe is associated
with a pressure change, the intensity of which is related to the rate of flow change. Slow
transients usually cause no 'problems to system operation, but fast transients, known as water
hammer, may cause serious problems.
The design process of a water supply system nonnally includes an appropriate hydraulic
analysis of all possible transient flow events and risk assessment of equipment damages and
system malfunctioning. Pumping stations have the main role in transient flow analysis.
Transients caused by them are usually the most severe, as boundary conditions in mathematical
modelling they are the most complex, and the equipment in pumping stations is that which is
hit initially by transients. The great variety of system layouts and pump arrangements in
pumping stations (single, parallel, series) do not allow general predictions to be made about
system behaviour in transients caused by pumps.
422
There are three classes of transient flow events that are important for a designer:
• Normal events, Le., events under normal operational conditions such as: correction of
flow or pressure by control valves, adjustment to changes in consumption, deliberate
switching on/off of a pump, etc.
• Emergency events, i.e. pump power failures, malfunction of control devices, pipe bursts
etc.
• Catastrophic events, i.e. earthquakes, strong explosions etc.
Normal events are not expected to cause significant pressure changes and to disturb the
normal operation of the system. They can be planned and expected. However, their
significance is usually overlooked especially in complex systems with several pumping stations.
The analysis of transient flow events in complex water supply systems with pumps,
valves, air chambers, sophisticated control devices, etc., can be carried out efficiently only by
computer simulations. Along with well established methods of transient flow analysis based on
the method of characteristics, there are some attempts to use methods developed in other areas
of Hydraulics and using new informatics tools. A method based on a high order compact
implicit scheme and object-oriented programming has been explained in one ofthe Sections of
this Chapter.
1. Pumping Stations and Transient Flow Analysis
Marko Ivetie
IRTCUW, Insitute of Hydraulic Engineering
YUGOSLAVIA
1.1. INTRODUCTION
The determination of the capacity, location and configuration of a pumping station has
to be based on a detailed analysis of the proposed system. Factors like forecasted demand,
available supply, function of the pumping station in the water supply system etc., have to be
thoroughly considered. Due to these requirements pumping stations in water supply systems
(WSS) almost exclusively have more than one pump, and lots of other devices and equipment
required for correct operation. In this analysis, however, only the schematic representation of
pumping stations will be used (for example, Figure \.) and equipment relevant for
mathematical modelling of transient flow will be indicated.
There is a lot of possible arrangements of pump units within a pumping station
designed to meet specific system requirements (parallel or series operation, variable speed units
etc.). The main task of a pumping station is to increase fluid pressure downstream, but also it
should be done at minimum cost and with an accepted level of reliability.
Regarding the suction side there are two main types of pumping stations. Pumps may
take water from an open surface reservoir, or from a relatively long suction line. In the first
case suction pressure varies in a narrow range regardless of the operating regime of the station.
If the pumping station and intake structure are located within a surface reservoir or in a well,
vertical turbine pumps with columns extending down into a suction reservoir are usually used.
Classical solutions are pumps that are taking water from a nearby reservoir through a relatively
short suction pipe. In this case, horizontal centrifugal pumps are usually used.
The other group of pumping stations is typical for use within distribution systems.
They are so-called booster pumping stations. Suction pressure may vary across a much wider
NATO ASI Series, Partnership Sub-Series, 2. Environment - Vol. IS

Edited by C. Maksimovic, F. Calamino, and J. Snoxell
424
range, which is dictated by the part of the system upstream of the pumping station.
Consequently, the question of efficient and reliable control becomes more difficult and
important than in other cases.
-!.l
I~ I HL (V"",·I=l
Pump
';--Curve
I ----IJ._ @
I ~1
IHp
I ZrZj
(J) I
1 Q
Figure 1. Schematic of a hypothetical pumping station, pump curve and flow control by throttling
Water supply systems are continuously growing in size and in complexity. New sources
of drinking water are more and more distant from water consumption area. Requirements for
efficient control and reliable operation become more restrictive all the time.
Pumping stations serve in water supply systems which are always in transition from one
steady state to another. Some transient flow situations (actually, the most dangerous ones) are
caused by pumps and it is necessary to be fully aware of their behaviour in some irregular flow
regimes and possible consequences.
1.2. PUMPS
The main elements in any pumping station are pumps and this is by itself, a wide topic
(the reader may refer to several textbooks, like Stepanoff, 1957, Karassik et a1., 1976,
Cherkassky, 1990, Dickenson, 1988, or those provided by big pump manufacturers).
425
The pump motor, through the torque applied on the pump shaft, delivers energy to the
impeller to cause flow through the pump and to increase pressure from the suction flange (s)
to the discharge flange (d) (Figure 1.).
In normal operation, the motor torque is balanced by reactive hydraulic torque of the
fluid pumped. The total dynamic head, H p' i.e., energy increase per unit weight of fluid is
equal to
v 2 Vs2
H =ITd -IT + i __ (1)
p s 2g 2g
where, ITd and ITs, are piezometric heads or elevations of hydraulic grade line on each side of
the pump. For a system with equal pipe area at (d) and (s), H p can be measured as a pressure
head difference by a differential manometer, provided that the suction and discharge flange are
at the same elevation. A typical head-discharge relation, or pump curve, is given in Figure 1.
For the reservoir-pump-valve-pipeline system in Figure 1. energy equation from
reservoir 1 to reservoir 2, yields the total pump head, H p in terms of pipe and valve
characteristics
(2)
The lift, or static head, is a difference in elevation between supply level and the discharge level,
(Z2 - ZI), and the rest is due to the system resistance. A. is the Darcy-Weissbach friction
coefficient, and !;,L is the valve head loss coefficient. For a constant value of, A. , Equation 2.
can be expressed as a parabolic function of Q
(3)
This is a traditional way to represent graphically the dependance of required pumping head on
system losses and static head, which make total system head.
Operating point of a pump or, of a pumping station, is at the intersection of pump
characteristic and system head curve.
426
260 l
\ I
I
240
-
:
~ fB-
220
In.' 1=10
100 200
1"'- ?
eJ~
~ ~ ~ ..........
I"
I
'"
180
1160 p,... ~~ I ';.~ r-
~ ~. ~
t:1 vl -
~
.c ~
-,-.r6(
f!Z /7 ~ ~ / I
~
l140
e ~ r--:2 ::...... I
~~ :
C120 '/ I
Ho=30
'#.
~ tr
100 rH. 10 -- -- -
~lr- 9j
~ ::",ll.r""JO
r-~
~r--:::::
40 80 0,
~ ~o f--
~~ ~~
I
~~ I
30 60 I f--
20 40 ~ ~s=3 I ':0
10- 20 it Y I '"
o o V/ I
o 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

% nonnal flow
240 I
220 "-'\, I
I
g. 200
.s 180 ""- I
~
I
] 160 .~
~ 140 ""- Po~"e I

<1JJQ I
8. 120 I!rgu n.F 0
1
,, n.~
100
r--- ns=9 - -..... =>95
".=6~
J...e!§!!z::v~I~ ~ ---~- :-
'--~=60 .......... llJ
d
'#.
80
60
-~..........:~ : >-..
~'" r.5ls
~;'~Iiii'!""'"' I
40 P1~ I r-
/rlt~'v I
I
20
o :
o 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
% normal flow
Figure 2. Standard pump characteristics

427
In addition to H p and Qp there are other quantities and performance characteristics

related to a pump: namely, shaft power P, rotational speed 11, efficiency 1], and some other,
non-dimensional parameters. Pump efficiency can be expressed as follows
pgQHp pgQHp
11=--=-- (4)
P TO)
where T is the shaft torque, and 0) is the angular velocity of the shaft (in rad/s). Instead of 0),
rotational speed, N (in rpm), is also used.

Various types of pumps are classified by their specific speed, lis,
N..fQ (30(2g)0.75
nS=HpO.75l=0 1[0.5 157.80
1 in SI units (5)
where, N is in (rpm), Q, pump discharge in (m3 / s) and, H p, pump head in (m) lis is an old,
dirtlensional definition, still widely used, while cr is non-dimensional.
Several researchers have provided standard Q-H-P diagrams for pumps depending on
their specific speed. Those diagrams result from the statistical analysis of several centrifugal
pumps. They cannot replace true pump characteristics, because they are not based on dynamic
similitude (so-called homologous relations, which are discussed in the following sections).
However, they do provide the basics for understanding pump behaviour in steady state
analysis. For example, radial-flow pumps (specific speeds lower than 50) develop relatively
small shut-off head and they are usually started with the downstream valve closed, while axial-
flow pumps (specific speeds, lis above 120) develop relatively high shut-off head and cannot
be started against a closed valve.
The most reliable information can be obtained by measurements in situ, but
unfortunately, this is not always possible.
1.3. MATCHING PUMPS TO SYSTEMS
Under constant flow conditions and where pressure requirements do not change, it is
easy to select appropriate pump(s) because the pump needs to match the system at a single
428
duty point. In complex systems, including all relevant regimes, water level fluctuations in
reservoirs, forecasted system development, roughness changes due to aging of pipes, a pump
is required to match the system at a number of different operating points. There is no unique
system head curve and no single duty point. It is the operating range of a pump, which has to
be defined.
Furthermore, there is a tendency towards local control of pumping stations (Record,
1988), which is based on the information available in the vicinity of the pumping station site.
Instead of the system head curve in the sense of Figure 1. we may speak about pressure head
required by the user at the control (monitoring, or, critical) point of the network (corresponds
to the level Z2) and about the discharge pressure head required from the pump due to system
losses between the pump and the monitoring point.
These analyses have to be done by computer simulation, though for basic
considerations the concept of system head curve will suffice.
NPSH (Net Positive Suction Head) requirements.

Also, it is necessary to check conditions on the suction side and to determine the net positive
suction head (NPSH) available. The required NPSH of the pump has to be smaller then NPSH
available, otherwise, it will result in very low pumping efficiency and in cavitation of the pump.
1.3.1. Operating Efficiency
The efficiency of the pump operating under normal conditions depends on how close
the system requirements are to the pump characteristic curve. It is important to select a pump
that will have its best efficiency within the operating range of the system and at the condition
the pump will operate most often.
In Figure 1. three parabolas are constructed, each corresponding to one valve position.
The vertical axis, zero flow condition, corresponds to the closed valve (I;L = 00). The
intersection of the parabolas with the pump curve (Hp - Q) represents the operating point for
each valve position. This is a common method of flow control, and probably the least
economical.
Typically for water supply systems (WSS), reduced flow is associated with reduced
pressure requirements. It is not acceptable to leave the consumers to control pressure in the
distribution network (high demand -+ low pressure, low demand -+ high pressure), because
excess pressures will generate increased leakage.
429
Pumps operating at lower flow rates (high pressure) consume more energy Oower
efficiencies) and their lifetime is shorter. Sometimes, a solution (not the cheap one) can be a
pump equipped with control valve and bypass line leading to the suction reservoir. There are
ways to overcome these difficulties and there is a price to be paid.
1.3.2. Reservoirs Serving as Buffers
With storage available in a downstream reservoir, pumps can be switched on and off so
that pumps always operate at the same fixed operating point. This is not always acceptable
because it requires larger reservoirs and may generate severe, and frequent, transients. Instead,
multi-pump schemes and/or variable speed units are used.
1.3.3. Parallel and Series Operation
The installation of multiple pumping units operating in parallel or in series, enables

more efficient use of pumps and allows finer, yet discrete, transitions in flow, whenever
required. Parallel installations are more common in water supply systems (Figure 3.).
discharge side
1 l. isolation valve
2. pump
3. check valve
~c=========================~ 4 control valve
suction side
Figure 3 . Four pumps in parallel operation
Pumping station characteristics can be determined by adding the individual pump

discharges associated With any particular head. Obviously, every additional pump in operation
increases the discharge of pumping station. However, the operation of two identical pumps in
parallel will not result in a discharge which is equal to doubled capacity of one pump, because
430
the system head curve will increase. In Figure 4. the operating points of one pump in a 4-unit
pumping station are indicated. Interval (a) is the operating range of the pump indicated as No.
1, while (d) is the operating point of a pump, when the fourth pump is switched on. Note that
the pump No.1, when operating alone, uses maximum power (this is true for centrifugal and
mixed flow pumps, which are more common in WSS) - and may operate in the zone of
dangerous cavitation (high NPSH required by pump).
100
Hp[m]
90
80
70
,, ,
.. -
,,'
!!------- -
p
60 80
[kW]
700
50 600
500
40 400
300
30 200
20
o
-
L -_ _ _ _ _ _ ~ ____ ~
1.0
____ ~
100
Q [m3/s]
_ _ _ _ _ __ L_ _ _ _ _ _ _ _ _ _ _ __ L_ _•
2.0 3.0
Figure 4. Operating range of unit no. 1 in a 4-unit pumping station
In determining pumping capacities of pumps in series operation (Figure 5.), pumping

heads are added (Figure 6.). They are used in systems with steep system head curves, and also
431
in booster pumping stations with a wide range of pumping heads. Two identical pumps with
capacity Ql at HI of system head, will, if working in series, discharge Ql at 2 x HI of system
head.
1.3.4. Variable Speed Applications
The flexibility and efficiency of a pumping station is significantly increased by adding a

variable speed pump. The major disadvantage is the high capital cost of the variable speed
device. The correct values of flow, head, power and efficiency could be determined only by
tests. However, this is an area where model similarity is used with the highest confidence. The
affinity laws give an acceptable answer to the question, through well known relations:
• flow: Q / Qo =N / No ,
• power: P / Po =(N / Not
1. isolation valve
2. pump
3. check valve
4. control valve
5. valve in by-pass line
pump "A"
"B"
Figure 5. Two pumps operating in a series
The efficiency remains much the same, regardless of the rotational speed N (provided
that it remains between 0.7 and 1. 0 of the rated speed No).
The efficiency of this kind of flow regulation can be seen in Figure 7. where flow
reduction from 55 Us at operating point A, to 25 Us is required. Pump characteristics are
shown for three speeds, 1200, 1500 and 1800 rpm. Dashed lines are affinity lines, and lines
432
with constant efficiency. v is non-dimensional flow rate (Q / Qo), and, a. = N / No, non-
dimensional rotational speed. The pump performance is compared with a fixed speed unit of
1500 rpm (although, the advantage over a higher speed unit is much more pronounced).
By throttling the system head curve can be changed, and the operating point would be
C. However, with a speed change to 1200 rpm, the new operating point would be B. The
difference between C and B is the energy loss caused by throttling and probably improved
pump efficiency.
Also, this pump can cover higher flow rates without any problems (there will be some
shift towards higher flow rates and lower efficiencies). This is a question of economy and
requires careful consideration, because variable speed units are more expensive than fixed
speed ones.
operating range
Figure 6. Operating characteristics of two pumps in series application

433
N=1800rpm
,, via = 0.5 N,.=1800 rpm.
QIf'O.0662 m Is
H,.=50m
60 T,.=200Nm
H[mJ
40
I
I
20
,,
I
I
I
I ,,'v/a = 1.0
,
,, ~
./
a ,'-
o 0.04
Figure 7. Operating characteristics of a pump for three rotational speeds
m~~~~~~~~~--~~~~~~~~--~--~--
Hp[m]
~ ~--~~~~~~--~---+~~~~~~~~~~~~~-+---
20~--+-~~~-+---4~--+---~---+---4----~--~---+---
10
o 100 200 300 400 500 600 700 800 900 1000 1100
Q [lis]
Figure 8. Operation of one variable speed pump in parallel with several fixed-speed units - Pumping
station "Schwechat", Austria
434
The best way to control the flow in a multi unit pumping station is to combine one
variable speed pump with several fixed speed ones, as in the case in Figure 8, where pump PI
is a variable speed unit.
1.4. TRANSIENT REGIMES
The analysis of transient regimes in systems with pumping stations is very important
because of the great destructive forces associated with these phenomena, known as water
hammer. This can not be achieved without reliable data on pumps in all possible operating
regimes. In transient regimes pumps interact with flow in a special way and also, they may
operate under very unfavorable conditions, which can hardly be identified in steady state
analysis.
According to the frequency of occurrence, transient events can be classified into three
groups:
1. normal, which occur very often, like pump start-up, flow adjustments etc.
2. irregular, which are expected, but uncontrolled, like pump power failure etc., and
3. catastrophic, with very low probability of occurrence.
The analysts usually overlook the risk associated with the first group of events and
concentrate on the second group of events.
1.4.1. Methods of Analysis
Basic equations for transient flow can be expressed in the following form (Wylie &
Streeter, 1978)
lOp DV
--+-+gsmcx+A.--=O
. vlvl (6)
pax Dt 2D
435
1 Dp 2 av
--+a -=0 (7)
pDt ax
Equation 6. is the equation of motion, and Equation 7. is the continuity equation. The
derivative ~ = .E. + V ~is for the slice of fluid having velocity V, p is centerline pressure,
Dt at Ox
and a, elastic wave speed in the pipe.
The bulk of the methods used in transient flow analysis are based on the method of
characteristics, although there are successful attempts at solving basic equations by finite
element methods and finite differences methods (Verway & Yu, 1993, Ingeduld et aI., 1994 )
as well.
After some simplifications related to the introduction of piezometric head, n, instead
of pressure, p, and, flow rate, Q, instead of velocity, V, through simple transformations, basic
equations for the method of characteristics can be written in the following form
~(n±~vJ=
dt g
+ AaVIVI
D2g
(8)
which are valid along lines
dx
-=V+a (9)
dt -
in the simulation domain (x,t).
These Equations are approximated by the finite difference method
[ n±~vI
g AlB
= + AM
D
VIVII
2g AlB
(10)
The expression in square brackets represents the difference between values of Riemann
quasi-invariants at point (P) and point (A), for positive characteristics, and point (B), for
436
negative characteristics. The right hand side is taken from the point A., for positive
characteristics, and from point B, for negative characteristics.
In Figure 9. two variants for the method of characteristics with interpolation are given.
The method with time-line interpolations has certain advantages over the other, although, it
requires more data to be stored.
If the right hand side of Equations 8. and 10. is dropped out, the well known
Joukovsky expression is obtained.
However, the most important in the simulation of transient flow is the way in which
boundary conditions, especially pumps, are presented.
P p
~+l ~~------~--------+-- tn+l~------~-----+--
A B
LlL
tn_1-t-------+------+--
Figure 9. Spatial and time-line interpolation method in the method of characteristics
Simplified analysis. As an example of the potential of simplified methods, the

trajectory of the operating point of two pumps caused by switching off the third unit, is given
in Figure 10. Before 'reaching the new working point C, the flow rates of other two units are
temporarily increased, moving down, almost vertically, to point B. Due to inertia of the water
column, pressure increases in the suction line and decreases in the discharge line. Due to
friction losses and after several wave reflections, the pumps will get to point C. The adverse
effects of this maneuver are more pronounced after switching off the second unit, which,
according to the NPSH curve, may cause transient cavitation and temporary motor
overloading.
The most interesting point is that this event belongs to so-called normal operating
conditions with very high frequency of occurrence. Detailed analysis using complete pump
characteristics can provide further information on the possible consequences.
437
Hp
~~---
O~--~~+-+--+~-+~----------'Q
NPSH
steady state dynamic
operating range
Figure 10. Steady and dynamic operating range during pump switch-<>ff
Almost generally, water hammer is more dangerous for systems with low head pumping
stations than for systems with high head pumping stations. Steady state velocities are of the
same order, as well as the changes of velocity at boundaries, which are directly related to the
pressure changes. Therefore, pressure changes of the same magnitude are not the same in
relative units scaled with pumping head or with steady state working pressure. Furthermore,
low-head high-capacity pumping systems are much more vulnerable to low pressures and
possible column separation, caused by pump power failure, than by high pressures.
The sequence of events following pump power failure is shown in Figure 11. The
system consists of 6 pumping stations connected to a common discharge pipeline with diameter
equal to 500 mm. The distance between PS6 and PS 1 is 2000 m. Pipe diameter is increased to
600 mm downstream of PS 1. The initial discharge is equal to 650 lis. The most upstream
pumping station, PS6, is switched off at discharge equal to 118 lis, causing a negative wave to
propagate downstream. Instantaneous elevations of Hydraulic Grade Lines (HGL) are given at
1 s interval, the first one related to 0.5 s after initiation of the pressure wave. According to
438
simplified analysis, one can conclude that the initial wave, equal to MI =-60m (in downstream
pipe All =-42m), will cause column separation in the whole of the downstream pipeline.
However, the envelopes of HGL in Figure 11., diamonds indicate maximum and
squares indicate minimum, show that nothing serious would happen. The initial wave is
reduced becuase of pump inertia and the effects explained in Figure 10., i.e. temporary
discharge increases in neighboring pumping stations. Therefore, in order to assess the risk
associated with flow transients, a more detailed analysis of pump behaviour in transient
regimes is required.
130
- pipe
120 O -to
- tl - te
llO o - II,.w,
a 100
c
...~
.9 90
ra 80
70
60
50
-1000 0 1000 2000 3000 4000 5000 6000
Distance [m)
Figure 11. Pressure wave propagation caused by pump power failure
1.4.2. Pump Characteristics in Four Quadrants
Simplified analysis may indicate problems in the operation of the pumps, but the
problems indicated will not be solved. More detailed data on pump behaviour in abnormal
situations are needed. Namely, after pump power failure, the pump may experience either
reversal in flow, rotational speed, or both. It is also possible that the head and torque may
reverse.
439
These data are usually arranged in dimensionless diagrams, as given in Figures 12., 13.
and 14. Basic variables are dimensionless discharge, v =Q/ Qo, and rotational speed,
a=N/No , while dependent variables are, pump head, h=Hp/Ho , and hydraulic torque,
(3 = T/ To, where subscript (0) indicates the values at the point of best efficiency. Similarly,
relative efficiency is obtained as rVTJo =vh/a(3.
\
\
\
3 \
\
\
2 \
\
h
a2 -£\
,
and 1 \
\
~
0
"-
--
Zone
H ,C
\
-1 \
Kannan-Knapp Quadrant \
\
IV
-2
-3 -2 -1 0 v 2 3 4
Ii"
Figure 12. Homologous head and torque characteristics for radial flow pump
(n s = 24.6, SI) for positive rotation ( a> 0)
These are usually called "four quadrant diagrams" because they can be represented in
the plane (v,a), with positive and negative values of v and a. Following the work of Knapp
(Knapp, 1937), which is widely accepted (Martin, 1982, StepanotT, 1957), eight possible
440
zones of operation, denoted by letters from A to H, will be analysed. Four of them belong to
the nonnal pump operation, when a>O (Figure 12.), and four to the negative rotational speed
(Figure 13.).
h
a2 Karman-Knapp Quadrant
1 - - - - II - - - 1 1 > 1 - - - - -
and 2-
O~----------+_--+_--~-----------------I
G
-1
-2~~----~--~--~--~--~------~----~
-2 -1 o 2 4
Figure 13. Homologous head and torque characteristics for radial flow pump
(ns = 24.6, SI) for positive rotation (a< 0)
Zone A (nonnal pumping) depicts a pump under nonnal operation for which all four
quantities, v,a,h, and ~ are positive. Also, 11 is positive.
441
Zone B (energy dissipation) is a condition of positive flow, positive rotation, and

positive torque, but negative head, giving 1'\<0. A machine could operate in zone B by a
sudden drop in head during a transient caused by power failure, or being overpowered by
another pump or by a reservoir.
In zone C (reverse turbine), it is possible for a pump to generate power. Flow and
rotation are positive, but head and torque are negative. The maximum efficiency in utilizing
this as a turbine would be relatively low because ofimproper entrance conditions.
Zone H, called energy dissipation zone, is often encountered after pump power failure.
In this zone, the combined inertia of the rotating parts of pump, motor and entrained water,
still maintain positive rotation, while flow has been reversed by positive head.
During power failure the pump will enter zone G (normal turbining) provided that
reverse rotation is not prevented by some mechanical devices. A pump rotating freely in a
negative direction is not generating power, but it is exactly the mode of operation of a
hydraulic turbine. Head and torque are positive as for normal pump operation while flow and
speed are negative.
Zone F (energy dissipation) can be encountered after continual operation ofa machine
earlier failed as a pump. The difference between zones FiG is that the torque has changed sign
for zone F, which causes a reduction in speed. In fact the real runaway condition is attained at
the boundary of the two zones, where torque, T = 13 = O.
The two remaining zones, D and E are very unusual in practical application (they can
be encountered in the test loop). They are possible in case of improper wiring of electric
motor.
Figure 14. integrates two diagrams from Figures 12. and 13. for the same pump.
Curves of h/a.2 for h = 1 and h = -1, and straight lines for h = 0, are represented by
continuous lines. Lines for h = 0 and 13 = 0 are used for separation of zones of operation.
Figures 12. and 13., with variables, h/a.2 and 13/a.2, as a function of vja, are very
useful for physical explanation of pump performance, but inappropriate for computation
because (vja) becomes infinite as the pump passes through, or remains at, zero speed (0.= 0).
Marchal, Flesh and Suter solved that problem by introducing a new variable as abscissa,
9= 1t= arctan{vja}, instead of, (vja), and also, hl(v 2 +0.2 ) and 13/(v 2 +0.2 ), instead of,
h/a2 and 13/0.2 . In this way complete pump characteristics are represented by two single
curves, as given for the same pump in Figure 15.
442
--h
2- ---- P
O~--~----~--------~-------+----~~r---;
-1 0/'\'1........
....
IV
-2 -1 -a. o +0. 1 2
Figure 14. Karman-Knapp circle diagram (ns =24.6, SI)
1.5
- - WH(x)
1.0 - - - WB(x)
0.5
0
/
-0.5 I
D C B A
-1.0 \ I
\ I
\ I
-1.5 .... /
-2.0
m IV I ll-
0 1t/2
Figure 15. Suter representation of head and torque characteristics for radial flow pump
(ns = 24.6, SI) in four quadrant
443
1.4.3. Pump Power Fllilure
This is one of so-called irregular, or uncontrolled, events because no adequate action

can be undertaken to prevent it. It is usually caused by pump power failure, but also by
automatic operation when necessary actions, like the closure of a discharge valve, are not
completed. All transient flow analyses include this event and with reliable data on the dynamic
characteristics of pumps and valves, and their condition, the results of simulation are
acceptable.
This problem will be illustrated by a case study, BVK Makis- M1adenovac, a booster
pumping station with a long suction line, the results of which are presented on separate pages.
In order to secure trouble-free operation of the system, several methods can be recommended
depending on the objectives of the analysis.
Characteristic flow rate is 1200 Us, and pipe diameter 1200 mm. In the beginning of the
analysis glass reinforced plastic (GRP) pipes of the nominasl pressure (NP) 16 bars were
envisaged, and the only problems expected were low pressures at high points of the pipeline.
Acceptable protection could be achieved by inexpensive surge relief valves. The results of the
analysis obtained by software package HYTRA (Ivetic, 1990) are given in Figure 16. The first
dip-gram represents the longitudinal profile of the pipeline and envelopes of extreme
piezometric heads. The numbers in smaIl boxes indicate the positions of sections where flow
rates and piezometric heads are monitored and given in the two remaining diagrams.
However, it was later decided to use reinforced concrete pipes ofNP 12 bars, which
are resistant to low pressures. Instead of a surge relief valve, an inclined stand pipe (more
expensive, and more reliable) had to be used (Figure 17.). The problem of pump power failure
may be further complicated by taking into consideration the dynamic behaviour of check
valves, which is analysed in the next subsection.
Check Valve Dynamics

This problem has been analysed on a high-head pipeline (length =14 km, diameter 330 mm,
flow rate 80 Us) for raw water (Figure 18.). The designer of the system was mainly concerned
with almost certain column separation which was about to happen shortly after pump power
failure about 9 km from the pumping station. An air chamber was proposed, and the analysis
had proved that the pipeline was protected from column separation. What is common for these
analyses is the rather unrealistic assumption that check valves close automatically as the flow
changes sign (used for results in Figure 18., as well). However, that is a characteristic of a
small number of check valves, let say, of nozzle check valves in this analysis. In this case, there
was a classical swing check valve, which is much slower in response, though admitting
444
significant backward flow. It turned out that an air chamber would worsen the situation and the
air chamber was dropped off the project. In this case column separation was not a problem.
There were no serious over-pressures following vapour cavity collapse. Pipe wall thickness
was detennined to sustain normal operating pressures.
Longitudinal profile; extreme values of II

- r-----------------------------T------,
- ~--------------------------~
~~ ~--~~~~~~~
1 u r-----~~~------------------~1r~
Jt,..
~ tIC
o
:;:; 1111
~ t~ ~--~~l-I~~~--------~~
ril no ~---!E&;¥'-I
Flow rates
1.2
-<I.
, .. , <I,
1.0
- - Q,
-<I,
1).8
..--
en
.§.
0.6
0.4
0' 0.2
0.0
.
-0.2
~
0 ~ ~ ~ ~ ~ m ~ ~ ~ ~
Time 16)
ELevations of HGL
~ ~ ~ ~ ~ ~ ~ tIC tIC ,..

Time Is)
Figure 16. Results of the simulation of pump power failure: BVK Makis-Mladenovac;
surge relief valve upstream of the booster pumping station
445

~ ,-------------------------------------.--------,
~ ~-----------------------------------;
~OOO r---------------------------------~--~--~
1n;)
~~ j---~;~~.~-•.,~~...~~~~~-j
.~ '10 1----R.-~-..,__ct-~!I_--------"'#'!~-___1
... UIO 1 - - --.!!a:YiI
! 140 1------eli:\'I':-I
riluoE~~~
lCO
6OCO '0000
Distance [m]
,- >OCOO tMICO 3COCO
Flow rates
1.2
- Q,
.... Q.
1.0 - - Q,
-Q,
0.8
00
......
"'s 0.6
c:t 0.4
0.2
0.0
0 20 40 60 80 100 120 140 160 180 200
Time[s)
Elevations of HGL
210
m
n;)
l'w
,,~
_2OIl
~
~
Ill!
,..
1110
• "" .. 10 10 ,co
Time [s]
1:11) 14. 110 ... >DO
Figure 17. Results of the simulation of pump power failure: BVK Maids -Mladenovac;
stand pipe (or exhaust pipe) upstream of the booster pumping station
446

MO r---------------------------------r-------, '_plswIItnIII
... , JIru:I
.-,.UiO ''''''''U .I II I II.IUII''''"IIU''111 1 •• "-m

1
... U ... III .' .. IIII .... ,,,, ............ III
Jt~ ~----------------------------------~~
!9
i~
4OO f----:::---"h'~'f'
Flow rates
(U
-Q,
It" Q,
0.08 -- Q,
-Q,
0.08
.
or
.......
~ 0.02
0.04
a
0.0
-0.02
-0.04
0 10 15
Time [s)
Elevations of HGL
~
-II,
•
IIUIII
"" ••,
I -- n.
-110
~,.,
••
1~ ••
••
6"480 •
..., - •, I
I
440
•
Gl
0 10 ~
Time [s)
..
Figure 18_ Results of the simulation of pump power failure: Bor;

usual assumption of instantaneous check valve closure
447
1.0
.&. /
!l-
.~ V
i / Swing
i~
.....
/ I
/
Ball
~
80.5
]
-
/
/
J
Nozzle
-
,..,- .4-
7 Split
Disk ~
~
\
~
-
~
§ fJ ' / ' V
~~
~
-
z In' ~ ~
o ~ r dt ~JO;
~ .JJ
,....- -v
o 2
Non-dimensional deceleration
Figure 19. Dynamic characteristics of check valves of different types and sizes (Thorley, 1991)
560 1\
1 , - theoretical
540 1 - \ - - - - - - - - - - - - - - - - 1 ..... bloclredc.v.
--- nozzle c.v.
520 H - - - - - - - - - - - - - - - - - I - swiDgc.v.
~r4--------------------------------~
1~~---------------------=~--~
~ 460 f--+-\------f-=--~--=--\+______!
t:: \' r-,.-=-- ---..::=::
4ID~~~~~--~~.
~r_~~(-~-~------------------~~~-~-~=--~-~j~=±==~
~ ".. . . . ............. ............--.t=: ...
>~
«Or----~~---~~-----------~
,----_.//
~L---~_ _~--~_ _~--~~~~_ _~--~--~--~
o 2 4 6 8 10 12 14 16 18 20
Time [8]
Figure 20. The influence of different presentation of a check valve behaviour

448
In Figure 20. several ways of presentation are compared concerning check valve behaviour in
the pumping station. through the changes of piezometric heads downstream of the pumping
station.
One of them, indicated as theoretical, is the same as in Figure 18.
The dotted line represents the case when the check valve remains open. with negative rotation
of pumps prevented. The dashed and thin lines represent nozzle and swing check valve,
respectively. Dynamic characteristics of check valves are taken form Figure 19. (Thorley,
1991).
1.4.4. Pump Start-Up
This is a regular event, neglected by many analysts. Usually, no problems are expected,
but, unfortunately, they do happen.
Starting Against a Closed Valve

This is a recommended operation for radial flow pumps, but not for axial flow pumps.
Starting Against a Check Valve

This is a usual mode of operation. Problems may arise due to the malfunction of the check
valve in high head systems.
Starting a Pump Running in Reverse

This is an unwanted but possible situation at unmanned pumping stations with automatic
controL Overloading and overheating of the motor are possible.
1.4.5. Available Surge Control Devices
It should be noted that there is no typical solution for surge protection. Every system is
unique and requires a unique approach in analysis. Some of the most common devices used in
WSS are shown in Figure 21.
449
Intermediate
r-reservoirAir vessel
t _
Reservoir
valve Pump and check val~e
Figure 21. Protection devices against pressure surge commonly used in water supply systems
1.4.6. Flow Control and Energy Recovery by Pumps
The problem of flow control in long gravity pipelines is closely related to energy
dissipation by control valves. It is accompanied by cavitation, vibrations, sometimes severe
transients etc.
An inexpensive solution represents a pump operating in turbine regime, more exactly in
zone F (Figure 14.). With small modifications to the motor it can work as a generator. Because
of inertia of the electrical system frequency control is usually not required.
One of the situations where the flow control in long gravity pipelines is required is
pumping over levees in flood risk zones with small or negative static heads. Relatively high
head is required during the priming of the system, while later, pumps are usually throttled to
prevent drifting of the operating point to the right end of pump curve. Compared to standard
micro-hydropower schemes this one is much cheaper (Apfelbacher and Etzold, 1989).
According to some analyses, recovered energy may exceed 50 % of the energy used for
pumping.
1.5. CONCLUSIONS
The problem of transient flow analysis related to the operation of pumping stations will
remain one of the most important in the future.
450
The emphasis will probably shift from irregular events (like pump power failure)
towards transients under normal operation conditions due to the ever-increasing complexity
and size of water supply systems. Apart from the standard schemes of pumping stations with
nearby suction reservoir there will be more booster (single-unit) stations, locally controlled.
In order to be able to predict the behaviour of real systems, it is necessary to acquire
full knowledge of equipment characteristics (control valves, pumps, check-valves etc.)
preferably under dynamic operating conditions.
New pipe materials and wider usage of some existing materials (PVC, PE, GRP etc.),
which have much lower elastic wave velocity, instead of traditional ones (AC, concrete, even
steel, etc.), will decrease the risk associated with transient phenomena.
As a solution for resolving severe operating conditions (cavitation, vibrations etc.) of
control valves in long gravity pipelines and to recover significant amount of energy dissipated
for the purposes of control, pumps operating in turbine mode can be used.
1.6. REFERENCES
Apfelbacher R, Etzold F. (1989) Energy-saving, shock-free throttling with the aid of a reverse
running centrifugal pump. KSB Technische Berichte 24e, Frankenthal.
Cherkassky V. (1980) Pumps, Fans, Compressors. Mir Publishers Moscow.
Dickenson C. (1988) Pumping Manual. The Trade and Technical Press Ltd., Morden, Surrey,
UK.
Ivetic M. (1990) HYTRA - Hydraulic Transient Analysis. IRTCUD Software.
Karassik 1.1., Krutzsch W.C., Fraser W.H., Messina J.P. (1976) Pump Handbook. McGraw-
Hill Book Company.
Knapp RT. (1937) Complete Characteristics of Centrifugal Pumps and Their Use in the
Prediction of Transient Behavior. Transactions ASME, Vol. 59, New York.
Martin e.S. (1982) Representation of Characteristics of Hydraulic Machinery. Intercambio
Internationat sobre Transientes Hidraulicos & Cavitacao, Sao Paulo.
Record R (1988) Local Intelligent Pumping - A Standard Approach, Computer Applications
in Water Supply. Volume 2, eds. B.
Stepanoff AJ. (1957) Centrifugal and Axial Flow Pumps. 2nd Edition, John Wiley & Sons.
Thorley ARD. (1991) Fluid Transients in Pipeline Systems. D. L. George Ltd.
Verway A, Yu J.H. (1993) A Space-Compact High-Order Implicit Scheme for Water Hammer
Simulations. Proc. XXVth IAHR Congress, Tokyo.
Wylie E.B, Streeter V.L. (1978) Fluid Transients. McGraw-Hili Inc.
2. Object-Oriented Tools for Transient Flow
Analysis
Petr Ingeduld
EvienZeman
CZECH REPUBLIC
Adrian Verwey
IHE, Delft
P.O. Box 3015, 2601 DA Delft
THE NETHERLANDS
2.1. INTRODUCTION
Today there is a need for an engineering software package for the simulation of steady
or unsteady-flow in pipe networks, providing all information about pressure distribution,
solving slow transients and water hammer (Abbott, 1991). An object oriented package can
provide a flexible description of a variety of pipes and hydraulic structures such as pumps,
valves, reservoirs, surge chambers, feeder tanks and air vessels. The modular structure of such
a tool may be based on graphical facilities for the definition and checking of input data,
including accurate and fast numerical solvers and an engineering approach to presenting
output data. A package structure may benefit from a database platform enabling query-by-
example and other modem data handling facilities.
The numerical modules may even cover eqUilibrium problems such as water distribution
in pipe networks and slow and fast transients including real time control simulation. The
complex and robust parts of the numerical code allow the user to describe a wide variety of
hydraulic components and their interconnections within any kind of network. The software
package may benefit from an object-oriented design of the code. The recently completed
HYPRESS package based on the above mentioned approach has been used to demonstrate
some of the characteristic features of new technologies.
NATO ASI Series, Partnership Sub-Series, 2. Environment - Vo1.I5

Edited by C. Maksimovic. F. Calomino, and J. Snoxell
452
2.2. THEORY
The solution domain of pipe network usually consists of branches connected one to
another by means of nodes. The authors take as an example the algorithm based on a finite
difference method. Grid points are generated along branches where they represent the place,
where the governing equations are computed (Figure 1.). Various hydraulic structures (such as
simple tanks, surge tanks, feeder tanks, air vessels, valves, pumps, etc.) can be included at any
selected place of the domain.
NODE BRANCH
Figure 1. Discretization of network
• Branches represent pipes of constant properties. In the pipe network, branches may
include such elements as hydraulic controls, e.g. valves, pumps,
• Nodes represent free branch ends, branch connections or specific storage.

Boundary conditions are defined at the nodes by setting up values of water pressure
or discharge as a constant value or as a function of time. At nodes connecting
several branches together, pressure compatibility is assumed. Simple tanks, surge
chambers, feeder tanks and air vessels are located at the nodes,
• Grid points are generated along the branches and they represent the computational
grid where the values of water pressure and discharge are solved, or specify points
where data input or output is required. The system requires a computational grid to
define small sub-sections in a branch over which the governing equations are
solved numerically.
453
The governing equations describing one-dimensional unsteady pressurised flow in a

pipe can be derived, for example, in the form of the continuity equation (Streeter and
Wylie, 1983; Chaudhry, 1987, Tullis, 1989):
oQ gA oH
- + - 2- = 0 (1)
OX a of
and the momentum equation:
oQ oH f
-+gA-+-QIQI=O (2)
of ox 2DA
where: Q discharge [m 3/s],

H piezometric head [m],
f Darcy-Weisbach friction factor [-],
D diameter of the pipe [m],
A area of the pipe [m2],
x, t space and time co-ordinates [m], [s].
1/2!J.x 112&
I' t t
2(1- 8)t!t
n- I L.2:::z::::::::::::.J
j j+1
1 6.x
I
Figure 2. Operator for space-compact implicit scheme
The governing Equations (1) and (2) are solved by the advanced fourth-order-accurate,
space-compact implicit finite difference scheme (Figure 2.),Verwey and Yu, 1993. The scheme
was developed by adding two Preissmann type operators (Cunge et al., 1980) over two
454
successive time levels and by subtracting all third-order derivative terms of the truncation
error. This was achieved by converting all third order derivative terms to a form a3tJaxat2 ,
which can be defined in finite difference form on a domain between two successive sections on
a branch and three time levels.
The space-compactness of the scheme assures the same formal accuracy of
computations all along the pipes, including the sections near the boundaries. The completely
centred scheme is stable for all Courant numbers less than or equal to unity and the scheme can
be stabilised for higher Courant numbers by increasing the weight parameter e.
The general implicit finite difference form of Equations (I) and (2) is:
n+1+ BI. Qn+1 + CI H n+1+ DI .Qn+1 = EI .

Al j H ,-I ,,-I " " , (3)
A2 ,,-I
.Hn+1 +B2 ,,-I " n+1+ D2 ".Qn+1 = E2 ,.
.Qn+1 +C2H (4)
where coefficients AI,BI,CI,DI,EI for the continuity equation and A2,B2,C2,D2, E2 for the
momentum equation are derived from the implicit high-order scheme.
The looped algorithm for the finite difference scheme is based on the fact that a looped
network contains elements known as nodes which represent the confluence of several flow
paths, some of which originate from other nodes, some from boundary points. A system of
simultaneous linear equations is developed where the piezometric head H at each node serve as
the only unknowns.
The solution of this system by a fast topology-based searching technique (Verwey,
1994) yields the H values at each node. Required H and Q values at all intermediate grid points
between the nodes are expressed in terms of variation in H at the nodes through a
transformation of the set of Equations (3) and (4).
For water distribution under steady state conditions and/or slow transient states, a
similar algorithm should be applied (Salgado et al., 1987). In this case, however, there is no
need for a refined grid or for a high-accuracy scheme.
These computations are based on a Preissmann-type scheme with grid points defined
only at the pipe ends, or at additional points around structures, wherever these are present
along a pipe.
455
2.3. OBJECT-ORIENTED DESIGN OF THE ALGORITHM
An object-oriented design brings the advantages to the developer to create more

efficient code that can be more easily shared by a team of programmers (Booch, 1991, Meyer,
1988). Combination of the traditional and the object-oriented way of programming can be
optionally used for the computational modules programming as discussed on the basis of a
following example (Ingeduld, 1993).
If a hydraulic structure is present in the domain of the solution, Equations (1) and (2)
applied by two succesive grid points must be replaced by other equations defining such a
hydraulic structure. Different hydraulic structures can be coupled together. The closing of one
valve can determine the operation of another valve, for example. In the case of the existence of
a link: between such hydraulic structures, communication must be maintained and controlled by
the main algorithm. This message has to be attached to the object in such a way that it
resembles the reality (Baugh and Rehak, 1992). The overall numerical algorithm, described
briefly in Section 2., can be schematised into the following steps:
Coefficients Routine
The coefficients of the implicit high-order scheme discretize the governing equations between
two successive grid points on a branch.
An object-oriented design is used for the derivation of the ABCDE representing hydraulics
components located on a branch.
Local Elimination Routine

This is used to express Q and H grid point values on each branch in terms of H at the branch
ends (nodes).
Node Routine
Applying the boundary conditions along the nodes lead to one equation for each node. This
system oflinear equations is solved by the matrix elimination method;
For hydraulics components located in nodes, again an object-oriented approach is used.
Recalculate Routine
Substitutions inside the branches yield the values of Q(i) and H(i) for all intermediate grid
points from the known values ofH at the branch ends.
An object-oriented design can be effectively used for the real-time connection of branch and
node components.
456
~? 1~1J-'~IClI~ Browsing Objects

HBRe-sPDns.Pump.
--...
...
.!.
JHBPumP-"'f.
HBSlmplePumpl -
HBCh.okV~".1
I. HBCon1roIV~v.1
Il HBDiff....nti..Lo··1
jHBval .... HHBISDIVal .... 1
/ ~ HBM ..... uringOevicel

\ HBOrHloeOuUlowl
HBRespanseVal,,_.
IHstructur.
HBlookK.,.e--ResDonsePumD I
.\ /). HF...d.r--Hoh.ok....Respom••Val ... 1
lHConneotionsl::-'1 HResponsePump-Re-sponsevaI.,·1
\lHR... pon .. ev~ye-Hch.ckl
lHResDonseVaive Gloheokl
HNAlrChamberl
HNFeederTankJ
HNSimDleTankl
HNSurgeTankl
r;
+1 I 1+
Figure 3. HStructure hierarchy
An example of HStructure hierarchy consisting of descendants of hydraulic controls is

presented in the following Figure 3. The hierarchy covers different classes that have different
function in the numerical algorithm.
HBValve: branch located hydraulic components. Replaces A,B,C,D,E coefficients of high-

order implicit scheme in Coefficient routine.
HBPump: branch located hydraulic components. Replaces A,B,C,D,E coefficients of the high-
order implicit scheme in Coefficient routine.
HNAirChamber, HNFeederTank, HNSimpleTank, HNSurgeTank: node located hydraulic

components. They modify parameters describing a type of boundary condition in a node.
HNAirChamber, for example, solves 6x6 equations each time it is invoked and these equations
457
operate on its private data. The complexity is hidden inside the object definition and methods
are encapulated with their data. A misuse is nearly impossible.
Hconnections: send messages to certain descendents of the classes mentioned above. These
messages, based on simple rules, change private data of corresponding objects.
The object-oriented features (inheritance, encapsulation, polymorphism) can be shown on an
example of a class HBValve. There are 3 main objects collections, BranchStructures,
NodeStructures, Connections, that contain descendants of corresponding classes. Using
ForEach@ iterators it is possible to evoke the virtual method such as for GetABCDE
coefficients and this message is passed along a hierarchy path as the proper descendant is
found. The same principle is applied to GetErrorMessage, GetReport functions common to all
HStructures or GetBoundary which are only valid for HNAirChamber, HNFeederTank,
HNSimpleTank and HNSurgeTank.
GetABCDE method in fact invokes GetA,GetB,GetC,GetD,GetE methods and returns a
pointer to a record containing A,B,C,D,E coefficients; this record is used in Coefficient routine
in the main numerical algorithm. Particular GetA,GetB,GetC,GetD,GetE methods are virtual
method.
Inheritance: a relationship among classes, wherein one class shares the structure or behaviour
in one (single inheritance) or more (multiple inheritance) other classes. Inheritance defines a
"kind of" hierarchy among classes in which a subclass inherits from one or more superclasses.
The following example is valid for all descandants ofHBValve.
GetE is inherited from HBValve and is unchanged, GetE=O for each.

GetB is inherited from HBValve and is unchanged, GetB=O for each.
GetD is inherited from HBValve and is unchanged, GetD=l for each.
GetA and GetC are inhented from HBValve and are overwriten.
For example, GetC=O for a closed IsolationValve.

Y
As another example GetC=Cd~ M for Orifice Outflow, (Cd is a flow factor of a valve and
MI the pressure difference).
Moreover, many report and error functions, functions allocating the valve on a pipe, functions
used for communication among objects are declared for HBValve and inherited by its
descendents.
458
Polymorphism: A concept in type theory, according to which a name (such as a variable

declaration) may denote objects of many different classes that are related by some common
superclass; thus any object denoted by this name is able to respond to some common set of
operations in different ways.
GetA method is overwritten in each descendent Report functions also have a typical
polymorphical behaviour. Different hydraulic objects, such as valves, pumps, measuring
devices create their own report during the computation.
Encapsulation: the process of hiding all of the details of an object that do not contribute to its
essential characteristics; typically, the structure of an object is hidden, as well as the
implementation of its methods.
Each function GetA for example when activated is applied to private data of a particular
object. Functions used for communication among objects send messages along the hierarchy
path. This message is recognized by corresponding descandants and sets the private data of this
object. Then, when the object is asked to GetABCDE coefficients it acts accordigly to these
changed data. Object-oriented programming helps a programmer to create a more safe
interface to the numerical algorithm, which low level operations that remain the same, are
hidden inside objects. This model brings the advantages of a modularity and hierarchy relations
into a numerical model. As a result, the model is more secure and can be easily changed in the
future.
On the other hand the main risk of the object-oriented approach in the form of start-up costs
might be easily declared now. If the particular object-based or object-oriented programming
language is used for the first time, the start-up costs associated with this object-oriented design
may prove to be a real barrier to adopting the method (Dubois-Pelerin et al., 1991a&b). The
creation of such a model, at least its object-oriented layer behind which the algorithm works,
takes time and requires extensive process of verification of the model, even if such a model
already exists expressed in a conventional programming language.
2.4. HYPRESS EXAMPLES
2.4.1. Single Pipe
Basic tests of water hammer simulation were made on a single pipe oflenght L=1524
meter, diameter D= O.76m with water hammer celerity c=1239m/s and carying a steady state
459
discharge Q= O.695m3/s. The pipe connects two adjacent reservoirs with water levels
Hl=202.80 and H2=198.20m. The valve is located at the end of the pipe and is closed linearly
in time over a period of 1.0 second.
The initial head loss in the valve is neglected. The results show water hammer during
the first 20 seconds in the form of water pressure history in front of the valve (Figure 4.).
H - value Grid p.=IO Branch Nl-':i> N2 dial 1524.00 m Pressure head time history
4OO~0
300.00
200.00
100.00
O.OO~~~ ____~-+__~~______~____________r-~________-+
0.00 5.00 10.00 15.00 20.00
Q - value Grid p.=IO Branch NI-':i> N2 dist. 1524.00 m Discharge time history
Linear closing of the valve
/
0.414
0.00 5.00 10.00 15.00 20.00
Figure 4. Pressure changes caused by linear valve closure
2.4.2. Surge Tank
This study shows the functioning of a surge tank as a water hammer protective device,
located on a pipe line between two reservoirs (Figure 5.). The network topology consists of
two pipe branches, Ll=ioo and L2=200 meters long, each with diameter D=1.25 m, the wave
celerity in these pipes is 1000 mls. Water levels in two adjacent reservoirs are Hl=960 and
H2=958.70 m. The initial discharge is Q=1.997m3/s. The surge tank is a rectangular basin with
dimensions 3x4m. The initial water level of this tank is 959.13 m. The water hammer
phenomena is caused by the instantaneous decrease of the discharge at the lower outlet.
The following figures describe the network topology, the graph of pressure changes
along the pipe branches during the first minute and the corresponding water level changes
inside the surge tank for the same time period.
460
The initial state was computed using the HYPRESS Initial Module. The subsequent
water hammer simulations were made with the Water Hammer Module with the a step Iit=O.05
seconds.
Surge tank
H
• I§I •
B
I I I I
0.00 100.00 200.00 300.00
water level inside the Surge tank

959.74
H - value Grid p.=4 Branch NI-+ N2 di'!l 200.00 m
955.00
~~~~~-+----------~--~~~--+-----~~~~~-+
0.00 30 sec 60 sec
Figure 5. Network topology and water level change in surge tank
2.4.3. In-Line Control Valve
In-line control cone valve is located in this example at the end of a pipe network. The
network consists of 10 pipe branches connecting two reservoirs (Figure 6).
The total length of the system L=6096m, the diameter D=1.0m and the water hammer
celerity C=974.90mls. The last pipe is 5457m long and the control valve is situated at its end.
The valve characteristics are described by the polynomial function, the valve is initially open at
13% giving a steady state discharge Q=O.2055 m3/s. The control valve is linearly closed over a
period of 14 seconds.
The water hammer phenomena is documented in the following figures. The layout of
the system is shown in the first picture. The next two figures describe the water pressure and
discharge time history in front of the control valve. The pressure profile in the last pipe branch
is shown on the last figure.
The initial state was computed by the HYPRESS Initial Module and the subsequent
water hammer by the Water Hammer Module.
461
in-line cone valve

I I •• ....
H H
1 I 1
0.00 10000.00 20000.00 30000.00 40000.00 50000.00 60000.00
H - value Grid p.=183 Branch NIO- NIl dist. 5457.01 m
175.00
150.00
0.00 200.00 400.00 600.00 800.00 seconds
Q -value Grid p.=183 Branch NIO- Nil dist. 5457.01 m

0.200
0.150 Discharge time history
O.IIK) the valve is closed in 14 sec.
0.050 /"
0.000 f.':::::;::::::;:::::::::;==:;::::::;::::::;:::::::::;==:;::::::;::::::;:::::::;==:;::::::;::::::;::::::::;==:;:::::;::::.
0.00 200.00 400.00 600.00 800.00 seconds
Profile H - value
Profile H on the last branch
58500.00 59000.00 59500.00 60000.00 60500.00 meters
Figure 6. Network topology and results of valve closure simulation
2.5. CONCLUSIONS
This software package has been designed on the basis of an interactive graphical
support to data editing through a multi document interface, during editing. All logical and
topological data types are processed immediately and, wherever appropriate, visualised. The
information system on pipe networks is based on storage data in standard database.
462
The numerical modules solve pressure and flow distributions in pipe networks as steady
state, or slow and fast transients including real time control emulation on a variety of hydraulic
components such as pipes, pumps, valves, reservoirs, surge chambers, feeder tanks and air
vessels.
2.6. REFERENCES
Abbott M.B. (1991) Hydroinformatics: Information Technology and the Aquatic Environment.
The Avebury technical, Alderschot, UK and Brookfield, USA
Baugh J.W.Jr, Rehak D.R. (1992) Data Abstraction in Engineering Software Development.
Journal of Computing in Civil Engineering, Vo1.6, No.3, July, 1992, 82-301ASCEE.
Booch G. (1991) Object Oriented Design with Applications. The Benjamin/Cummings
Publishing Company, Inc.
Cunge JA, Holly F.M.Jr, Verwey A (1980) Practical Aspects of Computational River
Hydraulics. Pitman Advanced Publishing Programm.
Dubois-Pelerin Y., Zimmermann T., Bomme P. (1991a) Object-Oriented Finite Element
Programming: I. Governing Principles, Computer Methods. In Apllied Mechanics and
Engineering, 98291-303.
Dubois-Pelerin Y., Zimmermann T., Bomme P. (1991b) Object-Oriented Finite Element
Programming: II. A Prototype Program in Smaltalk, Computer Methods in Apllied
Mechanics and Engineering, 98 (3) 361-397.
Chaudhry M.H. (1987) Applied Hydraulic Transients. Van Nostrad Reinhold, 1987.
Ingeduld P. (1993) Elements of Object-Oriented Simulators. M.Sc Thesis H.H 155, IHE,
Delft, The Netherlands.
Meyer B. (1988) Object-Oriented Software Construction. Prentice Hall.
Salgado R., Todini E., O'Connell P.E. (1987) Comparison of the Gradient Method With Some
Traditional Methods for the Analysis of Water Supply Distribution Networks. Leicester.
Streeter W.L., Wylie E.B. (1983) Fluid Transients. McGraw Hill, NewYork, 1983, 1978.
Tullis J.P. (1989) Hydraulics of Pipelines. John Wiley & Sons.
Verwey A, Yu J.H.. (1993) A Space-Compact High-Order Implicit Scheme for Water Hammer
Simulations. Proceedings ofXXVth IAHR Congress, Tokyo.
Verwey A (1994) Linkage of Physical and Numerical Apects of Models Applied in
Environmental Studies. Proceedings of the Conference on Hydraulics in Civil
Engineering, Brisbane.
Chapter VIII
Use of GIS and Remote Sensing

in Water Supply
Chapter VIII
Use of GIS and Remote Sensing in Water Supply
Remote sensing and Geographical Information Systems (GIS) are two complementary
technologies that can be applied to the problems of water supply.
The technologies of GIS and Remote Sensing that are relevant to water supply, water
body monitoring and water quality will be described in the first Section of this Chapter.
The second Section deals with applications and case studies using remote sensing and
the spatial modelling capabilities of GIS for applications as diverse as determining urban
irrigation demands and detecting leaks from buried pipelines.
The third Section describes a specific example of how the digital elevation model and
other geographic data can be used in design and operational management of water distribution
systems.
Finally, in the fourth Section the application of GIS and digital mapping to a major
water supply company covering a population of 1 million people will be discussed.
1. Relevant GIS and Remote Sensing Technology:
Theoretical Considerations
John Eigy
Thomas William Charnock
Peter David Hedges
Aston University
Aston Triangle, Birmingham B4 7ET
UK
1.1. INTRODUCTION
The Geographical Information System is well suited as a tool for assisting with
water supply problems. It not only carries out the inventory functions, as to what lies
where, but assists with the manipulation of the data and the extraction of information for
modelling and decision making. This section will describe the fundamental features of GIS
directly relevant to water supply projects, and the ways that data are stored in a GIS. The
Digital Elevation Model (DEM), which is of great importance to hydrology and water
supply, will be considered in some detail. In terms of cost for implementing a GIS
solution: manpower, data acquisition, software and hardware rank in about that order.
Manpower is a very difficult aspect to specifY for a project, so it will only be discussed
briefly. Manpower depends as much on existing skills, training requirements and
workloads for existing staff as on the GIS. The other three aspects are discussed with
respect to the current state of the art; however technology is likely to change in the future.
1.2. GEOGRAPIDCAL INFORMATION SYSTEMS
Geographical Information Systems (GIS) can be used to store, retrieve, manipulate

and display information having attributes that can be related to a position in a co-ordinate

Edited by C. Maksimovi~. F. Calomino. and J. Snoxell
468
system, specifically information relating to the Earth's surface. In this respect it is identical
to a Computer Aided Design (CAD) system. It is the nature of the manipulation that
distinguishes GIS from CAD. A definition of GIS is;
'an integrated system to capture, store, manage, analyse and display information
relative to concerns ofa geographic nature' Carter (1989).
With such a definition it can be argued that many organisations that systematically
deal with maps and mapping can be said to have a GIS, whether the process involves the
use of computers or not. It is the availability of inexpensive, powerful computers that have
made the full power of a GIS readily available. It is now financially possible to carry out
detailed studies that were impossible a few years ago.
GIS makes it possible for the engineer or resource planner to produce rapidly a
combination of maps and tables that show together "where" (location of resources),
"what" (kind of resources), and "how much" (quantityllimitation of resources). Resources
may be taken here to include: reservoirs, water supply network, land cover, percentage
impermeable area, contributing areas, etc.
The GIS is an institutional entity, reflecting an organisational structure that
integrates technology with a database, expertise and continuing financial support over
time. Thinking of the GIS in this way takes it beyond being just a package of technology -
hardware and software capable of performing GIS types of function.
There are few books and periodicals on GIS available at the moment and the reader
is recommended to consult the excellent book by Burrough (1986) on the principles of
GIS in land resource, which covers the theory of GIS very well. This chapter will only
consider theory where directly relevant to water supply applications.
Computing technology, hardware and software is changing rapidly (this is
particularly true for GIS software), new products are announced almost every year, thus
making it very difficult to make any recommendations as to software purchase, or indeed
what is possible. The reader must keep abreast of current developments by studying the
specialist press, for example the International Journal of Geographic Information Systems
or Mapping Awareness.
The function of any information system, including a GIS, is to improve a users'
ability to make decisions in research, planning and management. An information system
involves a chain of steps from the observation and collection of data, through analysis, to
the use in some decision making process. The distinction between data and information is
important; that which reduces uncertainty, and enables decision making, is information;
data is merely its transmission medium.
469
1.2.1. GIS Concepts
There are two distinct classes of GIS application and two differing technologies for
those two classes. The different technologies are explained in the next section on data
structures. The two main areas of application are as follows. Facilities Management
systems, which are aimed primarily at managing infrastructure such as roads, sewers,
customer information and other objects in the built environment. Another term for this
type of GIS is Automated MappinglFacilities Management (AMJFM). The vector based
systems as opposed to the pure raster based systems are generally preferred for this kind of
GIS. Land Information Systems (LIS), which are aimed primarily at natural resources
applications frequently use image based data and are good at manipulating thematic map
data. The raster based GIS are generally preferred for these systems, (the terms raster and
vector are discussed in the next section and illustrated in Figure 2).
The two different branches of GIS are not mutually exclusive; one GIS package
may be suitable for both types of application. There are also other branches of GIS
concerned with weather forecasting, transportation routing, navigation and modelling
systems.
Hydrography
Elevation
Roads
Landscape
Figure 1. The overlay concept, the landscape is simplified into layers, each one representing an
attribute of interest
470
Though digital mapping is an important concept in GIS, the principle that makes
GIS useful and exciting is that different themes can be extracted and stored as separate
map overlays, see Figure 1. The themes can be combined logically and mathematically to
form new themes of interest by:
(i) an AND combination of two other themes, for example, a well drained soil and
small surface slope,
(ii) an OR combination of two other themes, for example, a well drained soil or a
sandy soil,
(iii) generation of corridor information, for example, areas within 10m of a domestic
dwelling may be liable to urban irrigation,
(iv) use of mathematical models linking existing themes, such as erosion potential
modelling,
(v) calculation of areas, lengths, etc.
1.2.2. Data Structures in GIS
There are five kinds of data to be represented in a GIS.
(i) Point features, e.g. location of soil samples, boreholes, manholes, rain gauges,
burst water mains, pumping stations, trees, buildings.
(ii) Linear features, e.g. roads (on small scale maps), rivers, pipe lines, power lines,
elevation contours.
(iii) Areas (polygons) with common properties, e.g. pressure zones, catchments,
contributing areas, soil association mapping units, climate zones, administrative
district areas, buildings and other land cover.
(iv) Actual or p.otential surfaces, e.g. ground elevation, variation of mean annual
temperature, spatial distributions of rainfall, population densities.
(v) Temporal elements, e.g. changes in land use over time, changes to a pipe network.
There are two common generic data structures that are used by GIS to store these
data; the vector and the raster data structures, see Figure 2. Any given GIS may use one or
both of these data structures. Generally the vector structure is used when most of the data
can be structured as points, lines or polygons, and the raster structure is used if most of the
data represents areas of land, or if the data origin was an image. A third structure, the
tessellated structure, is used for representing surfaces in a vector system; in particular the
Triangulated Irregular Network (TIN) structure is favoured for elevation or surface data.
471
a b
2 r -_ _....!b'--_---.:;3~
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
a c 0 0 0 I I I I I I I I I 0 0 0 0 0 0 0
6 0 0 I I I I I I I I I 2 2 2 2 2 0 0 0
0 I I I I I I I I I 2 2 2 2 2 2 2 2 0
0 0 0 I I I I I I 2 2 2 2 2 f<r'o 0 0
0 0 0 0 0 0 0 I I 2 2 0 0 0 0 0 0
5 0 0 0 0 0 0 0 0 0 0 orr- '0-0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Link Right Pol. LeftPoI. Nodel Node 2
a 1 0 I 2 o No data
b 1 0 2 3 I Coniferous Forest
c 1 2 3 4 2 Mixed
d 1 2 4 5
e 0 1 I 5
f 2 0 3 6
g 2 0 6 5
node x y
I 23 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
2 25 7 0 0 0 I I I I I I I I I fo.- 0 0 0 0 0 0
3 34 7 0 0 I I I I I I I I I I I I I 11"1- Jl.o 0 0
4 33 6 o 19' I I I I I I I I I I I I I I I I I 0
0 0 I I I I I I I I I I I 0 0 0
5 34 4 0 0 0 0 0 0 o .... :"", I I I Ill-: :<,,0 0 0 0 0 0
6 41 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Polygon Owner Landuse

1 Forestry Coniferous
o Nodata
1 Forestry Commission
Commission Forest
2 Forestry Mixed
Commission
Figure 2. (a) Simple vector representation, using the topologic model presented by dangermond
(1982), more complex structures are used to improve access times,
(b) Raster representation, a raster layer is required for each attribute to be represented
1.3. DIGITAL ELEVATION MODEL
The common type of map used in GIS for inventory storage and land evaluation is
the choropleth map; within each region of the map the value of the attribute represented is
considered constant. However, in many situations we wish to use other data that varies
continuously over the land's surface. This representation of a surface is often called 21/2
dimension representation, or, since it is often used to represent elevation data, a Digital
Elevation Model (DEM). Sometimes the term 'digital terrain model' (DTM) is used to
describe this surface. Since the term 'terrain' is also used to describe the land cover as well
472
as the elevations, we will consistently use the term digital elevation model to describe the
surface alone and reserve 'terrain' for the combination of elevation and land cover.
DEM can be used to represent other surfaces than topography; peizometric
surfaces, cost surfaces (the expense incurred when moving across the landscape) or the
surfaces of geological units are just three examples.
In water supply and hydrology DEM's can be used; to calculate slopes and aspect, which
can in tum be used to calculate contributing areas; to calculate pressure zones; to
determine drainage networks and stream lines; to calculate erosion potential and to aid the
design of networks.
There are many ways of representing surface data, but only the two described
below are in common usage in GIS now. These are both local patch methods: one based
on a regular grid of height values and the other based on an irregular triangular
tessellation. Both methods have their own advantages and disadvantages.
1.3.1. Grid DEM
Perhaps the simplest way to represent the elevation is by way of a regular

rectangular patch called an altitude matrix. In this method the area under study, the
coverage, is divided into a number of rectangular cells like a raster, and each cell contains
an elevation. Some of the features inherent in altitude matrices can be seen in Table 1.
Table 1. A DEM altitude matrix
7.8 7.8 8.0 8.2 8.3 9.4 10.5 12.0 9.1 6.0
7.8 7.8 '8.1 8.3 9.3 9.5 11.1 13.2 12.8 9.1
7.9 7.8 8.0 8.2 8.3 9.8 12.2 14.5 13.0 12.1
7.8 7.8 8.1 8.5 10.4 13.1 16.2 13.3 12.7 8.1
7.8 7.8 8.0 8.2 10.8 14.1 17.5 15.2 12.1 6.0
7.8 7.8 8.1 8.1 10.0 12.2 22.4 14.8 9.2 5.8
7.9 7.8 8.2 8.2 9.7 11.2 18.6 12.6 15.4 7.1
7.8 7.8 8.1 8.8 9.1 14.3 14.9 16.0 17.4 8.8
7.7 7.8 8.1 8.3 9.2 14.2 15.6 17.2 17.2 7.9
7.9 7.8 8.0 8.7 8.8 13.9 16.2 18.1 19.4 5.1
473
The grid size is largely fixed by the most complex varying part of the model. In
Table 1. the grid spacing has been chosen in order to represent the rapidly varying right
hand side of the coverage, this means the left hand side has adjacent cells that only vary by
a decimal point or so. To represent the subtlety of the surface required for urban
catchments the grid size needs to be small and real numbers must be used to differentiate
height from cell to cell. Grid OEM data is difficult to compress using techniques such as
run length encoding, without losing information. To store the data required for a typical
water supply catchment of 10km by 20km at a grid spacing of 20m will require
(50Oxl000x4bytes) 2 Mbytes of storage.
Just what height should be allocated to a cell? Is it the mean altitude of the cell, the
height of the centre point of the cell or the highest or lowest point within the cell? The
height selected has important consequences for determining drainage basin boundaries in
almost flat terrain, especially with large (50m plus) cell sizes.
Oata derived from different sources will be of different qualities. The elevation data
can be readily produced using modem photogrammetric techniques. However, if the data
is extracted from existing maps, either as contour lines or as point elevations then spatial
interpolation techniques need to be used. False pits and hollows often arise from this
interpolation process.
Ideally, for hydrological studies, the height of each cell should be the lowest point
in the cell, but since much of the data is currently derived from interpolation of contour
data (isolines) the error of this interpolation exceeds the within cell error. It should also be
remembered that in many situations the contour data itself is an interpolation of spot
levels recorded by a surveyor.
Two of the most important products that can be derived from a OEM are slope and
aspect, Equations (1) and (2). Like the OEM these can be manipulated just as any other
layer in the raster GIS.
slope= (-) 8z 2/ az.

ax
(-)2
fJy
(1)
aspect =: / :
(2)
1.3.2. Triangular Irregular Network (TIN)
A common alternative to the altitude matrix is the triangJular irregular network

(TIN). Elevation is represented by a series of irregular triangular patches, Figure 3a. A set
474
of points can be triangulated in any number of ways, but most TINs are based on the
Delaunay Tessellation of irregularly spaced nodes (see Sibson, 1978, for a mathematical
treatment of the properties of the Delaunay Tessellation). Values at points interior to a
triangle are found by interpolation of the vertices, and the Delaunay Tessellation ensures
that these values use the most local vertices possible.
The Delaunay Tessellation requires that a circle passing through the vertices of any
triangle must not contain any other vertices, Figures 3b. and 3c. It can be seen from this
definition that the Delaunay tessellation is the graphical dual of the Thiessen Polygon with
the intersections of the bisecting lines between points being the centres of the circles,
Figure 3d.
Each vertice represents a spot height
b e d
~~
Figure 3. (a) A simple TIN. (b) A valid Oelaunay tessellation, each circle only contains three
vertices. (c) An invalid Delaunay tessellation, the dotted line shows the correct connection,
(d) Thiessen polygons the dual of the Oelaunay tessellation
In areas of gently varying topography the data points can be widely spaced,
whereas in rapidly varying, or areas of great interest, the nodes can more densely placed.
This means TIN's do not suffer from the data redundancy of the grid DEM. They can be
stored as vectors, but, because they are much more constrained than the general vector
representation, simpler structures can be used. A TIN list records the nodes in each
triangle; then a further list records the x, y, z co-ordinates of each node.
475
In some ways the TIN makes computation easier. Using linear interpolation;
contours consisting of straight lines can be readily drawn. Also as each triangle has a single
slope and aspect these can be readily calculated. The false pits and hollows that often arise
in interpolating the altitude matrix do not occur. However, like other vector structures,
there is considerable computational effort involved in linking the TIN and its derivatives
with other land cover information.
1.3.3. Applications of the DEM
Many useful themes can be derived from the DEM; drainage basin boundaries;
potential stream lines within the catchment; input parameters for hydrologic models;
erosion potential and possible sediment yields. Though these can all be calculated from the
TIN, only the methods used for the altitude matrices will be described here.
To determine the catchment area draining to a particular point consider Figure 4a.
which shows the direction of flow within each cell (the aspect). It can be readily seen that
water from the north and north-west cells will flow into the target cell and that water from
all the southern cells will flow away from it. But what about the north-east cell and the cell
to the West? The flow directions are determined by whether or not the slope direction falls
within the contributing angle e (Figure 4b.).
"~ I J r /e'\.. . ~
~ target cell
\ ~ }>
I \ \ L "Y
~
Figure 4. (a) Flows into the target cell (b) Contributing angles
Once a cell has been allocated to the catchment it becomes a target cell. This
process continues until no more cells can be added to the catchment. The contributing
angle is typically set at 90 degrees but may be increased to make more cells belong to the
catchment. If a cell has zero slope then it may be considered to flow in all directions, or
476
contribute to all cells surrounding it. In practice it is found that allowing flow in any
direction from any cell with a slope of less than 2 degrees alleviates some problems with
pits and hollows, see Elgy et al (1993).
Determination of potential stream networks can, conceptually, be carried out
simply by calculating the area of the drainage basin upstream from every cell and excluding
all areas less than a certain threshold from this accumulation map. The cells that define the
resulting stream lines show the area draining into them and hence indicate the relative size
of the resulting rivers. Problems arise from this simplistic approach:
• computational effort is extreme (unless more sophisticated algorithms are used),
• in valley bottoms that are wide, the slope perpendicular to the stream is zero and
parallel rivers are often calculated, which must be removed by suitable heuristics.
By inverting the rules for flow into a target cell to flows from a target cell,
irrigation command areas can be evaluated.
If the universal soil loss equation is employed, by considering slope landcover and
rainfall, the potential for erosion of each cell can be calculated. The sum of all these
potentials for a particular catchment gives the expected sediment yield for a river (see
Bradbury et al, 1993).
1.3.4. Hardware and Software of GIS Systems
Figure 5. shows the typical layout of a single user GIS system. In most installations
there will be more than one user and, in such cases, duplication of parts of the system will
occur, for example a number of workstations may be networked and the console and CPU
elements duplicated.
If the migration of source data is tracked through the GIS to the usable outputs,
the hardware requirements can be considered in tum.
Data arrives at the GIS in one of five forms: paper maps, digital maps, remotely
sensed images (either satellite or aerial photographs, digital or analogue), tabular data, and
from existing data bases.
Digital maps are becoming more readily available in the industrialised world but are
still not very common. The data when available is often very expensive. Transferring such
data into the GIS requires a suitable device to read the supplied medium, and software to
structure it into the database format. Paper maps are by far the most common source of
477
data and need to be digitised to be incorporated into the GIS. There is a proliferation of
defined and undefined standards for the transfer of digital data.
Tape or disk
Digitiser Data Logger
reader
Plotter Magnetic
Printer Media
Figure 5. Hardware components of a typical GIS
To input remote sensed data requires either a magnetic tape drive or some other
medium, such as an optical disk. Imagery in the form of photographs, or negatives should
be scanned, or video digitised. Airborne videography is becoming more common for
gathering certain types of imagery, and is especially useful for urban studies.
It is vital to record the source of the data as it is incorporated into the GIS, as well
as other information such as revision history, accuracy of source data, map projections,
scale of original maps, date of original survey and of digitisation, etc. Data, once
incorporated within a GIS, tends to stay there for very long periods and is often passed
from user to user with little quality control along the way.
Geographic Information Systems themselves run on almost all types of computer
and the selection of hardware rests more with the software selected than a simple choice of
the best computer hardware. Some GIS software requires the use of proprietary hardware;
these should be avoided if possible. Generally the vector based GIS systems run on mini-
computers or workstations, partly because of their memory and processor power
requirements and partly because they tend to be the older GIS systems developed before
micro-computers had sufficient power or networking facilities to share data. By running on
multi-user/multi-tasking systems one user can be inputting data whilst another user
manipulates other data in the system, and a third undertakes plotting.
If a micro-computer system based GIS is chosen then separate processor and
memory boards may have to be inserted to perform image processing. The special chips on
478
these boards allow for large amounts of infonnation to be displayed upon the screen and to
be processed very quickly. A mouse or tracker ball pointing device is an essential data
input facility for the operator. The software supplier will give a specification as to the
requirements of the host computer.
If, as recommended, a team of people is working on the GIS project it makes
considerable sense to link the machines together with a Local Area Network (LAN) so
that expensive peripherals and data can be shared.
Thought should be given to how the large amounts of data, which GIS generate,
are to be stored. As an example, for a raster based GIS without data compression and
assuming a 20m grid size, ~ themes, (land cover 1 byte, pipe network 1 byte, elevations,
slopes and aspect 12 bytes in total) for an area of 20km by 20km will require
approximately 14 Mbytes of hard disk storage. However, experience has shown that for
urban drainage studies about 3 times this amount is required as a working store. This
pushes the limits of micro-computer based GIS, even though with data compression
techniques reductions to 10010 or 20% can be achieved.
The hard disk storage of the computer provides the working store. Archival
storage should be provided from either: magnetic tape storage, streaming tape, WORM or
similar optical disk drive system
Output from the GIS takes a number of forms, not all of which are essential for the
system, but there must be a way to visualise at the computer monitor the results of actions
being performed and some way to produce hard copy. A colour display monitor for real
time manipulation and analysis of the data is essential and this can be part of the control
system of the computer or a separate monitor. Colour printers and plotters are useful for
producing hard copy output, though a standard 35mm camera pointed at the monitor
screen can in certain circumstances produce excellent (though not dimensionally correct)
illustrations for presentations, reports, etc. There is still, however, considerable difficulty in
printing colour reports and many publishers cannot handle raster images.
GIS soft~are comes in a variety of packages. The two main types, as already
described, are the vector based system and the raster based system. More modern systems
permit the total integration of raster and vector data, allowing the advantages of both
methods to be enjoyed, with few of the disadvantages.
Vector systems are often supported by traditional DataBase Management Systems
(DBMS). The most common conform to the relational model, see Avison (1992). Arc-
Info, the most widely used vector GIS package, follows this approach, Info being a
relational DBMS in its own right The relational model is the basis of most DBMS used in
organisations and businesses. This underlies the vector model's principle use as an asset or
resource inventory system. A DBMS should allow access to appropriate parts of the
database to different types of user, and prevent unauthorised viewing or changing. It
479
should also maintain data concurrency, provide archive facilities and present a simple
interface to the user for manipulating the data
Raster systems genera\ly do not employ such strict data management. They have
developed from image processing systems and are often used by a single user. Clearly
these are generalisations, and many packages will embody aspects of both systems.
The most up-to-date systems are described as 'object oriented'. The distinction of
object oriented systems is that all data items are described as being of one or more object
type; e.g. a linear feature, a point, a vector polygon, a regular raster, a raster cell, a TIN, a
DEM, etc. In addition to storing the description of the object, the methods of displaying,
plotting and general manipulation are also carried with the object type, this is known as
encapsulation.
Objects are hierarchical; rivers, roads and pipes will be objects that are descended
from the linear object, each will, therefore, have the properties, behaviour and methods
inherited from the linear feature, such as length. However they will each have behaviour
and properties that are distinct; roads will have classes (i.e. 'A' roads and motorways);
pipes and roads will not be able to connect to form a network.
The object oriented paradigm is currently of great interest to the computer science
community. Object oriented programming languages, databases and, of course, GIS are
under development, (Worboys et ai, 1990).
There are several advantages that are stressed by advocates of the object oriented
approach:
(i) it is intuitive as people naturally think in terms of objects,

(ii) by specifYing behaviour, inconsistencies in the database can be reduced, for
example sewers and water mains objects exhibit different behaviour and should not
be part of the same network,
(iii) developing applications is easy; by having a hierarchical structure new objects are
easily created ..
1.4. REMOTE SENSING
Remote sensing in the broadest sense is defined as:
, the science and art of obtaining information about an object, area or phenomenon
through the analysis of data acquired by a device that is not in contact with the
object, area or phenomena under investigation'. (Lillesand and Kiefer, 1987)
480
This section will concentrate on remote sensing systems that use electromagnetic
radiation for studying the Earth's surface.
1.4.1. Background
Remote sensing systems using electromagnetic radiation, have four basic

components (Figure 6.): an electromagnetic source, an interaction with the earth's surface,
an interaction with the atmosphere, and a sensor. The source of the electromagnetic
radiation may have one of three origins; radiation generated by the sun, thermal radiation
emitted from the earth and man-made radiation such as RADAR.
e
Source Sensor Sensor Sensor
COT 08 T ""
Source
IBack- I
, i scattered i
\ ! Radiatiori ......
~do
\ 1 .
Reflected Radi~lion ,. Atmosphere
1Emitted Incident Radiation
!Radiation
Earth's Surface
Source
Figure 6. Relationships between atmosphere, earth, sources and sensors
1.4.2. Electro-Magnetic Radiation
Visible light is only one form of electromagnetic energy; the full spectrum includes
radio waves, heat, ultraviolet rays and X-rays. Properties of electromagnetic waves are
linked by the equation:
c= A.f (3)
where c is the speed oflight, A. is the wavelength and/is the frequency.

481
In remote sensing it is customary to describe electromagnetic waves by their

wavelength. emote sensing is primarily concerned with the following bands:
the visible spectrum 0.4 to 0.7....m

with subdivisions into
blue 0.4 - 0.5 ....m
green 0.5 - 0.61J.m
red 0.6 - 0.71J.m
Adjoining the visible red is infrared (IR), which is divided into three categories:
near - IR 0.7 - 1.3 ....m
mid - IR 1.3 - 3 IJ.m
thermal - IR beyond 3 mm.
At much longer wavelengths are microwaves, lmm to 1m.
The amount of energy an black body can radiate is a function of the object's surface
temperature, and is expressed by the Stefan-Boltzman Law:
M=(j~ (4)
where M is the total radiant excitance from the surface of the material (in Watts/m2), T is
the absolute temperature of the emitting material in OJ<. and (j is the Stefan-Boltzman
constant (5.669 x 10-8 W m-2 OJ<.- 4). The sun emits radiation similar to a black body of
6000 oK, by comparison the earth emits radiation similar to a black body at 3000 K. Hence
it is harder to detect radiation emitted by the earth, and the thermal sensors which detect
this energy tend to have low spatial resolution.
1.4.3. Atmospheric Interactions
All radiation detected by remote sensors passes for some distance or "path length"
through the atmosphere. The path length will vary widely depending upon the sensor and
platform combination used.
As the radiation passes through the atmosphere specific bands of wavelength are
absorbed by particular gases. As a consequence there are certain wavelengths whose
intensity is significantly reduced or even missing when the solar radiation reaches the
earth's surface having passed through the atmosphere.
482
This has important implications for remote sensing; only those regions of the
spectrum where the radiation is not blocked out by absorption can be used. These regions
of the spectrum are called atmospheric windows, and sensors are designed specifically to
'look through' these windows.
1.4.4. Interactions at the Earth's Surface
When electromagnetic radiation interacts with the Earth's surface it can be

reflected, absorbed or transmitted through the surface it strikes. Different wavelengths will
interact with the same surface in different ways and thus the proportion of each
wavelength that is reflected back will vary according to the surface. These changes yield
specific spectral patterns, which are related to the surface with which the radiation has
interacted.
1.4.5. Spectral Reflectance
The reflectance characteristics of earth surface features can be quantified in terms

of the relationship between the incident radiation and the reflected radiation.
E R(A) energy of wavelength A. reflected from object

Spectral Reflectance = -E- x 100 = f I gth '··d b· x 100
I(A) energy 0 wave en II. mCI ent on 0 ~ect (5)
120 Blue Green Red

Iii
NlR Mid-IR
- - healthy
vegetation
100
---- roads
80
DN --water
60
40
20
0
0 0.5 1.5 2 2.5
Wavelength J.m
Figure 7. typical reflectance curves for three different surfaces, taken from an air-borne scanner
thematic scanner (ATM) image of an urban area
483
The spectral reflectance is expressed as a percentage and is determined for specific

wavelengths. Experience has shown that many of the earth's surface features can be
identified and studied on the basis of their spectral reflectance in different bands. Figure 7.
shows some typical reflectance curves for three features in different bands.
1.4.6. Thermal Infrared Radiation
All bodies above OaK (absolute zero) emit radiation (Equation 4). Devices sensitive
to the appropriate part of the electromagnetic spectrum (3 to 50 11m) can record some of
the energy emitted from the Earth's surface - the thermal infrared.
There are four factors that influence the amount of thermal infrared energy emitted
and hence received by a sensor:
Emissivity
All features differ in the amount of incident radiation that they re-emit. Thus objects may
be distinguished by their emissivity.
Kinetic Temperature (T)

The temperature as recorded by a thermometer or a direct thermal sensing device.
Thermal Properties of the body

The thermal capacity, conductivity, diffusivity and inertia together indicate the ability of a
body to retain, conduct or diffuse heat.
Rate of Heating
This may depend on aspect, as the rate of heating is a function of both the rate of
absorption of solar en~rgy, and the amount falling upon a body.
1.4.7. Spectral Signatures
The characteristic spectral response patterns of various bodies and features to

incident or emitted radiation are often referred to as Spectral Signatures.
Although it is true that many features exhibit very distinctive spectral signatures,
they must be regarded with caution and only used as guides in image interpretation, since
both temporal and spatial effects will influence the spectral response of individual features.
484
The spectral reflectance for healthy green vegetation usually shows the peak and
valley characteristics of Figure 7. These are caused by the different absorption and
reflectance characteristics of the different components of leaves.
In the visible spectrum the valleys are dictated by the pigments in the leaves
absorbing energy. Chlorophyll strongly absorbs energy around 0.45Jlm and 0.67Jlm - for
this reason vegetation is seen as green.
At about 0.7Jlm the reflectance of the near-infrared (NIR) increases dramatically,
with the leaf typically reflecting 40 to 50010 of the incident energy between 0.7Jlm and
1.3Jlm - most of the remaining energy is transmitted with only about 5% being absorbed.
This high reflectance results primarily from the internal structure of the plant leaves. Since
structures vary between plants, the reflectance varies; this enables differences between
plants to be distinguished.
It is often possible to tell if vegetation is stressed. Stress may be caused by water
shortage, water-borne pollution, air-borne pollution, restricted root growth, root damage,
or seasonal effects. This may result in the production of chlorophyll being decreased or
stopping entirely, or damage to the internal structure of the leaf, either shifting the
reflectance towards the red, or causing a change in near infrared reflectance. This is
illustrated in Figure 8.
~~HealthY
Reflectance ~---- Stressed
Red ------:»~ NIR

Wavelength
Figure 8 The difference of reflectance between red and near infrared (NIR)
of healthy and stressed vegetation
The difference between healthy and stressed vegetation, and between vegetated
and non-vegetative surfaces can be accentuated by using a ratio of the red and NIR bands,
to produce a new image, the normalised difference vegetation index (NDVI):
NDVI = R-IR (6)

R+IR
The NDVI is generally scaled to fit a suitable range of values, i.e. 0 to 255.
485
l.4.S. Photography
The technology of photography is well known. It should be noted that, as well as

the usual monochrome and colour film, the near infra-red part of the spectrum can be
captured to give a false colour image.
Video cameras can be used in much the same way as traditional colour films. The
images are, however, easier to capture in digital fonn for image processing.
1.4.9. Linescanners
A linescanner measures the radiance of the earth's surface along a line

perpendicular to the path of the platfonn - an aircraft or satellite. As the platfonn moves
forward successive lines are scanned enabling a picture to be built up. There are
essentially two types of linescanner; the optical-mechanical or 'whiskbroom' scanner,
Figure 9b. , and the 'pushbroom scanner', Figure 9a.
a b
Array of
111 1111 1111 1111 1111 ~ Detectors
Figure 9. (a) Pushbroom scanner (b) Whiskbroom scanner
Linescanners were developed to escape from the spectral limitations of

photographic and vidicon methods, essentially 0.3 to 0.9 /lm, since they also allow data to
be collected for the mid and thennal infrared parts of the spectrum.
The multispectral Scanner (MSS) is an example of a whisk broom scanner (Figure
9b.). It contains mechanical components that are subject to failure. Whiskbroom scanners
have other limitations too (eg poor signal to noise ratio) and they cannot have both high
486
spatial and high spectral resolution since the time spent receiving radiation from any
particular point during a sweep of the mirror is so short.
The 'push broom' Iinescanner gets round these problems. It uses tiny radiation
sensitive cells - called charge-coupled devices (CCDs). Currently these can operate up to
2.4~m, so the full visible and Near IR range can be monitored - however Thermal IR is
not yet accessible. The CCDs are arranged in thousands along a linear array. The radiation
is directed on to the CCDs by an optical system, so no moving parts are required to build
up an image (Figure 9a.). Each CCD of an array measures directly the radiation from an
individual segment of the ground swath as the platform moves over it.
1.4.10. Microwave Imaging Systems
The wavelengths used in microwave imaging systems are in the range Imm to 1m;
some thousand times longer then thermal infrared. There are two distinctive features that
characterise microwave energy from a remote sensing standpoint;
• microwaves can penetrate the atmosphere under almost all meteorological conditions;
haze, light rain and snow, clouds and smoke,
• microwave reflections from the earth's surface have different characteristics from their
counterparts in the visible or thermal portion of the electromagnetic spectrum; as a
result of the longer wavelengths involved they see roughness in a different way.
There are two forms of microwave imagery systems; passive microwave systems
and active microwave systems or radar.
Passive microwave sensors employ an antenna to receive the natural radiation
emitted by all earth's bodies. This is much weaker than within the infrared region but is
still a function of temperature.
A passive microwave system comprises an antenna to pick up the radiation, an
amplifier and a detector. The ground resolution (the detail that can be seen) depends on
the antenna size - the larger the antenna the better the resolution. This is, of course, a
problem when trying to carry the instruments aloft in an aircraft.
Radar is an acronym for Radio Detection And Ranging and was developed during
World War II for military purposes. Just as with passive microwave sensing, the spatial
resolution of a radar system is determined by the size of the antenna. To at least partially
get over the problem of having large antenna, most airborne radar remote sensing systems
487
use an antenna fixed below the aircraft and pointed to the side - it is therefore known as
side looking airborne radar or SLAR.
a .....
. -...•....
.......
...Radar Pulses
............ .....,
....
r==J0bject
7
Return reflection
Figure 10. The principles ofSLAR
The system operates by emitting discrete pulses that are microseconds long, which
travel to an object and are reflected back. Figure 10. illustrates the principles involved.
To maximise the resolution of an image, we need to minimise the radar beam width
so that only a narrow strip is illuminated. Beam width (13) is directly proportional to the
wavelength (A.) and inversely proportional to the antenna length (L).
A.
13 cx:;- (7)
L
It can be seen· that to minimise the beam width the antenna length must be
maximised. Systems that achieve this with real antenna are referred to as real-aperture
radar. However, the same effect can be achieved by electronically simulating long antenna.
Such systems are called synthetic-aperture radar (SAR) systems and are employed in high
altitude aeroplanes or in satellites.
1.4.11. Platforms
Platforms vary from ground level hand held instruments to geostationary satellites.
As the sensor moves further from the object increased spatial coverage is achieved at the
expense of spatial resolution.
488
Ground level observations enable the most detailed infonnation about an object or
area to be obtained; ground truth data for calibrating, classifYing and confirming airborne
or satellite imagery observations; detailed visual record at a specific time and place (i.e. a
photograph), or local spectral data using a radiometer.
By far the most common platform used in aerial photography is the Aeroplane.
Generally there is a deliberate overlap in images of about 60% along the flight path. This
allows two adjacent photos to be viewed through a pair of binoculars (stereoscope) to
obtain a 3D view which facilitates intepretation and allows ground height information to be
extracted.
To many people the term remote sensing is synonymous with satellite remote
sensing. Satellites are classified according the altitude of their orbits. If a satellite is
positioned at an altitude of 35,900 Ian above the earth's surface it will have an orbital
speed that matches the earth's rotation - termed geosynchronous. A satellite placed in such
an orbit will appear fixed above the same point on the earth - i.e. geostationary .
Geostationary satellites include METEOSAT (Europe), GOES (USA) and GMS (Japan).
These weather satellites provide Iinescanned images that have considerable application in
the provision of real time meteorological information and in estimating rainfall volumes
over large areas and time spans - catchments of hundreds of Ian2, and time periods of
months rather than days (see Siyyid, 1993).
A polar orbit takes a satellite over or near the North and South poles at an altitude
above the Earth between 500 and 1000 Ian. The satellite's orbit moves progressively
around the globe and, over a period of time, coverage of most of the earth's surface is
obtained. The repeat cycle is the time taken for the satellite to return to the same position
over the Earth's surface - expressed in days. Polar orbit satellites have spatial resolutions
within the range 10m2 to 100m2 . Polar orbit satellites include LANDSAT 1 - 5 and the
SPOT series, that supply linescanned data. Data from high resolution military satellite with
resolutions down to 2-5m is slowly coming onto the market, though data about sources,
etc. is difficult to come by.
1.5. REFERENCES
Avison D.E. (1992) Information Systems Development: A Database Approach. 2nd

Edition, Blackwell Scientific Publications.
Bradbury PA, Lea N.J., Bolton P. (1993) Estimating Catchment Yield: Development of
the GIS-based Calsite Model. Report 0D125, April 1993, HR Wallingford.
Burrough PA (1986) Principle of Geographical Infonnation Systems for Land Resources
Assessment. Clarendon Press Oxford.
489
Carter (1989) On Defining the Geographic Information Systems. Fundamentals of

Geographic Information Systems: A compendium, edited by Ripple WJ, pp3-6.
Dangermond J. (1982) A Classification of Software Components Used in Geographic
Information Systems. Proc. US - Australia Workshop on the Design and
Implementation of Computer Based Geographic Information Systems, Honolulu
Hawaii, pp70-91.
Elgy J., Maksimovic C., Prodanovic D. (1993) Using Geographical Information Systems
for Urban Hydrology. International Conference on Application of Geographical
Information Systems in Hydrology and Water Resources, Vienna, Austria.
Lillesand T.M., Kiefer R.W. (1987) Remote Sensing and Image Interpretation. 2nd
Edition, Wiley.
Sibson R. (1978) Locally Equiangular Triangulation. The Computer Journal, v21 n3,
pp243-245.
Siyyid A.N. (1993) The use of METEOSAT data for rainfalVrunoff modelling. PhD.
Thesis, Aston University, May, 1993.
Worboys M.F., Hearnshaw H.M., Maguire D.J. (1990) Object-Oriented Data Modelling
for Spatial Databases. Intention Journal of Geographical Information Systems, v4
n4, pp369-383.
2. Use of GIS and Remote Sensing in Water Supply:
Applications
John Eigy
Peter David Hedges
Aston University
UK
2.1. INTRODUCTION
This Section applies the theory of the preceding Section to the problems of water
supply systems. The first section looks at the use of GIS with environmental process models.
The second section covers the use of GIS and remote sensing techniques in demand evaluation
and forecasting. The third section looks at two current applications of remote sensing for water
supplies; the detection of buried objects and leaks, and the assessment of water quality. The
last section considers the issues of using GIS and remote sensing in a utilities environment.
2.2. ENVIRONMENTAL PROCESS MODELLING AND GIS
Environmental process models are now common in research environments, these

techniques have been applied to modelling erosion, fluvial systems, groundwater and ecological
systems. This section looks at the use of GIS with environmental process models for water
supply systems in general, it then examines two case studies in detail; the use of GIS with
urban drainage models and the linkage of a particular groundwater model with a GIS. These
serve as examples of the considerations involved in linking any model to a GIS.

492
If a model is to be used within an operational environment for decision support then the
administrative effort must be minimised. Linking a model or several models to a GIS can
provide this, particularly if a corporate GIS already exists within the organisation, (see the
section on GIS Within a Utilities Environment). The advantages of linking a GIS to
environmental process models are:
• data management functions are subsumed by the GIS,
• several models can be used together in an ad hoc manner, without reconstructing the data,
• different models draw on the same version of the data,
• datasets can be easily constructed,
• consistent user interface across models,
• output can be visualised in combination with other data within the GIS.
For a water company the operations that such a system might aid are:
• assessment of environmental impacts of water resources or supply system development or

operation,
• gaining information about the processes affecting the water resources upon which the
system draws and which might affect supplies, such as siltation of reservoirs, saline
intrusion into aquifers, etc.,
• modelling in-system processes such as flow processes, algal growth etc.
There has been an explosion in the use of environmental process models with GIS,
though these are mainly being developed and used in research environments. Chairat and
Delleur (1993), linked the hydrological model TOPMODEL with the GIS GRASS, to model
hillslope runoff production and gain an understanding of the effects of soil variability. On a
large scale De Lange and Van der Meij (1993) describe a national groundwater model for the
Netherlands, based on the custom GIS REGIS that manages a huge hydrogeological dataset of
Holland. Kern and Stednock (1993) use hydrochemical models to identitY surface waters that
exceed standards of heavy metal concentrations.
493
One of the strengths of this approaches is that models can be linked in an ad hoc
manner as they share similar input and output formats. Klaghofer et al. (1993) linked the GIS
Idrisi to the soil erosion model EPIC and to the Agricultural Non Point Source Pollution
Model (AGNPS), developed by the US Agricultural Research Service and the University of
Minnesota, to estimate sediment and nutrient produced by an alpine catchment.
Ross and Tara (1992) describe a system designed to be used in an operational
environment from regulatory authorities to smlill consultant companies. It involves the linkage
of several hydrological models; a groundwater model, surface water models and specililly
written code with the commercial GIS SPANS. They praise the GIS for performing the time
consuming and tedious data management tasks such as georeferencing, and claim considerable
savings in time and money:
"what would have taken many months of effort in terms of model parameter definition
previously can now be accomplished using digital data and the GIS".
Other processes are also amenable to modelling. McDonnell and Macmillan (1993) use
a set of different models not only to study the environmental aspects of large dam development
but also social and economic aspects of the system.
2.3. PRACTICAL IMPLEMENTATION OF GIS TO URBAN DRAINAGE MODELS
The development of general purpose GIS systems has urged and encouraged their
application in various fields of engineering, particularly in the area of complex urban
infrastructures. This application is spreading in two major directions:
i. generation of maps and inventories,
ii. links with simulation models and the graphical presentation of results.
Although it is difficult to draw the dividing line between them it could be generlilly
stated that the final product of the first type is a digital map and it's accompanying database.
The second group is more complex and it requires specific target oriented information
processing (modelling). It is the second application that is of interest here.
494
Current urban drainage models are well established and generally adequate for the
simulation and design of urban drainage networks. There is, however, a problem with the
acquisition, manipulation and storage of data for these models. The combination of the
technologies of remote sensing and geographical information systems allows a solution to these
problems. This section describes a project in which GIS are being used to provide the input
data for two drainage areas: one a pilot study of the Miljakovac catchment in Belgrade is
described in detail, and one a full scale study of Dresden is referred to. Smith (1992) presents a
different way of linking a GIS and an urban drainage model where he employs the drainage
model as an intrinsic part of a GIS. This presents new difficulties in the calibration of finely
grained distributed models that are beyond the scope of this text. Further descriptions of these
methods can be found in Eigy et aI. (1993), Eigy (1992) and Prodanovic (1992).
Many details of urban drainage networks are already stored in GIS systems of type (i)
above, but the most effective use of that data relies on type (ii) applications. This section
examines how four commercial GIS packages were used to create input suitable for two
standard urban drainage models. The packages concerned were: the raster-based GIS Idrisi,
and the vector based; ARC-INFO, INTEGRAPH, and the CAD package Autocad which, with
suitable macro programs, offers a sort of GIS capability. The outputs from each of these GIS
were then to be passed to two urban drainage models: WALRUS, from Hydraulics Research
UK, and BEMUS, from IRTCUD, the University of Belgrade. The University of Belgrade has
a well-instrumented and documented catchment at Miljakovac in Belgrade and this was used as
the study area to investigate the approach. The detailed objectives of the project were:
• to form the land cover and DEM for a catchment,
• from the DEM identify the contributing areas to each inlet to the sewer system, and
determine how these can be modelled in each GIS,
• investigate the effect offlow diversion structures, for example, walls, road chamber etc.,
• calculate the statistics of each subcatchment required for WALRUS (now replaced by
SPIDA) and BEMUS,
• compare the statistics from each method with traditional hand methods,
• prepare input :files for the models from the GIS,

495
• investigate the different perfonnance of the model for each method of deriving the input
files,
• consider the sensitivity of output to each method of generating input files,
• assess the resources required for each approach in tenns of manpower, training, computer
power and software costs.
However, due to time limitations many of these objectives have yet to be met. This
section describes the use of the Idrisi GIS system (Eastman, 1992), and how the data stored on
this system was used to create input files for the BEMUS and WALRUS urban drainage
models. Only the BEMUS model has produced output for analysis so far. The work continues.
The vector-based GIS's proved difficult and slow for data acquisition and modelling.
Their complex data structures meant that it was also difficult to write suitable programs to
extract the required data for the urban drainage models. Almost no progress was made using
ARC-INFO, and AUTOCAD proved useful only as a tool for digitisation and to move data in
the DXF fonnat to Idrisi. Digitising the maps for INTEGRAPH proved very slow and though
the accuracy of the maps produced was very high, failure to correctly brief the operators of the
system meant that many of the features required for urban drainage modelling, such as
topologically consistent polygons, were not included in the database. Correct briefing meant
that the correct maps were produced for the Dresden project (Maksimovic et al., 1993)
2.3.1. Choice of GIS
Idrisi is a low cost raster GIS available from Clarke University in the USA (Eastman,
1992). It is very popular as a teaching system, and its simple file structure means that new
modules can be readily written for the systeril. Two new modules were written for this project.
The first was to determine the contributing area to each inlet in the sewer system - similar
modules could be written to determine irrigation command areas and areas that have potential
to be fed by gravity from reservoir, (Chapter VIII. I.) The second was to take these
contributing areas and write the output files suitable for the urban drainage models, with
features such as percentage impermeable area, area of contributing area, length/width ratio,
average slope of catchment etc - again similar modules could be written to output data for pipe
network analysis.
In addition, the relevant Idrisi module, which fonns a digital elevation model from point
data, was rewritten to make use of a transputer parallel processing board added to the micro
496
computer system for faster execution. This improved the execution time from around 12 hours
to 20 minutes.
The only other modules from the original Idrisi package used for this project were: the
viewing module, COLOR; the slope and aspect module, SURFACE; the reclassifYing module,
RECLASS; the area calculating command, AREA; and the geometric correction module,
RESAMPLE. The cross tabulation command CROSSTAB was used to verify the output files
calculated from the post-processing program, generating direct inputs for the urban drainage
models. Other modules can be used for traditional GIS applications.
2.3.2. Results and Recommendations
Much work still needs to be undertaken to complete the objectives mention above and
only one successful run of a simulation program has actually taken place. This run (on the
BEMUS storm water drainage package) indicated that after tuning the percentage roofs
connected to the sewer system, the model accurately matched the measured outflow
hydrograph - as did the traditional hand method. Provided that most of the catchment had been
allocated to an inlet the exact boundaries seemed to matter very little. There were very large
discrepancies between the contributing areas to particular inlets derived from the hand and
from the automatic methods.
The lessons learnt from the GIS approach can be summarised as:
• initially the GIS approach is much more time consuming than drawing boundaries roughly
by eye and measuring areas by planimeter, but this may be largely due to the steep learning
curve of developing and using new technologies,
• it took 2 days to .digitise the 4 map !'heets covering the Miljakovac catchment, though if this
had only been one sheet half a day would have been sufficient - having gained experience
and a larger digitiser, it was faster for the Dresden catchment,
• using aerial photography took only a few hours to create the land cover map - digitisation
took 2 days, aerial remote sensing offers the advantages that land-cover data can be
obtained, that is up to date and easily ported into a GIS,
• formation of the DEM is easier from stereo pair aerial photography than by any other
method,
497
• the GIS approach is more objective than traditional approaches,
• the availability of digital maps makes the GIS approach faster;
• existing urban drainage models do not need rewriting to take account of GIS technology,
though direct interaction would be beneficial it would limit the choice of GIS,
• grid based GIS are easier to deal with than vector GIS for this purpose,
• once a GIS study has been carried out, changes to the catchment or to the sewer system can
be readily investigated,
• more work is still needed on subcatchment boundary delineation, particularly with regard to
flow diverting structures,
• levels of accuracy of the digitised maps and of the resulting output still need the judgement
of an engineer; completely automatic techniques still do not exist,
• if the source of the information is an existing paper map then this is still the weakest part of
the system,
• field visits are essential, if only to discover which roofs are directly connected to the sewer
system; this data could be directly entered into a portable computer,
• more work needs to be done on examining other GIS approaches and urban drainage
models,
• very flat catchments still prove a problem when determining contributing areas, see the
Section'1. in Chapter VIIt
In this example Idrisi proved a very flexible and useful tool for developing an objective
approach to obtaining the input files for the traditional urban drainage model, BEMUS. The
programs developed for calculating the contributing areas from the DEM and to create the
input files for BEMUS and WALRUS are very useful tools. With further development the GIS
approach could provide the standard method for gathering information for urban drainage
models.
498
2.4. GROUNDWATER MODELLING WITH A GIS
In the Seventies several large groundwater developments were initiated around the UK.
However, the effect of the drawdown on soil moisture and the resulting impacts on crops were
given little attention. Hedges (1989) investigated this issue with reference to the Tern River
Catchment within the Shropshire Groundwater Scheme.
Hedges found that in areas where the soil is in hydraulic continuity with the aquifer and
where the root zone is within a critical height from the water table, crops could experience a
loss in available water due to the drawdown of the water table. GIS technology gives a method
for defining areas that are vulnerable to this effect. Hedges produced maps of vulnerable areas,
this was achieved by combining map layers representing agriculture, critical height, solid and
drift geology, depth to water table and drawdown.
The current research effort is to produce a system that will dynamically model,
groundwater, soil moisture and crop yield. This section describes the linkage of a GIS
(GRASS) with a groundwater model (Modflow).
The Geographic Resources Analysis Support System (GRASS) was developed at the
US Army Construction Engineering Research Laboratory (USACERL), (Shapiro et aI., 1993).
The principle data structure used by GRASS is the raster, though it can handle vector data. It
was chosen for this project for several reasons:
• GRASS is public domain software, it IS freely available, along with extensive

documentation and, significantly, the source code is readily available,
• the standard C programming language is used as its implementation language, and GRASS
comes with large libraries of functions for developing new applications, which means that
any shortcomings can be programmed around and new modules written,
• a raster data structure is used, that is more suitable than the vector structure for problems
that involve the handling of features whose attributes can vary continuously over space,
such as the transmissivity in an aquifer,
• GRASS shares features with many other GIS; it is atemporal and attempts to handle vector
data as well as raster,
• there is a large co-operative user community, and easy contact with users and developers is
possible via active internet discussion groups.
499
Modflow was written by the US Geological Survey (McDonald and Harbaugh, 1988),
and models groundwater flow in three dimensions using a finite difference numerical method.
It is stable, well developed, has a large user community and is arguably the standard
groundwater model used throughout the world. The source code, in FORTRAN, is public
domain, well written and easy to understand and modify.
The issues that must be addressed, concern both the conceptual and practical features
of Modflow and GRASS. This section concentrates on the conceptual issues that are likely to
be common to most GIS, in particular reconciling the differences in spatial and temporal data
representation between the process model and the GIS.
2.4.1. Spatial Representation
The GIS and the model represent two-dimensional space in slightly, but significantly,
different ways. Both use a grid structure, but GRASS uses a raster structure that imposes
more constraints on the data. In GRASS the grid lines are normally oriented north-south east-
west, there are regular intervals between grid lines, but these intervals may differ between
north-south and east-west (Figure la.). Modflow has fewer constraints; like GRASS it uses a
rectilinear grid but this can be oriented in any direction and grid intervals may vary along either
direction (Figure lb.).
a b
Figure 1. (a) A raster grid (b) A typical modflow grid
Modflow allows variable grid interval and orientation for important reasons. Within the
study area there will be regions where the gradient of hydraulic head changes rapidly within a
few metres, such as around an operational borehole, conversely there are other areas where the
500
gradient is nearly constant through space over tens of kilometres. In areas where there is a
rapid change in gradient, error, between the model and reality, can be reduced by minimising
the size of grid elements, see Rushton and Turner (I 977). However, this is unnecessary in
those regions with a small change in gradient. A large grid interval can save considerable
computation time by reducing the number of calculations required. Simple functions were
therefore developed to sample GRASS data to the more irregular Modflow style grid, and to
resample the output back.
The resampled grid is still oriented north-south, east-west. Where flow is through an
anisotropic medium this may not give an adequate representation of reality. Several authors
have stressed the importance of having the directional axis of the equations upon which
Modflow is based, parallel to the principle axis of permeability (Rushton and Redshaw, 1979).
However, this problem will not be faced in the current project as the model will be applied
where isotropic conditions can be assumed in the x,y plane.
a b
Two rasters act as OEMs that

represent upper and lower
surfaces of the geological unit
Another raster is need for
the spatial variation of
each attribute
Figure 2. (a) Simply geological situation. (b) Upper and lower surfaces of each unit are represented by
a DEM, another raster is needed to carry the attribute information
For this project GRASS is required to manage data that can vary through three-
dimensional space, . in particular it must cope with geological data, Figure 2a. The raster
structure is fundamentally 2 dimensional, though it can be used to represent three dimensions
as an altitude matrix (see the Section 1. of the Chapter VIII., section on OEM). The method
chosen represents the surfaces of geological units with a raster DEM; each pixel holds the
value of the height of the surface at that point. Another raster is required to hold the value of
each attribute as it changes across the geological unit Figure 2b. The onus is on the user to use
such data correctly, as GRASS has no structure to logically link rasters used in this manner.
This method has several advantages. (i) It is easy to produce; as most of the raw data is in the
form of borehole logs, interpolation functions can be easily applied to estimate the surfaces of
the geological units. (ii) The use of functions such as kriging or splines to interpolate
501
parameters is valid across a geological unit but these functions do not necessarily hold across
geological boundaries. By explicitly separating the attributes of different units into different
rasters these functions can be applied more easily. (iii) Most importantly, this approach closely
resembles the way Modflow handles three-dimensional data. In Modflow the third dimension is
represented as layers, and for each layer attributes such as transmissivity and specific storage
are represented as Modflow grids. Layers are not necessarily flat but can be adjusted to follow
geological boundaries.
This kind of modelling, which is known as 21/2 dimensional surface modelling, has the
draw back that complex geology, such as extreme folding, cannot be easily represented.
2.4.2. Temporal Data Representation
A significant issue is how to deal with temporal data - particularly that which varies at
different time scales. In this study geology can be considered static. Agricultural land-use can
be considered static over a growing season. Weather can change from day to day.
Groundwater levels can vary within a few hours next to an operational well but remain nearly
constant over a season elsewhere.
Modflow is able to cope with transient flow situations; groundwater flow that varies
with time. Time is divided into stress-periods that are further divided into timesteps. However,
as yet GRASS has no explicit model of time (this is in common with most traditional database
models, spatial or otherwise, see Kowalczyk and Kemp, 1993). This makes the analysis of
temporal phenomenon cumbersome. The usual solution is to treat raster overlays as snap shots
in time, and give them appropriate names; e.g. watertablejune.84
As with three-dimensional data, the onus is on the user to manage these data correctly
and analyse them in meaningful ways. For example, if a parameter is required between
available snap shots, the user must devise a method of calculation using the raster manipulation
facilities available.
The research into temporal GIS is still largely at a theoretical stage, see for example;
Kowalczyk and Kemp (1993) or Wachowicz and Healey (1993). Research has generally been
aimed at solving the traditional database problems of inventory and asset management systems,
see the final section. Al-Taha et al. (1993), give a useful bibliography of current research. Only
a few of the authors have touched on the issues oflinking any kind of environmental model to
a temporal GIS, Hazelton (1991) is an example.
502
2.4.3. Remotely Sensed Data
Remotely sensed data on its own is oflittle value to a groundwater model. It requires
the functionality of a GIS to make use of this data source. For example; Modflow models
recharge to the aquifer simply as a grid of values that is added to a given layer at the start of
each stress period. It is up to the user to calculate these values. A GIS greatly facilitates this
process.
Land-use influences recharge to the aquifer by modifYing the infiltration rate and
evapotranspiration. Remote sensed images, such as landsat TM, can be classified to provide
the data source for land-use using the image processing modules of GRASS. This data set will
be further manipulated and combined with other datasets, such as geology, drift, soil and
rainfall, to produce maps of infiltration rate and ultimately aquifer recharge.
1-------11 1 km
Figure 3. Typical output from the GRASS-modflow system

Here two wells have been modelled giving, after 5 days, two areas of drawdown in a raster layer. The
river Tern has been overlaid. Where the river passes over the drawdown, some loss of surface water is
expected (depending on the conductivity of the riverbed) and the efficiency of the augmentation scheme
is impaired
503
2.4.4. Discussion
Linking environmental process models with GIS can provide a powerful tool for
environment managers. Ross et al. (1991) found that considerable savings in time and money
could be made when they linked a groundwater model with a GIS to investigate pollution from
a mine. However, the task is not trivial, particularly as the current generation of GIS is only
beginning to deal with the dimensions of height and time. These difficulties are likely to
become less as GIS become more sophisticated, Hazelton (1991) suggested that as a GIS
becomes more fully functional in 4 dimensions it "... subsumes dynamic modelling almost by
default" . Figures 3. and 4. demonstrate some of the output from the GRASS-Modtlow system.
f - - - - - - - - - 1 1 1 km
Figure 4. Typical output from the GRASS-modflow system. Vector contours have been derived from a
raster layer of draw down after 22 days pumping, again the river Tern is overlaid, as well as name
labels for the wells. After 22 days of pumping the areas of drawdown are considerably expanded, and
the efficiency will be further impaired.
504
However, users should not be insulated from the considerations of the model
construction, otherwise there is the risk of using the model in inappropriate ways. For
example, it would be inappropriate to use the GRASS-Modflow link in its current state in areas
where there is considerable anisotropy in the x,y plane. Reddi (1990) discussed some of the
pitfalls to be avoided when using groundwater models and lists the following issues;
• using an over complex model, "over-killing", that may not be warranted by the problem it
is being applied to or the quality of the data available,
• misconceptualizing the system,
• using inadequate parameters,
• misintepreting the results.
Groudin and Lite (1990) discuss the problem of scale with regard to groundwater
models; regional groundwater models can miss important local effects and local models can
miss regional trends. Groudin et aI. (1990) stress the importance of understanding the
principles and the limitations of modelling techniques, a necessity which may be lessened in
future with the application of expert systems that can aid the modeller construct the model and
interpret the results. Grayson et aI. (1993) warn that a user may have an ".. unrealistic sense of
reality of model accuracy" induced by the sophisticated nature of the user interface and data
handling and visualisation functions. Clearly these are considerations that can be applied to all
dynamic models, and should be kept in mind when linking these models to GIS.
2.5. GIS AS A TOOL TO PREDICT WATER DEMAND
In order to improve the design of water supply systems it is necessary to obtain an

accurate forecast of the demand the system would face. This prediction can be based either
upon extrapolation of the existing demand without reference to the components of that
demand, or upon demographic studies and economic models of water consumption.
Demographic methods of forecasting future water requirements have ranged from simple
projected population multiplied by projected per capita usage, to multiple regression analysis of
different types of user and their demands for water. The crudity of the first is obvious, but the
505
second suffers from the same problem in that without knowing how these users are distributed
it is little more than a better projection of the aggregate consumption of water. The detailed
spatial modelling of GIS allows for each small area to have an accurate demand profile
associated with it, and this demand can then be incorporated into the design of the water
distribution system.
For the best predictions of water demand we need to know three things:
i) the location of users,

ii) who the users are,
iii) what is the average water consumption of each type of user.
GIS can assist in determining i) and ii), and remote sensing can also assist in acquiring
much of the raw data. As for iii), the modelling of demand functions is beyond the scope of this
chapter, but is discussed in a paper by Hughes et al. (1994). Hughes' approach is to derive
demand functions in linear, semi-log etc. form, and to determine the parameters of these
equations by linear regression using as input parameters such items as housing density, lot size,
family size, etc. For example, Hughes presents the following equation for summer demand:
b bl PPH bLOT
Qs=O 2
b 3 AV b 4 WHC (1)
where Qs is the unit capita demand, PPH is people per household, LOT is lot size, WHC is soil
type and AV is the assessed value of the property. Interestingly they do not include household
income within their prediction model. Bishop and Hughes (1993), offer a more comprehensive
paper describing their GIS Water Planning Model (GIS-WPM).
In the UK small area statistics are available from the national census, and this data
includes such items as age, martial status, occupation, housing and car ownership. The data
can be spatially referenced in accordance with the national mapping agency (the Ordinance
Survey, in the UK) system to give the data for each enumeration district (typically a few tens
of households). Extracting information from this data is a complex process but with suitable
regression models an estimate of each enumeration district's water demand can be obtained.
The use of such data is, of course, widespread in marketing (particularly direct postal
mailshots) and in planning shopping and recreational facilities.
A number of commercial companies have come into existence to take census data and
other data, such as credit and debt data, share ownership data etc, to arrive at measures of
afiluence and demand for a particular type of product, for each enumeration district (Slieght
and Leventall, 1994). These companies could just as easily provide likely demands for water as
506
for soap powder. Similar geodemographics studies are available for the rest of western Europe
(Watts, 1994).
As an example the process of calculating demands for imaginary pressure zones in
Birmingham using a GIS will be illustrated. For illustrative purposes a much simplified
equation will be used. Assuming per capita consumption within a given region is 2m3 and each
household contributes an additional 4m 3 them demand per month is:
demand per month = (2 x population) + (4 x households) (2)
Figure 5. shows raw 1991 census data for Birmingham: (a) is the population and (b) the
number of households, by ward. Average population and households per cell is found by
dividing each ward by the number of cells in that ward, and to find the demand per cell
Equation (2) is applied. Figure 6a. shows imaginary pressure zones, these were created by
digiti sing around areas, on a DEM of Birmingham, where height is approximately equal,
(Figure 6a.). The domestic demand within a pressure zone is the sum of the demands of the
cells contained within it. Figure 6b. shows the results of this calculation.
_ \(JfnJl""I'i,
:7C1Ot)'.'0I)(()
Ll :.u-...:'!7uno
o ~H,,)I')...:...$(_)(J
Q I ( . 1I1 )':l or.,
Figure 5. (a) Population by ward (b) Households by ward in Birmingham
This method depends on the assumption that population and household density are
constant across a ward, which is seldom true. A more reliable result may be obtained by using
smaller units; in the UK census data is available at the enumeration district level, which may
contain as few as ten houses.
507
Once we have the domestic demand layer created it is necessary to create the industrial
layer. As with domestic consumption we take the available data and use that. Such data takes
the form of either existing consumption data or projections based upon the development plans
for the region and the location of the industrial consumer.
>200000 m'/month
150000-200000
D 100000-150000
D
r
-0000-100000
<50000
Figure 6. (a) Imaginary pressure (b) Domestic water demand by pressure zone
zones within Binningham
Hughes et al. (1994) identify a third layer of water demand: urban irrigation. In many
parts of the world the watering of gardens and parks is the significant part of the demand for
water. In Utah it is common for this irrigation water to be provided, untreated, in a dual system
by specialist irrigation companies. This contribution needs to be taken into account when
forecasting water demand.
In some areas, plot size (included in the regression model) gives a measure of the
demand for irrigation water, but a better measure can be obtained by remote sensing. Satellite
remote sensing (pixel resolution about 30m by 30m) has too coarse a resolution for accurate
measurement but use of a vegetation index (as described in Chapter VIII.I.) will give an
indication of the proportion of each pixel covered by vegetation. Airborne remote sensing has a
better resolution (about lxlm) than satellite imagery, and this can be used to map urban
vegetation, see Blagejevic et al. (1994). A vegetation index image is shown in Figure 7a: note
that the vegetated areas are considerably lighter in shade than the non-vegetated area (roads,
houses and factories etc). Figure 7b. shows the same image with a simple reclassification into
two classes; areas with and without vegetation. Knowing the average irrigation rates of the
508
principal vegetation, we can then calculate the water demand per pixel or polygon - though
there is evidence that smaller plots may be irrigated at higher rates than large expanses oflawn.
Urban irrigation is associated with domestic houses, and if information regarding the
spatial spread of houses is available, then an estimate of demand can be made. This information
can be extracted from maps, which is particularly easy if the map is in digital form.
Alternatively remote sensing provides a more timely source of information. Traditionally
remote sensing classification is based on spectral signatures, however increasingly other
information such as context and texture available from the image is being used to give a better
classification - this will be illustrated in the following example.
Figure 8a. is a thermal image of an urban area from an air-borne sensor which includes
residential and urban areas. In Figure 8b. the image has been classified into two classes; pixels
with a radiance less than 70DN (relatively cold pixels) are given a value of 1 and pixels with a
radiance greater than 70DN (hot pixels) are given the value O. As a result, as can be seen,
roads and buildings stand out clearly.
It is possible to further enhance Figure 8. by reclassifying it on the basis of size or
shape rather than on spectral properties. Figure. 9a. shows the same image further enhanced,
small specks and very large areas (which cannot be houses) have been removed. Another
reclassification and size filter was subsequently applied to produce a layer of roads, which was
used as a mask to remove the slivers of road that had been mis-classified as houses. In most
cities the difference between residential and commercial areas is sharp, cities are generally
zoned. Using this information or visual interpretation of the image, it is possible to cut out the
industrial areas which will not contribute to domestic urban irrigation, but which may irrigate
their own green space extensively.
A further rule can be applied to give a first estimate ofirrigable land, if we assume that
all land within 20m of a house is irrigated we get the raster layer in Figure 9b. and an estimate
of O.26km2 irrigable land within this O.7Skm by O.7Skm sample area. Clearly this estimate
could be improved; for example the intersection of this image and a classified image of
vegetation could be ~sed, or roads could be excluded.
The above approach relies on the distinctly different spectral characteristics of
vegetation and man made structures, and on the different shape and size characteristics
between residential buildings roads and other buildings. It works because each house in this
suburban zone is surrounded by vegetation and this allows each house to be automatically
recognised and treated as an object distinct from the road network.
In the absence of accurate census data, remote sensing may also be used to determine
the type of housing (or other land-use), and the housing densities within the distribution area.
This process is, however, far from automatic and will need further research.
(J1
o
<0
Figure 7. (a) Part of an image of the Nonnalised Difference Vegetation Index (NOv!) band, derived from air-borne thematic mapper red and infrared
bands, of an urban area. The printing process has degraded the image, in the original each pixel can potential take a value of 0 to 255
(b) A simple reclassification of the NOV! to highlight areas of vegetation, black areas show vegetation white areas little or no vegetation
" ...,.
....... ~.
' .... ' ...
.. . '",- ""0,,-
" ..... ' ...
~ ,... I - ..
........... - ...
--_
-- ,
~ "#-..
... . ~
-- . " ......... •. ............... i
.., -... .......--.....
,. .''-> ---- ,. I
--- ....
eo ..... _ _ .... : .,1 ,
'·~' I".
~
, --.
'-
~
.. o
, ,
":
.:
:"
~:~.
Figure 8. (a) A thennal image of Dudley, an Urban Area in the Midlands of UK. Again the printing process inevitably leads to a loss ofinfonnation, in the
original image each pixel may take a value of 0 to 255 depending on the surface emittance ofthermaI radiation
(b) A simple reclassification of the same image, black pixels are these with a radiance of less than 70DN and white pixels have radiance greater than 70DN.
I . ~
J • "-
I ,•• -"1<
I '
....
I
"'"
~.~
...
'"' ......"..;,'..........
I ~
, •.••
l
...
•• -""- ....... '-.,
------- ....
...~
-, - - - . I
...... ...... -....:.. ...... f
' ... ..""'-. I
..... ...---.
.....'...... --- ..........i J
....~~ .......... ---.:. ~ '\.
,
...... ...
--_., --:u.. __ . :
• .... '" ,.... .
.............. ~ I,a-. __ .-.._ • .,
'- __ .,', " .
·4 .... f...., ....
,
I. ~
~
. - ..
\ ..".',
.. •, #
• • I ~
f ,,
~~
l' ~
" f f:;'
"• ... ... ... - ...
-
" .- ~
. ....
Figure 9. (a) Cleaned-up layer of houses, small areas and very large areas have been removed with a size filter, the industrial zone has been remove by
cutting it out, roads have been removed by using a layer of roads as a mask.
(b) Areas within 20m ofa house, a rough estimate ofirrigable land of2.6km2
512
The combination of the different layers of demand is accomplished in a raster GIS. In

the resultant image each pixel will cany an estimate of the value of total water demand for it's
area. We have seen how pressure zones may be overlaid onto that total demand layer, to give
total water demand for each zone. This method is not presented as an alternative to demand
projections based purely on historical use, rather as a complimentary technique to reduce
uncertainty. It can be argued that the more information and knowledge you have the more
accurately you can predict the future or understand the present. Furthermore this approach will
allow areas where there is a discrepancy between the expected demand (given by the profile of
different users) and the actual demand (a large proportion of which may be due to unidentified
leaks or some other undesirable process) to be identified as areas which need further
investigation.
2.6. REMOTE SENSING APPLICATIONS IN LARGE WATER SUPPLY SYSTEMS
The discipline of remote sensing is in a period of rapid development. Sensors are

improving, and new problem domains are being explored. This section will cover the current
use of these techniques in applications that have relevance to water supply systems; the
detection of buried objects and the assessment of water quality. The user is referred to such
publications such as Remote Sensing oj the Environment or Photogrammetric Engineering
and Remote Sensing (PE&RS) to keep abreast with with future developments.
2.6.1. The Use of Remote Sensing to Detect Buried Pipelines and Leaks
The advantages of locating objects and leaks before excavation are clear - there is a
financial saving because of the increased speed and mininal effort of the operations. Remote
sensing systems can aid in: the location of pipes and other buried objects for maintenance, the
identification of possible leaks, the detection of free regions underground into which a new
pipe might be placed, the detection of voids that might affect construction or imply a cave-in,
and the mapping of networks.
As systems age, as the competition from different utilities for underground space
increases, particularly in urban areas, and as comprehensive GIS based inventories become
more common the use of remote sensing techniques are likely to become more important.
Current techniques use ground penetrating RADAR and thermal infrared sensors. Vehicle or
513
handheld platforms are generally used, the large spatial coverage provided by aircraft and
satellite platforms has yet to be regularly exploited because of the restrictions on spatial
resolution.
2.6.2. Thermal Infrared Techniques
Thermal infrared techniques rely on temperature differences at the ground surface as an

indication of what may lie underneath. For a pipe to be detectable it would have to have a
significant temperature difference to its surrounding material. Wei! and Graf (1991) report on
the use of thermal imaging to detect the presence of a hot water pipe and to detect voids. It is
in the detection of leaks that this technique would be most useful. Infrared techniques have
been used very successfully in the past to detect the discharge of springs into water bodies such
as streams or lakes (Brereton and Downing, 1975; Brereton and Hall, 1983). Brereton and
Downing (1975) listed two factors which determined whether a discharge can be identified:
• the temperature difference between the groundwater and the surface water,
• the dilution ratio.
Their work suggested a temperature difference of at least 20 C and a dilution ratio of

1:2 groundwater to surface water was required - indicating that a temperature difference is
10 C can be detected.
The temperature difference between the saturated soil (caused by a leak) and the
normal ground surface temperature can also be used to identifY either water leaking from
pipelines or natural springs. Weil (1993), reported on detecting leaks in buried pipelines using
an infra-red imaging camera - a plume of water flowing from a pressurised pipeline causes a
reduction in ground temperature.
Both Weil (1993) and Weil and Graf (1991) rely on either handheld or vehicle mounted
cameras and work at very high spatial resolution. A more useful system would involve an
aircraft and have a resolution of about 0.5m.
Deeply buried pipes with slow leaks where the water will not migrate to near the
surface are not likely to affect surface temperatures. The sensitivity of imaging thermographic
cameras is around 0.1 oC, so these methods are only suitable for large leaks. Their use is likely
to be in the detection of unreported large leaks as part of a leak reduction strategy, rather than
the pin-point locating of leaks that have been reported by other means. Tracy and Walton
(1989) report on the application of airborne thermal infrared imagery to monitor about 233 km
514
of a New York state canal and about 691 Ian of a major covered channel in California. The
method was to detect moisture along the channel edges, it was found to have particular
advantages in reaching isolated stretches.
Surrogate data may be used as an alternative to direct moisture monitoring, in
particular vegetation data. Within an area of water-stressed vegetation, a patch of lush
vegetation might indicate an abnormal source of soil water possibly a leak. The remote sensing
of vegetation was discussed in Chapter VIII. 1. and using a normalised difference vegatation
index (NDVI) the difference between stressed and non-stressed vegetation is clear. This
method requires both a knowledge of the position of the pipe and for that pipeline to be
traversing an area of homogenous stressed vegetation. Clearly it will take time for an effect to
be discernable so again this is a method that might be used as part of a leak minimisation
strategy to identifY pipes that might warrant further investigation.
2.6.3. RADAR Systems
A different approach to detecting buried phenomena is the use of ground penetrating

RADAR (GPR) systems. Unlike thermal techniques these are active systems, which emit
microwaves that are reflected by the ground surface and by other surface such as a pipe, a soil
layer or a water surface beneath the ground. The return signal must be processed to remove
noise and to construct an image that may be intepreted and used.
As with all radar systems resolution is effected by antenna length, this must be balanced
by the need for portabilty. The image quality is also affected by soil properties, particularly the
soil's conductivity and it's dielectric properties. These will vary in three-dimensions within the
soil, meaning that considerable processing may be required to obtain a usable image. Several
authors report strategies for dealing with this: see for example, Osumi and Ueno (1988) and
Ho et al. (1993). Kikuta and Tanaka (1989) give figures for the performance of their system;
they successfully detected 59 out of 64 gas pipes and state that "a 200 mm diameter pipe is
clearly detected at a depth of2.3 mil.
The different networks beneath the ground coexist in three dimensional space, there is a
need for the information to be presented to the user so that the spatial relationships between
objects, including depth, are explicit. Caldecott et al. (1988) describe a system which will give
vertical slices through the ground, as well as plan views.
Currently handheld or vehicle mounted platforms are used. The increasing
performance and decreasing size of computers allows them to be portable and enables
processing to be performed on the spot, giving real-time information. As yet airborne platforms
are not regularly used, however the potential for subsurface objects to be accurately detected
515
from an aeroplane has been demonstrated by Clark et al. (1993) who use RADAR to detect
buried landmines, both metal and plastic.
Weil et al. (1993) use a combined method; thermal infrared images with a large spatial
coverage are examined to locate and roughly gauge the position of subsurface objects and
regions of interest in contaminated waste sites. Ground penetrating RADAR was employed to
give the fine detail that is need for excavation, this proved to be an efficient strategy. Whilst
searching for suspected petrol tanks beneath a Petrol Station, they also found a void which led
to the discovery of the cracked sewer main which had caused it.
Ground Surface
t
Depth to
Groundwater
0
.. , .....-..... ..................
Figure 10. The distorted ground water table beneath a leaking pipe
For leak detection there are several approaches that may be employed. One method is
to monitor the depth to saturated surface water beneath the Earth's surface. The principle is to
look for anomalies in the vicinity of known buried pipelines - i.e. a raised phreatic surface,
Figure 10. A second method is to search for changes in soil properties such as conductivity,
which are detectable by RADAR and are influenced by the presence of water within the
unsaturated zone.
Again handheld or ground vehicle mounted sensors are the norm however there are
examples of air-borne techniques in the literature. In hydrological studies, air-borne sideways
looking RADAR (see previous Section of this Chapter) is frequently used for measuring the
soil moisture content in the top few centimetres of soil - anomolous soil moisture is another
indication of possible leaks.
Sutinen et al. (1994) were able to delineate a plume of water from a landfill site, using
low frequency radio waves mounted on an aeroplane flown at 40m. Delineation was possible
due to the presence of pollutants which caused a detectable change in conductivity between the
plume and the surrounding medium.
516
2.7. REMOTE SENSING AND SURFACE WATER QUALITY
Phenomena such as turbidity and algal blooms, which influence the quality of surface
source water, and the processes sedimentation can be monitored using remote sensing. This
has the advantage that isolated and large bodies of water can be easily and frequently assessed.
This approach provides a time-series of data which can aid the understanding of underlying
processes.
The approach of Dekker et aI. (1992) was to use specially selected bands for
monitoring water quality. They used different bands for the sensing of turbid or algal waters
than for clear waters, since the particulate matter causes a high degree of scattering and the
organic matter, particularly the algae, causes high absorption. Nine bands were selected, each
was 10-20~ wide within the spectral wavelength range of 500 to 800llm. This waveband was
identified as the range of "prime interest" when investigating turbid waters. A purpose built
sensor was developed and flown successfully in the Netherlands and it was found that optical
water parameters could be determined accurately.
Many investigators have attempted to use images from standard sources for water
quality monitoring. Lathrop (1992) reports on a study to relate standard Landsat thematic
mapper (TM) images with water quality, particularly turbidity. Ground truth, in the form of
total suspended solids and Secchi disk depth, was collected almost simultaneously with the
landsat images. Using regression techniques they where able to derive a model which proved
valuable in estimating total suspended solids and water transparency. However, the algorithms
derived could not be applied to other study areas.
In a study of the wetlands of Norfolk, UK, Baban (1994) used two TM bands (TMI
the blue band and TM6 a thermal band) to assess the influence of circulation on the distribution
of temperature and turbidity in large slow moving water bodies. Areas with significant
currents could be distinguished from those without flow from the distribution of these
parameters.
Marcia et aI. (1991) also collected ground truth simultaneously with a landsat pass.
Like Lathrop (1992) regression techniques were used, to relate temperature, Secchi depth,
total suspended solids, chlorophyll concentration and nutrient concentration to the landsat TM
bands 1, 2, 3 and 4. Significantly they showed a relationship between total phosphate, total
nitrogen, suspended solids and chlorophyll to the optical water quality parameter Secchi depth.
Sedimentation is a process closely linked with turbidity. Wagle (1985) was able to map turbid
water zones in coastal waters, using landsat. These zones correlate closely with the
distribution of seabed sediments.
517
Though these techniques are still largely under research, as they improve they will no
doubt enter the operational environment. The advantages are: the spatial coverage, the ability
to build up time series of data on important processes, the identification of trouble spots and,
the ability to collect data quickly in response to developing situation.
2.S. GIS IN A WATER UTn.JTIES ENVIRONMENT
Currently in the UK the widest application of GIS in water companies is as an

inventory system for asset management; sophisticated spatial modelling is usually a secondary
consideration. On first appraisal the problems of acquiring and maintaining an inventory of the
components of a water supply system may seem trivial. However, the shear amount of data
means this is not the case, Moutal and Bowen (1991) report that they had to digitise 6000
miles of water mains in New York city. For water companies the task is essential and many
are turning away from paper systems and putting computer based GIS at the centre of their
strategies.
A water company must aim to provide water 24 hours a day, the whole year round
They must maintain the current system, integrate new components into the existing system,
plan for disasters and satisty various statutory requirements. This clearly implies an extensive
and easily assessable database, describing not only location of pipes, valves, pumping stations
and reservoirs, but other facts such as depth, size, pressure and engineeririg specifications. For
an inventory, complex spatial analysis functions are not required, rather an effective database
that provides all those facilities that any database would provide along with the spatial
dimensions. Clearly the vector model, which allows each spatial object to be linked to a
database record, is appropriate, though in future object oriented GIS may be used, see
previous Section of this Chapter. The issues involved in inventory and asset management can
be split into technical and organisational considerations and the economics of the system.
Though operations management is the principle aim of GIS development at the moment it is
possible that decision support will become more important than at present.
2.S.1. Technical Considerations
Technical considerations include: allowing multiple users to access the data whilst
maintaining security; linking to other databases; coping with a very large dataset, and
518
protecting the data physically. The system must provide for these considerations whilst giving
an adequate system performance, which means careful system analysis and design including the
selection of hardware and software. The system design must recognise the problem of
uncertainty and incomplete data sets.
The inventory database will be used by a considerable number of users; some will
require only specific bits of the data, others will need to make updates, and managers may
make ad hoc queries. For some field operations it may be useful to have the data in the form of
a map. At the same time data that is not of concern to the user should be hidden from them,
not only does this present a simpler view of the data to the user, but it protects data that may
be sensitive. It is the DataBase Management System (DBMS) that must provide alI this.
Athough water supply networks cover large regions, users may require the data at
street scale. This will clearly require considerable storage. If the data is required on-line or
access is to be distributed around a number of sites, or even to mobile terminals, then the
demands are further increased. All these considerations are similar to any corporate database;
there is a whole body of literature concerning the engineering of databases and information
systems, see for example, Avison (1992).
Uncertainty may be a considerable problem. Where the responsibility for water supply
has changed hands, as in the UK, some information about the system may be lost, paper maps
may be damaged or may be compiled to differing standards and to different scales. Jacobs et
aI. (1993) describe the construction of an inventory. The main source of pipe specifications
came from observations made when the pipes were repaired, this process began in 1975 so any
pipes not repaired since then will not have their specifications recorded in the database.
The temporal dimension is one of the current areas of GIS research. A water supply
system will change over time, customers may change, pipes are· replaced or re-routed.
Computer scientists have developed a concept of time that differentiates between world time
and database time: world time concerns the actual time of an event; database time is the time
that the event was recorded in the database. The recording of database time allows the user to
find out when the change on some object was known - this may be as important as the time of
the change itself The 'snap-shot' model of time exploited in the section on Environmental
Process Modelling and GIS is not adequate to represent this multi-dimensional view oftime.
Another associated problem is dealing with future time; different designs and associated
data of future components must be held, but there must be no confusion, these are objects that
do not and may not exist. For further information about temporal databases and GIS see
Wachowicz and Healey (1993).
519
2.8.2. Organisational Considerations
Organisational considerations include: planning for the transitional period; providing

training; imposing the system on an existing system and on existing personnel, and bringing in
new personnel; establishing new routines and procedures. It is a common experience in
information system development, that failure in these considerations can cause the system to
fail, corporate GIS is no exception.
A database is more than a computer package and a set of data - it's use in an
organisation requires changes to organisational systems. It is clearly advantageous if each
department and individual have access to the same version of the data, and that the data is the
most up-to-date available. This implies consistent systems of updating data across the
organisation and the sharing of data across departments. This may go against corporate
culture, where different departments have different procedures and where data is viewed as the
property of the department, or even the individual.
Several authors have stressed the importance of the users in both system development
and in operation, see Avison (1992) or Senn (1992). If the system has been badly designed
then in operation users may adopt a behaviour of avoidance; the department or individuals find
ways of performing their functions without using the information system, so the new system
becomes an expensive waste of money and envisioned benefits do not happen. This can occur
for various reasons; in particular poor usebility or leamability of the system, and inadequate
training.
Davies and Medyckyl-Scott (1994) conducted a survey that uncovered some
interesting patterns specific to GIS use within various organisations. Vector systems, which are
.the most suitable for the inventory tasks, were found by users to be the least usable; single-user
PC based raster systems scored the highest usability. Raster based systems were "significantly
more likely than vector based systems to be used for complex analysis and modelling tasks".
Davies and Medyckyl-Scott found evidence of "gross underuse" as well as low usebility of
large corporate vector systems. This may reflect the users of each type of system; in the
corporate system the user is more likely to be performing a job that existed before the GIS
arrived, whereas the single user may have adopted GIS to perform a new operation.
To an extent corporate GIS have to be complex and so inevitably have a low usebility -
they must, after all, serve the needs of a large population of users with vastly varying needs.
Althought these corporate systems have the greatest potential for customisation, Davies and
Medyckyl-Scott found that customisation in general "seems to have not radically improved
usability". Despite these findings, Davies and Medyckyl-Scott, also found that utilities
reported a higher usability than any other type of organisation. Several conclusions can be
drawn.
520
• Corporate GIS have been oversold on the basis of sophisticated functions and analysis,
whereas much of this functionality is never used,
• The addition of more and more functions, whether through customisation or upgrade, does
not necessarily improve productivity. A more realistic approach might be to customise to a
crucial subset of functions, an approach adopted by certain software houses. For example,
Arc-view, which uses only a small subset of Arc-Info's functionality, has been developed
to allow users to explore data,
• Clearly the needs of users are a central consideration in systems analysis and design.
Ideally the user should be included in the development process. The reader is referred to
the standard information systems texts such as Avison (1992), Senn (1992) or
Hawryszkiewycz (1991). Papazian and Reddingius (1992) describe a process of human
factors analysis to produce user friendly GIS.
2.8.3. Economics of Corporate GIS
Any development must be economically justifiable. Traditionally this justification has

taken the form of a cost-benefit analysis and if the benefits outweigh the costs then the
development can go ahead. The risk must also be assessed. How likely is it that the proposed
system will fail? If the risk is high, then despite a positive cost-benefit analysis, it many not be
advisable to pursue the project. There are many standard texts that discuss these issues in
depth, see Hawryszkiewycz (1991) or Senn (1992) for an introduction with respect to
information systems.
With regard to a GIS in an operational environment several issues impinge on these
considerations.
In terms of costs, the cost of data is far greater than that of software and hardware. In
the UK to purchase a single tile of vector data for an urban area covering SOOm by SOOm costs
about £134 single payment, plus £10 maintenance per year. To acquire national coverage of
Great Britain with 1:S0,000 raster map tiles costs £39,SOO, copyright fees are extral . To get a
single 1:10,000 aerial photo covering 4km2, from an archive can cost about £IS; a flight for
airborne thematic mapper can cost around £20,000 and a single current LANDSAT S scene
with 7 bands covering 18S by 170km area can cost about US$4000.
1Prices taken from 1995 Ordnance Survey Price list

521
If the data do not exist then they need to be gathered by some form of survey. In order
to keep the database current this will need to be an ongoing process. There is also a
considerable administrative cost in handling large dataset and converting data from one form to
another. It is thus important to note that data is an ongoing cost.
The costs of personnel are hard to quantify, but are again likely to be considerably
more than hardware or software. A large corporate GIS will require a full time manager and
staff to develop and maintain it. The costs of personnel compound the costs of data. For
example to digitise a theme from a single map sheet requires approximately 1 man/day.
Remote sensing data must be processed; it takes 2-3 man/hr to extract a theme from a single
air photo or stereo-pair.
The benefits of moving to GIS are likely to be less tangible than the costs, this is a
feature of most information systems, as benefits generally derive from indirect improvements in
efficiency and service. Having a good street level GIS that enables buried pipes and valves to
be located more quickly has several advantages; a tangible but probably negligible saving in
waste water; a saving in man hours and energy in searching for the pipe or valve and, an
intangible perception of a better (or less bad) service in the minds of the customers affected by
the maintenance work.
Pay back times, the time taken for benefits to outweigh the costs, can be much longer
than is usually acceptable in business environments - and it may not even be possible to define
a payback time because of the intangible nature of benefits.
2.8.4. Decision Support
Currently what most organisations want from GIS is an asset management or inventory
system to assist with operations management. Decision support is a lesser priority. However,
as the issues ofinventory are resolved the advantages of using GIS for decision support may be
realised, particularly the possibilities of system modelling applications such as network analysis.
An important part of the water supply system will be the network of pipes. Network
analysis is a specialised field of GIS. Although pipe network analysis is considerably different
from the usual traffic network analysis, there are some examples in the literature, such as
Djokic and Maidment (1993). Network analysis uses the topology of the system rather than
spatial relationships; distances are functions of the resistance to flow in the pipes as well as the
length; resistance will depend on, amongst other parameters; diameter, roughness and slope.
Clearly effective network analysis can only be accomplished with a complete database and will
require the vector model or, in future, the object oriented model.
522
Rather than incorporating pipe network analysis into the GIS, an alternative approach
for existing GIS, is to design the GIS interface to output files for inclusion with existing
standard pipe network analysis programs (see the section on Environmental Process Modelling
with GIS in this Section).
2.S.5. Discussion
GIS provides solutions to the existing tasks of water-supply inventory that companies
have previously undertaken using paper systems. Silbovitz and Jolls (1989) found that the
availability of GIS led to an increase in the quality of service, better decisions by management
and a lowering of the cost.
GIS also affords a whole range of new operations that can increase operational
effectiveness throughout the company. French (I993) used the data from GIS inventory to
estimate damage to a network in an earthquake scenario. Wang et aI. (1991) demonstrated
that a GIS inventory can be used in earthquake contingency planning. It is not only in disaster
analysis that GIS can allow companies to take on a more pro-active stance, a temporal GIS
should allow managers to identify parts of the system that are getting old or that may be under
stress from damaging activities or increasing demand.
As with any corporate database, changes will need to be made throughout the
corporation to take full advantage of the GIS. The costs are high and there are risks, but the
GIS should not be allowed to become an expensive map drawing package. The issues are as
much managerial as technical.
2.9. REFERENCES
A1-Taha K.K., Snodgrass RT., Soo M.D. (1993) Bibliography on Spatiotemporal Databases.
SIGMOD ACM, v22 nl, pp59-67.
Avison D.E. (1992) Information Systems Development: A Database Approach. 2nd Edition,
Blackwell Scientific Publications.
Baban S.M.I. (1994) Mapping Turbidity Surface Temperature and Water Circulation Patterns
with the Aid of Satellite Imagery. Journal of the Institution of Water and Environment
Management, v 8 n 2 pp 197-204.
Bishop A.B., Hughes T. (1993) GIS Water Planning Model for Urban and Agricultural Areas.
Management ofIrrigation and Drainage System: Integrated Perspectives 1993, ASCE
New York.
523
Blogojevic B, Elgy J., Chen Z., Maksimovic C. (1994) Airborne Videography as Data Source
for an Unban Hydrological Model. Remote Sensing and GIS in Urban Waters, UDT
94.
Brereton N.R, Downing RA (1975) Some Applications of Thermal Infrared Linescan in
Water Resources Studies. Water Services v79, pp91-96.
Brereton N.R, Hall D.J. (1983) Groundwater Discharge Mapping at Altnabreac by Thermal
Infrared Linescan Surveying. IGS Report FLPU 83-7, British Geological Survey, UK.
Caldecott R, Poirer M., Scofea D., Svoboda D.E., Terzuoli A (1988) Underground Mapping
of Utility Lines Using Impulse RADAR. IEEE Proceedings Part F, Communications,
Radar and Signal Processing, v135-F n4 pp343-353.
Chairat S., Delleur J.W. (1993) Integrating a Physically Based Hydrological Model With
GRASS. Application of Geographic Information Systems in Hydrology and Water
Resources Management, Eds Kovar K and Nachtebel H.P., Proceedings of the
International Conference HydroGIS 93 Vienna 19-22 April 1993, ppI43-150.
Clark G.A., Sengupta S.K., Sailes K, Sherwood RJ, Hernandez J.E., Buhl M.R, Schnaich
P.C., Kane RJ., Barth M.J, Del Grande N.K. (1993) Sensor Feature Fusion for
detecting Buried Objects. Proceedings of SPIE - International Society for Optical
Engineering v1942, ppI78-188.
Davies C., Medyckyl-Scott D. (1994) GIS Usability; Recommendations based on the Users
View. International Journal of Geographic Information Systems v8, n 2 ppI75-189.
Dekker AG., Malthus T.J, Wijnen M.M. (1992) Spectral Band Location for Remote Sensing
of Trubid and/or Eutrophic Waters. Proceedings of the lst Thematic Conference on
Remote Sensing for Marine and Coastal Environments, New Orleans, USA
De Lange W.J, Van der Meij (1993) Application of Geographic Information Systems in
Hydrology and Water Resources Management. Eds Kovar K and Nachtebel HP,
Proceedings of the International Conference HydroGIS 93 Vienna 19-22 April 1993,
pp333-344.
Djokic D., Maidment DR (1993) Application of GIS Network Routines for Water Flow and
Transport. v119 n2, pp229-245.
Eastman JR (1992) IORISI User Manual. Clark University, Graduate School of Geography.
Elgy J (1992) TEMPUS Project, JEP 2424-91/1, Final Report on 1991192 Activities.
Department of Civil Engineering, Aston University.
Elgy J, Maksimovic C., Prodanovic D. (1993) Using Geographical Information Systems for
Urban Hydrology. Application of Geographic Information Systems in Hydrology and
Water Resources Management, Eds Kovar K and Nachtebel HP, Proceedings of the
International Conference HydroGIS 93 Vienna 19-22 April 1993, poster session.
French S.P. (1993) Infrastructure Modelling with GIS, Geographic Information Systems and
Their Application in Geotechnical Earthquake Engineering 1993, ASCE, pp45-48.
Grayson RB., Bloschl RD., Barling RD., Moore I.D. (1993) Process Scale and Constraints
to Hydrological Modelling in GIS. Application of Geographic Information Systems in
Hydrology and Water Resources Management, Eds Kovar K and Nachtebel HP,
Proceedings of the International Conference HydroGIS 93 Vienna 19-22 April 1993,
pp83-92.
Groudin G.H., Lite Jr KE. (1990) Numerical Model Uses and Limitations for Ground Water
Management. Transferring Models to Users, American Water Resources Association
1990, ppI21-129.
524
Groudin G.H., Gannett M., van der Heijde P.K.M., Patt RO. (1990) Critical Errors that
Hydrogeologic Professionals can Make with Computer Programs, Transferring Models
to Users. American Water Resources Association 1990, ppI49-157.
Hawryszkiewycz I.T. (1991) Introduction to Systems Analysis and Design. Prentice Hall
Australia.
Hazelton N.M.J. (1991) Extending GIS to Include Dynamic Modelling. Proceedings of the 3rd
Colloquium of the Spatial Research Centre.
Hedges P.J. (1989) The Impact on Agriculture of Drawdown of Shallow WaterTables. PhD
Thesis, Aston University UK.
Ho G., Kawanaka A, Takagi M. (1993) Image Quality Improvement for Underground Radar
by Block Migration. Proceedings of SPIE - The International Society for Optical
Engineering v1942 p105-116.
Hughes T.C., Bishop A.B., Fullerton H.H. (1994) Application of GIS and Remote Sensing to
Projecting Urban Water Demand. UDT-94 Moscow, Russia, Sept 12-16, 1994.
Jacobs P., Goulter I.C., Davidson I (1993) Water Distribution GIS from Fragmented and
Incomplete Information. Journal of Computing in Civil Engineering, v7 n3, pp372-386.
Kern T.I, Stednick ID. (1993) Identification of Heavy Metal Concentrations in Surface
Waters through Coupling of GIS and Hydrochemical Models. Application of
Geographic Information Systems in Hydrology and Water Resources Management, Eds
Kovar K. and Nachtebel H.P., Proceedings of the International Conference HydroGIS
93 Vienna 19-22 April 1993, pp559-567.
Kikuta T., Tanaka H. (1989) Ground Probing Radar System. IEEE Aerospace and Electronic
Systems Magazine v5 n6 pp 14-16.
Klaghofer E., Birnbaum W., Summer W. (1993) Linking Sediment and Nutrient Export
Models with a Geographic information System. Application of Geographic Information
Systems in Hydrology and Water Resources Management, Eds Kovar K and Nachtebel
HP, Proceedings of the International Conference HydroGIS 93 Vienna 19-22 April
1993, pp501-506.
Kowalczyk A and Kemp Z. (1993) Incorporating the Temporal Dimension into a GIS
Application, GIS Research UK Proceedings 1993.
Lathrop RG. Jr. (1992) Landsat Thematic Mapper Monitoring of Turbid Inland Water
Quality. Photogrammetric Engineering and Remote Sensing, v 58 n 4 pp465-470.
Maksimovic C., Fuchs L., Prodanovic D., Elgy I (1995) Full Scale Application of Standard
GIS Packages with Urban Storm Drainage Simulation Software, Kurier abwasserung.
Marcia E., de Moraes Novo L., Zuccar C., Braga F. (1991) Use of TMlLandsat Data to
Estimate ChloroPhyll Concentration and Turbidity in the Barra Bonita Reservoir.
Proceedings 24th lNternation Symposium Remote Sensinf of the Environemnt, Rio De
Janeiro, Brazil, v 1 pp389-401.
McDonnell RA, Macmillan W.D. (1993) A GIS -based Hierarchical Simulation Model for
Assessing the impacts of Large Dam Projects, Application of Geographic Information
Systems in Hydrology and Water Resources Management. Eds Kovar K and Nachtebel
HP, Proceedings of the International Conference HydroGIS 93 Vienna 19-22 April
1993, pp409-418.
Moutal H.P., Bowen D.R (1991) Updating New York's Sewer and Main Distribution
Systems: Practical Applications of GIS. Second National Speciality Conference on
Civil Engineering Applications of Remote Sensing and Geographic Information
Systems, 1991, ASCE, pI55-164.
525
Osumi N., Ueno K (1988) Detection of Buried Plant. IEEE Proceedings, Part F:
Communications, Radar and Signal Processing v135-F n4 p330-342.
Papazian B., Reddingius N.H. (1992) Human Factors Analyses of GIS use Patterns.
Geographic Information Systems: Proceedings of the Seventh Annual GRASS Users
Conference, March 16-19 Lakewood, Colorado, ppI95-202.
Prodanovic D. (1992) Automatic Delineation of Subcatchments. TEMPUS JEP 2424/91-1,
Institute of Hydraulic Engineering, Faculty of Civil Engineering, University of
Belgrade, 1992.
Reddi L.N. (1990) Potential Pitfalls in Using Groundwater models, Transferring Models to
Users. American Water Resources Association 1990, pp131-140.
Ross M.A., Schwartz M.A. Tara P.D. (1991) Groundwater Modelling with GIS: An Example.
Proceedings of the International Symposium on Ground Water in Practice. Nashville
Tennessee, pp244-249.
Ross M.A., Tara P.D. (1993) Integrated Hydrologic Modeling with Geographic Information
Systems. Journal of Water Resources Planning and Management, v119 n2 ppI29-140.
Rushton KR, Turner A. (1977) Permissible Mesh Spacing in Aquifer Problems Solved by
Finite Differences. Journal of Hydrology, v34, pp63-76.
Rushton and Redshaw (1977) Seepage and Groundwater Flow. Numerical Analysis by Analog
and Digital Methods, John Wiley and Sons.
Senn lA. (1992) Analysis and Design ofInformation Systems. 4nd Ed, McGraw-Hill Book
Co-Singapore.
Shapiro M., Westervelt l, Gerdes D., Larson M., Brownfield KR. (1993) GRASS 4.1
Programmers Manual. US Army Construction Engineering Research Laboratory.
Silbovitz A.M., Jolls J.D. (1989) Computerised Infrastructure Mapping and Management
Systems at Low Cost. Journal of the New England Water Works Association, vl03 n3,
ppI51-155.
Slieght P., Levenehall B. (1994) Meeting the Neighbours: a Review of Britain's New
Classifications Systems. GIS Europe v3 n4 May 1994.
Smith M. (1992) A Distributed Parameter Hydraulic Model for Urban Storm Water
Protection. Dr. Sc. dissertation, FAGG Ljubljana.
Sutinen R, Hannula P., Hyvonen E., Kontio M., Lanne E., Turunen P. (1994) Airborne and
Terrestrial EM Detection of Landfill Plume in Suburban Rovaniemi, Finland.
Proceedings of the International Airborne Remote Sensing Conference and Exhibition,
12-15 Sept 1994, Strasbourg, Vol 3 pp215-222.
Tracy lP.,Walton P.A.. (1989) Hydrologic Investigations in Canal and Aqueduct Systems
Using Airborne Thermal Infrared Linescanning. Digest - International Geoscience and
Remote Sensing Symposium (IGARSS) v5 pp2815-2819.
Wachowicz M., Healey RG. (1993) Towards Temporality in GIS. GIS Research UK
proceedings 1993.
Wagle B.G. (1985) Distribution and Dispersion of Turbid Waters on the Western Continental
Shelf of India - An Intepretation of LANDSAT images. International Journal or
Remote Sensing, v 6, nil, ppI681-1684.
Wang L.RL., Ishibashi I., Wang lC.C. (1991) Inventory and Seismic Loss Estimation of
Portland Water/Sewer System. GIS Application to Buried Pipelines. Proceedings of the
3rd US Conference on Lifeline Earthquake Engineering 1991, pp490-499.
Watts P. (1994) European Geodemographics on the up. GIS Europe, v3 n4, May 1994.
526
Wei! G.I. (1993) Non-contact, Remote Sensing of Buried Water Pipeline Leaks Using Infrared
Thermography. Proc. 20th Coll. on Water Management in the 90's, May 1-5 1993,
Seattle USA.
Wei! G.I., GrafR.I. (1991) Infrared Red Thermography based Pipeline Leak Detection. Proc.
2nd National Speciality Coll. on Civil Engineering Applications of Remote Sensing
and Geographic Information Systems. Washington May 1991.
Weil G.I., Graf R.I., Forister L.M. (1993) Non-Destrucitve Remote Sensing of Hazadous
Waste Sites. Proceedings of the Conference on Digital Image Processing: Techniques
and Applications in Civil Engineering, pp279-284.
3. Application of GIS in Analysis and Design of
Cedo Maksimovic it
Dusan Prodanovic it
IRTCUW, Faculty of Civil Engineering
YUGOSLAVIA
3.1. INTRODUCTION
In the preceding Section of this Chapter the application of basic principles of GIS
(Geographical Information Systems) have been presented along with the application of these
techniques to the analysis of the effect of drawdown of groundwater, water demand
consumption analysis and prediction of water demand. This Section deals with the particular
application of GIS matched to hydraulic analysis of the water supply network and for
interpretation of the results of hydrodynamic modelling. The examples used in this Section are
based on the application of an inexpensive GIS package IDRISI (Eastman, 1992). For large
scale systems and full scale projects more powerful GIS packages may be more appropriate.
3.2. SOURCES OF INFORMATION FOR APPLICATION OF GIS IN WATER

SUPPLY SYSTEMS
The use of GIS in Water Supply Systems is linked to:
• access to general purpose data bases (sources of data) dealing with spatial features,
* computer support: D. Strbac, RGZ, Belgrade, M. Draskovic, University of Belgrade, YUGOSLAVIA

528
• assessment of natural water resources for design and management,
• assessment of water demand,
• assessment of pollution sources and consequences,
• creation ofDEM for matching with simulation models (zonning, etc.),
• creation ofland use maps for population density, height of buildings, urban irrigation, etc.,
• layers ofinventory with infrastructure, network, structures, etc.,
• leakage detection and presentation of the results obtained by mains burst statistics,
• presentation and analysis of results of simulation,
• design (drafting, plotting),
• presentation of other results, publishing, etc.
The most widely distributed source of information that can be used for simulation
models are traditional, maps and drawings on paper that have to be either digitized and/or
scanned and processed.
However, there are other possible sources of information, such as:
• satellite images,
• images obtained. by photogrammetry and videography,
• automatic (total stations) field surveying,
• data collected by surface born platform and sensors (for example thermal infrared for
leakage detection).
To match these data to the commercially available GIS packages such as ArcINFO,
Intergraph, SQL, GRASS, etc., a series of user friendly interfaces; gluing routines (Elgy et al.,
529
1993), have to be developed and customized. Some examples of such routines have been
presented and discussed in the preceding Section.
3.3. INFORMATION NEEDED FOR MATCHING GIS WIm SIMULATION

PACKAGES
For practical reasons we will concentrate only on applying GIS to simulation packages
for hydraulic analysis of water supply networks.
One of the major advantages of matching simulation models with GIS is the ease in
which the results can be presented in human readable form and used for analysis and further
improvements of the system's design and operational management.
For this purpose three major groups of information are required:
• underground services (utilities) such as pipes, valves, meters, etc. and their position, status,
characteristics, etc.,
• information on water consumption in all network nodes,
• land use, digital elevation model, topography, surface devices.
It is assumed here that the information on the underground utilities required for
hydrodynamic modelling (pipe characteristics, valves, etc.) have to be obtained from the
relevant data bases and that they are available. Information on water consumption in nodes is
also assumed to be known. The means of obtaining and analysing the data are presented in the
preceding Section.
Information on topography (elevation) is handled in the form of digital elevation
models. In addition to the other applications they can be used for analysis of pressure changes
during the period of simulation or during daily operational management.
With the above sets of data one can create appropriate input files for running simulation
models. The results from simulation models are then superimposed over the digital terrain
model for obtaining spatial distribution of piezometric levels, pressure fields, etc.
530
3.4. MATCHING GIS WITH RESULTS OF MODELLING
Standard GIS packages can be used either for presentation of results or (more
preferably) for interaction with modelling. The authors favor the second approach
(Maksimovic et aI. 1994).
In what follows use will be made of the figures obtained by the interaction of the results
of hydrodynamic modelling with the IDRIDI GIS package. For this purpose a simple
hypothetical Water Supply network has been created. The network is shown overlain over the
contour map in Figure 1. and the DEM in Figure 2.
- - Mains ( Zone 1 )
- - Mains ( Zone 2 )
3
80
85
90
95
2 100
105
110
115
120
125
2 3km
Figure 1. Digital Elevation Model presented in the fonn of contour lines with a hypothetical, two zones
WS network layer superposed
531
A hypothetical network consisting of two pressure zones has been imposed on this
DEM. It is assumed that water is supplied to the first (lower) zone which contains one
reservoir. From the reservoir water is further pumped to the second zone by booster pumps. It
is assumed that there is no reservoir in the second zone. This is often the case in smaller
--
networks. In these cases pressure in the networks reacts more intensively as the consumption
in the network changes.
<10m
---
10-20 m
20-30m
3O - 40m
4O-5Om
5O - 60m
6O - 70m
70 - 00m D
00-90m 0
90 - 100 m D
lOO-110m D
110 -120 m CJ
>120m D
Figure 2. Plan and ortho view of the DEM
The network (containing two pressure zones) has been shown on the ortho view
(Figure 3.) and plan view (Figure 4.).
532
With the above mentioned data prepared in digital form a hydrodynamic model for
simulation of daily water consumption has been run.
Figure 3. WS network systems for two pressure zones
.,
~Ir - '.
",
)
.- '1:"
-
0
;:I.J
-c· 0
~. :....J
;(1 0
ID.
.. ., 0 t o'1 I
.;~ -l
f · •••• 0"
1.,. ,...,. .... ..... 1
, ..iI . '
'1 ... · '0'
If , '1'< " , "V
Figure 4. Digital Elevation Model with network system (plan view)

533
Some results of the hydrodynamic simulation (piezometric field) are shown in Figure 5.
The first results obtained from the hydrodynamic analysis are the piezometric levels in nodes.
They vary in both space, (from one node to another) and in time (as a result of variable
consumption). Thus they are very suitable for presentation by GIS tools.
The piezometric fields shown in Figure 5. are two snapshots (maximum and minumum
consumption).
By an overlay operation of GIS the piezometric field has been transformed into
pressure field as follows:
"'£"'=p-z
pg
where: p is pressure
p is water density
p is piezometric level
Z is elevation.
The result of this overlay operation is shown in Figure 6. for two snapshots (hours of
maximum and minimum consumption). The image has been created by using only the nodal
values of pressure heads.
For the sake of simplicity the nodal negative pressure heads have been allowed to
assume physically unrealistic values (-22.90 - for example).
This means that the part of network with such values is filled by ~r and the network
analysis should be altered accordingly.
The changes in the pressure field are illustrated in Figure 7. fot four different ~shots
(8 a.m., 2 p.m., 8.p.m. and 2. a.m.).In this Figure the whole range of pressure heads has been
subdivided into only four categories as follows:
• dry areas (air in mains),
• areas with insufficient pressure (pjpg< 20 m),
• areas with acceptable and favourable pressure heads (20 m <pjpg< 55 m),
• areas with high pressure (pjpg > 55 m).

534
Hour of maximum 55.04 D Hour of minimum

consumption 5978 _ consumption
64.52
92.97
97 .71
102.45 _
107.19 0
111 .93 D
116.67
121 .41
126.16 _
Figure S. Piezometric fields
Hour of maximum -29.25 D Hour of minimum

consumption -22.90 consumption
-16.55
66.03 _
Figure 6. Hypothetical pressure bead fields

535
8:00AM 2:00 PM 8:00 PM 2:00AM
Network elements
Dry Area N
A
Pipes - Zone I
Low « 20m) _ Pipes - Zone II
Acceptable (20 - S5m) Reservoir
o 1km
High (> S5m ) _ Pump
Figure 7. Daily dynamics of pressure head fields
This presentation of results enables an easy and reliable selection of reconstruction and
rehabilitation options, because it is apparent that areas continously suffering fron either low or
excessive pressure should undergo structural redesign, such as pipe replacement, distribution
relocation, installation 0f pressure reducing or pressure sustaining valves or new booster
stations, etc.
By making use of DISTANCE, a GIS operator, the shortest length from each pixel to
the nearest pipe in the map can be determined. This information can be used for energy losses
calculations in small diameter pipes, which are normally neglected during network analysis:
L v2
~=A-
D 2g
where: ~ is energy losses,

L is length obtained (Figure 7.),
536
v is velocity,
A. is friction factor,
D is diameter.
For this case all other values (A., D, v) are kept constant. In normal operations a user
can make connections between these values and cover information using other GIS tools. In
this way the pressure field obtained is more realistic and it is much easier to delineate the
zones subject to low and high pressure.
This information has been produced by ordinary GIS operators (Borroughs, 1986) and
does not require any special programming.
The coupling of the GIS and hydrodynamic simulation, shown in the above figures,
could be extended further with animation to describe the behavior of the system in an even
more friendly manner.
3.5. REFERENCES
Borroughs P.A. (1986) Principles of Geographical Information Systems for Land Resources
Assessment. Claredon Press, Oxford.
Eastman J.R. (1992) IDRISI User Manual. Clark University, Graduate School of Geography.
Elgy J, Maksimovic C., Prodanovic D. (1993) Using Geographical Information Systems for
Urban Hydrology. Proc. of Hydro GIS 93, lARS Publ. no 211.
Maksimovic C., Prodanovic D., Elgy J., Fuchs L. (1994), GIS (or GIM) in Water Projects-
Tools or Toys. Hydroinformatics '94. Proc. of the First Int. Conference on
Hydroinformatics, PubI. by Balkema, Delft.
4. Geographical Infomation Systems and Digital
Mapping for Water Supply Inventory

John Eigy
Aston University
John Snoxell
WESSEX WATER
BS2 OJQ Bristol
UK
4.1. INTRODUCTION
Water companies are complex organizations 'with many spatially distributed assets;
pipes, plants, reservoirs and sewers. Furthermore they must communicate and interact
effectively with a constantly changing base of hundreds of thousands of customers; industrial,
agricultural and domestic. The statutory framework is continually evolving and companies
must, more and more, consider such issues as the environment, liaise and report to regulatory
bodies and operate in an increasingly competitive market. An efficient and secure information
system is, essential. More than most businesses, water companies have to manage uncertainty.
In the UK in 1995, for example, a drought, followed by a cold snap caused many burst pipes
and consequently increased leakage.
With the increase in power and flexibility of computer technology, water companies are
turning to computer based information systems and because of the spatial nature of much of
the necessary data, companies are either implicitly or explicitly building or buying Geographic
Information Systems (GIS).
The activities that the information system must support can be divided between core
operations at the heart of the companies business, secondary concerns, such as statutory
duties and extraordinary operations caused, for example, by drought. Also the use of
NATO ASI Series, Partnership Sub-Series, 2.Environment- Vol. 15

538
computer technology makes possible many activities such as sophisticated data analysis and
system modeling which were not possible or practical with a manual system. The benefits of
these additional operations, in particular, are difficult to quantify and cost.
Water companies have the objectives of supplying water, handling sewage and charging
the customer. Supplying water involves collecting and treating water, maintaining the pipe
network, monitoring reserves and estimating future demand and supply. Handling sewage
involves maintaining the sewer network and treatment plants and monitoring the receiving
environment. Charging the customer involves estimating or monitoring customer usage and
sending out bills. Furthermore there are support operations such as the payroll or maintaining
the vehicle fleet. All these operations occur within a legislative framework and, increasingly, a
competitive environment. In order to support these core operations some form of information
system whether manual or computerized is required. Many of the core operations involve
information that is fundamentally spatial in nature; the distribution of the pipe network, the
customer base for example. Increasingly water companies in the UK are utilizing GIS to
provide the basis for an information system that will effectively support these operations.
Having a single store of information which is then moved to individual computers
ensures that every one in the organization is working to a consistent set of data. This prevents
the issue of out of date maps or data where a modification has been carried out by one part of
the organization but not everyone within the organization has the same information.
As well as the core operations, water companies have other activities that must be
performed. In the UK water companies must Iiase closely with the National Rivers Authority,
with the 'watchdog' bodies which regulates pricing, with local authorities that regulate access
to facilities buried in the highway and with other utilities. Water companies in the UK have
been privatized so there is greater need for business and investment planning based on sound
information.
Water supply is a crucial part of the infrastructure and the economic viability of a
region. Failure of supply for even a short period, such as a few hours can cause large loss of
revenue due to lost production. Here the information system can help in three ways. Firstly by
identifying vulnerable parts of the water supply system and allowing the company to be
proactive in dealing with problems, this chapter will briefly discuss how a water company uses
GIS to identify vulnerable parts of the network. Secondly by enabling the development of
contingency plans and the performance of scenario analysis. Thirdly a flexible system including
portable PC's can enable a very efficient reactive response to incidences.
In addition to the primary and secondary business needs of the organization which can,
usually be quantified, GIS offers a number of tertiary unquantifiable advantages. The first of
these is the idea of simply doing a job in the best possible manner. A leap of faith. It is always
reassuring to know that the quality of data being used in any part of design or maintenance is
539
as high as possible and a good GIS with documented and implemented quality control aspects
gives that reassurance. A word of caution is required here; just because the data is stored in a
GIS does not guarantee that the quality of the data is high. A method of quality control must
be implemented at the setting up of the GIS.
Once the data is stored in a portable form (in this sense portable means the ability to
move the data from one computer or application program to another) other methods of
analysis become available that were not thOUght of before. For example studies can be carried
out on the recorded incidence of disease, such as cancer, and the water treatment process used.
The introduction of any new information system, including GIS, to an organization
involves costs and risks. The costs include: the hardware, the software, data acquisition, data
translation, staff training and the costs associated with the change over. The risks are that the
new system will not deliver the intended benefits, that the costs will be greater than expected
(very common in any IT system) or that the system detrimentally affects the organization's
operations. The establishment of a corporate GIS within a water company has very tangible,
immediate and also ongoing costs; the benefits, however, are comparatively intangible
involving projected increases in the effectiveness of core operations, improvements in
secondary operations, possible commercial advantages and the benefits involved with more
sophisticated data analysis and system modeling. These benefits occur away from the GIS in
other parts of the organization and the GIS may be seen as a cost centre. It is becoming
common for companies to establish internal markets to deal with this problem, whereby
departments buy in the services of the information systems department and thus the benefits
become apparent. This can be taken further and the information services department may sell
services outside the business, indeed this chapter will describe the work of both Wessex Water
and the Dwr Cymru's GIS departments.
Many authors have dealt with the problems, issues and strategies of introducing
corporate information systems to organizations, see for example Senn (1992). But the issues
concerning data should be emphasized. It can not be overemphasized that data capture is the
over-riding cost of Utilities GIS, the co:;;t include; manual digitizing paper maps, buying
external sources of data in particular Ordnance Survey digital maps and the collection and
inputting of attribute data.
This Section focuses principally on Wessex Water, a water company in South west UK.
It also briefly describes the ways GIS is being utilized in Welsh Water (Dwr Cymru), a much
larger organisation and in South Staffordshire Water which is much smaller.
540
4.2. WESSEX WATER
The UK Water Industry underwent a major reorganisation in 1974 which resulted in 10

large, catchment-based Regional Authorities responsible for complete water-cycle management
(Le. Water Supply, Sewage Treatment, Land Drainage). Wessex Water is one of the smaller
pic's with a total workforce of around 1,200 supplying about 1 million water supply
customers and 2 millions sewage treatment customers (2 independent water companies exist
within the Wessex area, supplying around I million supply customers).
One advantage of the 1974 reorganisation was the possibility for large economies of
scale, including computing facilities, and the 1970's saw major investments in mainframe
provisions, telemetry systems and communications infrastructure. The legacy of record
drawings from the predecessor organisations (e.g. Water Boards, Local Government Sewage
Treatment agencies) was immense and it became apparent that some form of computerised
solution was necessary to control, rationalise and utilise the enormous amount of data involved
(10,000 Ion of water mains - 13000Ion sewers).
A further stimulus to the development of GIS systems was the perennial Water Industry
problem, shared by the other major utilities (Gas, Electricity and Telephone), of damaging each
others underground plant when digging holes in the highway. Computerised records were seen
as offering a means of reducing such damage through shared data and better communications.
A national working group, NJUG (National Joint Utilities Group) was formed in 1977 to
promote such a solution.
Thus the 1980's saw widespread investment in GIS systems, generally called at the time
"digital mapping" systems. Wessex Water acquired a Synercom system in 1985 following
several years of research and data collection and the Company now has over 15 years
experience in the use and value of such systems. The development of the system can be
divided into several di,stinct phases and these are examined in turn, reflecting on what the
original objectives w:ere with actual exp'!rience in each area; the lessons learned are described.
Wessex decided that Mapping was an inevitable long-term investment for managing
geographical information as early as 1978. The following potential benefits were identified:
• Rationalisation and verification of all existing records from the predecessor

organisations.
• Proper centralised control of all such information (Le. updating).

541
• Savings from improved handling and access of paper-based information for all
departments; avoidance of staff costs associated with manual record updating.
• Savings in design, planning and new works departments resulting from improved
information and rapid access.
• Greater understanding (and thus efficiency) of spatially-based information (frequency of

burst mains, for example).
• Potential for reduced plant damage through shared information with other utilities.
• Greatly improved Asset Management information resulting in improved investment

programmes.
• Wide variety of operational management improvements.
• Potential links to other systems (network analysis, customer databases, telemetry).
• Generally, progress has been made in all of these areas, but some of the perceived
benefits have not materialised as first thought.
Wessex started the process of computisation of records by establishing the key

workstations and methods of digitising records. Work started on water supply data in 1979
and, in 1991, around 95% of the distribution network has been digitised. 75% of the network
has been physically inspected to verify the data on the ground and was completed, updated and
verified by 1993. One early problem was that of background data. The UK National Mapping
Agency, the Ordnance Survey, had a long-term programme of digitising all background data
over the period 1980 - 2020, with the Wessex Water area scheduled for 2020. Wessex entered
into an agreement with the Ordnance Survey to bring this forward for the Wessex area and
currently 60% of background data is now digitised on the system. Progress in digitising
sewers has been slower and currently 20% of the network (of 13,000 km of sewers) has been
digitised with 4% verified on the ground. The situation is more complex in sewerage because
Wessex, in common with all the UK companies, has Agency agreements with Local
Government organisations who manage a number of the sewerage networks - this has
undoubtedly slowed down the development of GIS in sewerage. The remaining part of the
sewerage network is scheduled for completion by 1997.
542
The Company has 11 workstations around the Regional area, with 15 personnel
working full-time on GIS, (9 on Water Supply and 5 on Sewerage), plus an estimated 2
personnel on hardware/software support.
Tables 1. and 2. give a breakdown of the time spent on the major activities currently
ongoing in Water Supply and Sewerage.
Table 1. GIS activities - water supply
Water Supply GIS (9 Personnel)

Activity Time(%)
Routine updating of network 30
Ongoing amendments (new developments) 20
Map Production 12
Operational Support 4
Planning/New Works support 9
OS database maintenance/development 10
GIS development 15
Table 2. GIS activities - sewerage
Sewerage GIS (5 Personnel)

Activity Time(%)
Routine updating 5
Map Production 10
Development 50
Planning/New Works support 15
Contact supervisionladmin 20
Total annual running costs for the Regional system including depreciation of all
hardware and the acquisition of OS background is currently around £690K per year. This
includes external sales of OS background.
543
The experience gained in the early 1980's led to a number of realisations:
• Updating procedures needed careful management.
• GIS systems are specialised and this has consequences for training, development and
access.
• The original expectations of wide on-line access to geographical information did not
materialise.
• The sharing of data with other utilities did not generally materialise.
The major benefits during the period are now seen to be:
• GIS has imposed basic standard disciplines in validating asset data which can now be
used with confidence by all departments in the Company.
• The asset database has greatly improved the capacity for planning.
• The capability to produce bespoke maps of specific areas with selected information,
presented in colour, has proved very useful, in almost all areas ofthe Company.
A number of other factors affected the perception of GIS in the 1980's. The UK
Industry was privatised in 1989 and a key feature of this process was the price regulation
mechanism, which involved a factor (K), above the Retail Price Index by which prices to the
customer could be allowed to rise. The K factor depended crucially on an Asset Management
Plan (AMP) and GIS systems played a key role in the formulation of this plan.
The GIS effort on this exercise is now estimated at 1 man-year for Water Supply alone;
the impoitance of producing accurate information for the Asset Management Plan is reflected
in the fact that expenditure on water mains and sewers in Wessex over the next 5 years is likely
to approach £ 100 millions.
On a technical level the 1980's was the decade of the PC, and this explosion has created
a generation of "PC literate" users familiar with a variety of database and graphics packages.
Some of these are now overlapping with tasks originally envisaged for GIS systems and
generally there has been a need to converge GIS, CAD and general graphics packages in a
sensible way. There are now several PC-based mapping systems and GIS available.
544
In addition, other forms of "mapping" have appeared, notably video-disc based systems
which offer rapid access and simplicity of use for scanned records. As experience has been
gained, the potential applications of GIS systems have expanded.
It is now clear that many types of information are best understood when presented
geographically. In the Water Industry this applies to such things as pressure, or quality
variations across a network, burst mains, certain customer data (e.g. renal dialysis users) and
many more; the list of potential applications is enormous, limited only by the imagination of the
users. Unfortunately the basic cost-benefit equation for these applications is not always very
favourable and it is cost-effectiveness which will determine the future of GIS.
For those now considering GIS systems for the first time the choice of hardware,
software, and the multiplicity of potential applications is enormous, probably beyond the
understanding of anyone individual. At a conservative estimate there are now around 1000
companies offering GIS-related goods or services (Dangermond 1991). In this situation it is
worth examining the current situation in the UK Water Industry, the areas where GIS systems
have been successfully implemented and where they have not lived up to the original
expectations.
The current state of GIS system implementation can be summarised thus:
• Most of the larger PLC's already have some years experience with established GIS
systems, either on a pilot basis or full-scale implementation. A number of the smaller
Water Companies (former Statutory Water Cos) likewise now have GIS systems, though
some of the smaller companies have not embarked on such systems; others are evaluating
the options and costs. Wessex Water has a well-developed system, with over 15 years
experience in full-scale implementation.
• Of the original objectives identified for GIS systems, good progress has been made in
basic record m~agement (e.g. map production) and a sound foundation for future
development pas been provided. This has been particularly beneficial in Asset
Management exercises.
• The original expectations concerning the "paperless office", shared data with other
utilities, and widespread access for potential users have not been realised.
• The 1980's has seen an explosion of PC-based systems which overlap with conventional
GIS systems and a new generation of computer users familiar with graphics and database
concepts.
545
It is worth reflecting on why some of the expectations have not been realised. The
concept of shared data with other utilities has made little progress. The major motivation
behind the idea was to minimise the disruption caused by the major utilities damaging each
other's underground plant and the duplication of excavation in the "common highway", a
feature of urban life which attracts ridicule from all sides. In the UK there has long been a
requirement for any utility wishing to dig a hole in the "common highway" to notify all other
parties (highway authorities and utilities) by post of their intent. This has spawned a complex
paperwork bureaucracy, known in the UK by the acronym PUSWA (public Utilities Street
Works Act) involving the sending of over 15 millions messages to cover the estimated 3
millions such holes dug every year. It is proposed that a Computerised Street Works Register
(CSWR) will be created which will reference each hole at an accuracy of 20 millions initially,
and lmillion on reinstatement. The potential for GIS is clear. It is not clear currently how this
will interact with typical Water Company GIS systems but it may provide a new stimulus and
development path, though in recent years the development of trenchless technology and
improved methods of excavation has to some extent reduced the problem. Privatisation has
probably not aided the process of data sharing.
The reasons for the lack of progress in this area are probably twofold: managerial
(typically National Working Groups achieve little), and technical in that the lack of
standardisation among GIS manufacturers has made the process of data exchange much more
complicated than it should be. Wessex Water is currently involved with one electricity supplier
to instigate a simple exchange of GIS data as a trial using PC's networked over the public
telephone network.
A disappointing feature of current systems is their lack of accessibility to non-specialist
users. There are two still largely different uses and views of the GIS, one, that it is a modelling
tool for design and analysis and secondly, that it is a sophisticated digital map available for
ready everyday use. This second use of GIS, that it would be available for users to access
quickly, zoom into a particular area and manipulate data for a variety of everyday purposes has
not been realised. GIS is still a specialised area with access limited to relatively expensive
workstations. Computer users who have learnt to use PC's with common software packages
such as spreadsheets and wordprocessors are now used to a high standard of user interface
which current GIS systems simply do not provide. Most users have discovered that GIS
systems need considerable ongoing expenditure on development and data aquisition to open up
their potential for even relatively simple tasks. Even then the system may perform the task
once very efficiently but then lack the flexibility for further progress. It is quite common for
the end result of a development to be simply a paper copy of a map which is then stored for
everyday use in the field. The notion that geographical information would be stored, used and
transported electronically in the field is still some way off. Speed of access can still be
546
relatively slow (dependent on hardware and system loading) and in day-to-day use paper maps
(albeit updated through the GIS system) are still almost universal; indeed, one major task
carried out by Wessex' GIS department is that of producing updated paper maps for smaller
depots and/or field use.
It is interesting to reflect on this situation in the UK with that within the USA where
vast quantities of digital map data are readily accessible over the internet for anyone with a
telephone, modem and computer.
The fundamental problem is related to the way in which GIS data is stored and
manipulated. If the GIS is to be used only as a digital map then it is simply a picture which is
used quickly by staff to locate something in the real world in relation to something else (with
underground assets this is particularly important since much time and money can be wasted
searching or excavating to find the particular item). At this level it is not necessary to access a
database describing the detailed attributes of each item - doing so via a typical GIS system is
very time-consuming and is far less efficient than using a detailed paper map. Similarly, unless
the map is portable (in the sense that the data can be carried into the field) it is of little use at
this basic level. Most GIS systems do not satisfY these requirements of speed, portability, and
simplicity, and for this reason GIS information on computer has not, to date, found its way into
day-to-day operations in the way that was first envisaged and this has undoubtedly created
something of a negative image for GIS; the high running costs are not reflected by widespread
evidence of GIS in everyday use. It is better to design a flexible powerful digital mapping
system than carry the burden of the unused functions of the GIS. Wessex has approached some
of these problems as described in the following section
It is clear that the conventional GIS system of the 1980's as installed in Wessex, has
major strengths but a number of weaknesses. Its strengths are in the control and maintenance
of the central database, controlled updating mechanisms and Regional Standardisation of
records. The database provides very powerful tools for asset management. The major
weaknesses of the sYlltem are in access, speed of response and the user interface (in
comparison to PC sC?ftware standards). The cost of improving some of these shortcomings
using conventional GIS hardware is disproportionate to the benefits. Consequently, Wessex is
addressing some of the shortcomings through the development of WESNET, a PC-based
network modelling and management package which is making major advances in integrating a
variety of key data sources, one of which is spatial data. The fundamental concept of
WESNET is to integrate on a standard PC platform, the 3 key data sources for overall water
supply management: Network data, telemetry data and mapping data. These 3 key sources can
then be manipulated to provide true integration. The PC allows local processing, a high
standard of simple user interface, portability and the ever expanding possibilities of networking
using standard hardware which is continually reducing in price.
547
GIS information is made available through the production of CGM files (Computer
Graphics Metafile), the international standard for graphics information as an output option
from the Regional GIS system. Such files may be manipulated in WESNET (or such packages
as Freelance +), opening up the information to a wide variety of users for local information.
The recent arrival of high quality (VGA) colour graphics displays on portable PC's now means
that for the first time field staff can carry complete necessary GIS information in a van on site
to the location where there is a problem. Advances in networking mean that maps can be
downloaded via the public telephone network if necessary. This is one further step towards
making staff independent of reporting/administration centres thus reducing travelling and other
lost time.
These improvements in access and speed of processing have opened up many new
potential uses - it is possible to use GIS data in a fast moving Control Room environment as a
support to remote field staff, and in parallel to video disc based systems now commonly used in
Control Rooms for vehicle tracking.
The basic GIS remains as the means of updating and controlling records and access to
the database. A library of CGM format maps is available regionally over a PC network for
users to download as required. These maps can be used for a wide variety of purposes
according to the user but the crucial point is that for the first time GIS information is available
throughout the organisation via inexpensive PC hardware which is portable and easy to use.
As discussed in the previous section a number of trends have merged in the 1980's
which are likely to influence development of GIS in the 1990's. Two important factors in
particular are discussed in this section: systems integration and cost benefits.
System integration: This concept is now widely understood but has made little real
progress in the 1980's. GIS systems have long been recognised as potentially much more than
platforms for manipulating geographic information. It is obvious that a computerised map
might also be used to access and display data from a wide variety of sources; a few examples
are listed below:
• Telemetry data: Real time data on flow, level, pressure or quality could be displayed on a
GIS background, theoretically making it easier to relate such data to the real world,
rather than a simplified telemetry schematic.
• Modelling data: Results of network analysis simulations (both water supply and
sewerage) could be displayed at appropriate points on a GIS background as an aid to
understanding those results in relation to the real world.
548
• Customer related data: Customer contact data, on, for example, taste complaints, leaks
or burst mains, often held on mainframe systems, could be displayed on a GIS
background to aid understanding of these geographically related events. Some are
already held as an overlay on GIS systems and the results certainly illustrate the value for
such presentation methods. There are many other such customer-related examples, for
example billing, metering and water quality sampling zones.
While all these potential applications are attractive and can all be achieved with current
technology there are some obvious drawbacks with current hardware. Firstly, the required
links have to be made to the appropriate systems (e.g. telemetry) but more important,
considerable software development is required to make such integration possible in a practical
way for the user. Secondly, even if such data links are made, for such a system to be used
widely, access must be opened up to a much wider audience than at present and with current
hardware this is prohibitively expensive. It is quite common to see ideas or demonstrations of
applications for GIS which are very impressive technically but which miss the obvious point
that for everyday use the system must be accessible to the potential user. It is for this reason
that many of the potential applications are never developed for real-world use. In addition, the
real benefits have to be identified quite objectively and not in terms of doing innovative things
with new technology for its own sake. If we consider the potential for displaying telemetry
data on a GIS background, for example, the cost of a workstation for GIS access is many
times greater than that for telemetry access. The benefits of displaying such data in this way
are not so clear - it is sometimes much better to use a simple schematic to display such data
rapidly than to include much irrelevant map data; one skill which is at a premium in a modern
water company is the ability to simplifY and perceive important events in a mass of detail.
Wessex has approached the question of systems integration through the development of
WESNET; some practical examples are illustrated below. The underlying principle of
WESNET is to integrate water supply modelling, telemetry and mapping data on a PC
platform in a manner in which these data sources are complementary; there is no point in
integrating data unless the "whole" is greater than the sum of the parts. Data transfer is by
simple standard data files - no attempt is made to link telemetry hardware to GIS hardware or
modelling hardware of any particular type. The system is entirely PC-based, leading to a high
level of interactivity, access, portability and low cost thus opening up many new areas for truly
cost-effective operation.
Telemetry data is available for direct comparison with model simulation results, making
on-line calibration of network models very cost-effective - this is a genuine example of
integration actually leading to a new and more effective type of operational management style.
All of the major urban water supply networks in Wessex, around 15, are now modelled with a
549
complete integration of telemetry data such as water level in a reservoir, flow at a particular
point in the network and pressure at a node in the network. These are compared with the
results of a network analysis simulation.
With this constant supply of telemetry data the models can be updated and calibrated
with minimal resources over long periods leading to greatly improved understanding of the
network and cost savings which stem therefrom. WESNET can be used as a simple mapping
system in its own right (as described earlier) and also in combination with CGM files produced
by any other package. Within Wessex, Lotus Freelance (+) is used extensively for a wide
variety of general purpose graphics applications, and through the CGM file format any such
drawings can be integrated into WESNET. CGM files can be used to display results from a
network simulation in a variety of ways - thus results can be displayed on schematic or a GIS
background. For example, a CGM mapping file, with complete background, could be used to
display pressure and flow data from network analysis simulation; one disadvantage of this
approach is immediately apparent. The amount of detail in "complete" GIS sometimes makes
understanding more difficult than with a simplified schematic.
Cost Benefits: One major obstacle to the realisation for some of the potential
applications in GIS is that often cost-benefit equations are either difficult to compute, or very
unfavourable. There are two major reasons for this; the nature of the benefits and the high
costs of hardware and software with many current GIS systems.
It is generally true that the benefits of GIS are difficult to quantitY. The economies of
scale resulting from the Regionalisation of the UK Water Industry in 1974 enabled major
computer developments to be implemented, including GIS, but to estimate the financial
benefits requires information on the costs of record maintenance of the predecessor
organisations and estimates of wasted resources resulting from inaccurate or inaccessible
records. This type of information is notoriously difficult to track down. Often the benefits of
GIS are in hidden efficiencies through reorganisation - Wessex, for example, has progressively
rationalised and streamlined its management structure, yet has maintained access to the central
database of asset records throughout, irrespective of the location, of say, Planning and New
Works departments, and this has greatly improved the efficiency of those departments. Wessex
recently carried out a desk study to estimate the costs of providing a manual record-
maintenance system to replace GIS and estimated it to be around £0.5 million per annum; this
was simply to provide record maintenance, without any of the database facilities or specialised
map production capabilities now in use. The value of these features is very difficult to quantifY
but is highly significant. Prior to privatisation each Company had to produce an Asset
Management Plan upon which future investment levels would be based. Wessex' GIS system
not only enabled this to be carried out quickly, but also with a high degree of confidence,
essential for future business planning.
550
Under any criteria, GIS is expensive, both in initial investment (the immediately
apparent hardware and software costs are small compared to data, personel and training costs)
and running costs. Wessex' current system costs around £0.7 million per annum in running
costs. There is now a clear realisation that GIS is a long-term investment and that there must
be a commitment to make GIS work.
A major hindrance to GIS development which is now clear is the complete lack of
standardisation (other than items such as SDTS, NTF, DLG which are not universally adopted,
or ARCIINFO ungenerate format, which is appearing as a defacto satndard) and among the
major manufacturers; each pursues its own internal standards with a view to making customers
captives in the future. It is to be hoped that the trend elsewhere in the Computer Industry, of
Standards, and competition (and thereby reducing prices) will be mirrored in the GIS in the
future.
The major applications and uses which have evolved in Wessex' experience are broadly
as follows:
• Record updating, map production and asset records. These tasks are all carried out via
the GIS system and will continue. The support for Planning, New Works and Operations
for ad hoc applications will also be via this system.
• Access to GIS at depot level, portable digital maps in vans. This will be on a PC
(networked) base through the production of simple graphics files (e.g. CGM), and simple
user-friendly software; Wessex has developed WESNET to fulfil this task.
• Rapid access emergency site location GIS. This will be via PC-based optical disc
systems (e.g. AIMS).
• Data integration using GIS. This will be achieved at two levels: Via the standard GIS
database for customer-based information, and on PC for technical integration of
modelling, telemetry and mapping data using simple file exchange mechanisms and fully
utilising the portability, networking and graphics capabilities of the modem PC.
WESNET has developed these capabilities to a high degree.
551
4.3. OTHER EXAMPLES
Wessex water is not the only water company in the UK to implement computerized
assets management. Welsh Water (Dwr Cymru) and South Staffordshire Water Company have
also introduced similar schemes.
Dwr Cymru, Welsh Water, owns some 24,400 Ion of water mains and 16,200 Ion of
sewage pipes as well as 163 treatment works and 850 sewage treatment plants (Lindsay and
James 1995). In 1990 a 3 year deadline was set for the development of an information system
to manage all the data generated by these assets; the Asset Information System (AIS).
Lack of digitized Ordnance Survey (OS) maps presented a considerable problem, but in
agreement with OS Dwr Cymru digitized paper maps to OS standards, in all some 21260 large
scale maps. Lindsay and James report that "At the peak [of digitizing], around llO staff
worked in shifts 21 hours a day, to meet the two year deadline for data capture".
AIS supports Dwr Cymru's core operations. It provides maps for all water network
maintenance operations. It is important to note that operation procedure allows data within
AIS to be continually improved and updated by reentering amendments to the map made by
field workers back into the system.
AIS has a role as an executive information system (BIS). To rapidly provide managers
with the information necessary for business decisions. Lindsay and James (1995) report that
information provided by the AIS " was used to develop a plan for refurbishing water mains and
sewers throughout the companies operating area".
As well as supporting operations and business needs, the AIS enables useful analysis
that would not be practical with a manual system. For example Dwr Cymru have developed a
software package now used by other organizations, "Background Leakage and Burst Estimates
(BABE)" that uses pressure zone boundaries, OS digital elevation data together with address
data and zone pressure data from the AIS as part of a leakage control strategy. Lindsay and
James state that "AIS is. a key element in the company's [Dwr Cymru's] business strategy."
South Staffordshire Water are using a GIS in order to improve the effectiveness of their
pipe maintenance and replacement programme (Rogers, 1995). A Sysdeco GIS is used to
maintain an inventory of the pipe network and includes pipe specifications and installation
details. Laboratory analysis of pipe samples taken whenever a pipe is uncovered provides an
assessment of the risk of pipe failure for different pipe types, ages and under different
conditions such as soil type. Using this laboratory analysis in combination with the spatial
distribution of pipes; areas most prone to failure can be identified and treated before the
problem occurs. In this way South Staffordshire Water can adopt a proactive approach to pipe
network maintenance.
552
4.4. CONCLUSIONS
Because of the spatial nature of much of a water companies assets a GIS is an

appropriate response to the company's information needs. A GIS provides opportunities.
• To enhance the effectiveness of core operation such as pipe maintenance, for example
Dwr Cymru's generation of maps for each maintenance task or Wessex's links to
telemetry and hydraulic modeling.
• To support secondary activities such as statutory report generation or as in the case of

Dwr Cymru providing the information required for investment planning.
• To assist in extraordinary operations such as drought or contamination incidence

management.
A GIS also enables operations that utilizing a manual system would be impractical,
such as South Staffordshires approach to proactive maintenance of the mains network.
However there are risks, just as there are when introducing any corporate information system
to an organization, and the money involved is large. It appears that the large investment is in
providing a high quality data inventory and for the production of digital maps. The use of such
data in complex spatial analysis is yet to be fully realized.
4.5. REFERENCES
Dangermond (1991) The Commercial Setting of GIS Geographical Information Systems. Vol 1
- Longmans. PP 55-65
Lindsay P., James A. (1995) Running Water in Wales: GIS meets the Business Challenge.
Mapping Awareness, Sept. 1995 pp28-30.
Senn J.A. (1992) Analysis and Design ofInformation Systems. 4th Ed. McGraw-Hill Book CO
Singapore.
Rogers A. (1995) Mapping out the Battle Plan. Surveyor, Nov. ppI5-19.
WESNET User Manual (1991) Wessex Water Services Ltd.
WESNET Technical Reference Guide (1991) Wessex Water Services Ltd.
Chapter IX
Institutional Social and Economic Aspects

Chapter IX
Institutional Social and Economic Aspects
Very different are the topics connected with water supply systems, and those
concerning the institutional framework are normally neglected. Social and economic aspects
are relevant at various levels, and decision making procedures as well as technology transfer
and management methods should be considered. The success or failure of very big projects
very often depends on non-technological elements, and the co-operation of many specialists is
essental.
Even some fundamental concepts of water law principles as well as fundamental
historical regulations have to be highlighted, because water administration is an
interdisciplinary subject, which requires that engineers could usefully collaborate with lawyers,
whose task is to proVide institutional solutions without necessarily an understanding of the
technical issues.
In both developed and developing countries water supply management is rapidly
changing, so that goverriments have to solve the most crucial problem, which is to provide
good services at reasOnable prices. Some European countries' consumption data show how
water prices are strictly linked to specific consumption.
In order to make the decisions of new countries easier it seems useful to look at the
management options by briefly showing the solutions adopted in some countries, among which
England, France, Germany and Italy may have relevant structures.
The goal of the next years becomes the setting up of modern systems in which water
resources will be managed in terms of efficiency, environment and economy. This goal will be
hindered by municipal inertia especially present in many Eastern European countries, where a
special effort must be undertaken in order to find a balanced way to pass from uneconomic
management to public or private or mixed economic enterprises.
556
Another important task is to unify drinking water quality standards as well as the
quality of water services, both because of increasing water transfers among bordering
countries and the growing water world market.
European Nations will have to face a very big financial effort: joint ventures should be
set up and national laws should promote and encourage systems by which, on the one hand,
public interest and safety are protected and, on the other, private enterprise interests are
guaranteed as well.
Public or privately managed enterprises have to be encouraged in such a way as to
master many services, in the sense both of a drinking water only service, in large districts, and
of many water services, including more phases of the water cycle, starting from water supply
to sewage disposal and eflluent treatment.
The implementation of a modem legislative and administrative framework concerning
water services is often complicated both by existing laws on drinking water and other water
uses, and traditional, largely accepted practices. In other words, all water usages must be
harmonized, by taking into account all the needs and differences of the involved subjects. Any
changes will probably be resisted by those people, who have long benefited from customary
water rights, and by municipalities whose will to be independent often contrasts with an
efficient and economic management.
There is no pre-established form, since economic and social aims can be reached by
either direct public or private service.
Several elements are fundamental to correctly build and manage drinking water
services.
First of all, a wide water service organization has to be fostered, which comprises both
large scale territorial associations and all civil water cycle phases. It will be so easy to plan how
to exploit the resource and how to use it in different ways. Charging policy is crucial: external
funds should be avoided and at the same time investments promoted in order to modernize the
obsolete distribution network facilities, by replacing deteriorated pipes, installing water meters,
setting up remote controlled networks, building dual networks in the largest cities. Without
any specific evaluation, tariffs cannot be fixed, because many factors widely influence them.
The tendency is to work out normalized tariffs, in which water charges depend on the
following elements: water cost, if the resource is purchased from others, electricity cost, if
pumping facilities are involved, water losses, water resource strategy such as spring, river or
other, treatment methods, distribution layout, ratio between delivered volumes and
consumption network length, trend of recent investments. Balanced prices could also be
obtained by establishing tariff bands, with lower prices for the required domestic consumptions
and higher ones for exceeding consumptions.
557
Secondly, planning and control of water service performances are crucial, because it
is necessary to guarantee both the economic interests of managing companies in terms of
fullfilment of investment plans and the consumer requests in terms of service quality. The
quality of water services falls into the broad framework of quality of life. which is more and
more attracting the interest of the citizens. The so called levels of service consist both in the
management style of water companies and in the standard of the service provided to
customers. The discussion on quality of management is now accelerating due to the ongoing
water services privatisation. Particular attention has to be devoted towards efficient control
mechanisms, which guarantee good competition by assessing and comparing levels of service.
Water services should be processed as an enterprise, following the same principles adopted for
other market products. In order to get objective criteria some means are needed:
implementation of quality management instruments, such as regulation, standardisation,
metrology and certification; building of rules and rights which are mandatory for both
producers and consumers; creation of a quality system, which should cover all consumer
expectations. Even if the discussion on how to assess the performances of water suppliers is
very broad, some fundamental criteria can be assumed: it should be established that consumers
can have access to supplier documents in order to examine records of water distribution and so
on; companies should make the levels of service of their performance public, in order to
guarantee a minimum level in terms of service reliability. Good indicators for levels of service
are the following: quality of delivered water; water availabilty over a fixed minimum threshold;
pressure of mains water over a fixed reference level; number of customers subjected to water
restriction; number and duration of service interruptions due to failures; number of accidents
due to back flow from sewer system; response to billing queries. In any case it is necessary that
levels should be based on characteristics of g1obality, simplicity, independence of the type of
managing institution and objectivity, so that each customer can assess indicators which are
easily measured.
Fimilly, training of specialists has to be promoted at various levels in order to

coordinate the different subjects involved in water supply matters. At a national level
government and ministries often conflict with each other, causing contradictory regulations and
a lack of information. At the educational level the complexity of water supply systems requires
integration and coordination of traditional "curricula studiorum", which must comprise
programs of resources management, economy, environmental protection, sanitary engineering,
information processing, and so on. Urban water projects often fail because of the poor
coordination of the many specialized engineers (hydraulics, chemical, structural, electrical,
mechanical, automation) in the design phase and of the difficulty in understanding among the
same engineers and the operational team in the management phase. The role played by
558
professional associations should be strategic: they could promote thematic meetings, courses,
seminars to encourage innovation of knowledge and transfer of new technologies. Consulting
companies should adopt more reliable data and perform projects based on deeply studied
existing systems. Water supply companies are in a transition situation, since the change of
ownership has been experienced in many countries recently. Generally speaking a low
awareness of the importance of collecting data is common, and it is crucial to remove the
barrier placed by senior managers against the introduction of new technological tools.
1. Historical and Institutional Aspects
Paolo Veltri
87040, Montalto Uffugo (Cs)
ITALY
1.1. HISTORICAL ASPECTS
"Water is indispensable to human life, because it satisfies both pleasures and needs of
daily use" (Vitruvio, 1978).
The importance of water in all the human activities is unarguable. Starting from the
earliest human communities the more was water available the faster was the progress towards
civilisation. It is neither easy nor relevant to this Section to analyse the enormous historical
material dealing with the legislative aspects of water: it is however necessary to start from
socio-legal considerations in order to understand the complexity and unified nature of water
problems and to handle the principles of water administration.
First of all water has to be studied from a physical point of view, in terms of the
hydrological cycle, precipitation, surface and groundwater flow, flow control, collection and
transfers. When passing from these last ones to the possible water utilisations it becomes
difficult to distinguish between engineering, management and legal tasks: in other words water
becomes a highly interdisciplinary subject. In this field it becomes important to give engineers a
basic knowledge of the social and legal aspects, so that they could have a useful dialogue with
lawyers, whose task is to provide institutional solutions but who may not have a full
understanding of technical matters.
In order to link "natural context" and "human context" with each other water has to be
regarded both in terms of a single resource and as part of all the natural resources; moreover
from a political and geographical point of view it is very often not possible to separate the
origin and exploitation of water from one single country to another.
If we consider only freshwater as a single resource, the following potential usages must
be emphasized: water can be used for civil purposes, irrigation, hydroelectricity, industry,
NATO ASI Series. Partnership Sub-Series, 2. Environrnent - Vo1.IS

Edited by C. Maksimavic. F. Calamina. and J. Snoxell
560
transportation and recreational uses. Each of these usages might have been considered more
relevant than other ones in a specific country so that the historical context and, consequently,
laws and administrative regulations have reflected this situation.
On the other hand water uses could be in conflict with other natural resources, such as
land, atmosphere, subsoil. It becomes clear that some lawful social and economic interests
could favour or not some particular water usages and also other different resources. You may
think, for instance, of water misuse, both in quantitative and qualitative senses, in exploiting
caves and mines: the results may affect the downstream users, and so on.
Finally, national interests in the available resources might be in conflict with those of
bordering countries.
Starting from the earliest human settlements it appears that towns developed around
river valleys or in humid areas, so that it is common to classifY those ancient civilisations as
"fluvial or hydraulic civilisations": the history of Egyptians is the history of Nile river people,
that of Assyros and Babyloneses is that of Tigris and Euphrates river people, and so on for
Hindu, South Eastern Asian and Chinese people.
All these people had the need to regulate in some ways the use of water, both in
regions in which this was abundant, in order to prevent damage and control floods for
irrigation purposes, and in arid regions, in order to make better use of the precious resource.
The first water regulations reflected situations of single uses, that were implemented
just to solve conflicts or to penalize users responsible for offences. During the centuries the
argument has assumed more complex features, whereby each individual legal body rises from
the preceding one modified by custom, religion, policy as well as modem events.
The early Egyptian water regulations originated in a context in which the ownership of
water and land had a divine origin: so the Pharaoh or in tum the high social hierarchies had
control of the whole resource: but, in any case, they provided for controversies regulation and
collection of the revenues deriving from water taxes through a scattered system.
In Mesopotamiam Civilisations the dominion of water was not private and the use of
water by private o~ers was restricted by early regulations. In the Harnmurabi Code some
principles were clearly stated: these concerned the rights to use water as well as corresponding
sanctions against people who either did not maintain and repair the banks of irrigation channels
or were responsible for breaches and failures.
Also ancient Talmudic laws stated for Hebrew people that rivers and springs belonged
to every man.
Breaking off other equally important early regulations, it is necessary to focus our
attention on Roman water law principles. Almost all the existing legislative systems around the
world derive their bases from Roman law. Starting from the origins and passing through
Republican and Principate periods until the fall of the Westem Roman Empire an impressive
561
and evolving legislative body looked at water use. The sources can be found (Caponera, 1992)
in customs, statutes (leges), edicts (edicta), decisions of the Senate (Senatus Consulta),
imperial constitutions (edicta, decreta, rescripta, mandata), opinions of the jurists (responsa
prudentium). The earliest rules on drinking water supplies originated during the Roman
Republic.
In that very impressive legal matter some important principles can be summarized.
Regarding ownership, Roman law distinguished between common water, public water and
private water. Those resources which did not have any owner, either public, private, individual
or collective, belonged to the first category. Most water was common; only an eventual right
of use was allowed. Public institution water, which was also subjected to the right of use,
belonged to the second category. Finally private water, which usually was considered as a part
of the land and was restricted to a few cases, belonged to the third category. In contrast with
modem tendencies in the ecological field, it must be highlighted that Roman law did not
consider the potential negative effects of water use: in fact people who had the ownership as
well as the water rights on their land were free to use it without any consideration for
neighbours.
Another legislative body which influenced hundreds of millions of inhabitants in all the
Moslem countries is that deriving from Koran as well as Islamic tradition. The water is
considered as a property of all Moslems and a right explains in a clear manner the philosophy
of that law: the so-called thirst right statutes that everybody can take water to quench his
thirst. A specific water legislative system based on the Prophet Mohammed teachings is the
code Mejelle (1870), that was in force in all the countries of the Ottoman Empire. It is halfway
between the Koran and modem laws in Moslem countries, but excluding some which have only
recently re-established the original principles of Koran. For many aspects the Mejelle code
emphasizes even more the natural rights of water, in fact it extends the thirst right to animals.
According to water administration in Moslem countries, few cases of organised systems can be
found: in fact the administration is nowadays based on local customs, which are derived from
traditional rules as well as from colonial heritages, imposed by France, England, Italy etc.
The modem water legislative systems in Western industrialised countries result from
the differences created by the Napoleonic Code (1804). According to this Code we can in fact
distinguish between Civil Law Countries (France, Italy, Spain, Netherlands, etc.) and Common
Law Countries (England and former English colonies). The difference arises from water
classification: in civil law countries there are two water categories, that is private or public
waters; in common law countries the Roman principle of common water is maintained,
constituting the basis of the so-called riparian doctrine. Both these legislative principles deeply
influenced the incoming legislations of the colonies, even if the resulting codes took into
562
account local customs as well as religious and traditional codes. Just to give a remarkable
example, the water doctrine in the USA is not as severe as the original riparian one.
1.2. INSTITUTIONAL ASPECTS
According to the civil law the ownership of water can be either public or private, and in
particular water necessary for domestic purposes is classified as public; as a consequence the
rights to use public water have to be legally granted. These rights have a fixed duration and
have a corresponding price. Some restrictions exist also for private water which, according to
the recent trends in ecological matters, has not to be polluted or misused. As regards water
administration it is not possible to classifY the institutions which are in charge of the control
and management of the resource, this task belonging both to many national boards and local
institutions. Anyway, as evidenced again by Caponera (1992), the prevalent tendency is now to
regulate by rules, permits and other legal tools both public and private water. Moreover the
differences among procedures either in common law or in civil law countries are disappearing,
especially when we are dealing with problems of managing and distributing drinking water for
domestic purposes.
This is a relevant consequence of the always increasing importance of the ecological
issues, whose political evidence can be highlighted even by the replacing of legal terms: the
recent Italian "Legge Galli" on Water Resources states in fact that water is public, and also in
the French law the term private water has been changed into non-domainaI water. The
differences between civil law and common law countries are less and less evident, and in
particular the countries which belong to the EU share common regulations and directives. But
the reasons of an overall water policy arise from more and more frequent water transfers, so
that European coun~ries will have to make legislative and financial efforts in order both to
extend to countries out of the EU the standards, controls and regulations which are already
mandatory for the member states, and to get water supply systems comparable in terms of the
quality threshold of the delivered resource as well as efficient and economic distribution
services. Due to the events which have occurred in Eastern Europe after 1988 and because of
either lawful inertia and resistances or consolidated customs and rights, it will presumably not
be an easy task to build in a short time a real water market economy. On the other hand, it is
difficult to find the same features in all the EU countries and many different management
systems may co-exist within a single member state.
563
In order to better understand the effort which the Eastern European countries will be
dealing with, it must be considered that the water administration in those states were, and in
part still are, derived from socialist principles. Water was considered as a whole with land and
the State had on water either ownership or all the rights. If we look at the previous USSR
system in particular, even administrative aspects were to be seen in a framework of national
economy, with many agencies responsible for the service, both at national and regional levels.
The result is that many problems can now give rise to a social and economical context in which
the contemporary existence of central financial support and municipal monetary resources must
not hinder the development of private companies, among which the international ones are in a
profitable position.
1.3. REFERENCES
Caponera DA (1992) Principles of Water Law and Administration. Balkema, Rotterdam.

Vitruvio P. (1978) (ed.) Dell' architettura. Giardini Editori e Stampatori, Pisa.
2. Management Options
Paolo Veltri
ITALY
2.1. INTRODUCTION
The features of drinking water management systems extend as a logical consequence

from the general legislative principles. Into a general law framework many rules regulate the
water supply service. Very often the success of an economic and socially acceptable water
management system is based on the nature of the institutions which manage the resource and
then ensure, both at national and local levels, control and investments.
The first thing we have to emphasize is that almost everywhere and regardless of the
water legislation in force the uses of drinking water for domestic purposes have the top
priority over all the others.
The subjects we are dealing with are obviously those related to public drinking water
systems, which must be defined in terms of a minimum number of customers and which operate
for a minimum of days on each average year.
From a strictly economical point of view it appears that in many countries all around
the world the drinking water supply system has unbalanced. prices, which only partially take
into account market rules. Besides, referring to European countries, even in each state many
managing configurations are present. In all the countries, in fact, except for England, many and
different management systems co-exist.
The water prices are often affected by elements which are out of the market. We must
in fact recognize that some elements make the water system original. The drinking water
service can be seen as "a natural monopoly", in the sense that for an already managed area
(municipality, district, basin or others) there are no competitive elements among different
contractors, and moreover no comparisons are possible among distant water management
systems. In other words the situation is contradictory: if a management company holds water

566
rights, then it needs to get them for long periods of time, in order to program investments and
forecast credits; but in this case the plurality of many contractors is missing.
If, on the other hand, the rights last for a short period of time, then the contractor can
neither program investments nor forecast anything.
It must be highlighted that each water supply system is a prototype and the exchanges
between contiguous systems are at present not very common.
For these reasons many existing management systems do not have purely market-
regulated water prices: other social, political and cultural factors playa role. Some economists
talk of a "fictitious market", in order to justifY the rules adopted to determine water charges.
By means of a fictitious market it is possible to put together rules and procedures
which allow the comparison of the performances of different and distant water supply systems.
2.2. FEATURES OF DRINKING WATER MANAGEMENT SYSTEM
Referring to the European context and in particular to the EU member states

(EUREAU, 1993, Rillaerts, 1994, Zabel and others, 1994), they show a great diversity of
management systems, even within the same country. It is a very hard task to classifY the
different systems according to a few features.
In fact we should detail forms, procedures, structures, powers and models of both
central and local governments, etc. As for the forms, either public or private or mixed ones
exist; as regards procedures, they depend on adopted rules; structures can be created both to
manage water supply only and to administer also all the phases of water civil uses, including
the drinking water services and those concerning sewage disposal and wastewater treatment.
There are many cases in which also electricity and gas distribution services are run by
the same company..
Anyway, even if differences can be found still in the terminology of operating entities, a
broad classification of the possible management models can be established by reference to the
most relevant features of the service.
a) Except for England, a public body, which can be either the State (this is the case of
Northern Ireland) or a regionaVdistrictimunicipai body, is responsible for the water
provision service.
567
b) The management entity, which is responsible for the distribution service to the end-
users, can be either the same or a distinct body from the responsible one.
c) The organization and the form of the distribution systems fall (BUREAU, 1993) into
one of the following categories:
• public direct management,
• direct management by inter-municipaVregional structures,
• delegated public management,
• delegated management with mixed or private capital,
• private management.
In the case in which responsible and managing bodies are distinct, different
circumstances can occur in terms both of ownership of facilities and equipment and
investment financing. There may be, in fact, situations in which facilities are owned
either by the responsible body or by the management body, and investment is either
allocated by the management body itself or supported by the responsible body by means
of subsidies as well as external financing. In the following Tables 1. and 2. (BUREAU,
1993) the major management types of drinking water services and statistics of the same
types in the EU member states are reported. In order to better understand the meaning
of the data summarized in Tables 1. and 2. more information about some existing water
management systems in some selected European countries will be given in Section 4. of
this Chapter.
d) The water prices, recognising that the drinking water supply is a public utility service,
can be determined (Ieno, 1993) according to the following methods:
• method based on a cost recovery system, that is water prices must cover the total
service costs,
• method based on water prices initially fixed at the risk of the management body
(established in a contract and eventually subjected to an automatic revisional
formula),
• method based on the control of the rate of return on invested capital.

Table 1. Major management types of drinking water seIVices in EU member states (EVREAU, 1993)
Management Types Belgique Danmark France Deutschland Ireland ltalia Grand Ducha Nederland Portugal Espana United Kingdom
de Luxembourg England Northem Scotland
and Wales Ireland
1. Direct Public Management
Direct Management x x x x x x x x x x x
Autonomous Board x x x x x x
2. Direct Supramunicipal Management

Direct Management x x x x x x
Autonomous Board x x x x
(]1
Ol
(X)
3. Delegated Public Management
Cooperatives Companies x
Public trading Companies x x x x x x
4. Delegated Private or Mixed Companies

Private Companies x x x x x
Mixed Capital Companies x x x x
5. Direct Private Management

Private Companies x
Table 2. Statistics of management types of drinking water services in EU member states (EVREAU, 1993)
Direct
Direct Public StJpraIoonicipai Delegated Public Delegated Private a Direct Private a
Managemert ~ Managemert Mixed MMagement Mixed
~
Bel9QJe 5 00 5
DiI1ma1< 0/ 33
iJelisctjCl1d* 35 20 30 15
Esr:ma 48 11 12 2.9
France 23 2 75
Ganci CXIc:he de lLD<emboorg 100
Ireland 100
Itaia 72 23 1 4
Nederland 15 85
c.n
(j)
Portuga 92 8
<0
Uritedf<j~ 3 9 88
/w~ 'Mligued I¥ PopUation 37 11 15.5 20.5 16
lkIited Kingdom
ErgIlnllnl Wales: 100",4, Ored lJivate ~

ScotIlnl: 100",4, Ragonal dred ~
Northern Irellnl: 100",4, 0 red pJllic rrmagemerd:
*Rerrark: The ~ dstrilx.dion corresponds to the pop.llaion of the foorer Federal Rep.blic.
570
Morever, the prices can be either balanced, if they cover the total costs in the balance,
or unbalanced, that is water charges do not cover the total cost of the service, the
remaining cost being financed by either municipal budget or external subsidies.
The cost recovery system is adopted in Germany and Italy, while the well known
"price-cap method", adopted in England and France with some variations of the
original features, belongs to the method based on initially set prices. The third method
is not very common in European water supply services, whereas it is sometimes
adopted in the USA.
As for the Cost Recovery method, the managing entity depends in any case on
Government, which usually fixes both the cost analysis methodology and the tariff
structure. The costs include personnel, maitenance and operational costs, technical and
financial depreciation, passive interest on investment, reward of the invested capital (if
the management is carried out by private companies). The prices can be then
determined by referring to either the annual balance or the many-years balance, and to
either provisional or consolidated balances. The Cost Recovery method could fail if the
Government control establishes restrictive clauses in determining charges, in which
case the service quality and the water system improvements could suffer from negative
effects.
As for the Price-Cap method it is largely used in England in all public utility services
with a monopoly regime. The method is usually applied by a year by year up-dating of
the initially set water price Po by means of a simple formula:
PO+1 = Po ( 1 + i + x)
in which: PO+1 is the up-dated price,

is an inflation coefficient, being equal in England to the retail
prices index,
x is the productivity coefficient, being positive if the service, as in
the case of water supplies, is characterized by a growth of the
service quality with productivity loss (Meucci, 1993).
The initial price Po can be determined by taking separately into account operative
expenditure, capital maintenance expenditure and return of capital.
Water companies in England assume x = -1. 0 -+ 3.5 %, with the compliance of the
Office of Water Service (OFWAT), which has to set the water price limits. A contract
between company and customers fixes service quality standards, expected goals and
penalties in the case of non-compliance. The Price-Cap method gives many advantages:
571
the users understand very easily how much the service prices are and they are moreover
guaranteed against any shortcomings in the management company, the public authority
control and procedures are very simple, the surplus remains part of the enterprise
economy.
Some improvements in the Price-Cap method have been introduced in France, by
means of particular arrangements in the contracts between companies and public
authority.
e) The management control policies usually concern price control and management,
service quality and health standards conformity controls. The control entities can be
different and belong to the central Governments as well as local authorities.
2.3. REFERENCES
EUREAU (1993) Management Systems of Drinking Water Production and Distribution

Services in the EC Member States in 1992. EUREAU editor, Brussels.
leno E. (1993) Le tariffe qualita alia luce delle esperienze europee. Atti delle giomate di studio
sui tema La misura nella gestione delle infi"astrutture idrauliche. CUEN, Napoli.
Meucci F. (1993) Tariffe e price-cap. H20biettivo '90, Federgasacqua - Cispel Emilia
Romagna, Bologna, 29-31 marzo.
Rillaerts F. (1994) Structure and Management of the Drinking Water Service. Hydrotop 94,
12-15 avril, Marseille.
Zabel T.F., Rees Y., Correia F.N., Neves E.B. (1994) Institutional Arrangements for Water
Management in Selected EC Countries. Hydrotop 94, 12-15 avril, Marseille.
3. Quality Management in Water Supply
Mario J. N. Neves
Universidade do Porto
Rua Jeronimo Mendon~a 65
4200 Porto Codex
PORTUGAL
3.1. INTRODUCTION
In recent years, a growing concern about Quality has been evident at all levels. On one
hand, this is connected to the quality culture that is being developed among the public, which
determines a greater pressure with respect to suppliers. On the other hand the suppliers
themselves are becoming aware that quality is a decisive factor for success in highly
competitive societies, and that therefore they must develop their own mechanisms to assure the
quality of their products or services.
Following concerns with the quality of products, preoccupations are now also being
directed towards the quality of life. Environment is one of those manifestations, and the
interest of the public in environmental impact assessments is clear. The quality of services is
another topic that is attracting the interest of society, as is the case in health systems,
education, transport, etc. In some cases, comprehensive evaluation systems are becoming
common practice, for example for universities, quality assessment is now widely used.
Quality assurance in water supply systems, the discussion of which started in the late
eighties, is described by Neves (1994). Manifestations of this concern can be seen in the
management style of water companies, for example, and also in the quality standards of the
service rendered to the customers, the so called levels of service.
Privatisation of water services is another factor that is accelerating the discussion of
levels of service. Privatisation has been implemented in some countries, and it is being
considered in many others.
On one hand, authorities that have signed or are planning to sign contracts with private
companies, wish to be sure that the consumers are in "good hands". The enormous social
importance of water supply justifies a particular attention towards an efficient control system.
NATO AS] Series. Partnership Sub-Series, 2. Environment - Vol.t5

Edited by C. Maksimovit, F. Calomino, and 1. Snoxell
574
On the other hand, competition between companies is stimulated by the levels of

service evaluation. Besides the intrinsic reasons, the designation of the "best practice" company
is, of course, very important in marketing terms.
Also financial reasons should be highlighted. Good performance can be seen as an
argument for increasing water prices. On the contrary, low quality services may invoke
penalties, as occurs in England, for instance.
The feeling that international legislation in this area will be prepared soon is also a
reason to accelerate the discussion concerning levels of service. As in other cases, for example
both drinking and bathing water directives, it is natural to expect that the European Union
(EU) will introduce some rules concerning levels of service in water supply systems. Many
people, at various levels, wish to be prepared for the construction, discussion and
implementation of the legal framework. International authorities, national authorities, private
companies, universities and other institutions, will certainly be involved in that process.
3.2. INSTRUMENTS OF QUALITY MANAGEMENT
The European Standard EN 29 000 (1987) defines quality management as an aspect of

the general function of management which determines and implements the quality policy. That
responsibility belongs to the managers of the organisation, but all its members must be involved
in the process of quality attainment. Quality management includes strategic planning, resource
allocation and other systematic activities related to quality, such as operational activities and
assessments.
This methodology, firstly used in high technology domains such as nuclear and special
engineering, was adopted by industry and is now being extended to many areas, including
water services.
For instance, the Working Group on the Operation and Maintenance of Water Supply
and Sewerage Systems created by the World Health Organisation, recommends (Hueb, 1993)
that water services should be processed as an enterprise, following the same principles adopted
for other products available in the market. The same author reminds us that the functions
usually carried out in these institutions can be grouped in organisational systems, whose
proposed configuration is shown in Table 1. He emphasizes that permanent assessment by
means of management information systems is crucial for planning and correction of quality
deviations and the use of indicators for operational and management performance is
fundamental for that purpose.
575
Quality must be present in all phases and components of water supply systems,
including planning, design, construction, rehabilitation and exploitation. Some means to
achieve that objective include (Matos, 1993):
(i) Implementation of quality management instmments, such as regulation,

standardisation, metrology and certification.
Regulation is the production of documents with obligatory character, which set the
essential requirements of products and services. In the European Union they can be
community directives, which have to be transposed to the Member States law, and the
"technical regulations", which are specific to each Member State.
Standardisation is the production of mandatory reference documents, which describe
methodologies to render a service or the characteristics of a product, defining tests,
control proceedings and criteria for acceptance or rejection.
Metrology is an activity concerning the quality assurance of apparatus and measuring
equipment, involving the proceedings for their calibration and reliability.
Certification is the last stage of the quality cycle, involving inspection, quality control and
auditing, activities which presuppose the existence of reference documents and the
systematic use of measurements.
(ii) Definition of attributes and responsibilities of the principal agents, which are the
producers (products and services enterprises), the consumers (citizens at individual or
collective level, or their representatives) and the State, who must enforce the established
rules and preserve the rights and obligations of the parts involved.
(iii) Creation of a quality system, understood as the set of the organisational structure,
responsibilities, proceedings and resources to implement quality management (EN 29
000).
Since 1987, ISO, the International Standard Organisation, has published a set of
standards related to quality management, the famous 9 000 series. These standards have been
adopted by the European Committee for Standardisation under the form of European
Standards, the EN 29 000 series.
EN 29 000 (1987) "Quality Management and Quality Assurance Standards" defines
some concepts and indicates the guidelines for selection and use of International Standards. A
distinction is made between standards for the external quality assurance (EN 29 001, EN 29
002, EN 29 003 ) and standards for the internal quality management (EN 29 004).
576
Table 1. Organizational systems (Hueb, 1993)
INFORMATION SYSTEM
HUMAN RESOURCES ADMINISTRATION AND

~-------~~~;Ie'I\'"i"$'I$i'e'oG,
BILLING AND COLLECTING
COMMERCiAl CONSUMPTION MEASUREMENTS
SYSTEM CONSUMERS RECORD
COMMERCIALISATION
PROJECTS MANAGEMENT AND CONSTRUCTION

OPERATIONAL
OPERATION
SYSTEM
MAINTENANCE
PHYSICAL PLANNING
ECONOMIC AND FINANCIAL PLANNING

PLANNING
ORGANISATIONAl PLANNING
SYSTEM PROGRAMMING
CONTROL
SUPPLIES ADMINISTRATION
LOGISTIC
PATRIMONY ADMINISTRATION
ADMINISTRATIVE
TRANSPORTS ADMINISTRATION
SYSTEMS
SOCiAl COMMUNICATION
FINANCiAl FINANCiAl ADMINISTRATION

~-----------------------
SYSTEM ACCOUNTANCY
HUMAN RESOURCES ADMINISTRATION AND

DEVELOPMENT SYSTEM
INFORMATION SYSTEM
Emphasis is given to the requirement that quality systems should be documented and
demonstrable. Documentation may include manuals, proceedings, audit reports and other
records. Demonstration means the capability of proving the existence of an adequate quality
system, and the aptitude of obtaining products or services in conformity with the specified
requirements.
577
The titles of another group of standards are self-explanatory:
• EN 29 001 (1987), "Quality Systems. Model for quality assurance in

design/development, production, installation and servicing",
• EN 29 002 (1987), "Quality Systems. Model for quality assurance in production and
installation" ,
• EN 29003 (1987), "Quality Systems. Model for quality assurance in final inspection
and test".
Finally, EN 29 004 (1987), entitled "Quality management and quality system elements",
provides guidelines for the conception and implementation of quality management systems.
The role of quality manuals and periodic assessments is strongly emphasised and attention is
also devoted to economic aspects, marketing, products conception and specification,
provisioning, production control, tests, corrective measures, handling and post-production
activities, documentation and quality records, staff, security, responsibility and use of statistic
methods.
Also relevant are EN 30 011-1, EN 30 011-2 and EN 30 011-3 (1993), transposed
from ISO 10011 (1990), concerning Quality Systems Auditing. Guidelines for the preparation
of Quality Manuals can also be found in standardisation documents (e.g. NP-2732, 1986).
With respect to technical aspects related to water and sewerage services, the relevant
EU Directives are the following, divided by thematic areas (Matos, 1993):
i) Consumption water
8017771EEC Exploitation of natural mineral waters
8017781EEC Quality of water for human consumption
ii) Water for several uses

7Si440lEEC Quality of surface waters for the production of drinking water
761160lEEC Quality of bathing waters
78i6S9IEEC Quality offresh water for fish life
79i869IEEC Analysis methods of surface waters for the production of drinking water
79i923IEEC Quality of conchiferous waters
578
iii) Discharge of dangerous substances

76/464IEEC Directive on dangerous substances discharged into the water and
specific
Directives 82186IEEC, 83/513IEEC, 84/1561EEC, 84/4911EEC,
86/2801EEC, 88/347IEEC
80/68IEEC Pollution protection of underground waters
911676IEEC Protection of surface waters against nitrate pollution
iv) Industrial control

781176IEEC Residues from the titanium dioxide industry, altered by 83/29IEEC
82/883IEEC Proceedings for verification of residues from the titanium dioxide
industry
87/217IEEC Asbestos pollution prevention
v) Products control
73/404IEEC Detergents, altered by 82/2421EEC and 83/94IEEC
73/405IEEC Biodegradability of anionic surfactants
vi) Waste water and sludge

86/278IEEC Agricultural use of sludge from waste water treatment
91/271IEEC Municipal waste water treatment
vii) Information excha~ge

7717951EEC . Information exchange concerning fresh water quality, altered by
861189IEEC
831189IEEC Information concerning technical regulation
viii) Public market of works and products

71/3051EEC Public works contracts, altered by 89/4401EEC
77/621EEC Contracts with suppliers, altered by 88/295IEEC
90/531IEEC Proceedings for products acquisition
579
ix) Other Directives

85/337IEEC Environmental impact assessment of certain public and private projects
8513741EEC Responsibility for damages caused by defective products
8813201EEC Laboratory good practice
871216IEEC Serious natural accidents, altered by 82/501IEEC
89/1 06IEEC Construction products
x) Proposed Directives
COM (84)303 Titanium dioxide pollution reduction
COM (85)373 Waste discharges into the sea
COM (86)733 Objectives of water quality concerning chromium
COM (89)162 Security of products.
COM (89)303 Environmental European Agency
COM (89)478 Alterations of 75/4401EEC, 76/160IEEC, 79/869IEEC,
8017781EEC
COM (90)297 Application ofEC rules to the proceedings for products
acquisition
COM (91)145 Use ofEC conformity mark
COM (91)347 Proceedings for services acquisition.
3.3. ASSESSMENT OF WATER SUPPLIERS PERFORMANCE
3.3.1. Existing Systems
The preceding paragraphs were devoted to some instruments for quality management,
with particular emphasis on water services. Now, let us centre the discussion on the criteria for
the evaluation of the quality of their product, that is, the quality of the water supply,
considering the aspects usually seen as relevant by the customers.
The English system to evaluate the performance of water suppliers is very
comprehensive and one of the first to be implemented (early nineties). This, and the fact that
there is now some experience in its use, explains why it is so often quoted. Let us follow Miller
(1993) and Coelho (1993) for a brief summary.
580
Privatisation of the water industry in England and Wales during 1989 led to the
creation of 10 holdings, the Water Holding Companies, each one including a Water and
Sewerage Company (WaSC); besides these, 23 more companies were appointed for water
supply, the so called Water Only Companies (WoC). These Appointments run for a
minimum period of 25 years.
A clear distinction was introduced between the role of service supplier and those of
customer protection and quality regulation, which were attributed to:
• The Drinking Water Inspectorate (DWI) - Drinking Water Quality,
• The National Rivers Authority (NRA) - Environmental Regulation,
• The Office of Water Services (OFWAT) - Economic Regulation and Customer

Protection; OFWAT is also in charge of facilitating competition and promoting
efficiency.
One of the most innovative aspects of this structure is the importance given to the
consumer, and for his protection the 1989 Water Act establishes that:
• Consumers must have access to documents where their rights are presented in a
clear and simple way,
• Consumers have the right to examine the records of water distribution,
• Companies must make public their performance, that is, their behaviour in terms of
minimum levels of service legally established,
.; A Guaranteed Standards Scheme assures to the consumer a compensation of 5

pounds per day or per event, in the case that the minimum levels of service are not
achieved.
Consumers are directly represented by the Customer Service Committees, created by

OFWAT for each WaSC operating area. The Director General of Water Services, head of
OFWAT and responsible in the last analysis for the defense of the consumers' rights, has an
obligation to act when levels of service stay below the legal limits.
The Appointment stipulates that companies must provide an annual report to the
Director General, where, among other information, they must compare their performance with
581
certain levels of reference and forecast their future evolution. Designated by the Director
General (DG), the following indicators for levels of service have been set out in the 1991
version:
• DGI -Raw Water Availability

Indicates the population supplied by a company from water sources where the
reliable output in abnormally prolonged dry conditions is less than the peak demand
of the customers and there is a risk of water shortage.
• 002 - Pressure ofMains Water

Monitors the number of properties assessed as being at risk of experiencing a water
pressure below the reference level of service for some period during the year.
• DG3 - Interruptions to Supply

Monitors the number of properties that have experienced an interruption to the
water supply for more than 12 hours without prior notice.
• DG4 - Hosepipe restrictions

Shows the population actually subject to water restrictions in a reporting year.
• DG5 - Flooding Incidents from Sewers

Identifies the number of buildings at risk of flooding from sewage more than twice in
ten years because of inadequate sewers.
• DG6 - Response to Billing Queries

A measure of the speed with which a company responds to both written and
telephone billing queries.
• DG7 - Replies to Written Complaints

Deals with responses to written complaints.
582
Due to its subjectivity, DGS was retired. Meanwhile, further levels of service have been
proposed, mainly (:Onnected with quality and environmental improvement, rather than
customer service:
• DG9, drinking water quality compliance measured at the customers tap,
• DGlO (A), the volume of water entering distribution from treatment works affected
by Undertaking and Relaxations,
• DGIO (B), the number of properties affected by section 19 Undertakings for the
distribution system,
• DG II, the number of sewage treatment works whose performance is unsatisfactory,
• DGl2, the quantity of sludge disposed of in an unsatisfactory manner.
Information also requires to be supplied on combined sewer overflows, sea outfalls and
leakage control.
Miller (1993) emphasises that the Director General may use these cross company
comparisons to stimulate efficiency improvements in those companies exhibiting low levels of
performance when compared to best practice elsewhere. The importance to a company of
achieving a "best practice" designation and the influence this may have upon their subsequent
riegotiations with the Director General with respect to charging policy, is therefore
considerable.
3.3.2. Possible Criteri,a for Measuring Levels of Service
The system described is an example of something that already exists but, of course, it
could not be totally transferred to other countries, even if it could be considered as an excellent
model. It is not important to discuss now if a universal system is possible. However, a set of
desirable characteristics may be suggested (Neves, 1995), even if the system is confined to
internal use (Alegre 1993):
(i) Globality
The rules must cover all the main aspects that may affect the users.
583
(ii) Simplicity
Restrict the rules to those aspects that really have practical importance.
(iii) IIJdependence
The system must be independent of the kind ofinstitution, public, private, or mixed.
(iv) Objectivity
The system must assure only one interpretation of the results. Therefore, it should be
based on indicators that can be objectively measured.
A complete formulation of criteria can be seen in Alegre (1993). Here, only the main
aspects will be described, as well as some application examples. The analysis must be based on
periods of one year, and the results of each indicator classified in four levels:
- Level 4: very good service with respect to that indicator,
- Level 3: good service, but some improvements must be planned,
- Level 2: the service has some deficiencies which must be rapidly solved,
- Level 1: the service has too many deficiencies.
With respect to indicators, one group is proposed for technical aspects (e.g. availability
of water, quality of the water, pressure in the network) and another group for management
style (e.g. response to complaints). The total number of indicators is 20.
Evaluation criteria
Two examples will be shown in order to illustrate how the quality of the service can be
objectively measured.
584
Example 1
Indicator: Consumption restrictions.
Definition: Number of customers that, in the period of one year, were subject to
consumption restrictions, divided by the product (number of customers x 365).
Classification: - Level 4: less that 5%
- Level 3: from 5% to 10%
- Level 2: from 10% to 15%
- Levell: more than 15%.
Example 2
Indicator: Number offailures in the network.
Definition: Number of reported failures in the network, divided by its length.
Classification: - Level 4: less that lO/km
- Level 3: from 10lkm to 50/km
- Level 2: from 50/km to 150/km
- Levell: more than 150/km.
As a final observation, the central idea is to fix values, so that anyone can understand
what is a high standard of service, and what is a low service standard. This is important at
national level, but also at the international level, in order that citizens and suppliers become
objectively informed about the interpretation given to quality standards in other regions. This
may even constitute a stimulus to improve service towards the best practice.
3.3.3. Companies Policy on Levels Service
The· importance of the levels of service assessment in terms of public image and even
under the financial point of view has already been emphasised. Therefore, it is natural to expect
that some companies will develop their own evaluation systems, as happened in England,
where more exhaustive systems have been implemented in some cases, for self-defense. An
inquiry carried out by Coelho (1993) has shown that companies had a tendency to consider DG
levels as minimum standards, below which the performances should be considered worrying.
The responses have shown that the most relevant points for the companies were
the continuity of the supply, followed by the pressure in the network and the quality of the
water, probably the more sensitive for the consumers. For monitoring and data acquisition,
model simulation combined with occasional measurement campaigns and/or district metering
(macro flow measurements in zones of average size, usually for leakage control) and tele-
585
metering, were the most used information sources, complemented by the consumers
complaints.
Figure 1. Customer care strategy
Computerised systems to treat the complaints are very popular, and sophisticated
solutions have been implemented in some cases, based on the consumers perception about
some variables connected to the level of the service, such as the pressure, interruptions to
supply, taste and odour of the water.
Customer Care Strategy (CCS) is a very sensitive point for the companies. Figure 1.
shows the so-called "Seven S" as an example adopted by one of them (Miller 1993).
Let us follow the explanations of the company:
Standards
Besides external standards, internal ones must be refined to support internal customer
relationships. The skill courses reinforce personal standards.
Skills
From time to time staff receive 2 to 3 days training to develop face to face, telephone and
letter writing skills.
586
Systems
Implementation of integrated computerised systems eIiminates less efficient non-integrated
computer and manual systems. This leads to fewer delays, mistakes and reworking of
information.
Reward systems promote the linking of individual performance and achievement of personal
objectives to corporate objectives.
Structure
A customer centre providing access to all services via a single telephone number for all
customers is important. Customers will speak immediately to customer service staff who have
both the technology and the authority to resolve many queries immediately.
Style
Corporate image expressed through its logo, letterheads, uniforms, vehicles and
accommodation is also relevant.
Surveys
Customer surveys should be repeated at intervals to register the impact of changes.
Staff
An open management style is recommended to improve two way communication and
encourage initiative and motivation at all levels. Introduction of team briefing and an employee
suggestion schemes are two examples..
Finally, the company recognises that new methods of reinforcing customer care services have
to be continually found out.
3.4. REFERENCES
Alegre H. (1993) Instrumentos de Apoio it Gestio Tecnica de Sistemas de Distribui~io de

Agua. Teses e Programas de Investiga~io, Laborat6rio Nacional de Engenharia Civil
(LNEC), Lisbon, Portugal, ISBN 972-49-1608-1.
Coelho S. (1993) Avalia~io de Niveis de Servi~o: Introdu~io ao Modelo Britanico. LNEC
Seminar, Lisbon, Portugal.
Hueb J. (1993) Aspectos Operacionais Ligados com a Gestio dos Servi~os. LNEC Seminar,
Lisbon, Portugal.
587
Matos R. (1993) Regulamenta~o, Nonnalizayio e Certifica~o do Sector da Industria da

Agua no Quadro do Mercado Europeu. LNEC Seminar, Lisbon, Portugal.
Miller A. (1993) Levels of Services and Customer Care Strategy in a Major UK Water Service
Company. LNEC Seminar, Lisbon, Portugal.
Neves M.J. (1994) Levels of Service in Water Supply Systems. Proceedings of the NATO
Advanced Study Institute New Technologies for Large Water Supply Systems, A1bena,
Bulgaria.
Neves A. (1995) Niveis de Servi~ em Saneamento Bilsico. Relat6rio Intemo da Comissao de
Coordena~ao da Regiao do Norte, Porto, Portugal.
4. Examples of Existing Management Systems in
Europe
Paolo Veltri
ITALY
In most cases the existing services appear to be complex models, because more than
one form of water distribution service is present at the same time. The current arrangements
have been influenced by many cultural, economical, political, historical, social factors, so that
no system could be easily exported to other situations, but it can represent a useful tool of
comparison, in order to bring on new water management systems in the countries which now
adopt obsolete and uneconomic systems.
4.1,FRANCE
In France water resource management has been divided since 1964 into Six River Basin
Agencies, recently renamed Agences d'Eau, based on river catchment areas. The tasks of the
Agencies are now related· to many aspects of water management, including water pollution
control, within an administrative framework in which final responsibilities are the concern of
the Prefect of the Department. The Agencies collect two kinds of charges, for drinking and
polluted waters, which are then partially transferred to communes in order to construct and
upgrade treatment facilities. Each Agency is formed by a parliament. In 1992 a new law
decreed the creation of the Local Water Commission, whose tasks are to plan water
management and collect information on the ongoing water politics. The State is always
present, together with other partners, namely local municipalities or private enterprises. The
body responsible for water provision service is the commune, which can choose the
management system best in keeping with municipal needs. Communes themselves (through
"regies"), intermunicipal water associations, inter-municipal multi-purpose associations directly

Edited by C. Maksimovic, F. Calomino, and 1. Snoxell
590
manage about 30010 of communes mainly in rural areas. Mixed or private companies manage by
delegation the remaining 70010 in a sort of concession or leasing contract. The same body is
often charged with the water supply as well as sewage disposal services.
In the case of public or semi-public managements the investments are financed through
subsidies. Water charges have to take into account both operating and investment costs, they
are fixed by the commune (or intermunicipal association), which may also provide for extra-
subsidies when prices are not balanced. Many institutional subjects, such as central government
and regional administrations, control legal, sanitary and financial management.
Also in the case of private company management, the commune is always responsible
for the service; it also owns facilities and decides on the investment program on the proposals
of the contractor. The commune and the company freely contract duration and water charges,
including an automatic price revision formula as well as the possibility to renegotiate prices if
relevant changes take place. These control mechanisms also regulate private management.
There is a very great number of communes in France, and drinking water prices vary
greatly, ranging from 0.42 FF/m3 to 10.92 FF/m3•
4.2. ENGLAND
Drinking water and sewage disposal services were centrally organized in England by a
1974 law, which created 10 Water Authorities, corresponding to the main river catchment
areas. They collected charges for the services from the consumers, but also received subsidies
from the government to promote investments. So in 1989 the English water system was again
re-shaped, and the above 10 Authorities became privately-owned companies, which presently
manage both drinking and waste water services. In addition, there are now 23 statutory water
only companies, which manage drinking water services, while waste water disposal is carried
out by the main companies. The entity responsible for water provision are the companies
themselves, which obviously own all the facilities. Some of the companies are listed on the
Stock Exchange, and the most important French private companies have recently purchased
shares. As for charges, the English system is based on the well known price-cap method. Prices
are up-dated through formula mentioned previously. According to their nature many regulatory
constraints are provided by law in terms of economics, quality and environmental aspects. In
particular the management is subjected to the control of national regulatory organizations, such
as OFWAT. Since water meter installation does not cover all apartments, customers do not
pay charges for the volume of water actually delivered, but bills based on the value of the
591
houses. The annual average charges for the drinking water service vary in England (Zabel and
others, 1994) from 73 £/house to 115 £/house. Because the number of management companies
is limited and charges are generally the same within the same company no great price
differences are noticed.
A very similar criteria is common in the US water supply companies.
4.3. ITALY
Many legislative provisions have to be taken into account when dealing with water
resources in Italy. First of all Italy is divided in 20 Regions and, according to a 1989 law on
Soil Conservation, River Basins have been created in the whole territory which are based on
river catchments areas. Moreover other laws concern sewage disposal and treatment plants.
When dealing only with drinking water supply, the distribution services are at present rapidly
changing, because of two recent laws passed by Parliament: the first one dates back to 1990
and is a general re-shaping oflocal government authorities; the second one dates to 1994 and
deals with provisions in water resources matters. It is necessary to focus on the situation
existing before these new laws in order to look at the possible changes and the problems to
face. The new legislative framework inherited from the past different management systems can
be summarized as follows:
• in most great communes the service management is performed by Communal or

Intercommunal Association Companies ("Aziende municipalizzate") which are also often in
charge of other municipal services, such as sewage disposal, gas, electricity,
• in small communes the direct management ("by economy") is nearly always adopted,
• in some cases concessions to private companies have been implemented or public

commercial companies created;
• in three cases (Sicilia, Sardegna and Puglia) Regional State Bodies have been set up, which
are responsible for both water provision and management service.
In the case of management "by economy" the commune itself manages the distribution
service, being the water service included into the general municipal structure. Even if the water
592
charges have to be fixed by trying to balance costs and profits, it is very difficult to estimate
accurately expenses; frequently the water price does not cover the costs, so that the balance
has to be filled through subsidies. The prices fixed by the local authority have also to be
approved by both Provincial and National Price Committees, while local public health agencies
are responsible for health standards.
Among communal or intercommunal Companies, the Azienda Municipalizzata is a
financially independent structure of the municipality, which manages the service on behalf of
the commune itself, and has a separate budget. The investments are financed according to their
nature, but only the local authority can allocate funds for the company. Direct management and
management by Azienda Municipalizzata share the same rules to fix charges and the same
control entities.
Also in the case of Regional State Bodies price setting and controls are the same as the
other management systems, except for those made by Provincial Price Committee.
The Public Commercial Companies are share holding companies, created by the
commune, whose participants are public entities, namely departments and companies. The
managing company performs the service on the basis of a contract, it finances investments and
owns the facilities. Again, price setting and controls are the same as the other management
forms. The economic management is at the risk of the company.
Finally, according to private company management, the commune delegates the service
on the basis of a contract, which can last up to 30 years; investments can be financed according
to their nature; facilities are municipally-owned, except those directly financed by the company,
which, in any case, will return to the municipality when the contract expires, after assessing
their depreciated value. Prices are proposed by the company, but they must be approved by
both Provincial and National Price Committees. Controls are the same as mentioned above.
Also in this case the economic management is at the risk of the company.
As a result of the above management systems, water has not yet been administered as
an economic good. Prices, which have been determined by political and social elements, are
often unbalanced, the service quality is outside of any control, investments are often delayed
and they are allocated by external funds without influencing charge policy, the conditions of
facilities and, in particular, distribution networks are often obsolete. Users are now accustomed
to think that low tariffs mean low quality of service. High domestic consumptions and
corresponding low prices clearly reflect the way water is supplied in Italy.
The last law which recently passed the Italian Parliament has re-shaped water service
management policy. First of all, this law promotes the integrated management of all water
public services, from water supply to sewage disposal and wastewater treatment; secondly
water services must have a territorial basis, according to some optimum territorial areas
("Ambiti territoriali ottimali"), while now several hundreds of municipal services are spread
593
even at a regional scale (at a national level there are now about 6.000 drinking water entities
and about 4.000 wastewater treatment ones). Moreover, the Price Committees have been
abolished; the tariff, which has to be intended as a compensation for the service, must in any
case cover total investment and running costs. Once again, the commune is responsible for the
water service, and it can choose the most suitable management system among the following
forms:
• direct management (this form, however, even ifbased on economic criteria, will disappear),
• delegated management by concession,
• management by a special company ("Azienda Speciale"), which will likely take the place of
the former "Azienda Municipalizzata", with a new legal form,
• management by joint-stock companies, with public and private capitals.
The relationships between the local authority and the management entity, if separate,
will be regulated through a formal contract, which will establish the duration, the criteria of
economic and financial planning, control rules, warrants, penalties, and the kind of tariff. In any
event tariffs will be fixed by the local authority.
4.4. GERMANY
Germany is divided in 16 individual Lander, 5 of which are new Lander created after
the fall of the Berlin wall and 1990 German unification.
Provisions which allow municipal self-government date back to 1808, when municipal
structures for water supplies and sewerage were established. Over this period the municipali-
ties of the Federal Republic of Germany developed a wide range of provisions in order to
handle all the municipal services, holding in any case the service responsibility or a large degree
of control. On the contrary, in the former Eastern Democratic Republic the municipal water
management was abolished in favour of centralized planning, based on several river authorities
as well as on the so-called People-owned water utilities, one for each district. As for the
service organization, another important difference must be highlighted (Kraemer, 1994):
Western municipalities separated the water supply service from sewage disposal and
594
wastewater treatment, because water supply had traditionally been considered as a commercial
activity, including also eventually other public services, such as gas, electricity, urban heating;
on the contrary, the former Eastern Germany set up integrated models of water supply and
sewage disposal. Though the integrated management of both services, which anyhow are under
direct municipal control, is profitable, separate management organizations in Western Germany
have a historical origin, because sewage disposal has always been looked at as an essential
public service, including those citizens who can not pay service charges.
Following the fall of the former Eastern Germany, which accepted the West German
law and administrative regulations, a very fast process of water service re-organization is now
operating in Germany, whose main aims are (Kraemer, 1994):
• the restoration of municipal control and administration in the cities of the new Lander, by
assimilating West German management models,
• the creation of integrated systems.
Regarding the existing management models, the local government is responsible for
water supply and it can freely choose the best means to reach its aim, according to city size,
historical reasons, customs and so on.
Each municipality can choose its water management structure among many models,
such as direct municipal management, joint management with other local communities,
management with a third public or private body. A simplified classification includes:
• direct municipal management (Regiebetriebe),
• management by "Stadtwerke", which includes both forms of municipal enterprises, the so-
called "Eigenbetriebe", namely a semi-autonomous municipal agency, and other commercial
forms, such as the so-called "Eigengesellschaften", or mixed companies, or private
companies,
• management by intermunicipal associations (Zweckverbllnde), whose members can be either

local authorities or private persons,
• management by public companies with other public sector organizations.
The Stadtwerke model prevails in the biggest cities, whereas in rural areas and in
smaller communes the Zweckverbllnde are common.
595
When a third body is responsible for the management a delegation contract between
the municipality and the company rules the relationship. The delegation can be carried out by
contract or statute.
From an economic point of view, water charges must enable the service to break even,
whatever the adopted management form, being in any case the Cost Recovery method
adopted. The charges can be either prices, if the management system is under private law, or
fees when public law is applicable in the management system. Depending on the company
features, charge control is carried out by municipalfmtermunicipal authority or a Cartel Office.
Investment strategies as well as overall control are up to local authority.
According to their nature, other controls are finally carried out by several bodies.
4.5. REFERENCES
Kraemer R.A. (1994) Restructuring Water Institutions in Germany. Hydrotop 94, 12-15 avril,
Marseille.
Zabel T.F., Rees Y., Correia F.N., Neves E.B. (1994) Institutional Arrangements for Water
Management in Selected EC Countries. Hydrotop 94, 12-15 avril, Marseille.
5. Institutional Framework for Training of
Specialists for Water Supply Sysytems
*The Status and Problems to be Solved *
Cedo Maksimovic
IRTCUD, Institute of Hydraulic Engineering
YUGOSLAVIA
5.1. INTRODUCTION
In broad terms the institutional framework requires the analysis of a wide list of topics
such as economic and social conditions, legal framework and enforcement policies, hierarchy
of institutions at various levels, decision making procedures and, interaction between public
and private sector, technology transfer and training incentives, public awareness, etc. These
topics have been discussed at various conferences and symposia (such as Lisboa Conference,
1995, for the group of several EU countries and USA). All of the above factors affect the way
and means in which the introduction of new technological achievements find its way into Water
Supply Systems companies. Thus in this Section, an emphasis will be placed only on the
educational aspects ofinstitutionaI framework.
To achieve high standards or high levels of service in water supply systems (WSS) it is
essential that all specialists and other employees who take part in planning (design),
production, distribution, recycling (reuse) have an opportunity to learn how they can
contribute to the improvement of either water quality, safety of supply, efficiency in
production, reduction in consumption, environmental protection, etc. These tasks are not easy
to achieve unless there is an organized system of continuous training and/or occasional update
of knowledge, improvement of skills and mastering of new technologies.
Since the present book is targeted at the international market one should bear in mind
the big differences not only between developed and developing countries, but also in
NATO ASI Series, Partnership Sub·Series, 2.Environment- Vo1.IS

598
technological levels and in institutional frameworks in various regions of developed countries

(North America, i.e. USA and Canada, Western Europe, etc.).
Presentation of the current situation would require more space then is possible here.
However, there is no doubt that water supply projects have a high priority among the other
infrastructure and environmental protection projects. This should normally imply that for
successful development of this essential urban infrastructure system, an appropriate
organizational, social and human resources infrastructure should be built in order to support
the system. Unfortunately this is not often the case. Several levels of institutional framework
will be tackled here (Figure 1.).
Level Major shortcomings

(region and country dependent)
• lack of coordination and interference of
competence
Government and ministries • improper priorities in training
• conflicting regulations
• lack of updated information
• outdated curricula
Education at universities • lack of holistic approach neglected
• poor coordination between various
specialists
• too much local expertise

Professional associations • lack of international flavor
• low influence on companies politics
• lack of funding in developing countries

Water companies and water supply • mismanagement
enterprises • improper training of staff
• outdated technologies
• outdated technologies
Operational management units • poor maintenance
• low reliability of data
• lack of holistic approach

Consulting companies • low responsibility for the system's long
term performances
• slow response to new technologies
Figure 1. Major problems determing needs for institutional framework improvements in WSS
599
5.2. COORDINATION AT GOVERNMENTAL LEVEL
Coordination between ministries responsible for different aspects of decision making

affecting WS projects is generally low. It is the case in many countries that there is a lack of
coordination and cooperation between ministries of water resources, environment, public
health, city planning and construction, forestry and agriculture and ministry of finance. This
lack of cooperation causes inefficient planning. The result is often that water supply projects
are badly conceived, delayed, poorly managed, systems are badly maintained and
technologically outdated.
5.3. DISTRIBUTION OF TEACHING AT UNIVERSITY LEVEL
Water supply systems are complex and require either multidisciplinary teaching or a
well balanced curricula within a particular department. The departments of sanitary engineering
(in most countries) or as separate departments or even universities (Eastern Europe) have
usually been overloaded by traditional "design school" programs, which lack integration
covering resources management, treatment, environmental protection, modem operational
management and supported by a high level of information processing.
Traditionally those engineers were able to design and construct systems which
conveyed the required (predefined) amount of water. There was not much concern about the
amount of water being measured, stored and conveyed in real world water balance, operational
control, diagnosis and energy efficiency. Environmental concern and long term adaptability is
the weak point of a system designed in this way. In some countries they are being replaced by
departments or institutes of environmental science and engineering (Figure 2. - Denmark,
Sweden).
Until recently, the talented young engineers from developing countries (especially from
Central and Eastern Europe) have had a very good background in fundamental disciplines
(physics, mathematics, mechanics, hydraulics, etc.) and, because of the lack of access to
modem computer and information technologies, were rather unskilled in coping with advanced
modeling and design techniques. On the other hand, one often hears young engineers and
managers from developed countries complaining about being overwhelmed by computers and
powerful modeling and graphics facilities but being unable to sufficiently understand the
physical background of the processes involved, the phenomena being simulated and the
600
complexity of the water-envirorunent interaction. The term DRIPS (data rich information poor
system) complements to the situation in which a lot of data is being collected and stored poorly
or improperly processed with very little proper information being generated.
It is not only the low accuracy and reliability of data that contribute to this kind of
situation, but also the lack of proper conceptualization and right interaction between the
monitoring and data acquisition programs and data processing and information generation.
The author of this Section has often experienced the situation where complex urban
water projects fail or malfunction because of the poor coordination of various specialists
(hydraulic, mechanical, chemical, automation, electrical, software, structural engineers, etc.) in
the design phase and their counterparts in the operational management phase.
A large number of water supply systems exist in which the elementary information such
as annual or daily water balance is known only by a fair guess, with uncertainties of the order
of at least 50010, are rather the rule than the exception.
This is proof of the low awareness of the importance of the basic quantities which are
essential in several tasks such as:
• energy consumption management,
• rehabilitation strategies,
• resources management,
• human resources,
• water balance.
In this respect, the future of the education and training of urban water infrastructure
specialists has to be reexamined at both undergraduate (university) and on-the-job level.
The proper cO'ordination in modernizing the existing curricula for undergraduate
engineering schools and organization of the proper programs and training tools for continuous
education in the field is a strikingly important task that will need to be fulfilled (CALWARE,
1992-1994). Recent trials, such as for water supply (Mak:simovic, 1991) and for urban
drainage are just small efforts that have to be complemented by a number of other activities.
601
Target group Topic to be taught
• water = money (in real terms)

Top level politicians • new technologies
and • integrated (holistic) approach
design makers, bankers • human resources development
• improve before investing
• introduction of innovative methodology

University (departments of sanitary, • informatic impact
environmental engineering and • integrated approach (multidisciplinar)
science) • new curricula
• *
data reliable data
• new technologies (need to be introduced)
Top level managers • impact of informatics
• human resources development
• innovative technologies
Chef engineers • water balance
and • energy efficiency
technology managers • human resources development
• maintenance and reliability of data
• appropriate technology to be adjnsted

• *
new technologies appropriate ones
Professional associations • some do it better
• training of members
• training sites need support
• new technologies
Engineers in consulting • informatic tools can solve part of my problem
and • efficiency
water supply companies • reliability of data and improvement of system
performances
• what was wrong with my project
• how to prevent "garbage in garbage out" result

Software developers with my package
• reliability assessment
• development of educational versions
• test cases
• *
colorful reliable
Equipment manufacturers • operational reliability
Figure 2. Needs for training

602
5.4. PROFESSIONAL ASSOCIAnONS
The professional associations at national levels are supposed to playa major role in the
organization of professional meetings (seminars, symposia, training courses, etc.) for
innovation of knowledge and for transfer of new technologies. However, it is often the case
that these events tend to be much of a localized nature and that there is no sufficient
international interference and transfer of the appropriate problem solving tools.
The thematic meeting of international associations such as IWSA, although often
attended by a few hundred participants do not have much influence on the work of operational
managers of engineers working in treatment plants (the "end users" of new technologies),
especially in developing countries.
Restructuring and reorganization of the educational and training units within the
professional organization is needed such that it enables sufficient information on modern and
appropriate innovative technologies to hit the right target groups of specialists:
• top level managers for strategic planning and decision making,
• chief engineers in consulting companies for implementation at design levels,
• chiefs of operational management units for implementation of the new and appropriate
technology in the real world,
• information specialists for improvement in the reliability of data on the existing status and
characteristics of the system and for merging tools from other disciplines such as GIS,
CAD, etc. with water related projects.
The appropriate training tools have to be developed and implemented for training of all
of the above groups jointly or separately.
5.5. CONSULTING COMPANIES
The major shortcomings of the infrastructure projects being effected by the consulting
companies seems to be:
603
(i) lack of holistic approach. Project design and management is affected by ignoring the
integrated approach at the level of the city, region or river basin, and
(ii) poor quality or lack of reliable data on the characteristics and capacity of the existing
systems - followed by low responsibility for the long term performance of the systems
being designed.
The latter fact combined with poor maintenance of the WSS especially in developing
countries causes fast deteriora.tion and outdating of systems.
One possible way of coping with this problem could be an improvement in regulations
by introducing the obligation of water and consulting companies to perform full scale tests on
the existing systems before undergoing reconstruction, rehabilitation or upgrade.
An example of successful policy of this nature is the procedure for upgrade of urban
drainage systems in the UK.
5.6. WATER SUPPLY COMPANIES
There is an apparent difference in the companies of developed and developing

countries.
The characteristics of the transition in the companies undergoing privatization in the
EU (UK, France, Italy) is that the companies with relatively high levels of technological
management systems (UK) are being forced to shrink these systems as a direct consequence of
the request for higher profit. There is not enough experience in the long term consequences
and possible worsening of the operational safety of the system. On the other hand the WSS in
Italy are just being privatized and there is not enough experience on the effect of the change of
ownership on the change in the level of technological modernization and reliability of data
collected from the systems and in real time.
The situation in developing countries is much different because of the fact that "cheap
labor" replaces technology. Generally, the levels of reliability of data on the systems'
characteristics is very low. Outdated, poorly maintained and malfunctioning measuring,
regulation and operation equipment is almost a rule.
The instrumentation in control rooms of many companies is similar to that shown in
Figure 3.
604
There is generally a low level of awareness of the importance of data (as simple as
water balance). The system can often run with a high level of leakage (losses) without the
management being aware of it. From the point of view oflong term sustainability these systems
are often badly managed.
An important problem is a psychological barrier being placed by senior managers which
prevents the introduction of new technology even where the financial resources are "almost'
available.
lnstrument noJ Instrument no. 2 Instrument no. 3
Outofordcr Doesn't function Ncver activated

Inslrument no.4 Instrument no.5 Instrument no.6
.~.
Dead? Active, but incorrect Control to the

instrument no. 5
Recorder no. J Recorder no.2. Recorder nO. 3. Recorder no. 4.
I]Bad
It
Correct
~I
~..
Recording
Never
monted
recording recording? without pcn
Figure 3. Typical status of instrumentation on a control room of an 'average' WSS
The decisions about big investments on structures, pipelines, dams, etc. are made much
more easily than in modernization of equipment, transmission, introduction of informatic
support, etc.
605
There is a major need and room for training of specialists and for the introduction of
advanced management at all levels in those countries. However one should be cautions in
planning. Gradual, realistically achievable steps are recommended.
5.7. PUBLIC AWARENESS
It is becoming more important to water companies to establish and maintain a

permanent contact and cooperation with the water consumers and public in general. As the
environmental and ecological awareness and concern of the population increases, the
importance of providing the right information to the consumers increases as well.
The projects for new water intakes (especially reservoirs and dams) are becoming more
and more difficult to justify and defend. The public has been mobilized and motivated to
oppose them. Thus the environmental impact assessment has not only to be done correctly, but
all possible measurements have to be well documented and prepared for presentation in an
"easy to understand" and popular fashion so that the good projects can get a "green light".
Otherwise, good projects can easily be compromised and difficulties in approval experienced.
For the reason of preserving the cooperation of the public in rational and prudent use ofa vital
resource such as water, it is important to devote some intellectual and technological resources
for the development of appropriate tools (posters, ride, TV spots, booklets, flyers, software
demos, etc.) that can serve this purpose.
Although the local audience has to be addressed in the correct way, taking into account
social, cultural, religious, economic and all other aspects, a lot of fact and convincing material
is universal and can be shared between various countries.
5.S. REFERENCES
CALWARE (1992-1994) Computer Aided Learning in Water Resources - EU project under

the TECHWARE framework.
Lisboa Conference (1995) Proc. of the International Conference on Institutional Dimensions of
Water Resources Management. Lisboa, Portugal.
Maksimovic C. (1991) Hypothetical Catchment - An Educational Tool for Training in
Sensitivity Runoff Models. Proc. Urban Drainage and New Technologies, UDT '91,
Dubrovnik.
Index
-A- -c-
Acids, 174, 186,224,260,271,287 Calibration, 339, 385, 392, 394
Activated carbon filtration, 172, 175, 177, 181, California, 12,37,72,80,183,279,514
185, 190,199 Cancer, 66
Actuators, 154, 343 Carbon filtration, 172, 175, 177, 181, 190, 192,
Administrative regulations, 560 199
Aerial photography, 488, 496 Carrying capacity, 340
Agriculture, 7, 20 62, 73,172 Catastrophic, 373, 434
AJdehydes,262,287,292 Cavitation, 150, 428, 430, 449
AJdo- and ketoacids, 287 coefficient of, 150
AJluvial Aquifers, 28, 50, 63 Certification, 129, 557, 575
Alternative disinfection, 226, 232, 275 Cesspools, 74
Anaerobic bacteria, 78 Chemical oxidation, 165, 177, 178, 179
Ancylostomiasis (hookwonn), 75 advanced oxidation, 165, 171, 175, 178,229
Antiseptic, 80 by-products of, 179
Aqueducts, 114, ozonation, 177, 179, 190,200,234,290
Asbestos cement, 90, 94, 104 Chemical treatment, 166, 181
Ascariasis, 75 Chlorine 66, 75, 82, 122, 125, 165, 171, 176,
Asset database, 543 181, 187,213,219,224,260,263,271,275,
Asset infonnation system (AlS), 551 290,345,345
Asset management plan, 543, 549 dose, 224, 260, 263
Auto-regulation, 155 Chlorine dioxide, 66, 79, 174,219,224,229,
Automatic control, 144, 345, 398, 448 261,271,280,291
Average zone night pressure (AZNP), 358 Chlorite, 261, 271, 281
Cholera, 31, 75
Clarifying effect, 310, 313
Clean water act, 216
-B- Coagulation, 169, 173, 186, 189,223,231,234,
237,265,280,293,296,307,310,321
Benefit/cost ratio, 38, 384 Coagulation-flocculation, 312
Biodegradation, 165, 171, 190, 199,202 classical coagulation, 296, 299, 303
Biofilms and biological regrowth, 227 coagulant dose, 311
Biological activated carbon filtration, 179, 185, colloidal particles, 237, 293, 307, 311
189,199 flocculant, 312
Biological treatment, 232 flocculant dose, 313
BiologicaVchemical approach, 181, 182 hydrodynamic conditions of mixing, 294
BMP, 62 initial water turbidity, 298
BOD, 76, 77 mixers, 285, 293, 307
Booster pumping stations, 157,423 prolonged coagulation, 296, 301, 307, 310
Brackish water, 27, 32, 204, 207 rapid mixing duration, 293
Bromide concentration, 263 residual water turbidity, 297
BS5750,129 velocity gradient, 296, 301, 304, 310
Burst main location, 389 Coefficient of recuperation, 150
608
Coefficients routine, 455 preanunoniation, 276

Colifonn bacteria, 67, 76, 276, 278 prechlorination, 228, 276
Conunon water, 561 precursor removal, 227, 266, 269, 271
Conununal association companies, 591 removal ofDBPs, 268
Conununity water systems, 212 source water quality control, 264
Computer control systems, 232 Control policy, 365, 373, 379, 384
Computer graphic metafile (COM), 547 Control rooms, 346, 547
Computer system, 124,477,496 Control system, 232, 237, 384, 399, 478
Concession, 590, 593 Control valves, 145, 150, 154, 376, 378, 380,
Construction methods, 89, 95 385,449,450
backfilling, 96 Contunuous measurements, 328
bedrnmg,96, 102, 105 Conventional treatment scheme, 27, 29, 63, 169,
bedding factor, 103, 105 171, 173, 181
preliminary operations, 95 Core operations, 537, 538, 551
traditional methods, 95, 104,416,462 Corporateinfonnationsystems,539
trench excavation, 95 Correlators, 341, 355
trenchless installations, 97 Corrosion, 35, 43, 50, 65, 69, 73, 91, 168, 189,
Consulting companies, 558, 598, 602 197,219,227,230
Consumption, 8, 9, 19,22, 113, 130, 131, 135, corrosion control, 219, 227, 231
142,332,335,341,344,346,350,353,374, Cost recovery method, 570
376,384,390,403,504,529,531 Cross connection, 68
average, 10, 505 Cryptosporidium, 222, 224, 66
by metered consumers, 330 Crystalization, 194, 195
civil, 20 Customer complaint data, 389
daily, 23 Customer protection, 580
domestic, 19,21,507,592
economic models of, 504
individual, 20
maximum day, 13
-D-
maximum hour, 8, 14
maximum month, 13 Data acquisition, 327, 330, 345, 467, 495, 539,
maximum week, 13 584,600
monthly, 19,21,24,72, 78, 113 Data loggers, 132, 135, 345, 348, 351,354, 361,
own, 330 364,371,385,396
value of, 20 Data logging instruments, 385
Contact time, 171, 177, 179, 182,202,224,263, "dial-up" loggers, 352
266, '}.76,279, 289 Data logging system, 360
Contaminants, 61, 67,167,169,172,179,181, Data management, 127,384,479,492
203; 211, 215, 224, 228, 232, 234, 238, 261, Data bases, 415, 479, 489, 518, 522, 541
390 DBP regulations, 270
Control characteristics, 151, 153 best available technology (BAT), 271
Control node, 157,378 Dead ends, 69
Control of disinfection by-products, 264 Decision support, 396, 416, 492, 517, 521
addition ofanunonia and chlorine, 276 Degradation of surface waters, 212
air stripping, 219, 220, 235, 237, 268 Delaunay tessellation, 474
chloramines, 219, 260, 271, 275, 278, 292 Delegation contract, 595
chlorine dioxide decomposition, 282 Demand, 8, 13
enhanced coagulation, 265 monthly, 8, 13
OAC adsorption, 266, 268, 269, 271 weekly, 8, 13
membrane filtration, 165, 172, 175, 181, daily, 13
203,206,209,232,235,266 zonal, 350
609
"Demand" driven, 353 Emergency situation, 348, 372, 377, 384

Demand management, 353, 360, 363 Emerging technologies, 228, 231, 240
Demand zones, 350 Encapsulation, 458
Demographic studies, 374, 504 Energy consumption, 156, 189,206,288,344,
Dependence model, 15 600
Design aspects, 209 Energy recovery, 449
Diffused reinforcement concrete, 94 Enzymes, 50
Digital elevation model (OEM), 414, 467, 471, Epoxy lining, 357
529 Erosion, 43, 50, 65, 470, 472, 475, 491, 493
"Digital mapping", 540 Estuaries, 19,27,32
Digital terrain model (OTM), 471 Excessive pressure, 343, 535
Digitizing, 412, 539, 551 Expert system, 378, 398
Director general of water, 350, 390, 580 Externa1loads,99, 103, 106
Discharge coefficient, 148, 149
Disinfection by-products, 165, 174,217,224,
228,232,239,259,261,264,272,275,277,
291 -F-
best available technologies, (BAT's), 226, 271
chloroform, 259 Facilities management systems, 469
di- and trichloroacetic acids, 260 Faulty instrumentation, 363
trihalomethanes, 170, 177,217,218,226, Feeders, 114, 117, 122, 125
259,262,271,273,275,292 Fictitious market, 566
Maximum contaminant level goal (MCLG), Filtering extraneous noise, 355
259 Finite difference method, 402, 404, 435, 452
Disposal of waste, 165, 172 Fire fighting, 68, 341, 409
DISTANCE, 535 Fixed flow meters, 341
District metered area, 128,351,386 Fixed-head nodes, 375
Domestic metering, 357 Flexible pipes computation, 104
Drainage basin boundaries, 473, 475 Floc-blanket clarifier, 307
Drinking water quality, 167,272,580 clarifying effect, 305
aesthetic/organoleptic standards, 165, 167 cohesion coefficient, 315, 317
guidelines, 80, 84, 90, 93, 167,218,238,354 concentration of compacted sludge, 315, 319
microbiological standards, 165, 167 floc-blanket hight, 314
operational standards, 165, 167, 168 Flow control, 145, 147, 156,424,428,449,559
toxicological standards, 165, 167 Flow measurement, 330, 333, 335, 338, 340, 584
Dual water supply system, 47 Flow modulation, 359
Ductile iron, 90, 108 . Flow of raw water, 332
Dune infiltration, 185, 187, 192 Flow regime, 338, 340, 424
Duration of the transfer, 124, 125 Flow throttling, 156
Dynamic data, 361 Freshwater demand, 19
Fumes, 52
-E-
-G-
Earthquakes, 32, 35, 65
Easy water pollution, 34 Galleries, 29, 31, 51
Economic efficiency, 38 Giardia cysts, 170,223,239,240
Economic models, 374, 504 Giardia, 66, 170, 222, 239
Education, 373, 379, 573, 598, 600 giardia lamblia, 260, 273
Electromagnetic flow meter, 337
610
GIS, 127, 131-144,347,361-364,371-378,388, Hydrodynamic modelling, 347, 527, 529

394-398,412,467-476,488-491,512-529, Hydrological model, 492, 523
533,535,602 Hydrogen peroxide, 287
access to, 550 Hyperfiltration, 169, 177, 185, 192,204,206
activities, 542 Hypolinmion, 28
background, 362, 389,547,549 frrPFlESS, 401, 409, 416, 451, 460
emergency site location, 550 HYTRA, 443, 450
grid based, 497
hardware, 476
matching simulation models with, 529
operator, 535, 536 -1-
raster based, 469, 478
use within various organisation, 519 Industrial wastewaters, 31
vector 478,497 Information collection rule, 66
Gradient methods, 402 Information on topography, 529
global gradient method, 410 Infrastructure, 8,469, 523, 525, 528, 538, 540,
crrains, 194, 198,209,50 598,600
Granular activated carbon, 66, 219, 226, 230, Inheritance, 457
232,234,235 Initial calibration, 394
Ground penetrating radar, 514 Insertion flow meter, 340, 341, 342
Groundwater, 19,28,34,51,71,74,80,113, Integral flow measurement, 333, 335
125, 127, 166,267,277,292,491,498,501, Integral services digital network (ISDN), 352
513,523,525,527,559 Interconnection, 48
artificially recharged, 30 Interdependence, 53
contamination of sources, 212 Internal pressure, 93, 99, 103, 106, 1I8
drawn in, 29 International commission on large dams, 27
models, 498, 504 Inventory system for asset management, 517
juvenile groundwater aquifers, 30 Israel, 71, 80, 84, 159
quality of, 165, 167, 17l
volcanic rocks, 29
Groundwater treatment, 171, 219
acidification and overfertilization, 165, 172
-J-
conventional ground water treatment, 171
Joint-stock companies, 593
Glass reinforced plastic (GRP), 94, 96, 443, 450
Joukovsky expression, 436
Junction, 376, 405
Juvenile groundwater aquifers, 30
- H-
Haloacetic acids (HAAs), 262
Halogenated compounds, 170, 187
-K-
Halogenated DBPs, 262
Karstic, 29, 32, 34, 55, 116
Head loss coefficient, 146, 148, 151,426
Karstic water systems, 29
Health risks, 71, 74, 287
Heavy metals, 169, 172, 189
Helminth, 75, 81
Hepatitis virus, 68 -L-
Historical regulations, 555
Holy Cross College, 68 Lakes, 25, 27, 29, 32, 55, 62, 74, 81,166,216
Homologous relations, 427 Lamella settlement, 307
Hydraulic models, 373 Land information systems, 469
611
Landsat~,502,516 Merits, 185, 193, 196, 199,206,385

Laws, 68, 215,238,431,556,560, 591 Meter accuracy, 350
Lead and copper, 50, 219, 227, 240 Meter errors, 363
Leakage, 135, 144,351,353,358,360,365, Metering, 237, 285,349, 352, 357, 360,363,
398,551 365,548,584
control strategy, 551 Method of characteristics, 435, 436
detection, 340, 343, 528 Metrology, 557, 575
initiative, 351, 354, 365 Micro(organisms), 169
monitoring, 352, 392 Microfiltration, 204, 206, 208
rates, 353, 356, 358, 363 Micropollutants, 165, 174, 181, 185, 189, 192,
spotting, 343 199,208
National leakage initiative, 354 Minimum night flows, 386
Levels of service, 127,327,371,557,573,580, Mining related wastewaters, 31
581,584,587,597 Mixed companies, 594
Limit cost, 38 Model construction, 394, 504
Limitations, 158, 185, 197,201,206,259,391, Monitoring system, 65, 350, 397
396,485,495,504,523 Multiple barrier system, 59, 65
Lmk,127,354,355,388,393,402,406,416, Multiple network, 47, 57
455,478,49,500,504,548,559 Municipal enterprises, 594
Load factor method, 102
Local elimination routine, 455
Location of pipes, 512, 517
Long term observability, 348 -N-
Longevity of water supply sources, 33
Losses management, 343, 346 Nanofiltration, 171, 176, 182, 204, 206, 226,
232,239,267
National leakage initiative, 354
National rivers authority (NRA), 350, 580
-M- Negative effects, 73, 201, 561, 570
Network analysis, 360, 362, 371, 396, 495, 521,
Main bursts, 329, 344 522,533,535,541,547,549
Management, 347, 566, 594 Network modelling, 361, 371, 388, 390,393,546
models, 347, 348, 566, 594 New standard, 175, 181,260,49
direct 567,591,593 New technologies for water treatment, 165, 185
emergency, 396 Newton-Raphson method, 402
basic record, 544 Nitrification, 186, 277, 292, 77
losses, 343, 346 Node, 157,237,376,378,396,402,406,413,
nenvork,144,361,363,397,399 454,533,549
operational, 127, 134, 144,345,389,392, Node structures, 457
399,529,541,548,598,602 Nodal hydraulic components, 402
source management, 390 Node routine, 455
Market for water, 25, 35 Non-agresive soils, 35
Mastering of new technologies, 597 Non-regulation contaminants, 217
Maximum hourly flow, 8 NPDES,61
Measurements, 150, 183,218,223,284,327, NPSfI, 428, 430, 436
332,338,343,345,347,350,427,575,605 NPSfI curve, 436
accuracy of measuring system, 330
Measuring points, 346
Measuring principle, 330
Membrane staging, 208
Membrane filtration processes, 177, 206, 209
612
- 0- Polymorphism, 458
Polynomial trend model, 15
Ponds, 26,29, 35, 62, 76, 77,80
Object linking, 415 Position of discharge measuring, 331
Object-oriented progranuning, 458 Post code system, 131
branch structures, 457 Post-filtration, 198
Objects collections, 457 Precipitation, 14,66,72, 77, 170, 181, 193, 196,
Operating manuals, 364 208, 278, 289 28, 30, 559
Operating range of a pump, 428 Preferential water tariff, 73
Operation efficiency, 428 Pressure control, 343, 357, 359, 363, 397
Operational changes, 360, 362, 393 Pressure field, 533, 536
Operational management, 127, 134, 144,345, Pressure reduction, 152,343, 345, 354,359,361
389,392,397,399,529,541,548,598,602 Pressure relief, 145,68
Optimization models, 374 Pressure sustaining, 145, 158,359,363,535
Organic compounds, 66,168,170,174,176, Prestressed concrete, 90, 92, 95, 104, 117
199,204,235,237,287,290 Price of drinking water, 37
Overlay operation, 533 Price-Cap method, 570
Oxidation products, 287 Primary treatment, 76
Ozone, 66, 171, 177, 181, 189, 199,219,224, Private water, 561, 562
226,229,232,235,237,261,281,286 Process scheme, 181, 185, 187, 190, 192
ozone decomposition, 66, 287, 289 Professional associations, 558, 598, 601, 602
ozone dosage, 202, 290 Public access, 61, 67
ozone generators, 288, 290 Public awareness, 597, 605
Public water, 17, 20, 59, 212, 239, 260, 562
Public water system, 212, 239,260
- p- Pump control, 145, 156,363, 390, 394
Pump behaviour, 427, 438
four quadrant diagrams, 439
Parallel and series operation, 429
Pump control valves, 145
Pathogens, 74
Pump efficiency, 427, 432
PC-based mapping system, 543
Pump power failure, 434, 437, 441, 443, 450
Pellet reactor, 183, 185, 190, 192, 196, 198,210
Pump start-up, 448
Pellet softening, 187
Pumping costs, 390, 395
Periodicities, 15, 17
PVC, 90, 92, 98,107,450
Permanent measuring station, 328
Pesticides, 61, 63,165,167,170,172,174,181,
185, 190, 192, 199,208,217
pH, 29, 47, 66, 79,82,108,170,189,193,224, -Q-
230,263,275,281,287,297,306,312
Physiclil treatment process, 182 QBE,408
Piezometric fields, 533 Qualitymanagement, 129,557,573,577,579
Pilot-plant studies, 235 Quality of products, 573
Pipe deflection, 104, 107 Quality of surface source water, 516
Pipe material, 50, 90, 92, 94, 100, 102, 106,363 Quality of the services, 583
PlatlOrm,413, 415,451,481, 485,488, 528, 546 Quality regulation, 217, 272, 580
Point of chlorination, 226, 260, 283, 285 Quality system elements, 577
Point velocity measurement, 337 Quantification, 288, 329, 342, 348, 353
Poliovirus, 77
Pollutants, 30, 54, 61,116,165,167,178,208,
216,292,390,515
Pollution control, 240, 53, 59, 61
Polyethylene, 92, 107
613
-R- Singapore, 71, 78, 525, 552

Sisyphusian, 52
Slope aspect, 473
Radio-active compounds, 169 Snapshot, 328, 392
Rapid sand filtration, 66, 181, 186, 189, 192, Softening, 169, 170, 176, 181, 182, 183, 185,
214,268,283 187, 189, 190, 192, 193, 194, 195, 197, 198,
Raster data structure, 470, 498 206,207,210,231,271,277,288,289
Rating curve, 338 Source management, 390
Recalculate routine, 455 Source outputs, 130, 350
Recalibration, 338 Sources of water, 25, 28, 32, 37, 40, 44, 48, 54,
Record updating, 541, 550 72,83
Rectal cancer, 224, 66 Sources of raw water, 166
Recycled water, 31, 37, 54, 55, 73 Space-compact implicit finite difference scheme,
Reg-neg, 261, 264, 270, 272 453
Regulated contaminants, 217 Special works, 119
Regulation and control, 122 Specific flow, 149
Rehabilitation, 327, 343, 346, 349, 353, 356, Specific speed, 427
357,360,363,535,575,600,603 Spectral reflectance, 482
Relining, 353, 357, 361, 363 Spetral signatures, 483
Remote sensing platform, 487 Springs, 15,25,29,32,50,54,375,513,560
Removal of organic matter, 176 SQL, 408, 412, 414
Replacement rates, 356 Standardisation, 545, 550, 557, 575, 577
Report No 26, 354 Standards for drinking water, 175
Reservoirs, 8, 25, 30, 54, 60, 67, 122, 145, 173, Standards of service, 390, 396
265,277,330,345,375,381,403,428,451, Steel, 90, 94, 96, 98, 100, 106, 117, 120, 234,
492,517,605 285,289
"Resource-led", 353 Stochastic model, 15
Residence time, 186, 309, 319 Stochasticities, 17
Retail price index, 543 Stream lines, 472, 475
Reuse, 71, 78,83,597 Streams, 25, 27, 73, 513
Rigid pipes computation, 102 Stressed vegetation, 484, 514
Rock fissures water systems, 29 Structural requirements, 89, 98, 107
Rotational speed, 427, 431, 439 Substutution reactions of chlorine, 263
Ruhr,53 Suction pressure, 423
Surface sources, 26
-s- Surface water, 27, 29, 36, 54, 63, 113, 118, 125,
165,171,180,185,192,211,214,219,222,
260,268,270,277,493,502,513,577,578
Safe drinking water act, 216, 231, 239, 260
treatment, 170, 176,219,222,240,260,268
Saline water, 25, 27, 32, 54, 56
treatment scheme, 174
Salts, 169, 173, 182,208,265,278
Surge control devices, 448
Sea water, 32, 177,204, 207
Suspended solids, 76, 78, 82, 167, 186,286,516
Secondary drinking water standard, 218
Sustainable development, 19
Secondary operations, 539
Swamps, 28
Secondary treatment, 76
System integration, 547
Selection of water source, 25, 32, 39
System redesign, 360, 362
Separate system, 43, 45, 48
Septic tank system, 61
Service connection, 212, 214, 372, 376
Settlement curves, 308
Simulation models, 327, 345, 348, 374, 381,
493,528,529
614
-T- -v-
Target levels, 354 Valve chamber, 119, 121,355
Tariff selection, 363, 390, 395 Valve characteristics, 149, 158, 426, 460
Tariffs, 21, 357, 395, 556, 592 Valves, 145, 156, 158,343, 359, 460, 535
Teaching, 495, 599 back pressure, 145
Technology based, 61 check, 145, 156,443,447,448
Technology costs, 233, 264 electric remote control, 145
Telemetered flow meters, 350 flow control, 145, 156
Telemetry, 127-135, 144,352,360-365,371, in-line control, 460
378-368, 390-399, 540-552 pressure reducing, 145,343,359
data, 131,547,548 pressure sustaining, 145, 158, 343, 535
data transmission, 347 Variable speed, 423, 429, 431
Temperature, 56, 116, ISO, 166, 171, 196,224, Variable-head nodes, 375
236,263,275,287,297,304,312,340,470, Vector structure, 470, 475, 498
481,486,513,522 Vegetation stress, 484
Tertiary treatment, 77 Velocity distribution, 338
Tessellated structure, 470 Velocity, 124, 336
Thennal image, 508 average, 124, 336
Time-line interpolation, 436 Velocity profile, 337
TnUrdng,133,237,291,355,364,373,380, Ventury meter, 333
467,495,519,539,543,550,557,600,605 Viruses and protoviruses, 48
training of specialists, 557, 597, 605 Volcanic rocks groundwater, 29
Transfer point, 375
Transient regime, 60, 62, 377, 434, 438, 347
Travel time of sound, 337
Treatment goals, 215, 218 -w-
Treatment processes, 67, 75, 165, 169, 176, 193,
217,223,230,262,345 Wall roughness, 153, 339,
Treatment technologies, 70, 228, 233, 238, 292 Wastewater recycling, 71
Trends, IS, 17, 183,210,352,384,504,547, Wastewater reuse, 71, 79,80,83
562 Water
Trihalomethanes (TIIMs), 218, 226, 259, 262, private, 561
275 public, 17,20,59,212,239,260,561
Truncation error, 454 abnospheric,30,54,71
availability of, 33, 36, 39, 44, 583
Water balance, 329, 340, 348, 352, 361, 364,
-u- 599,604
Water charges, 21, 556, 566,590,595
Water demand, 7, 8, 12, 14, 17, 19,21,23,33,
lntraflltration, 79,171,204,208,219,230,232 37,44,54,57,67, 72, 78, 113, 136, 139,
lntrasonic flow meter, 336 374,378,387,504,507,512,524,527
Unaccounted for water, 351, 384 analysis of, 384
Unbalanced prices, 656 annual, 7, 8, 10
Underground plant, 540, 545 drinking, 21
Underground water, 28, 53, 63, 578 growth in, 37
Urban drainage models, 491, 493, 496 hourly, 149
Urban irrigation, 48, 52,470,507,528 seasonal, 12
Urban water demand, 7, IS, 17,524 urban, 7, 15, 17,524
US EPA, 71 Water hanuner, 118, 156,373,409,417,434,
Use of wastewater, 75 437,451,458,462
615
VVarerlos~,329,332,341,343,346,556
VVarer market economy, 562
VVaterprices, 21, 23, 555, 565, 574, 590
VVarer provision, 566, 589
VVarer law principles, 555, 560
VVarer level control, 155
VVarer quality, 26, 53, 61, 167,215,289,308,
327,329,345,373,397,512,548,578,597
based, 61
goals, 216, 218
improving, 215
VVater regulations, 217, 241, 260, 273, 560
VVarerrigh~, 7,25,33,35,41,44,556,561,566
VVater service, Ul, U8, 398, 523, 552, 556,
565-574,579,587,590,592,594
management, 592
organization, 556
performances, 557
VVatershortage,72,80,484,581
VVarer softening, 183, 193,210,289
VVarer sources, 25, 29, 32, 33, 40, 54, 71, Ill,
165,181,212,215,217,222,231,350,581
VVater storage capacities, 35
VVater suppliers, 557, 579
VVarer supply companies, 173,558,591,601,
603
VVarer supply performances, 579
VVarer tapping purposes, 34
VVarerbome outbreaks, 220, 270
VVES~T,361,365,389,391,395,546,552
VVetIands,28
VVithdrawal (abstraction) of raw warer, 332
-z-
Zero discharge, 217, 405
Zero risk, 80
Zonal demand, 350
The Partnership Sub-Series incorporates activities undertaken in collaboration with NATO's
Cooperation Partners, the countries of the CIS and Central and Eastern Europe, in Priority
Areas of concern to those countries.
The volumes published as a result of these activities are:

Vol. 1: Clean-up of Former Soviet Military Installations. Edited by R. C. Herndon, P. I. Richter,
J. E. Moerlins, J. M. Kuperberg, and I. L. Bicz6. 1995
Vol. 2: Cleaner Technologies and Cleaner Products for Sustainable Development. Edited by
H. M. Freeman, Z. Puskas, and R. Olbina. 1995
Vol. 3: Remediation and Management of Degraded River Basins. Edited by V. Novotny and
L. Somly6dy. 1995
Vol. 4: Earthquakes Induced by Underground Nuclear Explosions. Edited by R. Console and
A. Nikolaev. 1995
Vol. 5: Transportation Infrastructure. Edited by R. M. Gutkowski and J. Kmita. 1996
Vol. 6: Sustainable Development of the Lake Baikal Region. Edited by V. A. Koptyug and
M. Uppenbrink. 1996
Vol. 7: Transboundary Water Resources Management. Edited by J. Ganoulis, L. Duckstein,
P. Literathy, and I. Bogardi. 1996
Vol. 8: Urban Air Pollution. Edited by I. Allegrini and F. De Santis. 1996
Vol. 9: Advances in Groundwater Pollution Control and Remediation. Edited by M. M. Aral
(Kluwer Academic Publishers). 1996
Vol. 10: Assessing the Risks of Nuclear and Chemical Contamination in the Former Soviet
Union. Edited by E. S. Kirk (Kluwer Academic Publishers). 1996
Vol. 11: Ventilation and Indoor Air Quality in Hospitals. Edited by M. Maroni (Kluwer Academic
Publishers). 1996
Vol. 12: The Aral Sea Basin. Edited by P. P. Micklin and W. D. Williams. 1996
Vol. 13: Radioecology and the Restoration of Radioactive-Contaminated Sites. Edited by
F. F. Luykx and M. J. Frissel (Kluwer Academic Publishers). 1996
Vol. 14: Economics of Atmospheric Pollution. Edited by E. C. van lerland and K. G6rka. 1996
Vol. 15: Water Supply Systems. Edited by C. Maksimovi6, F. Calami no, and J. Snoxell. 1996
Vol. 16: Bioindicator Systems for Soil Pollution. Edited by N. M. van Straalen and
D. A. Krivolutsky (Kluwer Academic Publishers). 1996
Vol. 17: Remediation of Soil and Groundwater - Opportunities in Eastern Europe. Edited by
E. A. McBean, J. Balek, and B. Clegg (Kluwer Academic Publishers). 1996
Vol. 18: Environmental Engineering and Pollution Prevention - European Network on
Excellence and Partnership. Edited by J. Wotte, W. A. Halang, and B. J. Kraemer
Vol. 19: East-West Life Expectancy Gap in Europe - Environmental and Non-Environmental
Determinants. Edited by C. Hertzman, S. Kelly, and M. Bobak
Springer
and the
environment
At Springer we firmly believe that an
international science publisher has a
special obligation to the environment,
and our corporate policies consistently
reflect this conviction.
We also expect our business partners -
paper mills, printers, packaging
manufacturers, etc. - to commit
themselves to using materials and
production processes that do not harm
the environment. The paper in this
book is made from low- or no-chlorine
pulp and is acid free, in conformance
with international standards for paper
permanency.
Springer

Water Supply Systems 1996

Uploaded by

Copyright:

Available Formats

Water Supply Systems 1996

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

Water Supply Systems 1996

Uploaded by

Copyright:

Available Formats

NATO ASI Series

Advanced Science Institutes Series

The Series is published by an international board of publishers in conjunction with the

A Life Sciences Plenum Publishing Corporation

The Partnership Sub-Series incorporates activities undertaken in collaboration with

Library of Congress Cataloging-in-Publication Data

Water supply systems: new technologies / edited by Cedo Maksimovic.

ISBN-13: 978-3-642-64739-0 e-ISBN-13: 978-3-642-61187-2

• BERMAD Control Valves, ISRAEL

Pavel Jirou§ek Thomas G. Sanders

tedo Maksimovic Adrian Verwey

Introductory Statement ................................................................................................... 3

1. Urban Water Demand ................................................................................................................ 7

2. Water Demand Has a Threshold ............................................................................................... 19

4. New Concepts for Water Supply Systems: What Kind of Water

5. Pollution Control for Urban Water Supply Systems .................................................................. 59

6. Wastewater Reuse - a Worldwide Issue ..................................................................................... 71

Introductory Statement .................................................................................................. 87

1. Construction Methods and Pipe Structural Requirements .......................................................... 89

2. Water Transfer Within Brussels Water Services ................................................................................ 111

3. Transfer ofWaier and Data Management in Wessex Water - UK ..................................................... 127

4. Control Valves - Principles of Operation ................................................................................. 145

Introductory Statement ................................................................................................ 163

1. New Technologies for Water Treatment: General Aspects ................................................................. 165

2. New Technologies for Water Treatment: Case Studies ....................................................................... 185

Introductory Statement ................................................................................................ 257

3. Technological Aspects of Advanced Processes of Coagulation and

4. Study of Floc-Blanket Clarifier, Combined with Lamella Settlement

Introductory Statement ................................................................................................ 325

2. Diagnosis and Rehabilitation of Existing Systems .............................................................................. 349

Introductory Statement ................................................................................................ 369

1. Water Supply Networl<: Modelling and its Application in Design,

2. Object-Oriented Hydroinfonnatics Tools for Water Distribution Analysis .............................. .401

Introductory Statement ................................................................................................ 421

1. Pumping Stations and Transient Flow Analysis ...................................................................... .423

2. Object-Oriented Tools for Transient Flow Analysis ................................................................ .451

Introductory Statement ................................................................................................ 465

Introductory Statement ................................................................................................ 555

I. Historical and Institutional Aspects ......................................................................................... 559

2. Management Options .............................................................................................................. 565

3. Quality Management in Water Supply .................................................................................... 573

4. Examples of Existing Management Systems in Europe ............................................................ 589

5. Institutional Framework for Training of Specialists for Water Supply Systems:

Water Supply Systems on the Threshold

Water Supply Systems on the Threshold

Water demand generally referred to as future water requirements will be a function of

1.2. DEMAND VARIATIONS

City Population Population Percent Consump. Percent Consump.

1900 1890 1890 1890 1905 1905

Chippewa Falls, Wis. 8.094 7.4 6.6 13.8 ... 100 I

Table 2. Recorded rates of consumption in some american cities (McGhee, 1991)

City Average daily per Maximum l-day Maximum in

Ashtabula, Ohio 766 985 129

Figure 1. Variation of yearly rates of consumption (Tumeaure and Russell,1908)

1.2.3. Monthly and Weekly

Typical maximum day ~

Name Unit Order of Rei ;ression coefficients

Dependence model Distribution

2.2. WATER DEMAND - EUROPEAN EXPERIENCE