57451611 Palmer Leakage Detection Guide
Dec 26, 2015
Guide for leakage detection
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Leakage Detection and ManagementA comprehensive guide to technology
and practice in the water supply industry
Written byDavid Butler BSc MSc CEng MICE
Published byPalmer Environmental
2000
David Butler
David Butler has over 25 years’ experience in the UK water industry, specialising in analysis, appraisal,auditing, policy development and training in water distribution leakage management and controlstrategies. Internationally, he has experience of leakage policy development in major cities in India,Germany and the Netherlands. He has provided leakage management training courses in Hong Kongand Portugal, and has addressed and chaired international water engineering conferences on the subject.
Palmer Environmental
Palmer Environmental is the world’s largest designer and manufacturer of specialised water leak detectionequipment, with over 40 company years’ experience of meeting customers’ leak detection needs. PalmerEnvironmental supplies products for all aspects of leak detection, including the widest range of leak noisecorrelators, acoustic products, market leading noise loggers and the first economically justifiable,permanently installed leak detection system.
Palmers’ design and manufacturing facility in Cwmbran, UK produces the world’s most advanced,innovative and easy-to-use water leak detection equipment. This is backed up by an internationaldistributor network providing comprehensive service, support and training.
Published by
Palmer EnvironmentalTy Coch House
Llantarnam Park WayCwmbran
NP44 3AWTel: +44 (0) 1633 489479Fax: +44(0) 1633 877857
email: [email protected]: www.palmer.co.uk
ISBN 0-9538014-0-3
Page
1. INTRODUCTION TO LEAKAGE DETECTION
1.1 The Importance of Leakage Reduction ........................................................................................................ 7
1.2 Leakage and Usage ...................................................................................................................................... 8
1.3 Sources of Leakage .................................................................................................................................... 11
1.4 Factors Influencing Leakage ...................................................................................................................... 13
1.5 Basic Leakage Growth and Active Control ..................................................................................................15
1.6 Leakage Control Strategies ..........................................................................................................................16
1.7 Consumer Reported Leakage........................................................................................................................19
1.8 Future Considerations ..................................................................................................................................20
2. HYDRAULICS AND NETWORK ANALYSIS FOR LEAKAGE DETECTION
2.1 Introduction ..................................................................................................................................................27
2.2 Energy Principles ..........................................................................................................................................27
2.3 Head Loss in Pipelines..................................................................................................................................29
2.4 The Hydraulic Gradient ................................................................................................................................30
2.5 Pipe Flow Formulae......................................................................................................................................30
2.6 Pipe Friction Diagrams ................................................................................................................................32
2.7 Network Analysis..........................................................................................................................................32
3. DISTRICT METER AREA MANAGEMENT
3.1 Distribution Network Structure ....................................................................................................................37
3.2 District Meter Areas......................................................................................................................................37
3.3 Waste Meter Areas ........................................................................................................................................40
3.4 Links to Other Data Information Systems....................................................................................................40
3.5 Commissioning ............................................................................................................................................41
4. METERING FOR LEAKAGE DETECTION
4.1 Hierarchy of Metered Areas..........................................................................................................................43
4.2 District Metering ..........................................................................................................................................43
4.3 Waste Metering ............................................................................................................................................44
4.4 Recent Meter Improvements ........................................................................................................................44
4.5 Meter Site Selection......................................................................................................................................45
4.6 Meter Installation Design..............................................................................................................................45
4.7 Meter Selection Criteria ................................................................................................................................46
4.8 Mechanical Meters - Helix (Woltmann) ......................................................................................................48
4.9 Electromagnetic Flow Meters ......................................................................................................................49
4.10 Insertion Velocity Probes ..............................................................................................................................51
4.11 Domestic Revenue Meters ............................................................................................................................51
CONTENTS
5. DETECTION EQUIPMENT
5.1 Detection Principles ......................................................................................................................................55
5.2 Stethoscopes (‘Listening’ or ‘Sounding’ Sticks) ..........................................................................................55
5.3 Electronic Sounding Devices........................................................................................................................55
5.4 The Mobile Advanced Step Tester (MAST) ................................................................................................56
5.5 Leak Noise Correlator (LNC) ......................................................................................................................56
5.6 Leak Noise Loggers ......................................................................................................................................57
5.7 Non-Acoustic Equipment and Techniques....................................................................................................59
6. EQUIPMENT AND LEAKAGE DETECTION TECHNIQUES FOR TRUNK MAINS
6.1 Introduction ..................................................................................................................................................62
6.2 Meter on Bypass ..........................................................................................................................................62
6.3 Heat Pulse Flow Meter ................................................................................................................................63
6.4 Pairs of Insertion Turbine Meters ................................................................................................................63
6.5 Infra-Red Photography..................................................................................................................................64
6.6 Leak Noise Correlation ................................................................................................................................64
7. IDENTIFICATION OF MAINS, SERVICES AND VALVES
7.1 Introduction ..................................................................................................................................................65
7.2 Location for Mains in a 2m Footpath ..........................................................................................................65
7.3 Service Pipe Layouts ....................................................................................................................................65
7.4 Valve Identification ......................................................................................................................................66
7.5 Electronic Pipe Locators ..............................................................................................................................66
7.6 Other Location Methods ..............................................................................................................................67
8. LEAKAGE INDENTIFICATION AND LOCALISATION
8.1 Demand Patterns ..........................................................................................................................................70
8.2 Night Lines....................................................................................................................................................70
8.3 The Development of Continuous Monitoring ..............................................................................................70
8.4 Determination of Leakage from Night Flows ..............................................................................................71
8.5 Necessary Checks ........................................................................................................................................71
8.6 Large Area Sub-Division ..............................................................................................................................72
8.7 Waste Metering ............................................................................................................................................72
8.8 Step Testing ..................................................................................................................................................73
8.9 Acoustic (Noise) Logging ............................................................................................................................74
9. LEAKAGE LOCATION, CONFIRMATION AND REPAIR
9.1 Sounding ......................................................................................................................................................79
9.2 Leak Noise Correlation ................................................................................................................................79
9.3 Visual Evidence ............................................................................................................................................80
9.4 Other Practical Points ..................................................................................................................................80
9.5 Confirmation ................................................................................................................................................80
9.6 Repair, Follow-up and Records ....................................................................................................................81
9.7 Leakage Contracts ........................................................................................................................................81
10. PRESSURE MONITORING AND MANAGEMENT
10.1 Pressure Management ..................................................................................................................................82
10.2 Pressure Control Options ..............................................................................................................................82
10.3 Pressure Control Benefits ............................................................................................................................82
10.4 Pressure Reduction Problems ......................................................................................................................83
10.5 Pressure and Leakage....................................................................................................................................84
10.6 Statutory Requirements and Levels of Service .......................................................................................... 85
10.7 Identification of Areas for Pressure Reduction ............................................................................................85
10.8 Pressure Reducing Valves - General Overview ............................................................................................87
11. THE ECONOMICS OF LEAKAGE MANAGEMENT
11.1 Introduction ..................................................................................................................................................93
11.2 Policy Development through an Economic Approach ................................................................................94
11.3 The Unit Cost of Leakage ............................................................................................................................97
11.4 Cost of Leakage Detection............................................................................................................................98
11.5 Typical Total Leakage Costs ......................................................................................................................100
11.6 Environmental and Social Costs ................................................................................................................100
12. RECENT RESEARCH AND DEVELOPMENT IN THEINTERPRETATION AND USE OF NIGHT FLOW DATA
12.1 Introduction ................................................................................................................................................111
12.2 Bursts and Background Losses ..................................................................................................................111
12.3 Components of Night Flows ......................................................................................................................111
12.4 Night Flow and Customer Use....................................................................................................................113
12.5 Losses from Bursts......................................................................................................................................114
12.6 Estimating Background Night Flows in Individual DMAs ........................................................................115
12.7 Prioritising Unreported Burst Location Activities ......................................................................................118
ACKNOWLEDGMENTS ........................................................................................................................125
6
7
1.1 The Importance of Leakage Reduction
Traditionally, in the UK, leakage detection has been seen only as a small part of a much bigger picture.Water Suppliers had little incentive to find and repair leaks because there was seen to be a relativelyplentiful resource. This has all changed.
Firstly, there have been three significant droughts in a twenty-year period that have each led to severesupply restrictions in some areas. The climate is changing, with rising temperatures which are forecast tocontinue, and shifting, less dependable rainfall patterns. With droughts predicted more frequently, thecumulative effect on ground water and river supplies will be noticeable. Furthermore, applications todevelop new water sources are rigorously examined in the light of leakage reduction performance. Also,households, industry and agriculture are all likely to want more water due to this climate change, on topof the normal gradual increases.
Secondly, the privatisation of the Water Industry and the creation of regulatory bodies have forced WaterSuppliers to more accurately quantify leakage, and to devise better strategies to manage it. Leakage hasbeen brought sharply to the attention of the media, and hence to the public. It has become a political issue,capable of making headlines. Consequently, tough mandatory targets have been set for the Industry in thelast half of the 1990s, based on a five point criteria:- methodology, data quality, breadth of analysis,robustness of the water balance, and consistency of approach.
Thirdly, the Industry is having to take much more notice of its customers views. If customers are beingencouraged not to waste water, then the Suppliers must be seen to be doing more. Research has indicatedthat customers:
• don’t think there should be any leakage• do not understand the concept of economic levels of leakage• listen to the media more than Water Suppliers about leakage performance• have more emotional than rational views about leakage• perceive a lack of leakage reduction as a profit related issue• have mixed views on water conservation and supply pipe responsibility
Fourthly, a more rigorous cost/benefit assessment of economic levels of leakage has been demanded bythe Industry Regulators, with reporting procedures established. The performances of individual Suppliersare now readily compared in the public view, and comparisons are being sought internationally, with theintent of ‘motivating’ for further leakage reductions.
These four aspects are summarised below.
Quantity ConsiderationsClimate ChangeResource Development LimitationsIncreasing Demand
Regulatory ConsiderationsOFWATEnvironmental AgencyDrinking Water Inspectorate
Customer ConsiderationsEfficiency PerceptionShareholder ExpectationsLegal Responsibilities
Economic ConsiderationsAn Optimum LevelA Consistent Strategy
1. INTRODUCTION TO LEAKAGE DETECTION
8
1.2 Leakage and Usage
Definitions
Water loss may be defined as that water which having been obtained from a source and put into a supplyand distribution system is lost via leaks or is allowed to escape or is taken therefrom for no useful purpose.‘Water loss’ is usually considered as ‘leakage’ and ‘water loss reduction’ referred to as ‘leakage control’.
Water loss is usually quantified on the following basis: -
Water Loss = (Quantity of water put in to supply) -(Non-domestic usage + Domestic consumption)
Unaccounted for Water
To allow for leakage and quantities termed as ‘other water uses’, the term ‘unaccounted for water’ is used(called ‘U’). This is a good way of distinguishing it from the useful water supplied to both domestic andnon-domestic consumers, which is grouped together as ‘accounted for’ water.
The classic leakage control formula is:-
U = S - (M + A x P)
Where U = Unknown or unaccounted for quantities of water including leakage.
S = Sum of all water inputs into a system.
M = Sum of all water accounted for by measure (metered supplies, domestic and non-domestic)
A = Average domestic usage per capita of population.
P = Population supplied (non-metered).
Leakage is within the unaccounted for water value ‘U’.
An allowance is normally made for miscellaneous usage and non-domestic, non- metered consumption.
The calculation of ‘U’, its comparison area by area, and analysis of trends, is thus an important basis uponwhich to establish leakage control.
Water Suppliers use the above formula to calculate their annual overall leakage levels by what is knownas the Total Integrated Flow method (i.e. a ‘top-down’ approach).
The Total Integrated Flow (or Water Balance) Method is a useful means of assessing the overall leakageperformance for a system on an annual basis. It will take into account all leakage from the system,including that from reservoirs and trunk mains, and can be carried out based on data which all WaterSuppliers require, not just for leakage control, but for forward planning and financial management.However, the errors inherent in the data used make the absolute accuracy of the calculation questionable.Also, it can take some time (e.g. a month or two) to assemble and compute.
Quantity of Water into Supply
The quantity of water put into supply is normally obtained from physical measurements of the watersource outputs and therefore the reliability and accuracy of source water meters is of great importance. Inleakage calculations it is known as the quantity ‘S’.
9
Water delivered
Water delivered is defined as the amount of water delivered by the Water Supplier at the stop tap (i.e. atthe boundary of its apparatus and the beginning of the private customer’s pipework), plus somemiscellaneous and illegal use.
Water delivered therefore includes private leakage, and is often expressed as a percentage of waterproduced..
Figure 1.1 illustrates the water balance.
Legitimate Use
Domestic
Domestic demand is a term which includes:
- reasonable usage in households supplied (usually measured in the UK), but excluding meter optionhouseholds for calculation purposes
- unreasonable, excessive or extravagant use (e.g. garden sprinklers left on overnight)
Domestic demand in the UK is traditionally evaluated based on estimates of the population served (‘P’)and an estimate of the average daily demand, referred to as per capita demand (‘A’).
Water Suppliers work closely with population census authorities to ensure the accuracy of estimates used.
It has been difficult to calculate this per capita value figure reliably until recently. However, the advent ofthe micro-chip and the modern data logger has made very significant advances in this respect. Dataloggers are now light, rugged and reliable, and relatively cheap, and by using them in a statisticallystructured manner, it is possible to derive values for domestic per capita consumption with a high degreeof accuracy.
Some UK Water Suppliers operate Domestic Consumption Monitors by which volunteer households areselectively chosen by property age, size and type, and are continuously monitored. All water usage isrecorded by data logger, and data retrieval is now being automated for interrogation by telephone.
Based on this data and other national studies, it is considered that actual water used in an average UKhousehold is about 145 litres per head per day. This includes allowances for excessive use and isdesignated ‘A’.
Household metering in the UK is becoming increasingly significant as Water Suppliers are required toinstall more and more domestic meters. As at year 2000, 15% of UK households are metered. SomeSuppliers expect this figure to double in the near future.
Non-Domestic
This component of demand is almost totally metered in the UK (only a small component is unmetered)and is designated ‘M’. It includes leakage on the customer’s side of the meter which has thus beenmeasured and paid for.
This leakage beyond meters is a matter of concern, because even though the water is paid for, it is wastefulin resources, and if left unchecked it can lower the level of service to other customers.
Measurement in terms of population for this component is not really helpful because metered demand canrange from a small café to a large steel works.
Non-domestic metered demand can amount to about a quarter of all water supplied.
Other Uses
There are other quantities of water which do not reach the consumer, and these include:
- Water used for fire-fighting.- Water used to clean service reservoirs.- Water used for mains testing and flushing.- Water for cleaning sewers, streets and other public purposes (sometimes metered and charged for,
sometimes not).
Strictly speaking these water uses are legitimate, are accountable, and could be measured. In practice, theyare usually small in total compared to the amount going into supply, and hence they are not measured, andare counted as part of water losses.
Illegal, and therefore not charged, use can be a problem, although it is normally small.
Volume of Leakage
In round terms in the United Kingdom, about 25% of the water leaving the treatment works is unaccountedfor. However, this is a simplistic statement and circumstances vary widely. Often it is more useful toexpress leakage in demand per property, or per kilometre of main, with time.
Some water leakage rates are recorded at a rate of 30% and above (in mining areas for instance), whilst insome developing countries the water losses can vary from 50% to 70%, and the water mains system canonly be pressurised for a few hours a day.
On the other hand in parts of Northern Europe, on modern systems, leakage as low as 3% is claimed.Singapore makes similar claims.
Outsiders to the industry are often amazed to learn that a quarter of the water which has been gathered,impounded, treated and pumped should then be lost, and they take it as indicating poor practice andinefficiency. This is not necessarily so. Compared to other fluids such as petrol and natural gas wheretiny or no leakage at all is demanded, water is cheap and a less dangerous product. Furthermore, it is beingdistributed through old and very expensive apparatus whose replacement or rehabilitation capital cost ishuge.
Figure 1.2 illustrates the volume of leakage pertaining to the UK, and indicates the sources.
Errors
Leakage control must be approached in the realistic knowledge that water volume measurement is subjectto error. Huge volumes of water are measured and distributed through a vast, aged network of pipes, andwhilst calculation of leakage on a logical basis is essential, it must be accepted that errors exist and valuesare sometimes approximate.
Studies have shown that bulk meters often have significant error which invariably represents an under-recording of the true quantity. Even the small diameter positive displacement meters used for consumermetering struggle to record very low flows. Such flows occur frequently when filling roof tanks andcisterns via ball valves, and can give rise to significant under-recording of the quantity used.
Meter errors occur because of:
- an inherent error- poor maintenance- ageing- misreading- incorrect location- incorrect sizing
10
11
Where domestic consumption is unmeasured, errors can obviously occur because of incorrect assessmentsof customers served, per capita demand and seasonal variations - a variation of up to 40% can beexperienced between a winter’s day and a hot summer’s day.
The estimation of non-domestic, non-metered consumption is also a source of error.
It must be stressed that the existence of these errors does not in any way invalidate the need to calculateleakage as carefully as possible.
The Appendix to this section discusses further the potential for error contained within the use of the‘Unaccounted for Water’ formula.
1.3 Sources of Leakage
Treatment Works
At the beginning of the water works operation, up to 7% of water can be lost as part of the treatmentprocess. On account of good practice and recycling, this amount can be reduced to 2-3% and is not usuallyincluded when leakage control is referred to.
Trunk Mains and Aqueducts
Trunks mains can carry raw water into a treatment works, or treated water onwards into the distributionmains system. There is no definitive size range, but often pipes of 300m and above are considered in thiscategory in the UK.
Because of their importance and the need to preserve security of supply, aqueduct leakage is usuallydetected and repaired quickly. It usually forms only a small part of the overall leakage total.
Service Reservoirs and Water Towers
Again, leakage from these structures usually only represents a small proportion of overall water loss.
It is nevertheless necessary to maintain and monitor reservoirs and water towers carefully, fromconsiderations of their structural safety as well as for leakage. Cracked walls or floors can leak water intounderdrain systems, so it is necessary to regularly record these flows and losses. Cracks and jointdegradation can be checked for during a regular cleaning programme. Water entering the overflow systemis preventable and regrettably quite common. Inspection of reservoir telemetry data, particularly overnightlevels, can often be a useful indication of leakage.
Distribution Mains
Distribution mains (including trunk carrier mains within them) represent the major source of water leakagein a water supply system. Distribution mains are an inheritance of different pipe materials, age andcondition, each leaking at different rates.
Valves, Hydrants, Stop Taps etc
Valves, hydrants, stop taps etc need glands to operate effectively. The gland sealants will deteriorate withtime and sealing surfaces become worn. This creates leakage which must be controlled by the necessarymaintenance.
Periodic operation of a valve will help prevent the need for repairs. Maintenance of all valves on a setfrequency is not recognised as an economic activity. However, strategic valves and trunk main valves maywarrant this attention. Similarly, all flow control valves, including pressure reducing valves (PRVs),should have regular inspections and planned maintenance.
12
External corrosion of valves and hydrants is not generally a reason for failure. However, internal corrosioncan be a significant problem and may prevent a leak-tight shut off.
The types of failures experienced with stop taps are usually leakage from joints or the inability to operatethe valve through corrosion or washer failure.
Leaking mains’ ferrules will generally have to be replaced.
Pipe Joints
Pipe joints are a major source of leakage.
Joints may have been fabricated in a faulty manner and remain water-tight for a few years only. Groundmovement as a pipe trench settles can overload the joint and induce leakage.
Service pipe joints, especially of the mechanical type, are a common source of trouble, compounded bylarge numbers of very similar specialist parts. An ‘O’ ring may look to be the correct fitting but begin toleak the day after it has been back-filled.
Service Pipes
These too are a large source of leakage. Again, these Water Supplier owned pipes may comprise manyages, materials and conditions, and they can leak seriously.
Service pipes may be of lead, copper, galvanised iron or polyethylene, each of which can fail. Copper pipecan be subject to pin hole attack. Much galvanised iron is now nearing the end of its life and is in anadvanced state of corrosion, and polyethylene, although the best of the modern materials and givingexcellent performance, can crack when laid wrongly, or cranked through a tight radius.
Joins to the mains at one end of the communication pipe (mains ferrules) and to the stop tap at the otherare a particular source of trouble.
Underground Private Supply Pipes
These frequently leak seriously. Age, material and condition once more vary widely but also there is oftenthe added complication of shared responsibility on common or joint supply pipes.
Access to repair on private supply pipes, particularly common supplies, is notoriously difficult. Forinstance, over the years, out-buildings may have been erected over the pipes.
Repair of supply pipes is a private responsibility usually enforceable by the Water Supplier but it is nearlyalways a protracted process. Water Suppliers may use special notices to require leakage repair, and indefault may have to do the work themselves and recharge it to the customer.
Continued effort by the Water Supplier to pursue these leaks is needed, and it is the UK experience thatthe offer of a free repair service greatly helps in terms of time. There is also a suprising increase in thenumber of such leaks reported.
National UK investigations have shown that underground private leakage is greater than previouslythought but it is only significant on a few properties (approx. 1 in 300).
Private Pipework Above Ground
This leakage is considerable and varies greatly with the season. Pipes burst more often in the winter! Onnon-domestic premises, any leakage is metered and hence paid for. On domestic premises, dripping tapsand leaking ball valves accumulate to a significant component of the total.
13
Faulty fittings within a house can cause significant loss of water. One faulty ball valve alone can accountfor up to 50 l/day.
A Water Supplier may have statutory ‘bylaw’ provisions which it can enforce concerning the size, nature,materials strength and workmanship of waste fittings. They may be able to forbid the use of fittingsconducive to leakage or undue consumption, and may be able to enforce their maintenance. In practicethe enforcement of bylaws/regulations is very onerous, yet if it is not pursued leakage persists. Leakageoften occurs on old or neglected property where the occupiers do not understand their responsibilities, orcannot afford, or wish to afford, repairs. Continued effort and allocation of resources is needed to containthis type of leakage.
1.4 Factors Influencing Leakage
Types of Mains
Old iron mains still form the majority of mains and they are the worst culprits for leakage. They sufferfrom both external and internal corrosion attack which progressively weakens them. Iron mains can thencrack and leak, or holes form due to the corrosion process. Once the leakage occurs, which may be finallyprecipitated by an increase in pressure, flow, or temporary change, then it will worsen. This may occursteadily, or rapidly degenerate into a large burst. Cases of subterranean caverns beneath metalled roadwaysare known where the escaping water hollows out a void by its pressure jet.
Concrete lining of iron mains virtually stops internal corrosion but has no effect on external corrosion.
Asbestos cement mains normally fail by cracking, acting as a beam under load, and the subsequent collarrepair can be a source of future trouble.
UPVC pipes are not thought to contribute largely to the total water lost. Failures in the early plastic pipeshave been frequent in large diameter sizes, and the pipes usually fail by shattering. Joint ring failure issometimes a problem.
MDPE (polyethylene) pipes are still a relatively recent introduction but their performance to date isexcellent, provided they are jointed properly. Furthermore, polyethylene pipes are still beingimproved/developed, which can only be good for the future.
Steel mains only form a small proportion of mains and these are usually in aqueducts with cathodicprotection. Steel fails usually with pin holing, necessitating welded patch repairs.
Soil Conditions
Clearly soils influence corrosion and leakage rates. Some light soils scarcely affect the pipes whilst otherssuch as Lias clays or alluvium are very aggressive. Trench back-fill of sulphate-rich ashes is especiallycorrosive.
Aggressive Water
Water fed into supply should be carefully controlled for quality. It should be checked to ensure that it isnot plumbo-solvent. Certain natural waters have a higher rate of attack on iron pipe than others.
Climate and Ground Movement
Seasonal variations in climate have a marked effect upon leakage levels. For instance, a hard winterinduces ground movement in the “freeze/thaw” cycle and this causes a high number of bursts.
In a similar way a long drought causes ground movement, and again often results in an increase in thenumber of bursts.
14
Long periods of hot, dry weather bring about high demand and a disproportionately high domestic percapita consumption. This may distort leakage estimates for particular years.
Sudden saturation of dried out sub-soil can also cause problems through local ‘heave’.
Mining subsidence can create successive tension and compression of the pipework causing jointmovement or failure of the pipe.
Removal of support from thrust blocks can lead to excessive joint movement. It should be noted that thiscan also be created by excavation adjacent to the thrust block destroying passive ground pressure at thesupporting face.
Dissimilar Metals
Dissimilar metals between pipes and fittings (e.g. juxtaposition of copper and stainless steel) can causegalvanic corrosion. This must be avoided by reference to guidance in bylaws/regulations/standards etc.
Electrical Earthing
Electrical earthing of buildings to the water fittings has been prohibited in the UK since 1961. It wascommon before that time, and faulty electrical fittings can create a “to earth” potential onto water pipeswhich, in turn, will create corrosion and eventual leakage. It should also be noted that this now obsoletepractice can make service pipes (and mains to some extent) electrically live and dangerous. Temporaryearth loops must be used.
Network Design
A properly designed distribution system should prevent some vulnerability to leakage at the outset.
Such design would assess the need for cathodic protection of steel and ductile iron mains.
It would ensure that all mains with unrestrained flexible joints had appropriately sized and positionedconcrete thrust blocks at all changes of direction and blank ends.
All mains and services should be laid with the correct amount of cover to the surface, and appropriatelydistanced from other underground services. The use of marker tape sited 300mm above the main will alertexcavation to the presence of the main, thus preventing incidental damage and ensuing leakage. Whereplastic pipe is used, such tape should have a metallic strip incorporated to assist with location equipment.
Correct sizing of mains at the outset, considering such factors as peak flow, fire fighting requirements andfuture development, will prevent the temptation to ‘force’ more water through by increasing pressures ata later date. Oversized mains also need to be avoided, particularly from a water quality point of view.
Workmanship
There is no substitute for good workmanship of the initial installation in preventing future leakage. Pipehandling, bedding, laying, jointing and backfilling must be to a high standard.
Extra care should be given to repair work, as a repair does represent a potential weakness to the integrityof the system.
Quality of Materials
It is obvious that all materials used in the distribution system must comply with relevant standards for longterm usefulness, be of a high quality, be appropriate to the surrounding conditions, and be of the correctoperational capabilities.
15
It should also be ensured that the same standards apply to repair materials, and that poor substitutes arenot used for permanent repairs.
Pressure
High pressure equals high leakage. This factor is very important in leakage control, and will be discussedin more detail in a subsequent section.
The Size of the Hole
This may seem obvious but it is very important to remember that a small increase in the size of the leakhas a big effect in terms of volume leaked. Figure 1.3 illustrates this for a constant pressure of water.
The longer a leak is left to run, the bigger the hole will get.
Duration of the Leak
A speedy location and repair of leaks is essential to reduce waste levels.
A leak running for a long time can waste just as much water as catastrophic trunk main burst which isrepaired quickly. Time before discovery, time to detect, and time to repair are the major components.
Leakage will only be reduced by sustained, determined detection, and rapid repair. ‘Find and Fix Fast’ isan appropriate axiom.
Disturbance of the Distribution System
Severe pressures can be generated by the rapid operation of isolating valves, thus precipitating bursts andleakage. Ironically, rapid re-charging of a system following leakage repair work can cause further damageand leakage. Valve closures and mains re-charging work should therefore be carried out in a steady,controlled manner. This is particularly relevant when mains scraping and relining is taking place.
Similar care must be taken during mains flushing, swabbing and air scouring.
Age of System
The ageing process cannot be stopped, and increasing leakage is indicative of deteriorating structuralcondition. It should therefore be recognized that a realistic and consistent level of renewal of theinfrastructure is an essential part of leakage strategy development.
This may be achieved by targeted mains relining (where iron pipes are in use, and corrosion is mostlyinternal), or by targeted mains replacement. The former has little, if any, impact on leakage rates fromthose mains, whereas the latter should eradicate it for a substantial period of time if done well. The moderntechniques of mains replacement have substantially cut excavation and backfilling costs.
It is essential that the renewal of service pipes is included in such work for the greatest benefit.
1.5 Basics of Leakage Growth and Active Control
Natural Leakage Growth
Leakage grows with time, and without action to curb it, would grow to a point where supplies wereunsustainable. Passive control, that is, the repair of bursts and leakage showing on the surface, and theelimination of poor pressure and flow complaints, is the minimum possible response. This is required toprevent damage to persons and property, and to maintain supplies to customers.
16
Figure 1.4 illustrates the growth of leakage with time. This graph is illustrative only. Over a large enougharea, leakage will increase steadily with time if nothing is done to correct it.
The actual leakage level reached will depend on how quickly low pressure and flow will be experienced,and other factors, such as how quickly leakage appears on the surface and is reported.
Characteristic Growth Rate
For any given area in the distribution system there will be a characteristic growth rate. This characteristicgrowth rate will be affected by changes in the physical elements of the system, such as rehabilitation ofmains, renewal of service pipes, and changes in pressure. Sooner or later, leakage control must beassociated with a programme of mains renewal in order to maintain the supply/demand balance.
However, improvement of mains and services is expensive and clearly, for the system as whole, is verymuch a long-term strategy.
Reduction in pressure is also effective in reducing both the volume of leakage and its rate of growth,although there is some doubt whether the latter effect persists in the long term. The scope for pressurereduction is, of course, limited, given that adequate supplies to customers must be maintained.
Reduction in Leakage with Active Control
To reduce the natural level of leakage at any pressure, a programme of leakage detection must be planned,co-ordinated and implemented. The effects of the introduction of various levels or frequencies of leakagedetection are again illustrated in Figure 4. This shows clearly the need to maintain a consistent level ofeffort if the required leakage level is to be maintained. It is not sufficient to put in a high level of resourcefor a short period, as any slackening of effort will lead to an increase in leakage over a period of time.Given that no two water distribution systems are identical in terms of physical or economic characteristics,it is not possible to determine the most appropriate leakage control policy in a general manner. The bestpolicy for any given system will depend on its particular characteristics.
1.6 Leakage Control Strategies
The Economic Balance
The economic balance of searching for and repairing leakage, and of controlling it to an acceptable level,is a complex issue. Typically a leakage percentage of below 10 or even 15% may not be economic topursue. In other words the effect of hunting down, identifying and repairing the leakage costs more thanthe value of the water saved. These remarks need to be heavily qualified however. For instance a modernhousing estate could have a serious problem with 10% leakage whilst an old area with a stubborn leakageof 30% say, may require a mains renewal scheme. Each area will have its own intrinsic economic leakagelevel.
In the UK, historically, a quantity equivalent to 55 l/prop/day was deemed too expensive to find and repair,and was termed “acceptable leakage”. It was further suggested that of the “acceptable leakage”, thequantity of leakage which was undetectable was approximately 30 1/prop/day.
Figure 1.5 suggests the relative percentages of leaks caused by different types of bursts, and the possiblewater quantity lost through them.
In addition to the volume of water lost, its scarcity and marginal cost per megalitre are vital factors. In anarea of rising demand, needing to promote, build and commission a new source, intensive leakage controlactivity would be essential. In an area which relied upon pumped supplies with high electricity costs, ahigh degree of leakage control would make sense, and have priority over an area with plentiful suppliesfed by gravity.
In selecting the required leakage reduction approach, there are two policies which may be adopted. Theseare to adopt either a PASSIVE or an ACTIVE policy.
17
‘Passive’ Control
This is a procedure whereby water loss is only tackled when leakage is visible or when problems arereported from the public.
The adoption of this policy minimises the day to day operating costs of leakage detection, but increasesthe risk of water being wasted. It results in an ever increasing upward trend in the annual supply of water,since leaks can remain undetected for many years until they reach such a magnitude that urgent action hasto be taken due to customer complaints. It is however, a perfectly feasible policy to adopt, providing it is‘politically’ acceptable, and may be carried out with full instrumentation to allow rapid location of leaks,although this will give rise to modest maintenance costs.
Such a policy is only applicable if:
- The revenue costs of leakage detection are high.- The costs of production are low, and there is ample capacity to supply all foreseen demands.- Bursts are readily visible and easily repaired.
‘Active’ Control
It is increasingly accepted that an aaccttiivvee approach of searching for leakage is preferential in cost/benefitterms to a ppaassssiivvee approach of only reacting when the situation has deteriorated.
This relates not only to the Water Supplier’s distribution system but also to private pipework wherecustomers are encouraged to carry out repairs on any leakage detected.
Active control would usually involve the monitoring of flows in a distribution network by using a systemof permanently installed distribution meters. If unexpectedly high flows of water are observed, these areimmediately investigated; leakage detection teams being carefully directed to ensure that leakage ismaintained within defined criteria (such criteria being prepared using an acceptable cost/benefit basis). Itis obvious that monitoring which does not initiate further action is unproductive. It will also beunproductive if, when further action is worthwhile, resources are not available to proceed with location ofthe leakage.
An active policy requires expenditure on meter installations, etc. and the day-to-day operating costs ofleakage detection teams. The following benefits should be achieved:
- It minimises leakage, and hence reduces the loss of water in monetary terms.- It results in an overall reduction of water demand.- Limited water resources are conserved for legitimate use, and rationing etc. is avoided. - It reduces operating costs (savings on electrical power and chemical treatment costs).- Work is planned (rather than acting in response to emergency).- Dangerous leakage is minimised (e.g. freezing water on highway).- Customer perception is improved.- Capital expenditure requirements on treatment works, reservoirs and mains are reduced.
Because of their potential, it is worth noting that active leakage detection in the future is likely toincreasingly employ acoustic loggers, some permanently installed. This could result in larger meter areas,and hence fewer district meters.
A well managed active leakage detection policy ensures that the cost of the leakage detection teams andthe repayments of the capital necessary to establish the system is exceeded by the value of the water saved.It is applicable if:
- The cost of water production is high.- The sources of water have limited capacity and cannot meet normal and/or foreseen demand.- Bursts are ‘invisible’ due to the strata.- The quantity of water being put into supply is increasing at an unacceptable rate.
Strategy Development
Leakage reduction and control is a long-term activity, and should be regarded as a part of good distributionmanagement. Occasional short bursts of effort are unlikely to produce lasting results because distributionsystems continue to deteriorate for one reason or another. If only the obvious leaks are repaired, leakagelevels will still increase, as will consumer problems.
The development of a long term leakage control strategy is therefore essential if water supply anddistribution systems are to be effectively managed. Such development needs to be flexible, with occasionalreviews to ensure that the strategy adopted is the most appropriate one for the situation. Cost/benefitanalysis is important in this regard.
‘Active’ leakage control (i.e. finding and repairing leaks before their presence becomes obvious orgenerates problems) has been found to be a cost-effective method of reducing water supply deficiencies.A planned approach should result in lower complaint costs, and lower repair and maintenance costs.
The establishment of controllable, manageable areas (District Meter Areas - DMA’s) within a distributionsystem whose demands are easily monitored, has been found to be extremely helpful for effective leakagecontrol and supply management. It forces plans to be updated, locating mains and buried fittings. Itintroduces new valves to give better operational control. It locates illegal connections and identifiesmalfunctioning meters and public supplies. In the very process of this setting up work, leakages andwastages are found and repaired. It enforces good housekeeping. Regard has to be given however, to theminimisation of dead-ends and their associated quality problems.
Training and Data Use
Leakage reduction requires a dedicated core of highly trained, specialist personnel using appropriate ‘stateof the art’ equipment and techniques. Local knowledge is essential together with an understanding of theday-to-day operation of the distribution system and demand patterns. Support can be obtained fromspecialist agencies/contractors, given precise briefs and targets.
Personnel motivation, good communication and synchronization of activities, and continuous feedback ofdecisions/results cannot be over-emphasized. This is vital for understanding, efficiency and success.Everyone should be included, from planners to repair teams.
The organisation of leakage control personnel can vary widely. Distribution personnel can either beorganised as a specialist team, or be integrated into general distribution system operational duties, andspend only part of their time on leakage control. It is generally accepted that to properly pursue activeleakage control, and to meet agreed monitoring/detection frequencies, it is necessary to set up specialistteams. However, general operational duties cannot and should not be entirely divorced from leakagecontrol.
Technical support is required for design and modification of district metering, computer systems support,compilation of base data for DMAs, production of reports, overall performance monitoring, production ofdrawings, system records updating, and for problem solving. Clerical support is required for computerinput and administrative duties such as serving notices relating to private pipe repairs. Skilled andknowledgeable technical support is crucial if the mass of data now regularly available is to be handled andanalysed to the best advantage for the leakage reduction effort.
Good leakage control depends upon good and progressively improving data. To achieve this it isnecessary to establish and keep an audit trail of data, building from individual DMAs and their componentdata up to the regional total figures for the Water Supplier. These can be collected in the two data streamsof:
- Aggregated night-lines/’bottom up’ calculation- Total integrated flow/’top down’ calculation.
18
19
It is easy to dissipate effort and resources on leakage control unless the work is properly planned. Toachieve this effectively a good information system and audit trail is essential.
Leakage effort needs to be directed towards the areas of greatest need i.e., where most water is leaking andwhere water costs are highest. Traditionally this has been achieved by the study of total demand relativeto population and industry, by examining bulk meter flow records, and by interpreting recording charts onindividual waste meters. These principles remain sound, but the advent of the computer and data loggerhave provided the means for automatic information processing, and direction towards areas of priorityleakage.
Action Plan Overview
The following ten points comprise key actions within a leakage policy where sustained effort is applied: -
1. Locate and repair obvious leakage.
2. Sub-divide the distribution system into DMA’s and continuously monitor them for leakage control.Maintain DMA meters, loggers and boundary valves. (Apart from physical constraints, the size ofthese areas may be determined by the leakage detection and location policy, and the equipmentused within them.)
3. Employ “Active” leakage detection policies using modern methods and equipment.
4. Effect PROMPT repair of reported and detected leakage.
5. Provide a sound, reliable leakage information system to underpin leakage control activities. Buildconsistently improving data.
6. Direct leakage control efforts to areas with priority of need.
7. Reduce pressures where possible.
8. Relay mains and service pipes in modern non-corrodible materials (welded polyethylene preferred).
9. Drive down leakage towards a target value and then review target.
10. Reduce leakage to an economic (or ‘politically’ acceptable) minimum.
1.7 Customer Reported Leakage
Reported leakage comes from the public, the police, other utilities and public bodies. All these parties arethus offering a valuable service which should be respected and acted upon. This response will thenencourage further reports and hence rectify more leakage in turn. On the other hand, the converse is true– neglected leakage will discourage further reports.
Customer queries relative to leakage particularly require a rapid response. Leakage often has a high profilewith the public and the media, and it is sometimes mistakenly believed that leakage is a symptom ofmismanagement rather than a legacy of old, corroded mains and pipework. Prompt attention to leakagequeries can correct this misapprehension.
Local contact between distribution staff and the groups who might report leakage is well worthwhile andpromotes the importance and benefits of a timely response.
20
Reported leakage may comprise a significant proportion of all leakage discovered (i.e. with the balancebeing detected). As such, much of the expenditure of leakage detection is avoided and, apart from thepublic relations benefits, it behoves the Water Supplier to respond quickly to reported instances on thegrounds of economy.
1.8 Future Considerations
The end of the 1990s has seen a very significant reduction of leakage in the UK. The Industry hasresponded well to the targets it has been set. Some of the contributing factors have been:
• The attention given to the creation of new and smaller DMAs• Continuous, night-flow monitoring• The improved management of pressures• The improvements in detection technology and techniques• Greater investment in renewing the pipe network• Greater active leakage control, seeking to locate and repair leaks before they become visible• Assistance given to customers in reducing leakage from their own pipework• Better understanding of the water balance components and associated issues.
It is unlikely that leakage will ever again become a background issue in the UK. Lower leakage levels willbe expected to be achieved, as the supply/demand balance comes under further scrutiny. Customerexpectations will continue to rise, as will the encouragement for them to save water. Economic analysiswill be a continued emphasis because cash flow will continue to be tight, and further operating costreductions will be expected. Climate change and environmental considerations will continue to be adriving force. Ways will be sought to introduce ‘competition’.
Though the relationship between effort, cost and result is still not fully understood, the Industry will belooking for more efficient and effective methods to identify, locate and repair leaks more quickly.Assistance will be sought from further technological development, particularly permanently installedacoustic loggers, possibly with telemetry. Attention will further turn to issues concerning service pipeleakage, better assessment of legitimate night use and the effects of social changes, more reliablequantification of seasonal variations in night flows, and trunk mains monitoring for leakage. There will bemore focus on ‘grey’ water re-use in order to meet rising demand without jeopardising the ability to meetdemand in the future. Asset life extension will remain an issue, with more attention given to the dilemmaof reconditioning or renewing mains. Data quality will be expected to improve, with particular emphasison the use of real cost/benefit data in models, appropriate values for water saved, and environmental andsocial costs.
21
Appendix to Section 1
Water Balance Method of Computing Leakage
Total Integrated Flow Formula:
U = S - (M + A x P)
where U = Unaccounted for waterS = Total volume suppliedM = Metered useA = Per capita useP = Population supplied unmetered
Each of the above terms is subject to error. Pilot studies of bulk meters in one supply area indicated thaton average, readings were only 80% of true flow with some recording only 60%. If it were assumed thatthe meters associated with the terms S and M above were under-recording by only 5%, whilst A x p wasover-estimated by 10%, then leakage could be a third more than estimated. It has been argued that an errorof such magnitude is quite feasible.
Large variations in unaccounted for water from year to year would indicate a certain inaccuracy associatedwith the use of this formula. The expression of U as a percentage of total consumption is rightly criticisedas being misleading and unsuitable for comparisons. For instance, in a very warm spell, S will increasebecause A and M increase. If A is not adjusted (as is quite common), U appears to increase both in quantityand as a percentage. Similarly, in an industrial recession, M and S will go down and even though Uremains the same quantity, it will increase as a percentage of the volume supplied.
The formula is also subject to other inaccuracies. A, the per capita consumption was usually based onstudies of existing domestic meters. It is now known that small, conventional domestic meters recordnothing until the flow exceeds 5 l/hour and under record any flow between 5 and 22 l/hour. In propertieswhere most water is routed through a ball valve to a roof tank, substantial under recording will occur. Forthe same reason, the meters of measured consumers will under record at low flows whatever the metersize. This is particularly true at cattle troughs or in factories with substantial periods of low demand.
P, population served, is a figure derived from one definition of resident population, (there are several), withdeductions for consumers not supplied or supplied through a meter. Population served varies through theyear, whilst the deductions are usually un-audited estimates.
The measure of ‘net night flow per separately charged property’ as a means of assessing leakage is nowcommonly preferred. Theoretically, this measure is prone to less significant errors than the total integratedflow formula, and it implies that leakage is expected to increase in quantity as development takes place.The provision of accurate means of measuring night flows within the distribution system is fundamentalto this approach, but it is unrelated to pressure however, and therefore comparisons between areas need tobear in mind their relative average night pressures.
22
Figure 1.1 Components of Water Supplied
Volume per Day (Not to Scale)
DISTRIBUTION INPUT (DI)
WATER TAKEN (WT) DISTRIBUTION LOSSES (DL)
POINT OF DELIVERY TO CUSTOMERS
WATER DELIVERED (WD)
WATER DELIVERED THROUGH MINOR SUPPLY PIPES (WDS) COMPONENTS
Miscellaneous
Water Taken
SIMPLIFIED BREAKDOWN OF DISTRIBUTION INPUT (DI)
DISTRIBUTION INPUT (DI)
WATER TAKEN (WT) DISTRIBUTION LOSSES (DL)
WATER DELIVERED (WD) DOU DISTRIBUTION LOSSES (DL)
WATER DELIVERED MISCELLANEOUS WATER DISTRIBUTION LOSSES (DL)THROUGH SUPPLY PIPES (WDS) TAKEN (WTM)
MEASURED UNMEASURED UNMEASURED MISCELLANEOUS WATER DISTRIBUTION LOSSES (DL)(WDSM) USE (WDSU) SUPPLY PIPE TAKEN (WTM)
LOSSES (WDSL)
DIS
TRIB
UTI
ON
SYS
TEM
OPE
RATI
ON
AL
USE
(DO
U)
TRU
NK
MA
INS
SERV
ICE
RESE
VOIR
S
DIS
TRIB
UTI
ON
MA
INS
CO
MM
UN
ICAT
ION
PIP
ES
MEA
SURE
D
UN
MEA
SURE
D
USE
UN
MEA
SURE
D
SUPP
LY P
IPE
LOSS
ES
HYD
RAN
TS
ILLEG
AL
USE
23
Total Volume Total Total Leakage Total LeakageSupplied Unaccounted for
TOTALVOLUME OF
WATERSUPPLIED
TOTAL LEAKAGE
MAINSLEAKAGE
APPROX 1/2
DISTRIBUTIONMAINS
LEAKAGE
COMMUNICATIONPIPE LEAKAGE
PRIVATESUPPLY PIPELEAKAGE
Treatment losses
Losses due to firefighting flushing etc.
Leakage on trunk mains& service reservoirs
Leakage oninternal private
pipework
Figure 1.2 Histogram to show Assessment of the Volume of Leakage Components
24
Figure 1.3 Effect of Hole Size on Leakage
Discharge inLitres/day
49 090
20 945
17 454
2 945
1636
Discharges through circular holes in 0.5 inch diameter lead pipe.
The well-known old diagram from Liverpool Corporation tests which shows how leakage increasessharply with a small increase in hole size.
Figure 1.4 Graph to show growth of leakage with time
Experiments were carried out byLiverpool Corporation to determine therate of loss through various sized holesin 0.5 inch diameter lead pipes under apressure of 31.6m head. The results areshown in this diagram.
‘Eruptive’ BurstsPassive leakage
Detection and repair
Complaints
Level
LEA
KAG
E
Time
Average leakage rate(will vary between
extremes dependenton detection andrepair frequency)
Intrinsic leakage level(effectively the
minimum for a givenarea without further
rehabilitation orpressure reduction)
25
AVGFLOWRATE
TOTALQUANTITYLOSTPER BURST
% OF ALL LEAKS
% OF ALL LEAKS
a b c d
a b c d
Bursts which graduallybecome obvious
Small bursts-only foundby detection
Large bursts -immediately obvious
Very smallbursts-
impossible oruneconomic to
find andrepair
Figure 1.5 Likely Proportion of Bursts/Leakage within an Ageing System
26
2.1 Introduction
An understanding of basic hydraulics is essential if distribution data is to be assessed correctly. Thisparticularly applies to the interpretation of flow and pressure data, relating as it does to the internal sizeand condition of a pipe, as well as to legitimate demands and leakage.
It is neither efficient, or necessary, to depend on network analysis to resolve every uncertainty presentedby recorded data. Given an accurate knowledge of the pipe network, a basic grasp of the principles ofhydraulic gradients in particular is very valuable.
2.2 Energy Principles
2.2.1 Water flowing in a pipeline possesses energy in three forms: -
Potential energy due to elevationVelocity energy due to velocityPressure energy due to pressure
2.2.2 Consider a position along a pipeline:
pressure -pN/m2
Then:
Total energy = potential energy + pressure energy + velocity energy
TE = Z +pw + V2
2gwhere
pw is known as pressure head - unit metres
V2
2g is known as velocity head - unit metres
(These factors derive from basic hydraulic theory.)
Total energy (TE) is expressed in metres relative to a given datum.
27
2. HYDRAULICS AND NETWORK ANALYSIS FOR LEAKAGE DETECTION
z
Any referencedatum
Specific weight ofwater = w N / m_(i.e. 9810 N / m_)
Specific weight ofwater = w N / m3
(i.e. 9810 N / m3
It can be expressed graphically on a longitudinal section at a point:
Therefore, three different regimes of pressure can be identified as acting on a pipeline:
• Static pressure - pressure created during no-flow conditions• Working pressure - pressure dependent on the flow varying from static pressure at no-flow
to zero at ‘maximum’ flow• Surge pressure - caused by transient pressure waves
2.2.3 Consider a length of pipe:
Provided that a) the fluid is incompressible (as water is assumed to be) b) there are no energy losses (1) - (2)
then:
Z1 + P1 + V12 = Z2 + P2 + V2
2
w 2g w 2g
This is Bernoulli’s Equation.
It can be expressed graphically as:-
28
V2
2g
PW
Total Energy
ZDatum
P1
V1
Z1
P2
V2
Z2
��
V12/2g
P1w
Z1
V22/2g
P2w
Z2
TELHGL
CENTRE LINE OF PIPE
��
Datum
HGL
29
Where TEL is the Total Energy LineHGL is the Hydraulic Gradient
NB. TEL and HGL are unique to a particular flow in the pipe; change it and they both change.
2.3 Head Loss in Pipelines
2.3.1 Bernoulli’s Equation in the form:
Z1 + P1 + V12 = Z2 + P2 + V2
2
w 2g w 2g
assumes no loss of energy. Although for very short smooth pipes, this may be near enough true, in generalwe need to modify the formula to take losses into account, i.e.
Z1 + P1 + V12 = Z2 + P2 + V2
2 + Lossesw 2g w 2g
- in graphical form:
The effect is to tilt the TEL in the direction of the flow (and the same with the HGL) by an amountdepending on flow, roughness and pipe fittings.
2.3.2 Consider the pipe reservoir and pipeline described in Figure 2.1. At times of maximum flow on the pipeline, what will be the resulting pressure at properties ‘A’ and ‘B’?
The actual pressure will be the STATIC PRESSURE less the HEAD LOSS at each property.
A number of factors are responsible for the loss of head in the pipeline:
a. Losses in entry to pipe.b. Changes in flowc. Restrictions in pipelined. Friction lossese. Burstsf. Internal condition of pipeg. Pressure Reducing Valves
2.3.3 For every part of the distribution system, there will be a level below which the pressure must not bepermitted to drop if an adequate and efficient water supply is to be provided. The minimum pressure mustbe considered when the system is designed or extended. The difference in level between the bottom water
��
V12/2g
P1w
Z1
TEL
HGL
CENTRE LINE OF PIPE
Datum
Losses
V22/2g
P2w
Z2
level in the service reservoir, and minimum pressure point, must at no time be exceeded by the loss of headin the mains from the reservoir due to friction and other causes.
2.4 The Hydraulic Gradient
2.4.1 The pressure head in a pipeline at a point = PW
Where P = pressure N/m2
W = specific weight of fluid
This expression is the same as that for the pressure at the bottom of a column of fluid ‘h’ metres high.
Therefore the pressure in the main would, if there was such a tube connected, force the fluid up it to aheight of ‘h’ metres (which equals P ).
Wi.e. water would rise to the height of the hydraulic gradient.
2.4.2 The Hydraulic Gradient of a pipeline is the gradient of a line joining the fluid levels measured at verticalintervals along the pipe
At a particular constant flow ‘Q’, it is the line showing the pressure in the pipeline between two points.See Figure 2.2. If the flow ‘Q’ increases, there will be an increase in the friction losses and the hydraulicgradient line will steepen.
The friction head loss ‘H’ between two points can be calculated using the Hazen-Williams equation orsimilar types of formula.
The hydraulic gradient can indicate points within a pipeline system where pressure reduction or systemboosting will be required. Figure 2.3 gives an example of this theory.
2.4.3 Where the HGL runs above the pipe, pressures are positive. Where it runs below, pressures are negative(i.e. suction pressures).
The maximum suction lift of a pump, or the greatest syphonic head, is in theory 10 metres for water. Inpractice it is about 7 metres.
Water Mains can run above the HGL providing this height does not exceed about 7 metres, but:
i) Properties above the HGL at a particular flow cannot be supplied.
ii) Connections above HGL lead to back syphonage and contamination risks.
iii) Suction pressures may draw in contamination through hydrants, leaks and air valves.
iv) Suction pressures may disrupt poorly made joints.
Therefore the practice is to be avoided.
2.5 Pipe Flow Formulae
These flow formula are based on observation and experiment, and not on theory. They are, in general,applicable only to clean water, slime free at normal temperatures.
In pipes used for water distribution, the flow may be turbulent. The friction factor depends upon theroughness of the pipe and also the Reynolds Number, which depends in part upon the velocity in the pipeand its diameter.
30
Therefore a pipe flow formula should have a roughness coefficient which varies with velocity and pipesize. Various pipe flow formulae are available to determine head losses in relation to velocity in pipes.Two are described below, but there are others, including Colebrook-White, which is sometimes used fornetwork analysis modelling.
2.5.1 Hazen-Williams
This formula is often used in the design of water distribution systems. It is based on well-documentedrecords of experiments on the pipes ordinarily used in distribution practice, has fairly reliable values forco-efficients, and is easy to use.
Imperial Units
V = 1.318C (D)0.63 (H)0.54
(4) (L)
This can be arranged to:
H = RQ1.85
where R = 14.623 x LD4.87 x C1.85
H = Head loss in ft.L = Length in ft.Q = Flow in galls/min.C = Constant (expression of roughness)D = Diameter in ins.V = Velocity ft/s.
Metric Units
H = RQ1.85
where R = 11.9 x L x l09
D4.87 x C1.85
H = Head loss in metresL = Length in metresQ = Flow in litres/sec.D = Diameter in mmC = Constant Maximum 150
Average 100Minimum 60
V = Velocity in m/sec
2.5.2 Lamont Formula
For hydraulically smooth pipes for both mains and services, Lamont’s smooth pipe formula has beendeveloped. It is intended for use with new pipes carrying clean slime-free water at normal temperatures(55ºF).
V = 95.5d0.6935 (H) 0.5645
(L)
31
V = Average velocity of flow ft/secD = Diameter in ftH = Head loss in ftL = Length in ft
2.6 Pipe Friction Diagrams
The Hazen-Williams formula may be applied to all types of pipe, with careful selection of the value of theconstant C, from available tables. For rapid solution of the formula, a Pipe-Friction Diagram is alsoavailable.
With the increasing use of hydraulically smooth pipes for distribution and trunk mains, a Pipe-FrictionDiagram is also available for solutions using Lamont’s smooth-pipe formula. The chart is for use withnew pipes carrying clean, slime-free water at normal temperatures.
2.7 Network Analysis
2.7.1 Introduction
Network analysis is the term used to describe the ‘analysis of water flows and head losses in a pressuriseddistribution system under a given set of demand conditions on the system’.
A network is the collection of pipes, valves, booster pumps and service reservoirs forming the waterdistribution system. Due to the complexity of most distribution systems, it was normal to simplify thesystem by considering only the key mains. With the development in recent years of computer hardwareand software, it is now possible to include all reservoirs and mains in a distribution system, and all thevarious control features, with their operating constraints and regimes.
The demands and demand patterns on a network are also vital ingredients, and are made up of a numberof components:
a) Domestic demandb) Metered industrial/commercial demand.c) Unaccounted for water including leakage.
2.7.2 Type of Analysis
This is the process of calculating the flows and head losses in a network for a given set of demandconditions. Two types of analysis are normally used:
Snapshot - In a snapshot analysis, the flows and head losses are considered at only a single given set ofdemand conditions. This is frequently expressed as a single time interval.
Dynamic or Extended Time - In each dynamic analysis the flows and head losses are considered for aseries of varying demand conditions. This is frequently a 24 hour time period, and is the sort of analysisthat is now most commonly used.
The power and speed of computing for network analysis continues to improve.
2.7.3 Model Construction
A network model is basically an intelligent mains record drawing - allowing one to access hydraulic dataas well as the position of the mains in the ground. A model represents everything we know about aparticular distribution system. It will have been calibrated by the model builders to ensure that, withinreason, the model gives the same flows and pressures as the real system. This is done by comparing theresults from the model with huge amounts of data from field tests. It is essential to know the system of
32
33
configuration on the calibration day - i.e the day chosen as the most ‘typical’ from the field test. Thecalibration process will find any significant problems with the model’s representation of the distributionsystem, but not all of them. To calibrate a model it is necessary to get the pressures right within one metreat virtually all points in the system at all times of day.
Once the model is created it has to be converted to what is known as an Average Day Model. To do this,the model builder converts the demands on the model to average demands by comparing the demands forthat area with the test day
2.7.4 Network Balance
If there is a disagreement between the computed flows and the measured flows, a number of factors canbe involved. The more common are listed below:
• Incorrect estimates for model demands• Incorrect assumptions for hydraulic resistances• Wrong pipe lengths or diameters• Unsuspected network cross-connection• Closed valves/opened valves• By-passes around PRV or meters• Restrictions in mains• Pressure measurement on ‘rider’ main.
The process of model building can thus uncover many problems which may go unnoticed until a burstoccurs, often wasting time and money.
2.7.5 Model Application
Network analysis is a powerful tool for the effective management of distribution systems. Once a modelexists, it allows any user to experiment with system changes before they are tried out on the ground. Thesecould be such things as checking what reinforcements are needed to supply a new development, so thatlevels of service are not affected somewhere else, perhaps miles away. The model could help maximisethe utilisation of low cost supplies, and in pumped distribution systems, minimise the cost of pumping. Itcould also ensure that levels of service are achieved at customer taps by identifying areas of inadequate orexcessive pressures, and areas of high leakage; corrective measures could then be simulated. It might beused for planning a trunk main shut-off, with effects over wide areas, perhaps to see how long the reservoirstorage will last. It can be used to check on rehabilitation problems – re-line, renew or up-size. It can alsobe used to design pressure reduction, or to alter distribution areas. As the techniques improve it will alsobe used to investigate water quality problems.
Network models can already tell us how old the water is throughout a system and how that changes duringthe day. They can also be used to tell us how different source waters blend in the system at different timesof the day. All these might point to problem areas and show the results on water quality of system changes.
Network models give us a better picture of the system operation, and help improve levels of service. A lotof money can be saved on capital schemes by using models to find out what size mains are really needed,or to sometimes find ways of not laying new mains at all.
Network models may be useful in locating large leaks by comparing modelled pressures against actual.Large leaks cause a lowering of pressures.
Network models are not perfect, but they are the only tool available to provide such detailed hydraulicinformation. In the past we often had to guess about the behaviour of complex systems.
34
TW
L
BW
L
4m
70m30m
B
A
C
20m250m500m
1000m
1.W
hat will be the m
aximum
pressure at property ‘A’
at BW
Lat average flow
?2.
What w
ill be the maxim
um pressure at property ‘C
’at T
WL
at average flow?
Figure 2.1 Service Reservoir and Pipeline
35
Figu
re 2
.2 H
ydra
ulic
Gra
dien
t of a
pip
elin
e
FLO
W Q
TH
RO
’PIP
ELI
NE
LEN
GT
H O
F P
IPE
LIN
E =
L
PIP
ELI
NE
HA
S D
IAd
met
res
FR
ICT
ION
CO
EF
F .
C
WA
TE
R P
RE
SS
UR
EA
TT
HIS
PO
INT
INP
IPE
LIN
E F
LOW
Q
LEV
EL
OF
PIP
ELI
NE
A.O
.D
FR
ICT
ION
HE
AD
LO
SS
HT
HR
O’P
IPE
LIN
E B
ET
WE
EN
(Can
be
calc
ulat
ed
from
Haz
en-W
illia
ms
equa
tion)
GR
OU
ND
LE
VE
L
AB
BA
OR
DN
AN
CE
DA
TU
M
LLoonn
ggiittuu
ddiinnaa
ll SS
eeccttii
oonn oo
ff PP
iippeell
iinnee
The
Hyd
raul
ic G
radi
ent i
s ob
tain
ed b
y pl
ottin
g th
e su
m o
f th
e St
atic
Hea
d an
d th
e Pr
essu
re H
ead.
The
Hyd
raul
ic G
radi
ent s
how
s th
e le
vel t
o w
hich
the
liqui
d in
the
pipe
wou
ld r
ise
if a
ver
tical
sta
nd p
ipe
was
inse
rted
in th
e pi
pelin
eat
the
poin
t und
er c
onst
ruct
ion.
INV
ER
TLE
VE
LO
F P
IPE
LIN
E
36
NO
TE
S
1.D
istribution pressures will be less than Pm
ax (at lowest ground level) due to frictional losses in m
ains, fittings etc.2.
Reduce pressure in the valley bottom
(point A) to reduce leakage.
3.Pressure at point B
must satisfy L
evel of Service criteria.4.
Mains at point C
will need boosting.
P
GR
OU
ND
LEV
EL
RE
SE
RV
OIR
DA
TU
M P
OIN
T- A
.O.D
HY
DR
AU
LIC G
RA
DIE
NT
P
P
CB
A
Pm
ax
Figure 2.3 Hydraulic G
radient
37
3.1 Distribution Network Structure
A wide range of performance standards and regulations apply to the operation of a distribution network.These include statutory requirements, customer standards, water quality standards, public relations andaesthetic needs.
Distribution management is required to meet these standards while minimising pumping costs, controllingleakage, maintaining security of supplies and maximising the return on investment in the assets. Inaddition there is a need to respond to and solve customer complaints of poor or no supply, discolouration,taste and odour, noise and so on.
All these requirements need a geographic reference framework of a manageable size.
The shape, composition and arrangement of the distribution system is dictated as much by the localhistory, topography and Town and Country Planning as by good hydraulic design. If uncontrolled, theincremental growth and integration of local supply areas leads to an ‘open’ distribution system in which itmay be very difficult to meet the performance requirements.
In open systems, water will mix in an unpredictable fashion, pressures will vary, will be excessive in someareas, and costs will escalate. Customers may receive water of variable quality and taste, which can oftenresult in additional complaints.
The alternative approach is one in which the distribution system is separated into manageable units, orzones, each of which has definable characteristics which can be monitored and maintained. In practicezoning takes place at several levels.
The techniques of active leakage monitoring require the installation of flow meters at strategic pointsthroughout the distribution system, each meter recording flows into a discrete district which has a definedand permanent boundary. Typical district size currently in the UK varies between 1000 and 5000properties, although some districts designed around old ‘waste zones’ are smaller, <500-1000 properties.Others, designed around reservoir zones or bulk meter areas, are larger, 5000-10000 properties. In general,the size of zone should depend on the monitoring requirement and the follow-up leak detection techniquebeing employed.
3.2 District Meter Areas
3.2.1 Establishment and Design
Distribution management is an important activity which has considerable impact on customers. The costsof distribution operations are high. It is therefore vital that management decisions are taken in aframework of knowledge and understanding of how the system operates. The development of DMAs aspart of a structured operation of the distribution system allows the network to be operated in a plannedway. This planned approach inevitably leads to better understanding and control of the distributionsystem, updated and more comprehensive records, fewer consumer complaints, and closer control oflabour. In short, more efficient and informed management. Such an approach helps to ensure thatdistribution managers can meet the primary objectives to the maximum benefit of the customer and theWater Supplier.
District Meter Areas are the basic building blocks of a zoned distribution system. They provide amanageable unit by which the distribution customer and performance information can be linked to otheractivities and data systems.
Their fundamental characteristic is that their boundaries are closed except for defined, measured inputsand outputs. Ideally this should be a single metered input, but this is not always achievable in practice.
DMAs in the UK are generally between 1000 and 5000 properties in size, and they have similartopography with limited head loss within their area. This ensures that pressures throughout the DMA are
3. DISTRICT METER AREA MANAGEMENT
38
even, and allows pressure and leakage to be managed most effectively. Larger areas are possible from adetection point of view if acoustic logging is part of the detection policy employed.
The principles of DMA design and structure are very simple. Nevertheless, where possible, the designshould be checked using Network Analysis to ensure that pressures are sustained at all likely demands,that no unnecessarily long water retention periods are created and that water quality variations are withinan acceptable range - larger areas usually means less ‘dead ends’.
System record plans are required, preferably at a scale of 1:2500, together with property count data. Thisinformation is used, together with the local system operator’s knowledge, to define the boundaries of eachDMA. Other important considerations in this process are as:
i) to cross the fewest number of distribution mains (helped by using natural boundaries such asrailway lines, canals and major roads), thereby reducing the number of meters used and the numberof closed valves (which can lead to water quality problems).
ii) to avoid districts with high outflows (this leads to inaccuracy in calculation of district demand asany changes in demand will be a small proportion of the total flow measured).
Having defined the limits of a DMA, it will normally be necessary to trial the area in practice. It will benecessary to ensure that all stop (stand shut, boundary) valves perform correctly, and that satisfactoryflows and pressures are maintained throughout the DMA. In practice, DMAs often have to be checkedvery carefully during establishment. Unforeseen difficulties may be found, such as buried, or closedvalves, or even unknown pipes. These problems are often discovered when the DMA is first modelled andanomalies in the model are investigated..
Once satisfactorily piloted, the DMA can be fully established. This will require:
i. The installation of flow meters at all inlets and outlets.
ii. The closing and marking of all boundary valves.
iii. The installation of flushing, or ‘OXO’, points.
iv. The updating of plans, records and related information systems.
The simple checklist below can be used to ensure that all of these activities are performed before a DMAis commissioned.
District Meter Area Design
1000-5000 properties
Minimum number of boundary valves
Preferably single inlet meter
Avoid export meters if possible
Beware of low pressure (on peak demand)
Beware of quality problems at stop-ends
Avoid l50 mm mechanical (Helix) meters (1 rev = 1000 litres)
Typically downsize mechanical meters (not necessary for electromagnetic)
Install mechanical meters on a bypass
Fit ‘out-reader’ chamber for logger if meter access problems
39
3.2.2 Operation and Maintenance
Once established, DMAs need to be maintained. For two adjacent DMAs, the opening of a single boundarystop valve is sufficient to destroy the accuracy of DMA demand monitoring. A regular regime of meterreadings, boundary valve checks, and pressure monitoring must therefore be established for each DMA.
For leakage control purposes, it is necessary to establish the number of domestic properties, and thedemand of major industrial users within each DMA. This requires regular, usually weekly, reading ofDMA meters and loggers, preferably with the input of the information into a computer analysisprogramme.
Careful inspection of the meter and logger readings can quickly spot any unusual results. This can be usedto trigger leak detection follow-up work. Simple management procedures must be introduced to ensurethat the integrity of the DMA is maintained, otherwise the cost and effort of establishment and monitoringwill be wasted. The following details are worth noting for effective management:
• All boundary valves should be kept tight closed and a regular checking programme should be followed
• All boundary valves should be clearly marked and identified• Valves within the DMA should be fully open• Status quo should be re-established after bursts, rehabilitation or other operational necessity• High pressure DMAs should be examined for pressure reduction• Logger readings of low pressure should be investigated to determine whether leakage is indicated.• Leakage within the DMA, whether visible or not showing, should be repaired promptly• DMA meters should not be valved out• DMA meters and loggers should be operating normally• PRV areas should be properly isolated and operating• Poor quality mains should be fed forward into the capital programme as candidates for renewal• Plans should be up to date and show new property.
3.2.3 Benefits of DMAs
The principal benefit of DMAs is that the key characteristics (e.g. demand, quality, cost) of a well definedarea of the distribution system can be closely monitored. The results of this monitoring allow managementaction to be prioritised and targeted on where it is most cost effective.
Specifically DMA’s impact on:
1. Leakage control
2. Pressure management and levels of service.
3. Asset maintenance and renewal.
4. The monitoring and maintenance of water quality
5. The planning and programming of repair and maintenance work.
Perhaps the most important benefit of DMAs is a little less tangible. Together with a zoned approach todistribution management, they provide a better knowledge of how the system works and how water getsto the customers in an appropriate condition. This allows the Water Supplier to focus attention on thoseactivities which produce most benefits to customers – a pro-active rather than a reactive approach. Forexample, flow reversals and retention times can be minimised and more consistent pressures established.This results in a better knowledge of the system, improved demand management, better and moreconsistent service to customers, all at a lower, long-term cost to the Water Supplier.
40
3.2.4 Disadvantages of DMAsDMAs and zoned systems, in general, do have some disadvantages which must be considered andminimised:
• Less robust under failure conditions. Open systems automatically compensate (up to a point) forchanges in demand patterns. DMAs, on the other hand need to be managed to allow for mainsimprovement, peak demands, loss of supply etc.
• The costs of establishment can be considerable. Not only are meters and data loggers required, butnew and replacement valves may be needed. In some cases, fresh tracing and mapping of thesystem may be necessary. Some of this may be considered to be operational ‘good housekeeping’.
• Water quality may suffer because of the creation of closed systems. Certainly the number of dead-ends can increase considerably when DMAs are introduced. These may, particularly in areas withunlined mains, accumulate debris resulting in discolouration or even blockages. The installation offlushing points and programmes can overcome this problem at a cost. A better long-term solutionis to improve treatment works and mains to improve water quality.
• A substantial commitment is required from management and workforce. It is vital that valves arechecked and meters read regularly, otherwise the information obtained is misleading or useless.This too has a cost, which has to be accepted and budgeted for.
3.3 Waste Meter Areas
For the purpose of leakage monitoring and investigation, some DMAs are occasionally subdivided intowaste meter areas (WMAs) by closing defined valves and measuring flows using portable or fixed wastemeters. These waste meter areas are similar to small DMAs, but their boundaries are not permanent; whenthe leakage work is completed, the valves are left in the open position.
The design of the areas is similar in principle to that of DMAs, although there is not the same constrainton boundary valve closures, as these are only temporary.
3.4 Links to Other Data Information Systems
DMAs are the common link between distribution and other activities. For example:
i) CCuussttoommeerr SSeerrvviiccee
Customer calls which require a visit or follow-up job can be logged by the DMA in which the customerlives, and recorded with address and problem information. Work scheduling and planning proceduresallow appointments and repairs to be programmed by DMA to improve efficiency. Historic DMAcharacteristics or activities (e.g. a burst main) may help understand and explain the customer’s problem.
ii) LLeevveellss ooff SSeerrvviiccee
Levels of service registers can be compiled by DMA. Remedial work, whether operational or capital, canbe identified and programmed on a DMA basis.
iii) WWaatteerr QQuuaalliittyy
Water quality zones are aggregates of DMAs. Sampling and reporting programmes can be built up usingDMA information and characteristics, particularly the sources from which water is normally supplied. Alink to a quality information system would ensure that statutory sampling and reporting requirements canbe met, and that managers are aware of quality variations and problems.
41
3.5 Commissioning
3.5.1 General
For both waste and district meter areas, on fitting the meter the area can be commissioned by closing theboundary valves. Pressures at critical points should be monitored, together with the meter flows, and anyincrease in complaints should be noted. It is preferable to data log the meter flows during commissioningto confirm the range of design flows, and to aid the solving of any problems which are subsequentlyencountered. A typical demand pattern is shown in Figure 3.1. Once it has been established that the areais functioning satisfactorily, the boundary valves should be recorded on the system record drawings, andclearly marked on site.
3.5.2 Establishment of ‘Norms’
As soon as possible after commissioning the area should be surveyed throughout for leakage, and all leaksquickly repaired. Measurement of the minimum night flow should then take place to establish the ‘norm’for the area, against which subsequent readings can be judged. Depending on the resources and technologyavailable, it may take considerable time to achieve a complete leakage survey of each area. In the interimperiod it will be necessary to calculate the ‘norm’ based on the number of properties and an appropriateallowance per property.
42
6 a.m. Noon 6 p.m. Midnight 6 a.m.
Leakage flows Continuously
Night Line
Peak Demand
Legitimate Water Usage
Figure 3.1 Demand Patterns in a Typical Area
4.1 Hierarchy of Metered Areas
In common practice, there are different levels of metering as shown in Figure 4.1.
4.1.1 Supply Areas
Metering of a supply area will comprise metering of all source works outputs, and all imports and exportscrossing the boundaries, in order to give an accurate daily figure for demand.
Metering at this level is essential to judge overall performance as it includes all possible sources ofleakage. However, it is of no use for leakage detection, as any leaks which are not obvious will beswamped by the normal daily variations in consumption.
4.1.2 Zones
Zone metering breaks down a large supply area into several zones, typically varying between 20,000 and50,000 properties. Again all inflows and outflows are measured continuously, including the effect of anyincrease or decrease in storage.
Zones are too large to identify small leakages, as again these will be swamped by normal daily variations.However, they could possibly identify major leakage, especially if daily readings are collected.
Zone metering may also be useful for comparing the performance of different leakage control teams, orfor collecting together data for parts of this system with similar characteristics such as unit cost, age,urban/rural character.
4.1.3 Districts
Within each zone, there will be several district meter areas (DMAs) ranging in size typically from 1000 to5000 properties. In the UK, this would typically mean a population range of 2500 to 12500 and a dailydemand ranging from 0.7 to 3.5 Mld.
District metering may be considered as the first level of metering which can be used for leakage detection,the previous two levels being used for performance assessment and monitoring rather than detection.
4.2 District Metering
4.2.1 Original Concept
The original concept of district metering was to measure the total volume entering the DMA between thereading intervals, and hence to calculate the average daily demand. This would then be compared toprevious readings, and also to the readings for all other DMAs for that period to assess climatic effects. Asignificant increase in demand, not generally reflected across the system, would signify a likely increasein leakage. Normally, a second cycle of readings would be taken to confirm the result before further actionwas taken.
This procedure suffered from a number of disadvantages:
i) it was insensitive as leakage would not be identified until it exceeded a significant proportion ofthe daily demand, normally at least 10%.
ii) the time taken to identify the leakage and initiate further action would be two reading intervals.
iii) it was not possible to differentiate between increases in leakage and increases in meteredconsumption, except for very large consumers whose meters may have been read as district meters.
iv) elimination of climatic factors and holiday effects was difficult, and very much a matter ofjudgement and experience.
4. METERING FOR LEAKAGE DETECTION
43
44
4.2.2 Data Collection
Due to the large numbers of meters likely to be involved, it may not be economic for all these meters tobe on telemetry, in which case data must be collected by site visit.
The frequency of data collection and analysis may itself be limited by the amount of resources which canbe economically justified to undertake this activity. This can be varied with the leakage growthcharacteristics of the area.
However, district meters are now usually fitted with data loggers which will record, in addition to the totalflow, the night flow over a specified period for a number of nights. This immediately achieves a betterthan five fold improvement in the sensitivity of the method in the original concept, as night flows willnormally be less than 20% of the average daily flow and will suffer less variation due to demand. Thetime taken to identify leakage is reduced to one reading interval as the night flow readings will confirmthe leakage, unless it occurred at the end of the period. The effect of climatic variation is significantlyreduced, although care may be needed on occasions when garden sprinklers may be left on overnight.Differentiation between leakage and metered use is easier, as any increase in metered use is less likely totake place at night, particularly at weekends.
Logger manufacturers usually provide powerful software to analyse and manipulate recorded data.
The equipment and economics associated with data collection are changing. Some Water Suppliers arebeginning to move in favour of automated, remote, and centralised interrogation of intelligent data loggersat meters, monitoring pressure as well as flow.
4.3 Waste Metering
This is the fourth and final level of metering. A waste meter measures the total flow into a waste area.The waste meter area (WMA) is specially valved in for the purposes of the test so that it is supplied solelyby the waste meter. The area covered is normally in the range 500 to 4000 properties. Where used incombination with district metering, a single DMA may be divided into several WMAs, sometimes usingthe same meter revalved into different areas. Waste meters are used specifically to record the minimumnight flow rate, this measurement being used to judge whether there is significant level of leakage bycomparison with previous readings or WMA ‘norms’.
Waste metering is now rarely be used on its own in the UK. In conjunction with district metering, it wouldbe termed ‘combined metering’.
Unlike the methods of metering previously mentioned, waste metering is not run continuously. When usedon its own, the waste meters are run at a set frequency. Alternatively, and now more usually in the UK,when used with district metering, they are run only when the district meters indicate a significant level ofleakage.
If it is judged that action is required, waste meters can be used to perform step tests to further locate theleakage within a still smaller area.
4.4 Recent Metering Improvements
The chart below indicates the improvements in flow range that have taken place alongside the evolutionof data loggers. These improvements now enable one meter to read minimum night flows and maximumdaily flows with an accuracy that facilitates leakage monitoring and detection. Even greater accuracy canbe achieved by using electromagnetic meters which can now be obtained at smaller sizes. These metersare becoming more competitive for ordinary use, because they hold out the possibility of reducedinstallation and maintenance costs.
45
MMeetteerr SSiizzee MMiinniimmuumm aanndd MMaaxxiimmuumm FFllooww RRaatteess iinn mm33//hhrr
Meter Type Deacon Kent Gate Kent IM Kent 3000 Kent 4000 Electrommmm magnetic
80 Minimum 0.23 0.15 2.50 1.14 0.50 0.14
Maximum 46 46 55 170 200 181
100 Minimum 0.23 0.21 2.75 1.59 0.60 0.21
Maximum 46 64 86 284 250 283
150 Minimum 0.45 0.47 4.50 3.41 1.80 0.48
Maximum 90 140 209 568 600 640
These relatively recent technological advances have reduced the distinction between district and wastemetering. Previously, waste meters were far more sensitive at identifying leakage than district meters, andthis off-set the reduced monitoring frequency. DMA’s are now as sensitive as WMA’s in identifyingleakage, and often the district meters themselves can be used to perform step tests. Thus, the benefits ofcombined metering (District Meters plus Waste Meters) can be achieved at less capital cost in real termsthan was previously the case.
4.5 Meter Site Selection
Having defined the boundaries, and hence those mains in which the flow must be measured, the next stepis to select the meter site. A site survey is necessary to check the location of the main and other physicalobstructions or limitations. Information on the other Utilities’ apparatus should also be obtained at thisstage to avoid subsequent problems during excavation.
A site may already be committed where an existing district meter is installed, but otherwise a site shouldbe chosen on the pipeline such that access is practicable under all circumstances for meter reading and forrepair and maintenance. Location of the meter in either footpath or verge is preferable because of safetyand accessibility.
Dependent upon the type chosen, an electromagnetic flowmeter requires a mains power supply.
4.6 Meter Installation Design
4.6.1 Mechanical Meters
Mechanical DMA meters should normally be sited on a bypass main which provides the necessaryupstream/downstream lengths of straight pipe to avoid flow disturbance. However, in DMAs supplied byseveral meters, bypasses may not be cost-effective. The criteria should be based on the ability to maintainsupplies when a particular meter is shut out, or on the availability of alternate supplies.
Strainers are sometimes needed upstream of mechanical meters in dirty water areas to prevent meterblockage, but provision for the extra head loss and cleaning maintenance is needed.
Meter chambers should be fitted with vandal resistant lids.
If a mechanical flowmeter breaks down it may need to be removed from the pipeline in which it isinstalled. Depending on the meter type, the breakdown may prevent fluid passing through the faulty meter.There are ways of minimising the consequences of a breakdown by proper design such as providing:
46
• Isolating valves either side of the meter.
• A drain valve to empty the meter.
• Free draining meter pit
• Ease of access to the meter.
• Ease of removal of the meter from the pipe.
• A bypass loop to allow flow to continue during repair.
• Air bleeds to facilitate filling the meter when it has been replaced.
Depending on the type of installation, not all the above precautions for mechanical meters may be deemednecessary. Nevertheless, it is too often assumed that a flowmeter, once installed, will run forever and willperform according to its original calibration into perpetuity. This is certainly not true!
Mechanical flowmeters after periods of use do not retain their original calibration, but in most casescontinue to give a readout, which may be believed or disbelieved depending on the shrewdness andexperience of the observer. Means should be provided to make the checking of flowrate at regular intervalsa routine part of checking the plant efficiency. This may be done by the temporary installation of a ‘mastermeter’ in a specially provided bypass loop incorporating isolation and bleed valves, or alternatively by theprovision of connections for insertion meters, or the provision of access for the retrospective fitting of‘clamp on’ ultrasonic flowmeters.
4.6.2 Electromagnetic Flowmeters
It can be argued, that because of the maintenance-free nature of these meters, it is unnecessary to have abypass arrangement, though valves on either side are advantageous.
They do not need an upstream strainer, and are suitable for installation in meter pits which becomeflooded. More significantly, they can be installed without a chamber altogether. The meter can be buried,with the sensor cabled back to a transmitter/display unit sited in a convenient location. This can result insignificant cost savings, and offsets the higher costs of the meter compared to the mechanical type.
Verification of the calibration of electromagnetic meters can be very straightforward, one manufacturerclaiming it takes less than half an hour to evaluate the status of the complete system (i.e. transmitter, sensorand interconnecting cables), with no need to gain access to the pipe.
The need for re-calibration of good electromagnetic meters is a rarity.
4.7 Meter Selection Criteria
4.7.1 General Criteria
The type of meter which can be considered will depend on the type, size and configuration of the district.For instance, if the district is a separate supply area with its own service reservoir or water tower storage,the incoming meter flow is likely to be continuous at a fairly uniform rate. The meter, during the normalworking day, recording peak demand to the area supplemented by backflow from the reservoir, and atnight, providing much lower flow for use whilst at the same time replenishing the reservoir for the nextday. A meter on the reservoir inflow/outflow is also required.
The other extreme would be a meter reading the daily demand with no storage on the system when, in theabsence of substantial leakage, the demand in the early hours of the morning might be very low indeed.
47
In the latter case the prime considerations would be to choose a meter sensitive enough to record very lowrates of flow, which otherwise would be unmeasured. The use of network analysis to identity currentproblems, to check the effect of the flows at the proposed valve closures, and to give an indication of theflows at the proposed district metering points, is desirable but not essential.
If the existing flow information is inadequate, a flow survey must be carried out. This can be done eitherby the use of an insertion flow meter, or by installing a full bore meter which can subsequently be changedif found to be incorrectly sized. The method chosen will depend on the size of the main and thedimensions of the proposed meter chamber.
The internal condition of the pipeline may be a significant factor in the selection of a suitable meter. Thecondition may be known from previous records if recent repairs or alterations have been carried out. It ispracticable to examine the inside of the pipe with an endoscope but this may only be worthwhile if seriousdoubts about the internal conditions exist. Removing a section of pipe for inspection may be morevaluable where there is concern.
District meters of the size likely to be encountered (less than 300 mm, and usually uni-directional) aresomewhat easier to calibrate than larger meters. It is practicable to do this in a workshop with testingfacilities, or on site with the discharge from the pipe beyond the meter being registered through acalibrated check meter.
4.7.2 General Specification
All meters specified should conform to the following basic characteristics:
i. Will be specified by flow (Qn), class and type, not by diameter.
ii. Will be capable of providing a pulsed output to an agreed standard specification which will beavailable without disturbing certification seals.
iii. Larger meters will be maintainable in situ by the removal and replacement of the measuringelements.
In order to specify a meter for new installations it is necessary to establish:
i. The maximum flow required, either actual or assessed.
ii. The minimum flow.
iii. The average flow (m3/day) calculated from the periodic volume divided by the number of days.
Meters should be selected to measure at least 95% of flow at + 2.5% accuracy.
Meters should be selected to ensure that at least 95% of the flow is above Qt and preferably below Qmax.In general the average flow and most (60%) of the volume should be measured between Qt and Qnom.
All DMA meters should have the capability of electronic adaptation to provide logging, remote read-outand integration. All these requirements can be met by a battery, but it is worth considering mains powersupply if it is readily and cheaply available.
4.7.3 Types of District Meter
In theory, a wide variety of meters may appear to be suitable for use as district meters, but by definition,a district meter is usually situated where it is remote from a normal working base, possibly difficult foraccess (i.e. in the roadway), and usually without a supply of electricity.
For this reason, these meters have almost always, in the past, been of helical vane mechanical type. Given
that the environmental conditions are suitable and that these meters can be maintained adequately, they areideally suited to district meter use.
If the condition in the distribution system is not gritty and the water is clear and free of suspended matter,the helical vane type of meter is cheap and accurate over wide ranging flows and provides an easy andcheap solution to district metering.
Occasionally however, due to reverse flow characteristics or the presence of suspended matter, mechanicalmeters may not be suitable.
Electromagnetic flow meters are a viable alternative to the helical vane type, and require virtually nomaintenance. If a power supply is unavailable, they can be supplied to operate from a long life disposablebattery. At present, and assuming continuous full use, battery life is about three years. Provision must bemade for a routine battery replacement.
In exceptional circumstances, it may be impracticable to shut down the supply, and in this situation the useof a retrofit ‘time of flight’ ultrasonics meter, or an insertion-type point velocity probe, properly calibratedfor the local conditions, would provide an alternative, but at reduced accuracy.
4.8 Mechanical Meters - Helix (Woltmann)
4.8.1 Description
Inherently only suitable for water metering applications, meters in this category are most numerous,particularly in water distribution networks.
In this type of meter, the measuring element takes the form of a helical vane mounted centrally in ameasuring chamber with its axis along the direction of flow. The vane consists of a hollow cylinder withaccurately formed wings. Water flow directed evenly onto these vanes will cause rotation which istransmitted to the undergear of the meter register by means of a ceramic magnetic coupling. This rotationis proportional to the rate of flow. Electrical transmission of this information is normally achieved by asuitably positioned magnetic reed-switch actuated by a rotating magnet in the meter register.
This type of meter has now had many years of use and development in the water industry and has theadvantages of requiring no power, low capital cost and ease of maintenance, but has the disadvantage ofmoving mechanical parts which will wear, resulting in degraded low flow performance, and necessitatingregular maintenance/ repair. Grit and particulate in the line can cause deterioration in performance,particularly at the lower end of their range, by damaging bearings and seals. During mains rehabilitationin particular, abrasive material can be passed through the meter . The passage of air, also a feature of mainsrehabilitation, can cause these meters to over-speed, again with the potential to damage seals and bearings.
4.8.2 Operational Requirements
LLooccaattiioonn - The mechanical flow transducer should be installed in process pipework which is free ofvibration.
OOppeerraattiinngg HHeeaadd - Mechanical flow meters in liquid service should operate with sufficient head to preventcavitation and avoid the resulting errors or damage.
IInnssttaallllaattiioonn PPrraaccttiiccee - Accuracy and repeatability of mechanical meters is especially dependent uponupstream and downstream piping arrangements.
Often the bore of the pipework will be a size less then that of adjacent pipework, because of the low flowcharacteristics. If so, the pipeline configuration shown in Figure 4.2 should be adopted.
Mechanical meters should be installed so that they have a positive head of liquid upstream. This headshould be equivalent to at least twice the anticipated pressure drop through the meter.
48
Care should be exercised in the installation of flanged meters to see that the pipeline gaskets do notinterfere with the flow pattern by protruding into the flow stream.
The pipework should be carefully aligned before fitting the meter, and the meter fitted with the directionof flow arrow correct, as performance characteristics are different for forward and reverse flow.
The meter should also be installed horizontally, and with a ‘pulse unit’ attachment for data loggerpurposes, although this is not required with some types of logger.
4.8.3 Maintenance Considerations
CClleeaanniinngg aanndd PPuurrggiinngg - The meter should be installed only after the process pipework has been cleanedand flushed. In areas with a history of dirty water problems, consideration should be given to the use ofstrainers upstream to prevent foreign matter from damaging the device or blocking the flow passages. Ifstrainers are used, they should be cleaned after flushing, and periodically during operation.
BByyppaassss PPiippiinngg - The need for bypass piping is determined by the application. It may be necessary toisolate and disassemble the flowmeter for maintenance purposes. Some of the conditions which maynecessitate disassembly of the meter are (a) damage caused by foreign material, (b) wear, or (c) solidsbuild-up. In continuous service applications, where shutdown is considered undesirable, a bypass must beprovided to permit process operation while the meter is being changed.
If bypassed, the meter should be in the main run and the bypass should be line size and placed at least 10diameters upstream and 5 diameters downstream of the meter.
If a conventional bypass arrangement is impractical, consideration should be given to making provisionfor a temporary bypass. This would involve the insertion of two hydrants either side of the meter, with avalve in between. It should be noted that helical vane meter internals can be removed without taking themeter out of line.
4.9 Electromagnetic Flowmeters
4.9.1 Description
The basic principle of operation of this type of flowmeter is based on Faraday’s law of electromagneticinduction which states that if an electric conductor moves in a magnetic field, an EMF is induced whoseamplitude is dependent on the force of the magnetic field.
In this application, the conductor is the liquid being metered and it is the fluid velocity that is beingmeasured. As the pipe is full at the point of measurement, velocity and flow rate are directly related.
The meter is effectively non-intrusive, giving the advantage of negligible pressure drop, and is inherentlymaintenance free. Its major disadvantage has traditionally been its relatively high power requirement,which necessitates the provision of a mains power supply. However, this situation has now been resolvedwith the development of new technology. The meter is inherently bi-directional and suitable for both cleanand dirty water applications.
The initial disadvantages of this type of system, associated with interfering voltage pick-up, high powerassumption and ‘zero drift’, have now been largely overcome.
FFeeaattuurreess - Accuracy of the magnetic flowmeter is typically + 0.5 % of full scale although + 0.5 % of actualflowrate is available from some manufacturers.
Operational flow range of more than 1000:1
Since this type of meter tends to average the velocity profile between the electrodes, neither long runs ofpipe (up or downstream) nor flow straighteners are needed unless percent of rate of accuracy is required.
49
50
There is negligible pressure drop.
A large variety of sizes are available (25mm to 2500mm or even larger).
Battery powering is now possible.
4.9.2 Operational Requirements
Manufacturer’s instructions should be followed where maximum accuracy is required, but a minimumupstream straight length of 5 pipe diameters and a minimum downstream straight length of 2 pipediameters is recommended.
EElleeccttrriiccaall GGrroouunnddiinngg//EEaarrtthhiinngg:: The importance of proper grounding cannot be overemphasised It isnecessary for the safety of personnel and for satisfactory flow measurement. The manufacturer’sinstructions on grounding and jumper arrangement should be followed carefully. Piping should always begrounded. A continuous electrical contact to the same ground potential is necessary between the flowingliquid, the piping and the flowmeter. This continuous contact is especially important if the conductivity ofthe liquid is low. How this contact is achieved depends upon the meter construction and the type ofadjacent piping (unlined metal, lined metal, or non-metallic ). Jumpers from the meter body to the pipingare always required. If the meter is installed in non-metallic piping, it is always necessary to make agrounding connection to the liquid. This connection is achieved by means of a metallic grounding ringbetween the flanges, unless internal grounding has been provided in the transmitter. This groundingconnection is extremely important and must be done as recommended if the system is to operate properly.
As development has continued, some manufacturers’ meters now have an in-built grounding/earthingelectrode, thus eliminating the need for grounding/earthing flanges
CCaatthhooddiicc PPrrootteeccttiioonn:: If the detector head is installed in a system that is cathodically protected, specialprecautions should be taken to ensure that:
• current at supply frequency does not flow through the liquid in the detector head.• any current at supply frequency flowing through the body of the detector head does not exceed 10A
rms.
These precautions will limit the magnitude of any resultant spurious magnetic fields.
The magnetic flowmeter responds only to the velocity of the flow stream and, therefore, is independent ofdensity, viscosity and static pressure.
IInnssttaallllaattiioonn AArrrraannggeemmeenntt -- GGeenneerraall:: The magnetic flow transmitter tube may be installed in any position(vertical, horizontal, or at an angle), but it must run full of liquid to ensure accurate measurement. Ifmounted vertically, flow should be from bottom to top to assure a full pipe. When mounted horizontally,the electrode axis should not be in a vertical plane. (A small chain of bubbles moving along the top of theflow line could prevent the top electrode from contacting the liquid).
Stress on the flow tube must be avoided, so adjustable couplings are helpful, and flange nuts must only betightened to the specified torque.
EElleeccttrriiccaall IInnssttaallllaattiioonn - The measurement signal generated in most flow tubes is in the order of 1mV andsource impedance often exceeds 1Mohm. Consequently, care must be taken to minimise electricalinterference. In locations subject to high ambient electrical noise, consideration should be given to the useof an integral flow tube/transmitter.
Electrical connections between the flow transducer and a remote converter or receiver unit should notexceed the maximum distance permitted by the manufacturer.
51
4.10 Insertion Velocity Probes
4.10.1 Description
It is often desirable in flow studies and survey work to be able to measure the velocity at a point withinthe flow pattern to determine either mean velocity or flow profile.
In normal applications, such a device would be inserted through a gate valve assembly on the pipeline,hence it can be installed and withdrawn under pressure without disruption to supply, with minimum costs,and can be precisely located for carrying out a flow traverse. Changes in flow during the traverse, andincorrect positioning of the probe, are possible sources of error. However, it is possible, though notnecessarily recommended, to use such devices as low cost permanent flowmeters, in which case they mustbe positioned to monitor mean velocity.
The following types are most common:
• Electromagnetic• Turbine
The electromagnetic device is basically an inside-out version of the electromagnetic full-bore device. Thevelocity probe consists of a cylindrical sensor/probe shape which houses the field coil and twodiametrically opposed pick up electrodes.
The field coil develops an electromagnetic field in the region of the sensor and the electrodes pick up avoltage generated which is proportional to point velocity in the vicinity of the probe.
This electromagnetic probe is a solid state competitor for the insertion turbine, with the obvious advantageof no moving parts, and hence no wear or blockage problems.
The probe is bi-directional, and operates with an accuracy of plus or minus 2% of the flow, or plus orminus 2mm/sec, whichever is the greater.
The operating principle for the insertion turbine device is the same as for a full bore pipeline flowmeter.It consists of a rotor, approximately 20mm in diameter, which is housed in a protective rotor cage andmounted on the end of a supporting insertion rod.
More recently, an ultrasonic insertion probe has been developed. Two transducers are inserted into the pipethrough a single entry point, and mounted close to the pipe wall. Initial trials have been promising, withgood accuracy being demonstrated. This method reduces problems associated with blockage, andeliminates errors due to flow changes.
4.11 Domestic Revenue Meters
4.11.1 Positive Displacement
Positive displacement meters are used to determine totalised flow, and measure actual volume of waterpassed in a given time by dividing the flow into discrete volumes, summing the volumes as they passthrough the meter. A very accurate device in operation, it is best suited for use as a revenue meter oncustomer’s supplies. Its mechanical nature and inherent high pressure drop make it unsuitable for mostdistribution metering applications.
4.11.2 Fluidic Oscillation
A recent development in small revenue meters, without any moving parts. It uses a transducer to pick upelectronic pulse signals, which relate to the flow velocity of the oscillating water jet. The meter is immuneto particles, is unaffected when air is present, and, because it has no moving parts, it promises highaccuracy throughout a very long life.
52
Simple battery replacement (used to power the electronics) every ten years is the only maintenancerequired.
This meter, developed by a UK company within the Severn Trent Group, should have significant impactin the domestic/small revenue meter market.
The same company also offers an automatic, meter reading system, which will work with all commonlyavailable meters, both encoded and pulse output. Meters can be read visually, or by a touch pad, or via aradio signal, and the data downloaded directly into a computer.
53
M
M
M
M
Intake andTreatment
Works
Source Metermeasures Total
Output
Meter measuresflow into
Operational Contro lSystems
District metermeasures flow intoDistrict Meter Area(DMA) e.g. 2000
properties
Waste metermeasures flow intoWaste Meter Areae.g. 700 properties
M
Each Source isMetered to theSupply areawhich is
Subdivided intometered
Operational ControlSystems (O.C.S.) whichare each
Subdividedinto metered
District MeterAreas (DMAs)These are thekey ‘BuildingBlock’ units
(which may besubdivided intoWaste Meter Areas (WMAs))
In which theleak can finallybe located
Figure 4.1 Division of Water Distribution System
54
FLO
W
Valve
Strainer
Min 5 dia
Straight pipe
Min 2 dia
Straight pipe
Helix m
eter
Figure 4.2 Mechanical M
eter Installation
Main
Main
FLO
W
55
5.1 Detection Principles
Generally speaking, routine leak location is dependent upon sound being generated by water escapingfrom the pipe. Water leaking from a pressurised main emits sound over a range of frequencies andproduces a hissing noise. The particular distribution of frequencies produced by a leak is specific to thatone particular leak and will depend upon such factors as the nature of the leak, size of the orifice, pressure,pipe material, nature of the ground into which the leak is discharging, or whether that ground iswaterlogged. The sound so produced will travel through the pipe, at a velocity which depends upon boththe characteristics of water and the pipe material, and could also travel through the ground surrounding thepipe. As the sound travels away from the leak its character changes slightly, since higher frequencies willbe attenuated with distance, and other frequencies may be amplified due to the presence of cavities or otherburied underground equipment. The leak noise detected therefore, will depend upon the position at whicha sounding is made. Furthermore, the position of highest sound intensity is not necessarily the positionnearest to the leak.
Not all leaks produce a detectable noise, and some are inaudible to the human ear.
5.2 Stethoscopes (‘Listening’ or ‘Sounding’ Sticks)
This is the traditional method employed in the Industry. They are passive devices, transferring the leaknoise to the ear with a minimum of attenuation. They are widely used, easy to employ, but entailconsiderable skill in their use, and are limited by the performance of the human ear, which introduces moresubjectivity. Training and experience are needed to give the best results.
Stethoscopes can be used for either direct or indirect sounding, and are available in a variety of woods andmetals. The more professional-looking devices may be aluminium tube with an ebonite ear-piece. Someare collapsible with short stems, which enables them to be conveniently carried in the pocket and can alsobe used for sounding on top of valve keys. Many operatives, however, appear to get on just as well withtheir standard stopcock key which is an iron rod about 1 metre long. Experience, knowledge of the areaand consumers, and intuition, also play their part.
The object, of course, is to identify the position of maximum sound intensity. This may not be easy,especially if there are large lengths of pipe without fittings, leaving more distance for the more difficultsurface sounding. Interference from traffic noise, and non-metallic pipe materials also hinder location. Atworst, and fortunately not commonplace, an interactive procedure involving dry hole excavations mayensue before the leak is found.
5.3 Electronic Sounding Devices
Where leaks produce a sound that is inaudible to the human ear, or where the leak noise is low, or thebackground noise is high, electronics can help.
These devices usually consist of a microphone, amplifier and frequency filters, seeking to amplify the leaknoise whilst seeking to filter out the amplification of extraneous noise. Frequency filter selection facilitiescan be a great help in this regard, but sometimes the unwanted noise may have a similar frequency to thatof the leak, and not all ambient noise can be isolated out.
The output of the amplifier can be fed to headphones, to a loudspeaker, to an indicating meter, or to acombination of all three. An indicating meter will display a measure of the total sound intensity receivedby the microphone, and hence is a more sensitive method of determining the position of maximum noiseintensity than the stethoscope, especially if the latter’s operator is unskilled.
To further assist the electronic device, different microphones are normally available depending uponwhether the sounding is done on metallic pipe and fittings, or on non-metallic pipes, or on the groundsurface.
Electronic devices, with ground microphones, are sometimes used as routine tools to survey areas ofsuspected leakage, and also as a final confirmation of leak position detected by a correlator.
5. DETECTION EQUIPMENT
56
Experience is still required in the operation of electronic detectors, and because of this dependency onoperator skill and frequency of use, comparisons between individual detectors, and even between detectorsand stethoscopes are not readily made. However, from a purely theoretical point of view, the followingconclusions might be expected by comparing the performance of stethoscopes with electronic devices andcomparing one device with another.
i. All instruments give better response when direct sounding on metallic pipelines than on non-metallic pipes or the surface.
ii. Electronic devices are likely to be more effective than stethoscopes in situations when the leaknoise is low or where the background noise is high - they can discriminate signals inaudible to thehuman ear.
iii. Electronic devices with separate microphones for direct and surface sounding are likely to be moreeffective than devices with only a single microphone (normally for surface sounding).
5.4 The Mobile Advanced Step Tester (MAST)
The primary function of MAST is the rapid identification of high consumption areas within ‘waste zones’via the basic concept of traditional ‘step testing’ methods, but with reduced manpower and time.
It has been usual for ‘step tests’ to be performed using the ‘man at the meter’ method, with one manmonitoring the meter, whilst usually two men operate the valves. Changes in flow are relayed to theoperators via the voice radio network. The introduction of wax charts and then data loggers eliminatedthe necessity for the ‘man at the meter’, and provided greater accuracy. However, the operators were notaware of the flow changes until the end of a test, or the next day. A major disadvantage of this was that ifhigh consumptions had been identified during the first part of a test, the further time spent closing valveswas unnecessary, expensive and caused extensive disruption to customers.
The next development was to have electronic on-site flow readout at the meter, but that again necessitateda man back at the meter to judge immediate response.
MAST combines these methods and relays the information constantly to the valve operator (as with the‘man at the meter’ method) whilst providing the accuracy of a data logger.
The 2-part system comprises a logger/transmitter and display/receiver. The logger/ transmitter is attachedto the incoming zone meter. This unit collects data at a user selectable time interval (typically 1 minute)and transmits this information to the receiver which displays the flow rate immediately for operationinspection.
By moving through the zone and operating the test valves, the operator is always aware of the flow status,and the loss associated with these valves is immediately indicated by the changes in flow.
5.5 Leak Noise Correlator (LNC)
This instrument does not directly seek the point of highest sound intensity, but a consistent noise source.It is relatively unaffected by background noise. It has two sensors, or accelerometers (transducers), whichare placed on fittings such as hydrants or valves on either side of the leak, and uses the technique of cross-correlation to determine the difference in time between the leak noise reaching the two sensors. The noiseis converted into electrical signals, and a comparison made of the signals, searching for similarity. Whenthe correlation is achieved, a display shows a typical high amplitude peak. The time delay for this peakto be produced is measured by means of a calibrated time trace. Figure 5.1 explains the simplemathematics involved in the correlator calculation.
One of the parameters used in the calculation is the velocity of sound within the particular pipe material.The apparent velocity is affected by the pipe material. As a pipe becomes ‘softer’, the apparent speed ofsound within it becomes slower. It is also affected by the diameter of the pipe.
As leak noise travels away from the leak, down the pipe wall in both directions, it is attenuated by thematerial of the pipe. In the case of softer, plastic materials, this attenuation results in the absorption of the
57
higher leak noise frequencies, such that the further away from the leak that the noise travels, the more itis dominated by its lower frequency component. In early correlators, correlation was difficult on non-metallic pipes because sensor technology was inadeqaute to correlate at low frequencies. The latestcorrelators incorporate sensors with an extended low frequency response, which has enabled correlationto be performed on all types of pipe material.
Current instruments now do all the mathematics themselves, and automatically give the leak position, alsocalculating the actual speed of sound in that particular run of pipework if necessary, thus enabling theoperator to build up a picture of the actual pipe velocities in a particular area, and to work where pipedetails are not known.
The development of the LNC has continued. The new generation are smaller, easier to use, have extendedoperation range due to more sensitive sensors, and some have the signals digitised at source. They do notrequire a dedicated vehicle, and can be operated by one man because the signals are transmitted by radio.This has made the LNC more flexible, and it has become the ‘workhorse’ of leak pinpointing.
The use of hydrophones (sensors coming into direct contact with the water via standpipes, hydrant outletsetc.) enables correlation on sounds of lower frequency and lower intensity, and at a greater distance. Thisis of particular help with plastic pipes and in rural areas.
Figure 5.2 indicates how important it can be to know the actual route of a main between the sensorpositions.
In conclusion, it may be said that the LNC offers:
• Accurate leak location in high ambient noise• Location of leakage with relatively low acoustic output• Location within systems containing few fittings for direct sounding• About 90% trial hole accuracy with consequent excavation/backfilling/ reinstatement cost
reductions• No need for an educated ear (though machine use training is very important).
5.6 Leak Noise Loggers
5.6.1 Introduction
Noise (or acoustic) loggers are probably the most significant innovation in leak detection since thecorrelator. They help to reduce leakage levels and operating costs simultaneously by facilitating themonitoring, or surveying, of large areas quickly and effectively, and with much reduced manpower,compared with traditional sounding stick use (sometimes known as ‘stop tap bashing’).
Noise loggers usually operate during the night, at the time of lowest background noise and highestpressure, and need only receive a leak signal at one sensor. This makes them a more effective survey toolthan the correlator. In addition, they are considerably more sensitive than the human ear, and facilitateidentification of the low level noise often associated with leaks on larger diameter mains, on plastic pipes,and in low-pressure areas.
They have been proven in urban areas where leakage is difficult to detect even by correlator use, and alsowhere the local mains system is complicated, making it difficult to track down leakage sources. They arenow an integral part of leak detection methodology, and promise greater cost-effectiveness in the furtherlowering of leakage levels now being pursued by the UK Water Suppliers.
5.6.2 Use as a Temporary Survey Tool
This version is used to survey an area, or particular ‘trouble-spot’, and is usually used in clusters of 6.More recently, economies of scale have lead to accumulation of 5 or even 10 sets together, enabling 30 or60 to be deployed at one time. Coupling to metal fittings (valves, hydrants, etc.) is magnetic, althoughhydrophones can be used to increase sensitivity if a ‘wet’ connection is available. They collect up to 2
58
hours of noise data at one-second intervals, and can be moved around to other locations to obtain severalnight’s data, if required, before analysis.
They are analysed by powerful electronic software, the presence of a leak being indicated graphically bya well-defined, consistent noise peak. Whilst correct interpretation of the presented data is crucial, theiruse does remove some of the individual, subjective assessment of a noise.
They do not precisely locate a leak, but they do ‘localise’ it, and are particularly helpful where the age andcondition of valves, or the general network condition/status, rules out step-testing as a leak localisationmethod. Once removed from the fitting, they can be immmediately interogated by a portable PC. Theyare an alternative to step testing, if used by skilled and fully trained personnel, and with the addedadvantage that they do not disturb the operational system.
Unlike correlators, these loggers do not rely on the same noise frequency arriving at two points with ashort time delay. They listen over a much longer period for the constant source of noise generated by aleak. By comparing the sound level and spread recorded at each logger, the user can identify theapproximate location of the leak and then focus attention on this section. By recording over a two-hourperiod, they are suited to busy night-time areas where traffic noise remains a problem, and wherelegitimate water demand continues. They can be easily deployed, being pre-programmed to operateautomatically, requiring no specialist labour or night work, and without the risk of equipment theft that canbe associated with correlator use in busy areas.
5.6.3 Use as a Permanent Monitoring Tool
This latest development of the version and principles described above is another major leap forward indetection technology. It has the potential to cause a major re-think about how to monitor areas for leakage,as well as localise it. Already it is proving that it is possible to lower leakage levels further than everbefore, and at reduced operating costs.
These loggers are again installed at fittings via a simple magnetic coupling, taking about 5 minutes to doso, but are battery powered for up to ten years, with no maintenance requirement, and no problems frombeing immersed in water. Their installation and function do not interrupt supplies, or affect the customerin any way, as they are non-invasive. Proximity to loud noise sources, such as PRVs, or continuous systemdraw-offs, is best avoided. The separation distance between loggers depends primarily on the pipematerial, with plastic materials requiring closer spacing than metallic.
Each unit is ‘intelligent’, and adapts itself to its environment. If no leak is present, a radio signal istransmitted to indicate normal background conditions. However, as soon as a possible leak is detected, theunit enters an alarm state and transmits a radio signal to indicate a ‘leak condition’. Signals are receivedby a module which can be mounted in a patrolling vehicle, or can be easily hand-held. This ‘receivingmodule’ analyses and ‘homes in’ on signals to identify the location of units indicating a ‘leak condition’,and thus the approximate position of a likely leak. Data reception is confirmed audibly, and an LCD screendisplays the leak characteristics against the logger identification number and location. The information isstored in the module’s memory, and can be printed out or downloaded to a PC, enabling correlation workand precise location to concentrate solely on suspect areas.
Once a leak has been repaired, the logger(s) that identified and localised it will recalibrate automaticallyso that, the next time that stretch of main is patrolled, it will not flag up that particular leak. There is noneed to re-programme a logger once it is permanently installed.
No night work is required for this method of monitoring/surveying, since patrols can be done during theday by one person. Providing a vehicle can pass within about 50 metres of the logger, whilst not exceeding30 miles per hour, the driver need not leave his transport, or stop, since the module will receive data whilstmoving. This is obviously helpful when ‘patrolling’ some areas, and saves a lot of time in the datagathering process. Additional surveys to check the effect of repair work are equally simple and quick.
59
It can be seen that this piece of technology offers the possibility of continuous, permanent monitoring forleakage for 100% of a distribution system, or, alternatively, just for those parts that are known problemareas.
5.7 Non-Acoustic Equipment and Techniques
5.7.1 Introduction
The techniques described below do not depend upon leak noise. These techniques are seldom used, the‘cut and cap’ method in particular being very inefficient and expensive. It is only by good fortune that aprecise leak location can be made using these methods.
5.7.2 Ground Probing Radar
A new development with little track record in the UK. It uses radar signals and electronic imaging of thereflected signals to locate underground leakage. It will not work with water saturated soil, and awaitsvalidation as an every day, cost effective alternative to acoustic methods.
5.7.3 Gas Tracer Technique
In this method, a non-toxic water soluble gas is added to the water supply in the area of suspected leakage.Bar holes are then made along the line of the main at regular intervals and a hand-held detector, sensitiveto the gas, inspects each hole for the presence of the gas which will have come back out of solution as itescaped from the leak. The gas used, has in times past, been nitrous oxide, but it is now preferable to usesulphur hexafluoride, or hydrogen. The technique is more suitable for rural mains and trunk mains wherethe absence of fittings prevents the use of normal sounding techniques, and bar holes can be easily made.
5.7.4 Cut and Cap Method
This ‘last resort’ technique requires no special equipment. The suspected main is isolated from all otherconnections, and water is supplied through a meter. The main is cut and capped in the centre. If the flowstill continues, the leak lies between the meter and the end cap; if the flow ceases, the leak is downstreamof the centre capping.
The process is repeated by subsequent section division. Its expense is obvious.
60
Figure 5.1 Leak NoiseCorrelator Calculation
The figure shows a length of main which contains a noise producing leak with the microphones, A & B, placed eitherside. The unknown distance of the leak from microphone A is small ‘a’ and the total distance between themicrophones, ‘L’.
The time taken for the leak noise to reach A = av
and the time taken for it to reach B = (L - a)where v=velocity of sound in the pipe. v
The difference in time to reach the two microphones (t) = (2a - L)v
Re-arranging, the position of the leak is given by a = (tv + L)2
A B
DelayLine
Leak noisesource
L-aa
Correlatorinput A
B
av
L-av
Correlatoroutput
t
t
Increasing time delay
Time
Figure 5.2 Sources of Error in Distance Measurement
Leak noise travels along the length of the pipe. When using the correlator, it is important to know exactlyhow the pipe runs, in order to measure the total length of the pipe. Here are two situations in whichmistakes are often made in assessing pipe length.
a) Pipe with many Bends
b) ‘T’ Sections
61
SENSOR
SENSOR
PAVEMENT
FITTING
FITTING
TRUE PIPE SHAPE
SENSORSENSOR
PAVEMENT
FITTINGFITTING
PAVEMENT
ASSUMED PIPE LENGTH
ACTUAL PIPE LENGTH
62
6.1 Introduction
Monitoring of leakage on trunk mains is notoriously difficult and inconvenient, and the precise locationof such can be very time consuming. It is much smaller in total volume than that occurring on thedistribution system, often taking the form of sudden eruptive bursts due to the high operating pressures,demanding prompt attention.
However, because trunk mains and aqueducts are vital and expensive assets, proper preventativemaintenance and inspection procedures are essential. These programmes will include:
• Surface inspection of the line of aqueduct.• Regular checking of cathodic protection systems.• Comparisons of input and output meters on the aqueduct and a mass balance assessment of water
flows (‘Trunk main DMAs’ are sometimes feasible).• Check operation of all key valves (and inspection for gland packing etc.).• Internal inspection of the aqueduct, using CCTV cameras as necessary, to check internal corrosion.• Excavation down to the pipeline every few years to check for graphitisation and other external
corrosion.
The simplest and most common way of looking for trunk main leakage is to walk the length of the main,looking for signs of water, changes in vegetation growth, illegal connections, or any other tell-tale signs.Marshy ground, permeable ground and roadways do not help such a search.
The reality of the situation is that in general a ‘passive’ policy is often applied to trunk main leakage,relying on the usual high pressures to make a significant leakage obvious to farmers or the public who willthen notify the Water Supplier.
Trunk mains usually have a number of large meters associated with them, at treatment works, pumpingstations, reservoirs, offtakes, but even so may not be comprehensively covered. Additionally, the metersmay not be sufficiently accurate to give confidence regarding discrepancies caused by leakage, especiallywhen a number of measurements have to be aggregated. However, such an approach can, by a study oftrends, indicate likely leaks as they occur.
Despite the drawbacks involved in all the methods described, it is very important to keep monitoring trunkmains and their leakage. Neglect can mean wasteful and dangerous leaks go unnoticed until they becomecatastrophic and threaten supplies to large areas. UK experience shows that lack of identification andmaintenance of ageing valves and fittings can cause serious embarassment when things go wrong.
The first three methods of measurement referred to below only give the approximate positions of leaks.Sounding, gas tracer techniques, infra-red photography and leak noise correlation are all possible meansof more precise location. Only the last two of these are mentioned since the other methods have previouslybeen described.
6.2 Meter on Bypass
The easiest method of measuring trunk main leakage is to close two valves on the line, one upstream andone downstream. 25mm tappings are made on either side of the upstream valve, and a small meter,typically a 25mm semi-positive displacement flow meter, is connected between the two tappings. Anyleakage on the section under test will registered on the meter. The advantage of this method is that itutilises equipment that will be available. Disadvantages are, firstly, that the main has to be taken out ofservice, and secondly, if either the upstream or the downstream valve is letting-by a false measurement ofleakage will be obtained.
The approximate position of any leakage measured can be determined by the successive closing of anysluice valves along the main, in the manner of a step test. This method has the disadvantage that trunkmains normally have few sluice valves along their length and that an accurate measurement depends uponthe drop tight closure of these valves.
6. EQUIPMENT AND LEAKAGE DETECTION TECHNIQUES FOR TRUNK MAINS
63
6.3 Heat Pulse Flow Meter
The second method of measuring the trunk main leakage is to use the heat pulse flow meter. This is aninsertion type meter and will pass through any 25mm clear, straight tapping. The meter is used by isolatingthe trunk main at some downstream point, and inserting the meter through a tapping made at someupstream point. Any flow registered by the meter will be leakage along that trunk main plus any waterwhich is letting-by the shut valve. Inserting the meter through a tapping adjacent to, and just upstream of,the shut valve will provide a measure of water passing the valve and the difference between the tworeadings is the leakage along the length of the main. If it is found that leakage does exist along the main,its approximate position can be determined by closing valves. In view of the disadvantages of this method,a better technique is to insert the meter through additional tappings made along the length of the main todetermine whether the leakage is upstream or downstream of this additional tapping. This is equivalent tocutting and capping, but of course is very much cheaper.
The meter is capable of measuring velocities in the range of 2 to 25mm per second with an accuracy ofabout + lmm per second. The meter will, however, indicate a much higher velocity than 25mm per second,but with less accuracy. For leak detection purposes this is adequate, since leakage velocities greater than25mm per second will usually warrant further investigation. The advantages of this method of trunk mainleakage measurement are:
a) The method accounts for any water which is ‘let by’ by the valves.
b) It can be used to determine roughly the position of the leak.
The disadvantage is that the trunk main has to be isolated from supply, albeit for a short period of time.
6.4 Pairs of Insertion Turbine Meters
It is well known that in situations where flow meters are installed on the inlet and outlet of the trunk main,only very large leaks can be detected because of errors in the meters themselves, and sometimes thedifference between two measurements indicates a net gain along the length of the main. At first sight, thethought of using two flow measurements made with insertion meters would appear to add to this problem,since problems of integrating the velocity profile to obtain mean velocity, and uncertainties about the exactcross-sectional area of the pipeline at the point of measurement, could increase the errors, since the twometers are not used to make flow measurements as such.
As the rate of flow through a trunk main varies, the velocity at the two measurement points also varies.Differences in the velocity at the two measurement points caused by differences in the velocity profile orcross-sectional area will vary in a velocity proportional manner, whereas the velocity differences betweenthe two measurement points due to leakage will be independent of velocity. Consequently, by comparingupstream flow with downstream flow, or upstream velocity with downstream velocity over a range offlows, it is possible to determine the degree of leakage between the two measurement points. Manualanalysis of the flow data obtained is sufficient to detect leaks producing velocities in the main equivalentto l0mm per second or greater. For detection of leaks below l0mm per second, and down to a minimum of3mm per second, it is necessary to repeat the measurement with the meters exchanged end for end, and touse a more sophisticated analysis of the data, which will involve the use of a computer programme.
The advantage of this method of trunk main leak measurement is that it is not necessary to take the trunkmain out of supply in order to make the measurement. The meters can also be inserted through additionaltappings made along the length of the main to determine roughly the position of the leak.
The disadvantage is that it requires two site visits, the first to install the meters and the second,approximately 24 hours later, to remove the meters and to collect the data.
64
6.5 Infra-Red Photography
This is a technique which has been applied in some parts of the UK, but from reports seen, it has not yetproved its worth. It relies on changes in the temperature of the ground caused by the presence of moisture(hopefully leakage) to be identified by an infra-red camera carried by an aircraft flying along the route ofthe main, which obviously has to be clearly identifiable. Development work continues, and it may yet findits place in rural mains situations, particularly where access is difficult, and where the climate is hot and/ordry.
6.6 Leak Noise Correlation
This technique has been previously explained in its use in the distribution system. The LNC has beenrestricted in its use on trunk mains largely because of the few access points generally available. However,developments with computer hardware enable correlations to take place utilising hydrophone sensorsplaced significant distances apart. Successful tests have been reported with such sensors 5kms apart. This‘long distance’ correlator therefore looks to be a promising technique. If the cost of the extra hardware andancillaries is favourable, leakage surveys of trunk mains may well be around the corner, moving trunkmains detection from a largely ‘passive’ to an ‘active’ policy.
65
7.1 Introduction
Being able to easily identify the location of mains and services is an obvious requirement if leakage is tobe located and repaired. Valve and hydrant chambers help, but vigilance is required to prevent their coversfrom being obscured, by tarmac in particular. Roadside markers for such fittings are extremely helpful, andmust be replaced if they go missing.
To avoid accidental damage (and asscoiated leakage), highly visible plastic marker tape should be laid inthe pipe trench, 300mm above the pipe soffit, to serve as a warning to excavators. If non-metallic pipesare in the trench, this tape should contain a metallised mesh to aid subsequent location by detectors.
Standard positioning of mains in footpaths, and service connections to houses, can not only help thelocation procedure, but also help identify other underground pipes and services when they appear in theexcavation process. Needless to say, other Utilities should always be contacted to ensure there is anawareness of such plant, before any excavation commences.
7.2 Location for Mains in a 2 Metre Footpath
Mains routes involving all Utilities normally require in the UK a minimum clear width of 2.0 metres inthe straight sections. The disposition of mains involving all Utilities would normally be as indicated inFigure 7.1. However, where all Utilities are not involved, a reduction in the route width may beacceptable, but it must have a clear width of at least 1.0 metre and the prior agreement of all Utilitiesconcerned.
The recommended dispositions as illustrated in Figure 7.2 result from a fresh analysis of Utility needs inthe UK. The relative depths of lay required for the various mains argue powerfully in favour of the lateraldispositions illustrated, and are therefore recommended as standard locations. The lateral clearancesbetween adjacent Utility mains are considered as the minimum, and represent the best use of the limitedspace available.
7.3 Service Pipe Layouts
7.3.1 General Arrangement of a Service Pipe
Figure 7.3 illustrates the normal arrangement in the UK for a service pipe where the water main and theservice pipe are located in the same footpath (short-sided services).
7.3.2 Depth of Service Pipe
All service pipes in the UK should be laid with a minimum cover of 750 mm to the final finished groundlevel. Under no circumstances should the cover be greater than 1350 mm.
7.3.3 Provision of Ducting
A duct should be provided for all service pipes located under the carriageway (usually on long-sidedservices).
The duct should be a minimum of 40 mm diameter, coloured blue, and clearly marked ‘Water ServicePipe’ at one metre intervals along its top.
No joints of the service pipe should be contained within the length of the duct.
7.3.4 Proximity to Other Services
The usual location of the service pipe in regard to other Utilities’ services in the UK is indicated in Figure7.4.
7. IDENTIFICATION OF MAINS, SERVICES AND VALVES
66
The dimensions indicated are generally regarded as the minimum distances allowed between the varioussections.
Under no circumstances, should the water service pipe be located nearer the gas service, or problems mayoccur with impregnation of gas through the walls of the polyethylene water pipe.
7.3.5 Alternative Layouts for Crossing Carriageway
Generally, under normal circumstances every supply pipe should have a separate communication pipe anda separate ferrule connecting it to the water main.
However, under certain exceptional circumstances, more than one property may be connected to onecommunication pipe. It must be stressed however that these alternative arrangements are not regarded asgood practice and should only be used when the provision of separate communication pipes is notpracticable.
7.4 Valve Identification
Valves should be identified on site by, for example, the installation of a coloured sleeve over the valve cap.
The direction of valve closure could be indicated as follows by the background colour of the sleeve:-
CClloocckkwwiissee cclloossiinngg Blue
AAnnttii--cclloocckkwwiissee cclloossiinngg Black
The function of the valve could be indicated by coloured bands on the sleeve e.g.
ZZoonnee//BBoouunnddaarryy vvaallvvee Yellow
DDiissttrriicctt ((DDMMAA)) vvaallvvee Red
VVaallvvee ccoonnttrroolllliinngg aa pprriivvaattee mmaaiinn Green
VVaallvvee ccoonnttrroolllliinngg aa ddiiaallyyssiiss uunniitt White
7.5 Electronic Pipe Locators
7.5.1 IInnttrroodduuccttiioonn
All existing pipe locators used by the water industry come under the general heading of ‘low radiofrequency instruments’ and can only be used for locating metallic pipelines. All of the locators work bycausing an alternating current to flow in the pipe and detecting the magnetic field thus produced. Thealternating current may be caused to flow in the pipeline by either induction or conduction.
Induction is probably the most convenient method, because it is not necessary to have access to thepipeline, nor to run out lengths of cable, or drive in earthstakes. Pipe locating equipment used inductivelyhas the transmitter placed on the ground above the line of the pipe, and the receiver used separately.
Equipment used to make the current flow in the pipeline by conduction can be used in two forms:
a) Earthstake coupling - the transmitter is directly connected between a fitting on the pipeline and anearthstake driven into the ground at some distance from the pipe to be located.
b) Direct coupling - the transmitter is connected by wires to two access points on the pipeline, thesection between which, is to be traced.
67
7.5.2 DDiissccrriimmiinnaattiioonn
The major problem, shared by all pipe location equipment, is that of discrimination between closelyspaced mains and services. The degree of discrimination will depend on the mode of operation. Forinstruments used inductively, it is likely that discrimination between two parallel pipelines will only beachieved where the separation exceeds 1.25 to 1.5 times the depth. With conductive coupling,discrimination is likely to be better than 1 times depth if direct coupling is used.
7.5.3 AAccccuurraaccyy ooff llooccaattiioonn
The sharpness of response of pipe location instruments to buried pipelines will depend on the depth andthe mode of operation. The location accuracy obtained with instruments used in the conductive mode willin most cases be better than instruments used inductively.
7.5.4 PPllaassttiicc PPiippee LLooccaattoorrss
Since plastic is non-conductive, current-based methods can not be used to locate such pipes unlessmetallised marker tape is laid at the same time as the pipe. Plastic pipe locators therefore rely on the audio-tracing of a noise genrated into the pipe or water column. They are generally less effective, but at leastoffer the possibility of tracing the pipe.
7.6 Other Pipe Location Methods
There are two other possibilities being considered for the location of pipes, plastic ones in particular. Theyare both very much at a developmental stage.
The first is to use a traceable, chemical coating, similar to substances at use in the food industry, to linethe inside of the pipe.
The second is to implant micro radio transmitters in the pipe wall that can be detected when in closeproximity above the main.
Figure 7.1 Water Main - Position in 2m Footpath
Figure 7.2 Recommended Arrangement of Main in a 2m FootpathDimensions in mm
68
Footpath
Water
Telecomms
Gas
Cable TV
Electricity
Road
300mm max
Meter &StopValve
2000
9601255
1550
690430
450 295 295 270 260 430
Road
300max
Ele
c
Pro
pert
y B
ound
ary
Pro
pert
y B
ound
ary
Cab
le T
V
Gas
Wat
er
Tele
com
s
mm
Figure 7.3 Typical Layout of Service PipeDimensions in mm
Figure 7.4 Usual Location of Service Pipe to PropertyDimensions in mm
69
25mmPolythene tube
(Provide duct onroad crossing)
Service Pipe
Communication P ipe
25mmnombore
900mm1350mm max750mm min
Propertyboundaryline
Plastic Tube
Supply PipeMaintained by consumer
300mm max
Water Main
Boundary box /Stop valve
750mm
Ferruleconnect ion
200 min 100 100
450
450
750Min
WaterService
Pipe
Gas
Telecomms
Electrici ty
70
8.1 Demand Patterns
The Total Integrated Flow (TIF), or ‘top down’ approach is of little use in assessing leakage on less thanan annual basis, so another method is required for operational leakage detection purposes.
Leakage is continuous, whereas legitimate demand varies with time. It is this difference which forms thekey to leakage identification, for if ways can be found of separating legitimate usage from leakage, or ifthere are times (even temporarily) when normal usage ceases, then the water still flowing can be identifiedas leakage. Fortunately there are methods by which this can be done.
Legitimate usage takes place mostly during daytime. Although some commercial undertakings work shiftsystems and some households work at nights, the large majority of demand occurs in daytime and resultsin the classic and well known demand pattern shown in Figure 8.1. This is true for large conurbations andis surprisingly still accurate for quite small demand areas within the overall area.
In all cases the leakage is running to waste continuously, although its volume varies with pressure. It canreadily be seen that since most of the legitimate demand does not occur at night-time, most of the flow atnight will be leakage.
8.2 Night Lines
The flow of water at night is thus a very important factor in leakage control and detection. It is known asthe Night-Line in the UK, and is usually the flow through a DMA meter for 1 hour between 3 am and 4am (times may vary).
A high night line is a good first indicator of high leakage levels, but it is not a leakage level itself, becauseimportant deductions from it (such as for factories working shifts etc.,) have to be made to arrive at a ‘net’Night Flow (NNF).
NNF is normally expressed in litres per property per hour. This enables comparison to be made betweenareas and against set targets. It is subject to less significant errors than the TIF method, and eliminates theextraneous factors included in ‘unaccounted for water’. It is known as the ‘bottom up’ approach. It is,however, more difficult to obtain, as it requires specific measurements to be taken, rather than usinggenerally available data.
In systems with either district or waste metering it is possible to aggregate the results of night flowmeasurements to produce an overall figure, provided that meter coverage is complete. However, suchfigures will not include leakage from service reservoirs or trunk mains, and therefore cannot be used toassess overall performance.
It should also be remembered that the Net Night Flow also contains some legitimate domesticconsumption, and that the rate of leakage at night is higher than the average daily rate because the pressureis at its highest at night. To convert night time leakage rate to total daily leakage, tests have yielded thefollowing approximation:
(night time leakage rate) x 20 hrs = total daily leakage
The multiplier of 20 instead of 24 hours takes into account the reduced day time pressure. This is knownas the ‘20 hour rule’.
8.3 The Development of Continuous Monitoring
Studies throughout the world have shown that continual monitoring for leakage control is cost effectiveon almost all distribution networks. The success of the method can be attributed to two major influences.
Firstly, the rapid advances in metering technology have expanded the flow range of the well establishedmechanical meters, and have led to the introduction of other meter types, such as electromagnetic , in thesize and flow range, and at a cost suitable for leakage measurement.
8. LEAKAGE IDENTIFICATION AND LOCALISATION
71
Secondly, data capture has become increasingly sophisticated, the techniques ranging from simple remotereading devices to programmable data loggers and telemetry.
Together these advances have encouraged a trend away from those leakage control methods requiring aroutine survey (the inefficient regular sounding or the labour intensive regular waste metering) to thosewhich utilise continual monitoring (district metering/combined metering).
The latest development is that of the permanent acoustic (or ‘noise’) loggers. This, of course, is differentin that they monitor directly for leakage, not for flow into an area. Furthermore, they automatically localisethe approximate position of a leak such that a leak noise correlator can be immediately employed toprecisely locate it. Even so, they are not a replacement for flow monitoring.
8.4 Determination of Leakage from Night Flows
The primary use of net night flow data is to provide operational data on which to decide on the need forfurther action. The minimum night flow (MNF) can be readily measured with reasonable accuracy for bothdistrict and waste meter areas, allowing small changes in flow volumes to be observed.
Determination of the night metered consumption is more difficult. In many areas it will be negligible andcan be ignored. Where it is not deemed negligible, the alternative methods available to determine it are asfollows:-
i) Use a percentage of average daily consumption. This is satisfactory where the total non-domesticconsumption is relatively small.
ii) Measure MNF immediately prior to and during a ‘bank holiday’ period. The difference will givethe night consumption of industrial users who shutdown for the holiday. Some allowance will stillbe required for commercial users with an element of domestic type consumption, and for industrialusers with continuous processes.
iii) Do a telephone survey of major consumers to determine whether there is significant night usagee.g. replenishment of factory storage tanks. Some users may be able to supply night consumptiondata. In addition, on large complex sites there is a possibility of misuse of water (e.g. unauthoriseduse of fire mains), and it may be prudent to check such connections before embarking on leaklocation work.
iv) Take night meter readings of the major non-domestic users. Use data loggers where the meters arelogger compatible - consider changing/converting old meters on major users where this is not thecase.
v) Trade effluent data may provide useful information.
It must also be remembered, however that whilst domestic consumption is reduced to a minimum bymeasuring flows at night, it is not eliminated entirely. Research in the UK suggests an allowance of about21/prop/hr, which includes minor undetectable leakage such as dripping taps and passing ball cocks. Thisconsumption is included in the net night flow figure. The increasing use of domestic appliances overnightusing economy electricity tariffs is also a factor which may need consideration.
Having determined the leakage, this can be compared to previous readings and ‘norms’ for the area, anda decision made on whether further action is required. If the night flow in a district exceeds somethreshold value, further investigation should be undertaken to locate the source of the extra losses, whichmay be unreported bursts.
8.5 Necessary Checks
Having identified an area with indicated high unaccounted for levels at the meter(s), the first step shouldbe to verify the data and check DMA meters, loggers and boundary valves. This can be the source of bigerrors, and, if overlooked, a lot of time can be wasted.
72
‘Pressure Zero Testing’ should be carried out in order to confirm the integrity of the area. This will notonly prove that the boundary valves shut tight, but will also disclose any unknown connections to adjacentareas. In operating at night, the water supply is turned off at the meters, and the system allowed to draindown through consumption and leakage. If the area is ‘tight’, then pressures monitored at key points in thearea will approach zero. The supply is then slowly re-introduced, and the situation monitored untilpressures return to normal.
Then, the search for leakage needs to concentrate upon the reasons for a high night- line. It may be thata meter registering water flow into a rural area suddenly shows an increase in demand. If this correspondsto a new large metered customer being connected to the mains, then it is possible to correlate the two andshow that the increased flow is not leakage. If the demand gradually builds up in an area and thus matchesthe number of new properties being built, then a high increase of leakage is probably not indicated.
On the other hand, where sudden increases in demand occur for no apparent reason, or flows creep upwhilst housing stock is constant, then the pointers are towards increasing leakage, and the need is forfurther investigations.
This demonstrates the importance of monitoring leakage regularly and taking account of TREND in flowpatterns. A realistic calculation of actual leakage is also necessary for each DMA so that an aggregated‘bottom up’ assessment of each District can be made, and performance monitored against the target levelset for each DMA.
On a weekly basis, using district meter information, comparisons may be helpful to confirm leakagebetween ‘bottom-up’ night flow calculations and ‘top down’ bulk consumption calculations. Regulators ofthe UK Water Industry now expect such comparisons of leakage as part of leakage assessment.
8.6 Large Area Sub-Division
Detection methods should be employed to progressively narrow the search for leakage (‘localisation’)using the most appropriate method to ‘home in’ and finally locate it.
If leakage is suspected in a large DMA, it may be necessary to subdivide it so that each sub-area containsless than about 1000 properties. This can be done using established Waste Meter Areas within the DMA,or by judicious valving. In either case, the flow into the sub-area will only be monitored temporarily.
Areas of less than about 1000 properties are more manageable for the final leak location. This is done byvalving-off the area, so that each district meter feeds a district area. Figure 8.2 illustrates this exercise.
If the installation of further meters or valves is necessary, although it is more capital expense, it has thepermanent advantage that future investigation of the area is made easier.
8.7 Waste Metering
8.7.1 General
This technique may still be necessary even within a district metering strategy, where a large area has to bekept ‘open’. Waste runs will be performed on each of the WMAs at an agreed frequency.
To carry out waste meter runs, it is necessary to close all the predetermined boundary valves, and recordthe flow through the meter, preferably with a data logger. The time taken to do this will, of course, dependon the number of valve shuts required and also on whether the meter is fixed or mobile. Having recordedthe flows in several such WMAs, the decision can be made as to whether further action is required.
Some WMAs may not be capable of being run for 24 hours due to low pressure in the area itself, or inadjacent areas downstream. In such cases additional costs for night-time overtime working will beincurred, but these can be minimised by shutting most of the valves during the day, just leaving the criticalvalves to be closed.
Waste meter areas are now sometimes referred to as district meter sub-areas.
8.7.2 Checking of Boundary Valves
It is clearly essential that the boundary valves shut tight, and each valve should be sounded after closingto check this, although this will not always indicate valves which are passing. If it is suspected that a valveis letting by, but no noise can be heard, it will be necessary to test the water tightness of the valve. Thiscan be done by installing a hydrant flow meter near to the suspect valve and within the valved-in WMA.Water is run to waste through the hydrant, and the flow rate is measured. This additional flow shouldappear at the waste meter if the valve is water tight, provided that pressures are not significantly reduced.If this is not the case, the valve will require repair or replacement.
8.7.3 Effect of Reduction in Pressures
If due to the valve closure pressures are significantly reduced during the night period, it may be necessaryto adjust the measured MNF to obtain a figure which is truly representative of the normal leakage level.
8.7.4 Combination with District Metering
If waste metering is used in combination with district metering (and this is now much more usual in theUK than waste metering on its own), provided that Minimum Night Flows (MNF) are being recorded bythe district meters, it is no longer necessary to maintain a set pattern or frequency of waste testing. Thelocation and frequency of the waste tests and step tests will be determined by the district meter readings.
8.8 Step Testing
8.8.1 General
If further detection action is decided upon as a result of a night-time increase at the district meter or thewaste meter run, it is necessary to decide on whether a step test should be carried out prior tosounding/correlating. This will depend on the size of the area and past experience.
Step testing is the process of successively closing valves to reduce the size of the area being metered. Theresultant drop in flow rate monitored at the meter by chart and/or logger, following the closure of aparticular valve, represents the MNF in that small section isolated by that valve. Thus the leakage in thatsection can be separately assessed. An undiscovered leak would show a disproportionate drop in flow. Inthis way it is possible to identify those small sections where the leakage is occurring, and this normallymeans that a much smaller part of the area needs subsequent location work. For best results, each step ofthe test should contain about l00 houses in a UK situation.
Step testing is usually carried out at night and thus carries penalties in overtime payments and disruptionof work routine, but it can be very effective and may be the only option in busy city centre areas if noacoustic loggers are available.
Since water flow is interrupted, checks of important water supplies and dialysis patients are necessarybefore work begins.
With technological developments, step tests can now be performed using radio, saving time and money.
As soon as possible after the step test, water supplies are restored.
Figure 8.3 illustrates the principle of step testing.
8.8.2 Methods of Step Testing
There are various procedures for actually performing step tests, all of which depend on valves shuttingtight:
73
i) The isolation method, where the area downstream of the closed valves is left without water for theduration of the test.
ii) The close and open method, in which each valve is only closed for a period long enough for thedrop in flow rate to be recorded.
iii) The back feed method, where each time a valve is closed a corresponding valve is opened behindit.
8.8.3 Night Tests
Step testing must normally be carried out at night, and each test is likely to take a two (or sometimes three)man team 6 hours to carry it out. Actual time taken will depend on the number of shuts, and on the durationof the night period when flows are at a minimum. This period may be influenced by electricity tariffs andsocial habits. The period will be indicated by the pattern of the MNFs previously recorded, and may varywith the day of the week. Often only 3 or 4 hours will be available. The number of shuts planned must betailored to fit within the time available. For large areas, steps larger than the ideal 100 houses will probablybe required.
8.8.4 Day Tests
Trials of afternoon step testing may be carried out with success in areas where many of the properties areunoccupied during the day. This technique may be useful in identifying some steps where leakage is notrunning, but positive readings are not able to differentiate between leakage and normal consumption.
8.8.5 Flow Recording
Where the Mobile Advanced Step Tester (MAST - see section 5.4) is not used, an on-site readout of flowrate, preferably in graphical form, is required. The advantages of this are that:
i) Each step can be noted on the graph, thus confirming the water tightness of the valvesii) If all the leakage is accounted for prior to the end of the test, this can be identified and the test
completed early.
8.9 Acoustic (Noise) Logging
8.9.1 General
The localisation procedures of waste metering, step testing, and initial sounding surveys are all manpowerintensive, and therefore relatively costly in operational expenditure.
The introduction of the first ‘temporary’ noise loggers (see section 5.6.2) has improved this situation. Theyhave proved a cost-effective survey tool, enabling better targeting of leak location resources. Their use inclusters has greatly helped to locate leaks in problem areas where it was already known that somethingwas wrong, but previous sounding and correlation had been inconclusive - in busy built-up areas, or wheremains intersections are complex (e.g. urban crossroads), a certain ‘conviction’ is required before holes canbe confidently dug, with all the interruption and expense that this entails. Experience needs to be gained,however, in the correct interpretation of results from these loggers.
However, it is the ‘permanent’ device (see section 5.6.3) which is causing the most excitement, carrying,as it does, the possibility of helping Water Suppliers achieve further leakage reductions without thespiralling operational costs that might have been anticipated. In fact, it threatens to reduce existingoperational costs.
In the UK, most of the ‘easy’ leakage has now been detected and repaired. In many areas, the leakagebenefit of pressure reduction has already been felt, and cannot be repeated. Reduced pressures have maderemaining leakage more difficult to find, and detection is also becoming more difficult because
74
background noise at night is increasing. This has meant that leakage reductions have started to ‘tail off’with existing methods, and detection costs are increasing. Furthermore, the time taken to identify, detectand repair an average leak is still relatively long.
It is obvious that labour-intensive methods cannot continue to deliver ongoing reductions in leakage andcost. The permanent noise logger offers the potential to reduce the economic level of leakage, and tochange the ‘thinking’ about what levels of leakage are ‘tolerable’ and ‘inevitable’.
Also, by facilitating further savings, it may afford the postponement of some capital schemes for resourceenhancement, and cause the postponement of some mains renewal schemes in areas where this was seento be the only remaining solution to reduce leakage.
8.9.2 Operational Aspects
Preparation prior to the installation of these loggers is minimal. The area to be monitored has to be checkedfor the availability of fittings and for the presence of loud noise sources on the mains, whose proximity isto be avoided. The spacing of the loggers will also be dependent on the pipe material. The planning of apatrol route, (to be within 50 metres of each logger), may also be a factor regarding where they are actuallydeployed.
Once deployed and initially ‘patrolled’, an initial list of leaks will be generated, with no reliance onsubjective, human interpretation factors. One person can survey several DMAs in a day, and skilleddetection staff are focussed on finding ‘known’ leaks with a correlator, thus avoiding wasted time lookingfor leaks in areas where there are none. This is obviously motivating for the workforce.
Once installed, the loggers allow more and more leaks to be identified, at the minimal extra cost of anotherpatrol. Because the noisiest leaks may not be the biggest, and repair may initially cause the breakout ofmore leaks that were ‘waiting to happen’, more than one patrol may be necessary to significantly reducethe leakage level. Management has control over this process, so by monitoring the night flow into theDMA, leakage reduction efforts can be stopped at any point once an acceptable level has been reached.
The reading of a DMA meter can easily include the monitoring of the loggers within it, so that new leaksare ‘localised’ at exactly the same time as increases in the night flow are noticed. This means a prescribedleakage level can be easily maintained, because the detection time is greatly reduced.
In ‘stable’ areas where leakage increase is slow, once the prescribed level is reached, the loggers could beremoved and re-deployed to another area. When there is an unacceptable increase in the night flow in theoriginal area, they could be taken back again to assist in the localisation of the new leaks.
These loggers may be the only solution in areas where it is not possible to set up DMAs, or where it ispreferred not to do so.
Where leakage control is the only reason for setting up small DMAs, they enable a move in policy to largerones, with consequently reduced setting-up costs.
Hence they provide greater flexibility in the development and operation of a distribution system.
8.9.3 Results
The ‘permanent’ version of the noise logger was launched in the UK in June 1999, following substantialfield trials. Many operational benefits have been confirmed. These include:
• Leaks being localised faster than before, with the obvious savings in water• More leaks being found than was thought possible with previous methodologies - some of these
were thought to have been ‘masked’, or inaudible• Increased repair efficiency, with leaks being dealt with in clusters, rather than in a widely dispersed
manner
75
• ‘Lowest ever’ leakage levels being attained, and maintained, in established DMAs• A marked reduction in leak detection operating costs
The evidence suggests that a new era is dawning for leak detection with this piece of technology. Morefigures are awaited regarding overall costs, but it is looking likely that initial capital investments willeasily be repaid within the life expectancy of the loggers. It remains for individual Water Suppliers todecide what coverage to deploy them at, and whether to use some in a rotational way between differentareas. Such ‘best use’ data will accumulate as experience is gained.
76
77
Figure 8.1 Demand Pattern in a Typical Area
Figure 8.2 Temporary Sub-Division of DMAs to Help Locate Leakage
A
B
METERM1
METERM2
MAINS
PERMANENTLY CLOSEDDMA
BOUNDARY VALVES
TEMPORARY CLOSURE AT VALVES ‘A’ & ‘B’ TO DIVIDE THE‘DMA’ INTO 2 PARTS WILL ENABLE ASSESSMENTS OFDEMAND PER PROPERTY IN EACH TO BE MADE. THISMETHOD SOMETIMES HAS APPLICATION IN LOCATINGLEAKAGE (ESSENTIALLY 2 TEMPORARY ‘WASTE METERAREAS’ ARE FORMED.)
6 a.m. Noon 6 p.m. Midnight 6 a.m.
Leakage flows Continuously
Night Line
Peak Demand
Legitimate Water Usage
Leakage Flows Continuously
78
Figure 8.3 Diagram to Illustrate the Principle of ‘Step Testing’ for Leakage Control
A
B
C
DE
F
GRANGE AV ENUE
MAINS
BETA GROVE
ALPHA AVENUE
METER & ‘CHART’ /ELECTRONICRECORDER
G
CLOSED BOUNDARYVALVES
VALVES CLOSED AT INTERVALSDURING STEP TEST
A
B
C
DE
F
G
1.00 1.15 1.30 1.45 2.00 2.15 2.30 2.45
TIME (a.m.)
DISPROPORTIONATE DROP IN FLOWWHEN VALVES ‘C’ & ‘E’ ARE CLOSED,INDICATING SUSPECTED LEAKAGEFOR FURTHER INVESTIGATION
FLOW
9.1 Sounding
Sounding can be used as a detection method in its own right (very labour intensive), or more usually as ameans of finally locating a leak previously identified and localised by other methods. Sounding seeks toidentify aurally the point of maximum intensity of the characteristic hissing noise of a pressurised leak.Sounding can either be done directly, that is by making direct contact with fittings on the main, or bylistening on the ground surface above the line of the main.
Surface sounding can be successful where there is a hard surface above the main, but generally it is lesseffective and reliable than direct sounding on fittings, i.e. boundary stop taps, valves and hydrants. It canbe particularly unreliable where recent excavations have been made and backfilled with imported material,or where other underground apparatus is very close. However, the more recent ground microphones withprobe attachments are a significant improvement.
The procedure for locating a leak is as follows. On the first pass a note is made of those fittings on whicha noise is found. Those fittings are then sounded again. If the noise is on a boundary stop tap, it isnecessary to ascertain whether this is due to use within the house. The stop tap is closed and then soundedonce more. If the noise ceases, it indicates leakage on the consumer’s pipework and a notification of suchis made. If the noise continues, it indicates a leak on the communication pipe, or more probably on themain if the noise can be heard on adjacent fittings. In this case surface sounding is carried out to locatethe position of the leak more precisely.
The advantages of sounding are:
a) The equipment is relatively simple and inexpensive;b) In conditions of low background noise, a large number of fittings can be inspected fairly rapidly;c) It is possible to detect leakage within the premises when used in conjunction with the turning off
of the consumer’s stopcocks.
The main disadvantages of both direct and surface soundings are:
a) It is sometimes difficult to determine the precise position of highest sound intensity;b) The position of highest sound intensity does not always coincide with the position of the leak or
the fitting nearest the leak;c) Sounding can be very difficult in areas with high background noise such as that produced by traffic
in busy streets, or by boosters and certain fittings such as control valves;d) Successful sounding is dependent upon operator skill;e) Some leaks are inaudible to the human ear, and some produce insufficient sound to be detected by
any sounding technique.
The disadvantages can all be overcome by use of the Leak Noise Correlator.
9.2 Leak Noise Correlation
Leak Noise Correlators measure the time taken for the leak noise to travel from the leak to sensors placedat different points on the mains system. Automatic correlation by the machine then indicates the positionof the leak. Properly used, leak noise correlation can identify leakage typically to within a metre of itslocation, which is usually more precise than ‘sounding’.
Leak noise correlation can be carried out either in place of sounding, or in conjunction with it. Where itis carried out together with sounding, the correlator sensors are attached to those fittings noted asproducing a noise, and the location of the leak determined.
Correlators can be used to carry out the initial sound survey by attaching the sensors to fittings on the mainat suitable intervals and seeing whether a correlation can be found. Where there are long lengths of mainwithout any access points, holes can be drilled through the ground, allowing access via an iron bar.
79
9. LEAKAGE LOCATION, CONFIRMATION AND REPAIR
80
The maximum distance between the sensors must be set according to the pipe materials encountered, andthe sensitivity of the particular correlator and sensor. Typical operating distances are as follows:
cast iron/ductile iron 600masbestos cement 220muPVC 150mmedium density polyethylene 100m
An improvement in sensitivity, and hence an increase in the maximum distance between the sensors, canbe achieved by use of hydrophone sensors. These sensors are fitted to hydrants, or other ‘wet’ fittings, andare in direct contact with the water. Thus, water is the sound transmission medium rather than the pipewall. Using these sensors sounds can be picked up over distances of 1 km or more.
Later correlator models facilitate the sound survey technique by incorporating a survey mode, in which themachine will search for a correlation without the need for the operator to enter the normal informationrequired for a leak location (e.g. length between sensors etc).
This technique requires less manpower than conventional sounding, and leak noise correlators do have amajor advantage in busy urban areas in that they are less affected by background noise. To sound suchareas manually usually requires that the work is carried out at night, and hence additional costs forovertime working are involved.
Reliability, portability and performance of Leak Noise Correlators has continuously improved over recentyears and they are now employed as an essential part of a leakage control programme.
They must be used by trained and skilled personnel, and preferably by people who are using the equipmentregularly.
Leak noise correlators continue to be the principal method of leak location. However, their use as a surveytool is declining due to the emergence of noise loggers. The combination of the two methods, together withelectronic listening devices, provides the most efficient and effective means of leak localisation, location,and confirmation.
9.3 Visual Evidence
Apart from the obvious emergence of trickling/running water, leakage location can be helped by othersigns such as increased growth of vegetation, moss on ground or walls, wet or damp patches, melted snowor frost, and water entering gullies or manholes. Adjacent premises with cellars may also provide clues.
9.4 Other Practical Points
The answers to the following and similar questions may have some bearing on the method chosen to locatethe leakage:
• Is the leakage likely to be a burst main or several leaking fittings or pipes?• Does the area contain industry? Will customers be affected by a valve inspection? Can meters be
logged?• Can the area be worked in normal working hours or does traffic density and noise necessitate night-
time or weekend working?• What are the age and pipe materials in the area? What is the previous burst history?
9.5 Confirmation
Where correlators locate leakage without previous use of sounding equipment, it would be normal to seeksome confirmation by use of a ground microphone. This is a helpful check before the expensivecommitment of digging a hole! In a verge or field, the insertion of a rod or bar may immediately revealleaking water.
81
9.6 Repair, Follow-Up and Records
9.6.1 General
Having located leakage, it is necessary to ensure that it is quickly and efficiently repaired. The totalleakage volume is directly related to the length of time leaks are left running. Leakage detection activitieswill only retain their credibility if leaks are repaired quickly.
Many of the leaks will, of course, be on the consumers’ pipework, in which case it may take very muchlonger to get the leak repaired. This can be a difficult process, particularly in the case of joint supply pipeswhere questions of responsibility have to be resolved.
It is sometimes overlooked that leaking pipes and fittings can give rise to water ingress during negativepressure incidents (e.g. in the vicinity of a large burst). Ingress brings the risk of pollution and hencereinforces the need for prompt repair of all known leaks.
9.6.2 Follow-up After Repair
Having repaired the located leaks, it is good practice to re-sound in the immediate vicinity of the repair tocheck that the previous leak was not masking other leakage. If permanent noise loggers are installed,another ‘patrol’ will reveal this information.
Night flows should be immediately checked after repair work, and compared to DMA ‘norms’. Theseshould be adjusted downwards where appropriate.
9.6.3 Repair Records
It is essential that all repairs of bursts and leaks are accurately and comprehensively recorded, preferablyon a database. This information should include:
• Location• DMA reference• Date• Size• Type of burst• Mains/service pipe material• Type of repair• Was burst reported or detected?
9.7 Leakage Contracts
The use of Contractors in leakage work is relatively new in the UK, but is likely to develop. Various formsof contract have been tried, from simple ‘Detection’ to ‘Payment by Results’. There have been limitationsfrom both the Employer’s and the Contractor’s point of view.
Work to produce a Model Form of Contract is ongoing. Basic sections on Conditions and TechnicalSpecification will be supplemented by specialist sections for different types, and further supplemented byadditional sections to cover the needs of individual Employers. A number of different payment options willbe included.
Employers and Contractors are working together on this, so it is expected there will be a move from anadversarial style of contract to a more co-operative approach. ‘Partnering’ will be a key word.
10.1 Pressure Measurement
Pressure is one of the most frequently measured parameters in the Water Industry, often being measuredalongside flow. Many methods of measurement are in usage, but pressure transducers have become themost common means in distribution systems. They operate by converting fluid pressures into electricalsignals.
Pressure measurement typically takes place for:
• General monitoring of the distribution system• Specific monitoring at critical points (levels of service)• Particular consumer problems of inadequate pressure• Co-ordination with particular flow tests e.g. new housing estates, high rise flats, industrial
consumers, fire fighting installations and fire hydrants• Network analysis calibration
10.2 Pressure Control Options
Pressure management is a major element in a leakage management strategy. Pressure reduction isprobably the simplest and most immediate way of reducing leakage within the distribution system. Itsbenefits are immediate. Even where already practised, it is likely to be worthwhile to re-examine and re-set equipment and schemes to take advantage of progressive technical developments, and local systemalterations.
Pressure management can be accomplished in a number of ways and not just via the installation of a newpressure reduction valve (PRV). In fact, the generation of pressure almost always costs money, so reducingpressure by means of a PRV is intrinsically inefficient. The following options should be considered first:
• Re-zoning the area supplied to match input head to topography and minimise system losses. Thismay include boosting to a smaller, critical area, reinforcing or reconditioning mains to allow lowpressure zones to be extended, or transferring demand zones to an alternative source with a loweroverall head. Network analysis could greatly facilitate this investigation.
• Matching pump output curves to closely match distribution demands. This could include resizingpumps to match known demands, or staged or variable speed pumping, or closed loop control usingflow or pressure signals.
• Installation of break pressure tanks. These generally have a higher capital cost and are a potentialcontamination risk. On account of this they are no longer used in the UK.
Having considered these three options, mechanical pressure control devices, typically PRV’s, provide thenext stage in a pressure control strategy.
10.3 Pressure Control Benefits
Pressure control can:• Reduce leakage• Reduce pressure-related consumption such as hand-washing, car washing etc• Reduce the frequency of bursts, at least in the immediate future – subsequent savings in repair costs
can exceed those due to reduced leakage• Stabilise pressure, decreasing the possibility of pipework movement and fatigue type failures, and
possibly eliminating certain household plumbing problems.• Provide a more constant service to customers – large diurnal pressure variations may give
customers an impression of a poorly managed service, and unnecessarily high pressures raisecustomers expectations and perceptions of what is adequate.
• Enable a company to standardise on pipes and fittings which have a lower pressure rating, and aretherefore cheaper.
• Assist demand management when flow restriction is necessary i.e. during drought.
82
10. PRESSURE MONITORING AND MANAGEMENT
10.4 Pressure Reduction Problems
Some examples of the problems that can potentially arise, and their consequences, are listed below. Someof these can be designed out of the system.
• PPoooorr PP rreessssuurreeIn correctly configured systems this is typically a result of restrictions and blockages of individualsupplies. Flow and pressure tests at the property affected will reveal the location of the problemwhich can then be dealt with in the normal way. Partly closed stop taps and valves are a typicalproblem.
Poor pressures may also be the result of pipework simply being undersized, perhaps throughcorrosion. They may also occur by the setting up of the PRV area severing the normal interlinkingof the system. This should be assessed beforehand at the area design stage.
• NNooiisseeNoise can be a problem close to PRV installations. Noise is usually associated with small valveopenings and may be associated with cavitation problems. Attention to pipework detail and valvesettings can reduce noise levels but it is best avoided by correct selection and siting. Noise througha PRV does create difficulties for leak detection work in the vicinity because of its interference.
• BBlloocckkaaggeessBlockages can occur as a result of mains material becoming trapped in the PRV. This may result infailure of the control and actuating mechanism and loss of pressure control, leading to excessivelyhigh or low pressures. Attention to the maintenance of filters and correct flushing are necessary toavoid blockages in distribution systems which are prone to solids contamination. It is generallyrecommended that planned preventative maintenance be carried out on a six- monthly basis.
Valves without close mechanical tolerances are less susceptible to this type of failure. Strainersupstream of the PRV will also help
• VVaallvvee OOppeerraattiioonnClosing of valves between the PRV and a remote pressure monitoring point will result in the PRVattempting to rectify the apparent loss of pressure at the remote point. Typically this occurs whenvalves are shut in the course of a routine repair.
The results of exposing the system to maximum pressures at moderate flows will usually be a seriesof burst mains. This situation should be avoided by ensuring that Inspectors, in particular, are awareof pressure control systems and follow appropriate procedures before closing critical valves.Network models can also be used to simulate valve closures prior to operation on site to helpunderstand how the system will react.
• PP rreessssuurree aanndd FFllooww SSuurrggeessUnder certain circumstances surges of pressure and flow can cause unpredictable PRV behaviour.This can result, with certain valves, in the piston exceeding its travel and jamming in the fully openor closed position.
Usually, the surges which cause this type of failure result from valve or pump operations whichshould be examined to minimise the risk. In addition, the provision of ‘stops’ to limit the travel inmechanical systems can be helpful.
• HHiigghh--RRiissee BBuuiillddiinnggssOrdnance survey data alone is insufficient in planning an area from a topographical point of view– a tall building survey should be undertaken. In areas where existing flats rely upon a high-pressure mains supply, pressure reduction may only be possible if the Supplier is willing to bearcosts of pumping and plumbing modifications. Where small boosters are already feeding multi-storey buildings, the lowering of pressures may cause the boosters to operate more regularly.
83
• SSppeecciiaalliisstt CCoonnssuummeerrssParticular care must be taken in consideration of pressure reduction on the effect on home dialysisunits (they may simply be able to be adjusted) and industrial consumers who use processesdependent on existing mains pressures, or who have sprinkler systems requiring pressure in excessof that required.
The Fire Service function will also be affected by pressure reduction.
Technical staff can usefully discuss such problems with consumers beforehand and find individualsolutions. Time must then be allowed to make the required changes.
• CCoossttssPRVs work best as a single feed to a DMA. The installation of a PRV may therefore incur extraexpenditure to convert the area being supplied to a single feed system.
• AAddddiittiioonnaall AAccttiivvee LLeeaakkaaggee DDeetteeccttiioonnIt should be noted that in the long run, the lower the pressure, the less leakage, but the greater theneed, frequency and cost of leak detection. Obviously, less ‘spare’ pressure exists before consumerscomplain. The prime reason for leak detection effort in such a case becomes one of responding tolow pressure complaints caused by leakage rather than to save water and money directly. Anappropriate balance must thus be found for sensitive areas.
Lower pressures also mean leakage is more difficult to actually detect because the noise ofescaping water is less intense. This suggests that intensive leakage detection should be carried outbefore pressure reduction is implemented, otherwise leakage which could be found could insteadbe rendered undetectable (or more costly to find), thus negating potential benefit.
10.5 Pressure and Leakage
Consideration of a basic hydraulic map indicates that the residual pressure at any point in the systemdepends on:
• The input hydraulic head of the zone resulting from either: a) the supplying reservoir level ingravity systems, or b) the pump outlet pressure in pumped systems
• The difference in level between the source of supply and the point of delivery• The frictional losses between the system inlet and outlet
There is often confusion between absolute and residual head when discussing pressure. Figure 10.1illustrates the relationships between these points in the distribution system.
The pressure within a DMA (for example) will therefore vary geographically with elevation and pipeconfiguration (longer, smaller pipes generate more frictional head loss). It will also vary with demand ashigher flows and velocities also result in higher pressure losses. The significance of these two effects willvary with the topography of the area and the DMA, the variability of demand, and the size and conditionof the mains.
Theoretically, leakage is expected to be related to the square root of the pressure at the leakage point. Fieldwork undertaken in the UK at the end of the 1970s, co-ordinated by the Water Research Centre, indicatedthat, in practice, the effects of pressure were greater than this and approached a linear relationship. Thismeans that the actual benefit achieved from a particular pressure reduction can be considerably greaterthan predicted. See Figures 10.2 and 10.3, which attempt to define the relationship between pressure andleakage, and are used to estimate the potential leakage savings from the introduction of pressure reduction..However, this relationship has produced mixed results, leaving the conclusion that there is no universalpressure/leakage relationship, each unique system having its own. It has understandably been argued thatthere are two types of leak aperture; one that keeps its size (e.g. holes and cracks in metal pipes), and onethat changes with pressure (e.g. leaks at joints and fittings, and possibly in some plastic pipes).
84
85
Another reason for greater than expected leakage at higher pressures may be due to a change in the effectof the surrounding backfill material.
Clearly, more research work is needed to better understand the pressure/leakage relationship.
It should also be noted that pressure is generally at a maximum overnight, when flow and friction lossesare at a minimum. It follows therefore that proportionally more leakage reduction will occur at night whenpressure control is implemented. It is for this reason that the previously mentioned ‘20 hour rule’ is used,whereby measured savings per hour at night should only be multiplied by 20, not 24 to derive a daily total.
10.6 Statutory Requirements and Levels of Service
In the UK, under the terms of the Water Industry Act 1991, it is the duty of a Water Supplier to ‘cause thewater in its mains and other pipes to be laid on constantly, and at such a pressure as will cause the waterto reach the top of the topmost storey of every building within the Supplier’s area’.
This specifically refers to:
i) Supplies of water for domestic purposesii) Mains which have hydrants fixed to them
The Water Supplier’s duty however is limited to supplying water to a height no greater than that to whichit will flow by gravity from the service reservoir or tank, and the Supplier is free to select which reservoiror tank is used.
If any house requires water to be delivered at a height greater than l0.5m below the draw-off level of thereservoir, the Supplier may require the installation of a cistern capable of holding up to 24 hrs storage.
The above requirements mean that, in the UK, a Water Supplier does not legally have to supply water toevery building regardless of elevation. He would normally do so, however, but would re-charge for allnecessary expenditure in these exceptional circumstances.
A reference level of service (LoS) has to be provided of 10 metres head, at the customer’s boundary, at aflow of 9 litres/minute for a single property, measured on the customers side of any metre, boundary boxor other fitting.
Checking compliance against this standard could require excavation etc. and is clearly impractical forwidespread compliance testing. Many UK Suppliers have therefore adopted a surrogate pressure reference.This is the pressure in an adjacent distribution system i.e. the nearest hydrant which can be shownstatistically to deliver 9 litres/minute without the pressure falling below 10m at the stop tap. The actualsurrogate pressure used varies between Suppliers, but whatever the target minimum is, it needs to beavailable at the critical point in the area, and should allow for future increases in demand, and deteriorationof the network.
10.7 Identification of Areas for Pressure Reduction
10.7.1 General
Before schemes for pressure are implemented, it is important to collect and keep data on uncontrollednetworks. Without this data it will not be possible to appraise completed schemes, and it will then be moredifficult to design and justify future projects. Pressure data over a seven-day period is needed preferablyin graphical and digital form.
Actual pressures occurring in a potential PRV area should be determined by deploying temporary pressureloggers. Pressure loggers should be sited at critical points within DMA areas which are often but notalways at the highest point AOD. The peak and minimum flows should be determined.
25 metres head is commonly considered in the UK as a normal desirable maximum at the target point.However, practical and physical requirements, for example topographical features, may dictate thatpressures as high as 75m must be tolerated at some properties.
Mains pressure reduction should be investigated for areas where night time pressure can be reduced by atleast l0m.
Where local pressures exceed 75m and cannot be reduced, then pressure management should beconsidered for individual properties or sub-groups of properties.
Use should be made of available network analysis models and information to assist in the identification,planning and design of prospective areas for proposed pressure reduction. These will normally be wherethe pressure always exceeds 30m at the critical point in the DMA at maximum demand. Account must betaken of the most sensitive customer location and the stability of the PRV likely to be installed at low flows.
Peak week and seasonal demands also need to be allowed for.
It should be remembered also that some areas do contract in demand as industry and population movelocation.
Other ‘signposts’ which may indicate areas worthy of further investigation are:
• Areas suffering from pressure bursts or high pressure complaints• Reservoir outlets, even though the scope may be limited• Uncontrolled branches on trunk mains• Multi-feed areas with some or all feeds not pressure controlled• Areas requiring high day pressure but low night pressure, for which flow modulation may be the
most appropriate• New developments/extensions to existing system• Local knowledge.
10.7.2 Existing Pressure Reduced Areas
If local pressure management already exists, it is possible that the application of latest generationequipment can optimise savings. Such applications range from the complete replacement of an existingPRV to the addition of retrofit devices to enhance performance.
For example, replacing conventional fixed outlet PRVs with flow modulated equipment presents theopportunity to optimise district pressures across changing demand profiles. As the following illustrationshows, pressure profiles can be achieved which reduce pressures for most of a typical day, but allowincreased pressure at peak demand.
86
MNT 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 MNTTime Hours
Levelof
Service(LOS)(m)
LOS at target consumer
Local peak demand period
Flow modulated PRV
RReessuullttss ooff rreeppllaacciinngg aa ccoonnvveennttiioonnaall PPRRVV wwiitthh aa ffllooww mmoodduullaatteedd PPRRVV
This example is taken from an actual installation. Benefits are twofold. Average daily metered throughputto the district decreased by 20%, and pressures at the target point were improved at peak demand. Result:rapid payback of investment, reduction in leakage, reduced incidence of burst mains and reduced customercomplaints.
It is worth noting that existing pressure management installations may well be lacking in maintenancewhich can impair and even nullify their performance. Items which require regular checks, and, ifnecessary, corresponding alterations in PRV settings are:
• Status of boundary valves• Extensions at margins of PRV area, i.e. additional properties/roads/streets, above design setting.• Additions/changes to consumption profile within defined area, i.e. new housing site or changing
industrial consumption.• Regular checks on PRV inlet/outlet settings to confirm profile against design settings.
10.8 Pressure Reducing Valves - General Overview
A pressure reducing valve (PRV) can be defined as a mechanical device which will give a reduced outlet(downstream) pressure for a range of flow rates and upstream pressures.
All PRVs have certain features in common. These are a means of controlling the flow (the valve), a meansof sensing the pressure differential between the inlet and the outlet, and a means of actuating the valve. Avariety of more or less sophisticated means of providing these features have been developed bymanufacturers.
The two principal categories of PRV are fixed outlet and flow-modulated, each with several variations.
Generally, fixed outlet characteristics maintain approximately the same value of downstream pressure overa range of flowrates. The pressure has to be set so that level of service (LOS) pressure is maintained at thetarget point for the maximum design flowrate. The resultant average zone night pressure (AZNP) will beat a higher value than a flow modulated pressure in a similar system since in the latter case pressures canbe optimised for minimum demand.
In reality, some fixed outlet valves are not always capable of maintaining a constant outlet pressure,particularly at low flow when some rise in outlet pressure can be experienced. A ‘pilot’ can assist inproviding the necessary variable throttling effect to keep a constant outlet pressure as inlet pressures andflows vary. Two pilots with a timed changeover can give a ‘day’ and ‘night’ setting of outlet pressure.
Flow-modulated PRV’s vary the outlet pressure in such a manner that a constant head can be maintainedat a target point in the distribution system for a range of flow rates and inlet pressures. The activatingmechanism responsible for regulating the outlet pressure may be mechanical or electronic, or acombination of both. ‘Look-up’ tables or telemetry may be involved in the outlet pressure control.
Figures 10.4 and 10.5 demonstrate the effect of the two basic types of valve on a critical point in thedownstream distribution system.
Generally speaking, where head losses across the target area exceed l0m (night time/no flow pressureminus day-time peak flow pressure) flow-modulated devices will provide greater net benefit (in spite ofthe extra cost), and are to be preferred.
Because of advances in control practice and communications, control systems for PRVs are becomingmore complex and more effective. The valves are now fitted primarily to reduce leakage and to someextent pressure dependent consumption, rather than the traditional reason of protecting the downstreaminfrastructure. It should be noted that in some cases where old mains systems are combined with highpressures, leakage reductions cannot be maintained until pressure is reduced, because the effect of
87
88
repairing a burst is to increase the pressure in the system, possibly causing further bursts.
The effectiveness and accuracy of the PRV’s will normally increase as the control system becomes moresophisticated. More sophisticated control systems are also better able to respond to unexpected demands.
PRV technology is still developing, and whilst the most common method of control is still local hydraulicoperation, controlled operation by intelligent process units is becoming more economic, even for smallerareas. These units do not necessarily need pre-designed pressure profiles to follow, but will ‘learn’ one in-situ from the real, dynamic network they are operating, always assuming they are in contact with signalsfrom the critical pressure point(s).
89
NO
TE
S
1.D
istr
ibut
ion
pres
sure
s w
ill b
e le
ss th
an P
max
(at
low
est g
roun
d le
vel)
due
to f
rict
iona
l los
ses
in m
ains
, fitt
ings
etc
.2.
Red
uce
pres
sure
in th
e va
lley
botto
m (
poin
t A)
to r
educ
e le
akag
e.3.
Pres
sure
at p
oint
B m
ust s
atis
fy L
evel
of
Serv
ice
crite
ria.
4.M
ains
at p
oint
C w
ill n
eed
boos
ting.
P
GR
OU
ND
LE
VE
L
RE
SE
RV
OIR DA
TU
M P
OIN
T-
A.O
.D
HY
DR
AU
LIC
GR
AD
IEN
T
P
P
CB
A
Pm
ax
Figu
re 1
0.1
Hyd
raul
ic G
radi
ent
90
Figure 10.2 Leakage Increase with Pressure
020
4060
8010
0
PR
ES
SU
RE
(m)
91
Figure 10.3 Relationship between Leakage and Pressure
100
90
80
70
60
50
40
30
20
10
00 20 40 60 80 100
AVERAGE ZONE NIGHT PRESSURE (m)
LEA
KA
GE
IN
DE
X
Figure 10.4 Diurnal Pressure Variation at Critical Point
Figure 10.5 Pressure Gradient to Critical Point
92
12 14 16 18 20 22 MNT 2 4 6 8 10 12 TIME (24hrs)
No PRV fitted
Fixed outlet PRV
Flow modulated
Gradient at minimum flow (fixed outlet PRV)
Gradient at minimum flow (flow modulated PRV’s)
Excesslevel ofservice
Level ofservice
GroundLevel
CriticalheadAODGround Level
Ordnance Datum
Distance from PRV
HeadAOD
AtPRVoutlet
Gradient at minimum flow (all PRV’s)
11.1 Introduction
Water Suppliers must be seen to be operating efficiently and effectively, and must be able to cost justifytheir level of leakage and works designed to manage leakage, particularly to their customers who want tosee their costs minimised. Leakage is often seen as synonymous with waste, and reducing leakage is seenas a means of saving money.
Water lost through leakage has a value and so reducing the level of leakage offers benefits. However,eliminating leakage completely is impracticable and the cost of reducing it to low levels may exceed thecost of producing the water saved. Conversely, when little effort is expended on active leakage control,leakage levels will rise to levels where the cost of the water lost predominates.
Water Suppliers must therefore strike a balance between the cost of reducing leakage and the value of thewater saved. The level of leakage at which it would cost more to make further reductions than to producethe water from another source is what is known as the economic level of leakage (ELL). Operating ateconomic levels of leakage means that the total cost to the customer of supplying water is minimised, andSuppliers are operating efficiently.
This means that leakage reduction should be pursued to the point where the long run marginal cost ofleakage control is equal to the long run marginal benefit of the water saved. The latter depends on the longrun marginal costs of augmenting supplies by alternative means, including an assessment of theenvironmental benefits.
The ELL is not fixed for all time. It depends on a wide range of factors, which will vary over time. Forexample, the cost of detecting and repairing leaks will fall as new technology is introduced. This will causethe ELL to fall. Conversely, if total demand falls to a point where there is a large surplus of water, it maynot be economic to reduce leakage, unless the water can be sold to other Suppliers.
Once the economic optimum level is known, this can be compared to the present level of leakage, and theSupplier can then set targets for leakage control in conjunction with other corporate policies on customermetering, mains rehabilitation, resource development and pressure control. To do so they will need toappraise the investment required for these various different supply and demand management solutions, andthe benefits which are expected to accrue.
Due to the complexity of the issues, it is not possible to generalise to provide standardised formulae forsetting leakage targets. Even if the same leakage policy is pursued, it is likely to be uneconomical to setthe same target leakage levels for areas of differing characteristics. Thus there is a need to examine eachsystem to determine the most appropriate method of leakage control and to plan the required capitalinvestment, manpower and revenue resource. However, any Supplier who is prepared to commit resourcesto collecting the required data, and to carry out the analysis and appraisals, will develop a greaterunderstanding of the factors which are important to target setting. They will also be less likely to haveunrealistic or uneconomic targets imposed on them from outside, or fall into the trap of setting leakagetargets themselves without full consideration of the practicalities of achieving them, or the economicconsequences.
There are many possible ways of setting a leakage target. These can include targets based on minimumnight flows, areas with excess pressure, areas with expensive water or the most urbanised areas. Thesetting of economic targets, i.e. a level of leakage which provides the most economic mix of leakagerelated costs, is independent of variations in physical factors such as property density, pressure, etc. andcan provide clear information upon which sound management decisions may be based.
However, it is recognised that there may be social, environmental and political factors which dictate thetarget leakage level, as well as economic ones related to the Suppliers’ own operating environment. Thishas given rise to a broader concept of ‘the most appropriate leakage target’, being described as “that levelof leakage which, over a long term planning horizon, provides the least cost combination of demandmanagement and resource development, whilst adequately providing a low risk of security of supply tocustomers, and not unduly over-abstracting water from the environment.”
93
11. THE ECONOMICS OF LEAKAGE MANAGEMENT
94
NN..BB.. The Tables and Figures in this section are only illustrative of the sort of procedures, data gatheringand analysis that a Water Supplier needs to give attention to in order to move towards optimum economicefficiency in leakage control. They should not be used in a ‘literal’ way, since they need ‘re-configuring’for the local cost environment , and the technology employed. They are themselves now out of date for theUK, but nevertheless demonstrate the principles and progression needed for informed, cost-relateddecisions.
11.2 Policy Development Through An Economic Approach
11.2.1 Economic Target
The economic target is considered to be that level of leakage that provides the lowest overall annual cost.An investment appraisal is necessary, therefore, to examine the costs of operating the water supply systemwith the present level of leakage, the potential savings and other benefits of moving to a different level ofleakage, and the cost of the measures proposed to effect the change.
The frequent monitoring of all costs, including overheads, is an essential factor in assessing real progressas the implementation of the leakage detection and control policy proceeds.
11.2.2 Assessment of Current Leakage Level and Leakage Control Costs
The precise nature of the method used to identify the level of leakage in the water supply zone is ofsecondary importance. The critical aspects are consistency between areas for comparative purposes andan understanding of the accuracy of measurement.
Economically, it is necessary to use total losses in the water supply zone. The cost of water lost willinclude both those elements of losses that occur on the distribution system and that from customers’ supplypipes. The cost to the Supplier of control and repair will be different, depending where leakage occurs,but both these elements need to be considered.
The cost of active leakage control is made up of costs associated with leakage monitoring, detection andlocation. Costs for these activities should be made up of appropriate staff costs, including on-costs andoverhead allowances.
Costs should not include other distribution activities such as levels of service requirements and waterquality related work. As before, a precise definition is not necessary, but to enable comparisons to be madebetween water supply zones, a consistent approach is essential.
Repair costs are not directly required for an economic evaluation but are used to calculate the cost ofreducing leakage when changing leakage control policy. A unit cost of repair, preferably for the watersupply zone, should be developed for repairs on mains, communication pipes and supply pipes. Repaircosts on supply pipes may be recovered from the customer, depending on current policy.
11.2.3 The Optimum Level of Leakage
The optimum level of leakage may be defined as that level of leakage where the marginal cost of activeleakage control (ALC) equals the marginal cost of the leaking water. In other words where the cost ofreducing leakage by one m3of water equals the value of that m3 of water.
Operating a water supply area at the optimum will result in a mix of costs for leakage control and the valueof lost water that gives the lowest possible cost. Operating at any other combination will be moreexpensive.
Repair costs are largely driven by the rate at which bursts occur. They will be similar for any approach toleakage control and therefore do not need to be considered in a calculation of the optimum level ofleakage.
There is not one global optimum level of leakage. Each water supply zone will have its own optimum leveldepending on the cost of water, the cost of labour, operating pressure, the age and condition of the mainsnetwork and the location of bursts whether on the mains, communication pipes or supplies.
Additionally, in any one global water supply zone there will be different optimum levels depending on themethod of active leakage control and the efficiency with which the method is implemented. Methods withlower operating costs, such as telemetered district metering, will generate a lower optimum level thanmore intensive methods such as regular sounding.
The concept of an optimum level of leakage can be presented graphically in terms of costs against levelof leakage (see Figure 11.1, which is illustrative only). On the first graph, the curves show that as leakageincreases the value of water lost increases and the effort spent on ALC is small. Conversely where leakageactivities (and costs) increase the level of leakage will reduce. Adding these two curves identifies aminimum point of expenditure which is referred to as the optimum level of leakage.
The second graph shows the same curves in terms of their marginal costs. The marginal cost of water isfixed by the most expensive source of water and is the cost that would be saved by reducing the watersupplied by, for example, one cubic metre. The marginal cost of active leakage control is the cost, at agiven level of leakage, of reducing leakage by one more cubic metre. The optimum level of leakage in anyscenario is where the two marginal costs are equal.
Having collected data on the cost of water and the cost of active leakage control for a level of leakage, itis possible to prepare graphs as presented in Figure 11.1. The optimum level of leakage and the optimumspend on leakage control can be defined through an intensification of the current method of control.
11.2.4 Calculation of Optimum Level of Leakage
The marginal cost of leakage control is, therefore, the additional cost required to reduce leakage levels inan area by one unit. For example, if the leakage level in an area is 40m3/property/year, and the cost ofactive leakage control is £2.00/property/year, then if the cost increases to £2.50 to reduce leakage to39m3/property/year, then the marginal cost of the active leakage control is 50 pence/m3/property at thatlevel of leakage.
The difficulty in calculating the actual or marginal cost of leakage control is that only one point is known,namely the current operating conditions. Total and marginal cost follow the curves shown in Figure 1where the costs increase in some exponential form as the level of leakage is reduced. Before any economicoptimum can be derived, a method to estimate costs away from the current level must be established.
A possible approach is to assume that in any water supply zone, levels of leakage could range between twoextremes:
• A base level of leakage where all bursts are repaired, and the only leaks running are those whichcannot be detected by the current method of active leakage control. This base, or intrinsic level,can be approximated to by measuring the level attained following an intensive programme ofdetection and repair in a specific area.
• At the other extreme, if no money were spent on ALC, the level of leakage would be that controlledby customer-reported bursts. This is the passive level of leakage and can be obtained, given anassessment of the intrinsic leakage level, from a Passive leakage control curve like that illustratedby Figure 11.2. (This Figure should be taken as illustrative only, since it based on UK data fromthe late 1970s, at an unspecified pressure, and 24 hour supply.)
Between these two extremes is the actual level of leakage and the cost of ALC in the water supply area.An equation can be produced based on the two extremes and the actual data point, to give a form of ALCcost curve as illustrated in Figure 11.3, which is also based upon UK data..
Measurement of the current leakage level and cost will give sufficient information to use the followingequation, which typifies this curve, and is taken from the UKWIR ‘Managing Leakage’ Reports:
95
Total cost of leakage control = C = (-1/d) 1n ((L - Lb)/(Lp - Lb))
Where d is a constant, and = (-1/Ca) 1n ((La - Lb)/(Lp - Lb))The other terms are as follows:
L = level of leakage, m3/prop/yrC = cost of leakage control, £/prop/yrLa = actual level of leakage for the area, m3/prop/yrCa = actual cost of leakage control for the area, £/prop/yrLb = base level of leakage, m3/prop/yrLp = passive level of leakage, m3/prop/yr
Once a graph has been drawn, it is possible to estimate the level of leakage, corresponding to any givenlevel of resource input, and hence calculate the optimum level of control activity. This process willestablish the optimum economic level of leakage for the chosen control method.
To move to the optimum level of leakage will require a one-off additional expenditure on burst repairs.Costs associated with burst repairs will remain constant for any area in steady state conditions, but it islikely that by intensifying the method of active leakage control, an increasing number of background leakswill be found and include in the repairs. This ‘one-off’ cost of repairs should be included in any projectappraisal study. Assuming that burst occurrences remain constant over time, then once this ‘backlog’ ofleak repairs has been made, overall repair costs will return to the level that existed before. The possibleexception to this is that pressure reduction could produce a new, lower level of burst occurrence.
The calculation of the above optimum level assumes that the method of active leakage control and systempressure is held constant. It is equally possible to conduct similar studies to investigate the effect ofpressure control or changing to a new method of active leakage control if there is a need to reduce leakagefurther. In each case there will be a series of one off costs to establish the new approach, includinginstallation of PRV’s or district meters, staff training, district reconfiguration and backlog of repairs. Theassessment of the optimum level of leakage then will reflect the cost of water and the new shape of therelationship between the cost of active leakage control and the leakage level.
The marginal cost of active leakage control, at any level of leakage, can only be confirmed when the newpolicy is implemented.
11.2.5 Implementation and Performance Monitoring
The final element in developing a new leakage control strategy is implementation and performancemonitoring. Performance should be monitored through a programme of robust data collection. Datacollection and recording systems should be established for each water supply zone covering the followinginformation:
• Man hours spent on active leakage control, broken down between detection, location andequipment maintenance
• Level of leakage• Information on bursts• Repair costs and frequencies on mains, communication pipes and supply pipes.• Cost of new equipment installation, zone configuration and staff training• Power and chemicals (pumping and treatment)
Once any new policy is implemented it will take a period of time before its operation settles down. Thetiming of any new operational stability will be dependent on the completion of the repairs to the backlogof leaks. For each water supply zone subject to a change of policy, it is recommended that an economicoptimum level of leakage review should be conducted at this time (or annually, whichever is the shortertime). This will investigate whether the theoretical optimum has been achieved in reality and may providevaluable information for future leakage policy development.
96
97
It will be evident from the above that performance monitoring in itself carries its own costs, and certainlyrequires persistence and discipline to achieve and keep up to date the level of information required.Nonetheless, good project management demands good post-project appraisal if the right conclusions areto be reached for future planning.
11.3 The Unit Cost of Leakage
11.3.1 General
The unit cost of leakage is made up of two principal factors – operating cost and capital cost.
The normal requirement is to calculate the value of water likely to be saved by improved leakage control.Thus it is necessary firstly to estimate the likely extent of the savings so that for a given system thosesupplies and sources whose output will be reduced in the event of a reduction in leakage can be identified.Usually the reduction will be made in the most expensive source, but the extent to which this is possiblewill depend on the particular supply and distribution system configuration.
11.3.2 Unit Operating Cost
The unit operating cost, insofar as leakage is concerned, is the marginal cost of pumping (source, sourceworks, and any distribution) and treatment, or in the case of bulk purchases, the unit cost. The last unit ofleakage is likely to be more expensive than the first as in general it will be supplied by the most expensivesource.
Having determined the sources to be reduced, the unit operating cost can be calculated as follows:-
Unit Unit Unit UnitOperating = Pumping + Treatment /// PurchaseCost Cost Cost Cost for bulk supplies
This operating cost does not include any allowance for the following:
• Other source operation costs, including wages - some of these are flow related• Damage payments made by the Water Supplier as a result of leakage • Costs of dealing with leakage related customer complaints
It may be considered reasonable to add 5 to 10% for these elements.
If the operating cost, and all other costs associated with the determination of leakage control policies, arecalculated as an annual unit cost, it simplifies subsequent cost comparisons.
11.3.3 Unit Capital Cost
The capital cost element is determined by estimating the cost of future capital works required to satisfygrowth in demand. These costs are discounted to produce a total current cost equivalent. A reduction inleakage equivalent to one year’s growth will enable all these costs to be deferred by one year. Whilst notrelevant in every case, such a deferment would be a major element in arguing the case for more leakagecontrol development.
In order to most easily compare the costs of the various leakage control methods available, costs areconverted to an annual unit cost. Thus the net present value of one year’s deferment of capital costs ismultiplied by the discount rate to convert it to an annual equivalent, and then divided by the value of oneyear’s growth to produce the unit capital cost.
98
The formula is:
Unit Capital Cost = (TDCC x r2) ÷ [(l + r) x 365 x d]In £/m3
Where TDCC = Total discounted capital cost, £r = discount rated = yearly growth in demand, m3/day
In addition to the capital cost, it is also necessary to include fixed operating costs when calculating thetotal capital cost (TDCC), as these costs will also be deferred if schemes are put back in the programme.
(In simplified terms, given the present economic climate in the UK, the value of deferral for one yearmight be considered to sensibly be 5% of the capital value divided by the first year forecast growth indemand.)
Figure 11.4 illustrates the way in which reduction in leakage achieves deferment of capital expenditure.
11.4 Cost of Leakage Detection
11.4.1 General
Having determined the unit cost of the water, it is also of course necessary to calculate how much of it canbe saved and how much it will cost to do so. These factors depend upon the method of leakage detectionemployed and the frequency at which it is employed.
11.4.2 Leakage Estimates Associated with Detection Methods
It is possible to estimate the resulting level of leakage applicable to each detection method in a simple wayby the use of a graph like Figure 11.2, which is based on experimental data gathered from across the UKin the late 1970s. Water Suppliers (or national groupings) need their own local, reliable data to compile asimilar graph for their own situation.
Technological advances mean that Figure 11.2 now requires amendment anyway. There is less distinctionnow between the various methods of metering, and in practice a more modern and sensitive version ofDistrict Metering is normally employed, as previously explained. LNCs are used much more now, and aremuch improved. Permanent acoustic logging is another new dimension. Lower levels of leakage are alsoattainable now because of the improved sensitivity achieved by modern meters and loggers.
Where passive control is currently exercised, use of Figure 11.2 is straightforward. The average net nightflow for the system is measured and a horizontal line is drawn on the graph at this flow. At the point wherethis line crosses the passive control line, a vertical line is drawn. The intersection of this vertical line withthe lines representing the other methods of control gives the net night flow applicable to each of thosemethods.
If some form of active control is being carried out, the level of activity, or frequency of existing control,is likely to have a significant effect. In such cases some judgement is required in assessing the true intrinsicleakage level.
Clearly, the intrinsic level of any area will also depend on the physical characteristics of the system andon the average system pressure. If much of the system is relatively new and ground conditions are non-aggressive then the intrinsic level will be low. Alternatively, if the system consists predominantly of oldunlined iron mains in aggressive soil conditions, the intrinsic level will be high.
The average burst rate will also provide a clue to the correct level.
It must also be noted that:
- Unusual weather conditions may affect the burst rate, and must be taken into account
- The average system pressure will affect the intrinsic leakage level - thus a medium burst rate andhigh average system pressure will give rise to relatively high intrinsic leakage levels
It should be again emphasised that the figures in Figure 11.2 are not related to pressure, and could bemisleading if they are not carefully considered in the light of the ‘local’ conditions and more moderntechnology.
11.4.3 Annual Cost of Leakage Estimates
Having determined the likely net night flow for a leakage control method, the annual cost of leakage foreach can be calculated as follows bearing in mind that net night flow on the graph includes legitimatedomestic use, estimated in the UK at 21/property/hr:
Annual Cost = (leakage x 20 x 365 x C) /100,000 (£ per prop)
Where leakage is in 1/prop/hr and C is the unit annual cost of leakage in pence/m3
Note that a factor of 20 is used to convert from a night flow to a daily average figure. As previouslymentioned, this takes account of the lower pressure, and hence reduced leakage, which normally appliesduring the day.
11.4.4 Frequency of Detection Activity
Leakage estimates and annual costs of methods will be dependent on the frequencies of detectionactivities. Recommended frequencies from UK research in 1980 are reproduced in Table 11.1. Clearly, thefrequencies adopted will have an impact on the net night flow achieved. However, the law of diminishingreturns will apply, and this can be conveniently illustrated by Figures 11.5 and 11.6. The volume of watersaved is shown by the shaded portion of the graph in Figure 11.5. A doubling of the detection frequencydoes not double the leakage saved. In this case the additional saving in leakage is shown hatched and isequal to half the previous saving.
11.4.5 Annual Cost of Detection Methods
These costs will comprise of three elements, namely:
(a) Initial setting up costs, which include purchase of equipment and installation of meters(b) The initial cost of applying the chosen method throughout the system(c) The annual operating cost, which will include manpower costs, maintenance, and replacement of
equipment
Table 11.2 shows typical UK costs for these three elements. The annual operating cost will, as noted above,depend on the frequency with which the various leakage control tasks are carried out, and also on therequired maintenance frequencies. In the compilation of Table 11.2, as noted, the recommendedfrequencies of Table 11.1 have been assumed.
Table 11.3 details the component parts of the annual costs in terms of man-hours per task.
Having determined the costs of the above components, it is necessary to calculate the total annual cost ofeach method. This is done by converting the initial costs to annual costs by multiplying by the chosendiscount rate, and then adding the annual operating cost.
Table 11.4 shows typical total annual costs calculated for each method based on the data in the previoustables, and assuming a discount rate of 5%.
99
100
11.4.6 Annual Cost of Repairing Leaks
The cost of repairing leaks is not included in the annual operating cost calculation for choice of methodsbecause, for a given distribution system, the rate at which leaks occur will not be affected by the chosenmethod. Thus once the method has been fully implemented and the corresponding leakage levelestablished, the number of repairs required will be the same for each method, and therefore the cost ofrepairs can be ignored when comparing methods.
In moving from the current leakage level to a reduced level, however, it will be necessary to locate andrepair an increased number of leaks. Whilst the cost of repairing this backlog will vary according to boththe starting leakage level and the control method being considered, the difference between the variousmethods of active control is likely to be small and can be ignored. The major difference is between passiveand active control, and an average figure, based on the 1980 UK research, has been included in Table 11.2as part of the initial costs.
11.5 Typical Total Leakage Costs
Typical total leakage costs for each of the methods are shown in Table 11.5.
In compiling this table the following assumptions were made:-
• The intrinsic leakage level is ‘medium’ as defined in Figure 11.2 • Initial costs of each method are as shown in Table 11.2• The annual detection costs are based on the recommended frequencies in Table 11.1
The figures in Table 11.5 are indicative only. However, it is clear that passive control is only economic atlow unit leakage costs or where intrinsic leakage levels are low.
The assumptions made in performing the calculations mean that differences in total leakage cost of lessthan 10% are not significant. Thus the choice between two methods with cost differences of less than 10%must be based on other factors.
11.6 Environmental and Social Costs
These costs are so specific to individual situations, and so open to interpretation and influence by the local‘political’ environment, that they are not explored in this section. They may, however, have a significantbearing on a leakage level target.
101
Table 11.1 Recommended Frequencies of Leakage Control Activities
Recommended Acceptable RangeFrequency (No/Year) (No/Year)
Regular sounding 1 0.5 to 2
Leak Noise Correlation 1.5 1 to 4
District Metering - Read meters 25 12 to 50- Inspection 1 0.5 to 2
Waste Meteringa) Areas up to 1500 props - Monitoring 4 2 to 6
- Inspection 1.5 1.25 to 2b) Areas over 1500 props - Monitoring 5 3 to 12
- Inspection 2.5 1.75 to 3
Combined District and Waste Metering- Read district meters 25 12 to 50- Inspection 2.25 1.75 to 3
Footnotes to Table 11.1
1. Recommended frequencies are generally taken from UK research except as noted below.
2. District meter reading rates have been adjusted to reflect some improvements in technology.
3. Inspection frequencies given for waste metering include the use of step tests. Thus, if for example, aWater Supplier has 100 WMAs, and the inspection frequency is 1.5, then 150 WMAs should be inspectedper year. The areas to be inspected will be determined by meter readings. Inspection will normally consistof a step test followed by inspection of between 45% and 65% of the area.
4. Inspection frequencies given for combined metering include recording MNFs using waste meters andsubsequent step tests and sounding as per note 3.
102
Table 11.2 Typical Component Costs for Leakage Control
ACTIVITY MEAN COST TYPICAL RANGE£ OF COSTS £
a) Initial Setting-Up Costs 4,500 3,000-9,000
Install PRV and set up pressure reduced area
Install District meters and set up DMA 5,000 3,500-10,000
(ave. 1.6 meters per DMA)
Install Waste Meter 3,000 2,000-5,000
Purchase Leak Noise Correlator 12,000 5,000-15,000
b) Initial Cost of Applying the Chosen MethodPer 1,000 PropertiesPassive NIL NIL
Regular sounding - Inspect properties 1,100 750-2,200
- Repair backlog of leaks 850 550- 1,400
Leak Noise Correlation - Inspect properties 800 550-1,700
- Repair backlog of leaks 850 550-1,400
District Metering - Read and record 2.3 1.8-2.8
meter readings
(DMA size) Inspect District 1,100 750-2,200
(assumed to be) Repair backlog of leaks 850 550-1,400
(4,000 properties) Read and record meter 2.3 1.8-2.8
readings
Waste Metering - Record MNF 85 45-140
(WMA size) Perform Step Test 170 90-220
(assumed to be) Inspect WMA 1,100 750-2,200
(2,000 properties) Repair backlog of leaks 850 550-1,400
Record MNF 85 45-140
Combined District and Waste Metering
- Read and record district meters 2.3 1.8-2.8
- Record MNFs of Waste areas 85 45-140
- Perform Step Tests 170 90-220
- Inspect whole District 1,100 750-2,200
- Repair backlog of leaks 850 550-1,400
- Record MNFs of waste areas 85 45-140
- Read and record district meters 2.3 1.8-2.8
103
ACTIVITY MEAN COST TYPICAL RANGE£ OF COSTS £
c) Annual Operating Costs (per 1,000 properties)based on recomm’d frequencies in Table 11.1
Passive 430 280-2,200
Regular soundi ng 1,100 750-2,200
Leak Noise Correlation Operation 1,200 850-2,500
Maintenance of 9 4-19
Equipment
District Metering Monitoring 55 45-75
Inspection 1,100 750-2,200
Maintenance of meters 35 28-55
Waste Metering Record MNFs 420 230-700
Perform Step Tests 370 250-500
Inspection 2,400 1,700-5,000
Maintenance of meters 55 35-75
Combined Metering Monitoring 55 45-75
Record MNFs 190 110-170
Perform Step Tests 370 250-500
Inspection 2,400 1,700-5,000
Maintenance of meters 75 55-110
Footnotes to Table 11.2
1. Initial costs do not include design.2. Labour rate assumed at £18.60 per hour, inclusive of overheads.3. Vehicle rate assumed at £3.80 per hour, inclusive of fuel.4. Assumed to be 1.6 meters per DMA5. Setting up of DMAs assumed to take 30 man hours.6. Setting up of PRV areas assumed to take 40 man hours.
104
Table 11.3 Typical Resource Requirements for the Regular Operating Components of LeakageControl
Activity Resource Requirement
Locate reported leaks when operating passive control. 20 man hours per 1000 properties
Sound 1000 houses, including locating detected leaks. 50 man hours
Carry out leakage survey of 1000 houses with leak noise correlator. 40 man hours
Locate leak using leak noise correlator. 4 man hours
Record the minimum night flow in a WMA using a fixed 8 man hoursmeter.
Record the minimum night flow in a WMA using a mobile meter. 10 man hours
Perform a step test. 12 man hours
Meter maintenance 3 man hours per annum per meter +materials at say £40 per meter
PRV maintenance 8 man hours per annum per PRV + materials at say £50 per PRV
Footnotes to Table 11.3
1. Figures for locating leaks with passive control or sounding are average figures based on UKresearch in 1980.
2. Figures quoted are actual man hours to carry out the activity, no allowance having been made forovertime.
3. Reading district meter includes manual recording of total flow and previous seven MNFs.
105
Table 11.4 Typical Annual Costs per Property of Different Methods of Leakage Control
METHOD OF LEAKAGE INITIAL COST ANNUAL COST TOTAL ANNUALCONTROL (£/prop) (£/prop/year) COST
(£/prop/year)
Passive NIL 0.43 0.43
Regular Sounding 1.95 1.1 1.2
Leak Noise Correlation 1.95 1.21 1.31
District Metering 3.2 1.19 1.35
Waste Metering 3.79 3.25 3.43
Combined District and Waste Metering 4.29 3.09 3.3
Footnotes to Table 11.4
1. A discount rate of 5% has been assumed.2. Costs quoted are based on the mean costs given in Table 11.2.3. For calculating initial set up costs for leak noise correlation, it has been assumed that one correlator
will be required for every 40,000 properties.4. Assumed that some district meters are also used as waste meters when considering combined
metering. Thus, allowance made for one waste meter for every 4,000 properties in addition to the1.6 district meters per 4,000 properties.
Table 11.5 Typical Total Annual Costs of Leakage and Leakage Control (£ / prop / year)
LEAKAGE CONTROL UNIT COST OF LEAKAGE (p/m3)
METHOD 2 4 6 8 10 12
Passive 3.12 5.8 8.49 11.18 13.9 16.6
Regular Sounding 2.66 4.12 5.58 7.04 8.5 9.96
Leak Noise Correlation 2.77 4.23 5.69 7.15 8.61 10.1
District Metering 2.52 3.69 4.85 6.02 7.19 8.36
Waste Metering 4.31 5.18 6.06 6.93 7.81 8.69
Combined District Metering and Waster Metering 4.18 5.05 5.93 6.8 7.68 8.56
Footnotes to Table 11.5
1. The total annual cost of detection is taken from Table 11.4.
106
Figure 11.1 The Optimal Level of Leakage
2 4 6 8 10 12 14
Leakage (l/prop/hr)
Total
WaterLost
Cost
Active Leakage Control
15
10
5
0
Cost(£/prop/yr)
Optimum Level
Marginal Cost(£/prop/yr)
2 4 6 8 10 12 14
Leakage (l/prop/hr)
Most Expensive Source Water
Marginal Cost
Active Leakage Control
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
107
Figure 11.2 Graph for Prediction of Net Night Flows
Low Medium High
INTRINSIC LEAGAGE LEVELS
Passiveleakagecontrol
RegularSounding
DistrictMetering
Waste Meteringand combinedmetering
NE
TN
IGH
TF
LOW
(l/p
rop/
hr)
108
Figure 11.3
TToottaall CCoosstt
ooff LLeeaakkaaggeeCC
oonn
ttrrooll
((££//pprroopp//yyrr))
LLeevveell ooff LLeeaakkaaggee ((mm33//pprroopp//yyrr))
001100
22003300
44005500
66007700
88009900
110000111100
112200113300
Base
Level of Leakage
Passive Level of Leakage
1144112211008866442200
Figure 11.4 Diagrammatic Representation of the Deferment of Capital
Figure 11.5 The Effect of Increased Detection Frequency on Leakage Level
109
Present capacity or Yield
TIME (years)
Passiveleakagecontrol
Activeleakagecontrol
Reductionin leakage
Consequentdeferment of
capital scheme
ACTIVECONTROL
PASSIVECONTROL
COMPLAINTSLEVEL
t
P
A
M
C
TIME
Extra volume saved by doubling Extra volume saved by againthe detection frequency doubling the detection frequency
LEA
KA
GE
RA
TE
Passive
Active
Minimum
110
Figure 11.6 Marginal Benefits of Increased Detection
0 1 2 3 4 5 6 7Passive
LEAKAGESAVED
t x P-M 2
No. of DETECTION EXERCISES in TIME t
111
12.1 Introduction
Measured minimum night flows into well-defined District Meter Areas (DMAs) of moderate size are avery effective means by which leakage control teams can become aware of unreported bursts, andprioritise activities to locate and repair them.
Thanks to advances in new technology, flows and pressures in DMAs can now be routinely recordedrelatively cheaply, and Night Flow Monitoring has become the predominant leakage control method in theUK.
This section seeks to collate some of the key elements of recent research and thinking, co-ordinated by theWater Research Centre, in a format which can be used and developed by leakage practitioners workingwith the measured night flow data now being routinely collected for leakage management in the UK andelsewhere.
Appropriate methods of analysis and interpretation are needed for this data.
12.2 Bursts and Background Losses
Flow rates of individual leaks, bursts and overflows are pressure dependent and can range from less than1 l/hr to more than 100,000 l/hr. For comparison, a fully open tap or hosepipe can run at around 1000 l/hr.
For the purpose of analysis all individual sources of losses from pipes and fittings can be classified (byflow rate) into two categories:
• BURSTS - individually more than 500 l/hr at 50 m pressure; these may be ‘reported’ (not,therefore, requiring detection), or ‘unreported’ (awaiting detection)
• BACKGROUND LOSSES - individually less than 500 l/hr at 50 m pressure.
12.3 Components of Night Flows
Figure 12.1 shows the components of minimum night flow. This diagram relates to UK practice, where,at the ‘point of delivery’, the responsibility for the service pipe changes from the Supplier’s undergroundcommunication pipe to the customer’s supply pipe.
Minimum Night Flows (MNFs) consist of differently sized components of customer night use from alimited proportion of properties which are ‘active’ on any given night, and losses (leaks, bursts, overflows)from a relatively small number of points on the Supplier’s distribution system and customers’ supplypipes.
Measured MNFs provide awareness of unreported bursts, which could otherwise run for long periods andaccumulate considerable annual volumes of losses.
The variation of MNFs in a DMA over a longer period of time can be considered to consist of thefollowing components (Figure 12.2):-
• Exceptional Individual Night Use exceeding 500 l/hr• Assessed Normal Customer Household and Non-Household Night Use• Background Losses on Mains, Communications Pipes and Supply Pipes• Bursts (of finite duration), both reported and unreported, on Mains, Communication Pipes and
Supply Pipes.
The components of minimum night flow from Figure 12.1 can be further subdivided, based on bursts andbackground losses concepts, as shown in Table 12.1. The values shown in Table 12.1 have been assessedfrom recent UK research into night flow measurements.
12. RECENT RESEARCH AND DEVELOPMENT IN THE INTERPRETATIONAND USE OF NIGHT FLOW DATA
112
Table 12.1 Best estimates of average values for components of 1 hour minimum nightflow at 50 m AZNP
Component Value or Method Assessment
Distribution (Pressure dependent)
Distribution mains Background losses* 40 l/hr x length of mains (km)
Distribution mains Reported bursts Number x flow rate (depends on pipe size)
Distribution mains Unreported bursts Number x flow rate (depends on pipe size)
Communication pipes Background losses* 3 l/hr x number of props
Communication pipes Reported and unreported Number x 1.6 m_/hrbursts
Supply pipe (Pressure dependent)
Underground supply Background losses* 0.5 l/hr x number of propspipe
Underground supply Reported and unreported Number x 1.6 m_/hrpipe bursts
Plumbing Background losses* 0.5 l/hr x number of props
Customer Night Use (Not normally pressure dependent)
Households Normal night use 1.7 l/hr x no of households (or 0.6 l/hr x no of persons)
Non-households - Group A 0.9 l/hr x number in Group Anormal use Group B 6.2 l/hr x number in Group B
Group C 12.6 l/hr x number in Group CGroup D 20.5 l/hr x number in Group DGroup E 60 l/hr x number in Group E
Households and non- >500 l/hr individually Sum of individual usershouseholds
exceptional night use
* These are ‘average’ values for background losses. ‘High’ and ‘Low’ values (respectively +/-50%of the average) can be expected depending upon infrastructure condition.
12.4 Night Flow and Customer Use
12.4.1 Household Night Use
From a variety of tests involving 8847 UK households, it was concluded that the average household nightuse is 1.7 l/prop/hr (0.6 1/person/hr); this does not include exceptional use for hosepipes or purposesequivalent to a fully open tap.
This average use is generated principally by a small percentage of ‘active’ properties (17 %).
Average use is also sensitive to small changes in numbers of properties using washingmachines/dishwashers overnight.
12.4.2 Non-Household Night Use
Non-household night use is highly variable. Table 12.2 shows data from a sample of 3000 external meters,categorised into 5 groups for the purpose of estimating assessed non-household night use. These figuresexclude individual non-household night use of more than 0.5m3/hr.
Table 12.2 Average values of night flow delivered to different types of non-household,grouped by similar averages
Group Sample Number % Average per Averagesize active Active Active for all
Property Propertiesl/prop/hr l/prop/hr
A: Unmanned fire/police stations,telephone exchanges, banks,church/chapels, gardens/ 123 16 13 7 0.9allotments, market gardens,water/sewage treatment works
B: Shops, offices, craft centres, launderettes, depots, large domestic properties, guest 2013 606 30 20.5 6.2houses, garage/filling stations, touring caravan sites, farms, small holdings and cattle troughs
C: Hotels, schools/colleges,cafes/restaurants, public houses, 505 244 48 26 12.6social halls, residential caravansites, livery stables
D:* Hospitals, factories (food andmanufacturers), public toilets, 205 79 39 53 20.5works sites
E: Old peoples’ homes, mines 33 25 76 80 60.6and quarries
* An alternative for larger nursing homes and hospitals is 2.5 l/resident/hr
A simplified estimation of 81/non-household/hour can be used where the property information has notbeen classified as in the above table.
113
12.5 Losses from Bursts
12.5.1 Losses from Bursts on Underground Services
Losses from a relatively small number of unreported bursts on underground service pipes (as few a 2 perthousand properties per year) can constitute a significant and highly variable component of night flowdelivered and annual water delivered. The volume lost from a single burst is the product of average flowrate x duration.
Burst flow data shows that service pipe bursts have a median (50 percentile) flow rate of 1,600 l/hr at 50mAZNP (average zonal night pressure) when located, and that a large proportion of supply pipe bursts maynot be reported. Only one such burst on an underground supply pipe (equivalent to a large tap running fullyopen) will:
• Increase the minimum night flow of a 1000 property DMA by 1.6 l/prop/hr• Lose 32 m3/day, or (if allowed to run a year) around 11,700 m3/yr (the annual consumption of 100
typical households).
Using this methodology (background and bursts) it can be demonstrated that annual average losses fromunderground service pipes can vary from 10 to 100 l/prop/day, depending upon local circumstances,notably:
• Pressure• Reported burst frequencies• Unreported burst frequencies• Leakage control or external metering policy (awareness of unreported bursts)• Economic justification for sending a leakage control team out to try to locate only one or two
suspected unreported service pipe bursts (equivalent to single open taps) in a DMA• Policy for enforcement of private supply pipe repairs.
12.5.2 Losses from Bursts on Mains
Figure 12.3 gives some indication of loss rates that may be expected from mains bursts, both reported andunreported, the difference being that reported bursts are brought to the attention of the Water Supplier,regardless of monitoring. The information should be used cautiously and only in the initial assessment ofthe likelihood of leakage, since mains bursts in particular can be highly variable.
12.5.3 Effect of Burst Duration
It is obvious that the average MNF over a year, for any particular DMA, will be influenced by the WaterSupplier’s policies for locating and repairing unreported bursts. This is easily seen from Figure 12.2; if theunreported burst in August had been allowed to run another 4 months before location and repair, the annualaverage MNF would have been higher.
Figure 12.4 shows the effect that duration can have on the volume of losses. The extra volume lost on thesupply pipe burst illustrates the procedural difficulties associated with the customer’s ownership of thatpart of the service pipe. The free repair service now offered by Water Suppliers in the UK, after some‘political’ pressure, is the Industry’s response to reduce the repair time and to save water which, in theabsence of domestic metering at the property boundary, is not paid for directly by the customer, eventhough it is his responsibility.
It should not be ignored either that reducing awareness and location times as well as repair times can havea significant impact on the overall quantity wasted. Improving administration and communicationprocedures is a part of this.
The annual average MNF is therefore influenced by the active leakage control methods, and the timescalesfor awareness/location/repair of unreported bursts, and the repair times for reported bursts.
114
12.6 Estimating Background Night Flows in Individual DMA’s
12.6.1 Introduction
It is not helpful to set the same net night flow targets for all DMAs in a supply area, irrespective ofpressure, mains length per property, percentage of non-households, and infrastructure condition.
If night flow targets for individual DMAs are not realistic, manpower resources can be uneconomicallydeployed to look for unreported bursts in DMAs where there are only minor ‘background’ seepages.
There is need therefore to incorporate other parameters (notably pressure, mains length and non-householdnight use) in setting realistic night flow targets for individual DMAs.The ‘Background Night Flows’ approach can be used to estimate, at any time, the excess night flowattributable to unreported bursts. The DMAs with excess losses can then be identified and prioritised (forscheduling burst location activities) by cost of losses, or other relevant parameters.
12.6.2 Methodology
A methodology to estimate background minimum night flows in individual DMAs, given all relevant localcharacteristics (mains length, number of households and non-households, pressure) is potentially ofsignificant value. It could indicate the night flow at which it is no longer appropriate to allocate resourcesto try to locate significant unreported bursts in that DMA.
Such a methodology also provides an independent check on the minimum night flow achieved when aDMA is initially set up, after the ‘best practice’ of thoroughly checking the DMA for leaks by step-testingand sounding has been carried out.
The background night flow losses (when no bursts exist in a DMA) can be calculated for any DMA (givenL (length of mains in km), N (number of properties), AZNP (average zonal night pressure in m)) from theequation:-
NFLB(l/hr) = (C1 x L + (C2 + C3) x N) x PCF
using the following values of C1, C2 and C3 from Table 12.3, and the pressure correction factor (PCF)from Table 12.4, based on the UK research of the 1980s.
Table 12.3 Background Night Flow Losses
Background Losses Units Low Average HighComponent
C1: Dist mains l/km/hr 20 40 60
C2 : Commun pipes l/prop/hr 1.5 3.0 4.5
C3 : Supply pipes l/prop/hr 0.5 1.0 1.5
115
Table 12.4 Pressure Correction Factors
AZNP 20 30 40 50 60 70 80 90 100(metres)
PCF .329 .529 .753 1.00 1.271 1.565 1.884 2.226 2.592
The background minimum night flow for an individual DMA can be estimated by using the indicativevalues for background losses (in Table 12.3), and adding the following components for night flowdelivered:
• 1.7 l/prop/hr for normal household night use (excluding individual use > 500 1/hr)• an appropriate allowance for normal non-household night use (see Section 12.4.2)• exceptional individual customer use (household or non-household) >500 1/hr.
Table 12.5 shows the average values of background minimum night flow, at 50m AZNP, for DMAs withdifferent values of L/N, with an overall allowance of 8 l/prop/hr for non-household night use.
Table 12.5 Background net night flows in l/prop/hr at 50 m AZNP, from ComponentAnalysis (Assuming no exceptional customer use > 500 l/hr, and 10% ofproperties are non-households using 8 l/prop/hr)
Components Average Net night flow in l/prop/hrValues Mains length per property
L/N = 10 L/N = 20 L/N = 50 L/N = 100
Distribution Mains 40 l/km/hr 0.4 0.8 2.0 4.0Losses Commun 3.0 l/prop/hr 3.0 3.0 3.0 3.0
pipes
Supply Pipe Underground 1.0 l/prop/hr 1.0 1.0 1.0 1.0Losses + Plumbing
Assessed Households 1.7 l/prop/hr 1.53 1.53 1.53 1.53Customer (90%)Night Use Non 8 l/prop/hr 0.8 0.8 0.8 0.8
Households(10%)
TOTAL for ‘average’ background losses 6.73 7.13 8.33 10.33l/prop/hr l/prop/hr l/prop/hr l/prop/hr
Allowing for ‘high’ background losses 8.93 9.53 11.33 14.33l/prop/hr l/prop/hr l/prop/hr l/prop/hr
Allowing for ‘low’ background losses 4.53 4.73 5.33 6.33l/prop/hr l/prop/hr l/prop/hr l/prop/hr
116
117
Table 12.6 Example Calculation for estimating background 1 hour minimum nightflows for any district meter area
CUSTOMER NIGHT USE COMPONENTS SUB TOTAL TOTALl/hr l/hr
SUM OF EXCEPTIONAL NIGHT USERS HOUSEHOLDS 0>500 l/hr individually NON HOUSEHOLDS 1000 1000
ASSESSED HOUSEHOLD NIGHT USE:1.7 l / household / hr x No of properties (NH) or 1.7 x 900 15300.6 l / resident / hr x No of residents (n)
ASSESSED NON - HOUSEHOLD NIGHT USE: 8 x 100 800A
SIMPLIFIED: 8 l / non-household / hour x BNo of non-households (NNH) or C
DDETAILED: Classified by Groups A to E E = 3330
BACKGROUND LOSSES AT 50m AZNPCONDITION L/KM/HR LENGTH (KM) SUB TOTAL
DISTRIBUTION Good 20 10 200MAINS Average 40
Fair 60
CONDITION L/PROP/HR NO OF PROPSSERVICES Good 2.0 1000 2000
Average 4.0Fair 6.0
2200AZNP (meters) 60 x PCF x 1.27 2794PRESSURE CORRECTION FACTOR (PCF)
AZNP(m) 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120
PCF 0.33 0.43 0.53 0.64 0.75 0.87 1.00 1.13 1.27 1.42 1.57 1.72 1.88 2.05 2.23 2.41 2.59 2.78 2.98 3.18 3.39
TOTAL BACKGROUND MINIMUM NIGHT FLOW (l / hr) 6124
STANDARD DEVIATION OF SUB TOTAL TOTAL‘ONE-OFF’ MEASUREMENT l / hr l / hr
ASSESSED HOUSEHOLD 3.8 √ NH or 2.4 √ n 3.8 x √900 114 224NIGHT USE
ASSESSED NON - HOUSEHOLD 11 √ NNH 11 x √100 110NIGHT USE
The sample calculation (Table 12.6) suggests that a 1000 property urban DMA at 60m AZNP with 10kmof mains, 10% non-households, ‘good’ infrastructure condition, and 1 m3/hr exceptional night use wouldhave a background night flow of around 6.1 m3/hr.
It must be remembered that these are only estimates for the purpose of repetitive calculations which areintended to identify DMAs with unreported bursts, and to prioritise activities to locate them.
Table 12.6 also shows that an individual one-off measurement in this DMA might have a standarddeviation of 0.22 m3/hr arising from variability in assessed customer night use. This implies that (ifexceptional night use does not vary) at the 95% confidence level (+/- 2 standard deviations) a ‘one-off’measured background night flow could be anywhere between 5.6 and 6.6 m3/hr.
In this DMA, at 60 m AZNP, measured night flows of around 13 m3/hr would be indicative of the existenceof a typical unreported l00mm mains burst (7m3/hr at 60m AZNP) or 3 to 4 service pipe bursts (typically2m3/hr at 60m AZNP). However, night flows consistently running at around 8 to 9 m3/hr could include asingle service pipe burst, or be due to infrastructure in ‘average’ (rather than ‘good’) condition,
The smaller the DMA size, the easier it is to differentiate unreported bursts from the background nightflows; 1000 properties is near the upper limit of size for awareness of a single unreported service pipeburst. Continuous recording of night flows (rather than one-off measurements) helps considerably toreduce uncertainty, particularly if exceptional non-household night use is very variable from night to night.
Where populations vary seasonally (e.g. DMAs in holiday areas), or the number of billed properties is nota good indicator of numbers of people (e.g. in city centres with large blocks of flats), it is preferable to usenumbers of residents to estimate assessed household night use (and its variability) in Table 12.6.
The methodology permits approximate (rather than precisely auditable) figures to be used to completeTable 12.6. The key elements are numbers of properties, AZNP and exceptional individual users. Numberof non-households can initially be estimated from the overall percentage for the Supplier.
The principal ‘unknown’ factor is the general infrastructure condition, insofar as it affects backgroundlosses. The methodology in Figure 12.5 proposes that an initial calculation of ‘excess’ night flow is made,using the assumption that background losses are ‘low’, and that large anomalies are then investigated (bysounding/step testing), to identify unreported bursts and unsuspected exceptional night users. Anyremaining anomaly can then be attributed to infrastructure, and the assumed value of background lossesin subsequent calculations adjusted to reflect this.
12.6.3 Units for Expressing Night Flows
The ‘background losses’ approach implies that it is preferable to set targets and define ‘background nightflows’ (and their viability) in m3/hr, based on a number of DMA local characteristics.
For example, Figure 12.6 shows that, if background net night flows are expressed in 1/prop/hr, forinfrastructure in ‘good’ condition, the NNFs would vary between 3 1/prop/hr (20m AZNP, urban L/N l0m)and 12 1/prop/hr (l00m AZNP, rural L/N l00m). (L/N is an expression for the length of main per property).
It is also important to remember that the average net night flows over 12 months would always be higherthan the background net night flows in Figure 12.6 (which excludes the volume lost through reported andunreported bursts during the year).
12.7 Prioritising Unreported Burst Location Activities
12.7.1 Introduction
In any group of DMAs, there will always be some bursts temporarily running at any time. It is thepractitioner’s job to assess, from the night flow data, which DMAs are most likely (at any particular time)to have high levels of economically recoverable losses. As skilled manpower resources are always limited,it is necessary to prioritise activities systematically (both within and between DMAs), in order to locate(by step testing/acoustic survey) the leaks and bursts which are causing these losses, many of them beingunreported (assuming that the suspected burst has not caused a failure in standards of service tocustomers).
This can be easily carried out using spreadsheets.
12.7.2 Prioritising by Excess Volume
The most basic form of prioritisation is by excess volume - that is, the amount by which the measurednight flow exceeds the target night flow in m3/hr. The target night flow may be based on either:
118
119
METHOD 1: The lowest achievable night flow established following intensive step tests/sounding andrepairs.
METHOD 2: The ‘background minimum night flows’ approach.
In practice, the methods are complementary to each other; Method 2 can be used to check that no leakshave been missed in using Method 1, and Method 1 is needed to check the cause of apparent excess nightflows from Method 2.
12.7.3 Alternative Prioritisation Concepts
Excess volume in m3/hr can be multiplied by marginal costs in each DMA (to prioritise on the basis ofexcess costs) or used to estimate the probable number of unreported bursts in each DMA.
Either priority may be over-ridden where unreported bursts cause failure of Levels of Service tocustomers, through low pressure, lack of water, or through shortage of resources during droughts.
12.7.4 The Equivalent Service Pipe Bursts (ESPB) Concept
Recent research work into typical burst flow rates permits consideration of the merits of expressing lossesin terms of ‘Equivalent Service Pipe Bursts’.
With this concept, the median service pipe burst flow rate at 50m AZNP (1.6m3/hr) is converted to a flowrate at the AZNP in the DMA using the square root relationship for individual bursts, i.e.
1.6 x √ (AZNP/50)
The calculated excess night flow for each DMA is then expressed as an equivalent number of service pipebursts. This gives an immediate indication of the maximum number of bursts which are to be looked forin each DMA (recognising that a mains burst is equivalent to three or more service pipe bursts).
A week-by-week addition of the calculated ESPB in all DMAs is likely to be a sensitive indicator of trendsin leakage control performance.
12.7.5 Economics of Unreported Burst Location Activities
The manpower resources which are allocated to location of unreported bursts will depend on the overalllocal economics of leakage control.
However, given any particular level of manpower resources for unreported burst location, it is preferableto use them in the most economic way, i.e. to find the biggest bursts fastest.
If the number of properties is divided by the number of equivalent service pipe bursts, the resulting figuregives an indication of the likely speed of location; working in a DMA with 200 properties per ESPB islikely to be more efficient than in a DMA with 1000 properties per ESPB.
120
Figure 12.1 Components of Night Flows (not to scale)
121
Figure 12.2 Components of Minimum Night Flows
Figure 12.3 Median Flow Rate of Bursts (l/sec) at 50m AZNP vs Mains Diameter (mm)
Bursts
Backgroundlosses
Use
Household
Non-household
Household
Non-household
Mains
Comm Pipes
Supply Pipes
Exceptional night use (individual > 500 l / hour)
Normal night use
Background losses
B R
BR
BR
B R
B R
Reported bursts B = Onset of burst
Unreported bursts R = Repair of burst
Minimumnight flow(m3.hr)
(Sum ofcomponentsrepresentedby thickblack line)
Jan Time Dec
MMiinniimmuummnniigghhtt ffllooww
((mm33//hhrr))
(Sum of componentsrepresented
by thickblack line)
122
Figure 12.4 Estimated Durations and Flow Rates of Reported Bursts
32m_/dayat 50mAZNP
16 Days
A RL
Reported communication pipe burst:Volume lost = 512m_
32m_/dayat 50mAZNP
46 Days
A RL
Reported supply pipe burst:Volume lost = 1472m_
216
m_/dayat 50mAZNP
1.1 Days
Reported 100mm distribution mains burst:Volume lost = 238m_
A = Awareness L = Location R = Repair
m3/dayat 50mAZNP
m3/dayat 50mAZNP
m3/dayat 50mAZNP
238m3
512m3
1472m3
123
Figure 12.5 Methodology for Assessing Likely Excess Night Flow from Unreported Bursts in aDistrict Meter Area (DMA)
Investigate and Identify Burstsand Exceptional Night Use;Attribute remainder to Background
Losses and categorise as ‘Average’or ‘High’
Positive difference (Excess)may be due to Unreported Bursts,Unknown Exceptional Night Uses,
or Infrastructure in ‘Average’ or‘Poor’ Condition
+ve Difference?
Deduct Background Night Flow toobtain difference (l/hr)
No Further Action at Present
Calculate StandardDeviation of AssessedNight Use, and Deductfrom Measured Value*
‘One-Off’measurement
Average ofSeries of
measurements
Obtain measured Minimum NightFlow over 1 hour in l/hr
Estimate Background MinimumNight flow over 1 hour, using best
estimates of Exceptional Night Use,simplified assessment of Non-
Household Night Use, and‘Background Losses’ for assessed
‘Good’ infrastructure condition
START
* This r educes the chance to 1 in 6that the adjusted value is morethan the average value from aseries of measurements.
124
Figure 12.6 Net Night Flows with ‘Low’ Background Losses (i.e. infrastructure in ‘good’condition)
The progress made in leakage management in the UK has been considerably assisted by the Water Research Centre,working with the Water Industry. There have been two major reports published of this important work. They are:
• “Leakage Control Policy and Practice”, Standing Technical Committee Report No. 26, July 1980, ISBN 0904561 95 X
• UKWIR “Managing Leakage” series of Reports, 1994
Figures 10.2, 10.3 and 11.2 are taken from the first Report.Figures 1.1, 11.1, 11.2, 11.3, 12.1, 12.2, 12.3, 12.4, 12.5, 12.6, andTables 12.1, 12.2, 12/3. 12.4, 12.5 and 12.6 are taken from the second Report.
(The first number in each of the references above relates to the relevant section of this document)
125
ACKNOWLEDGMENTS
126
Published by
Palmer EnvironmentalTy Coch House
Llantarnam Park WayCwmbran
NP44 3AWTel: +44 (0) 1633 489479Fax: +44(0) 1633 877857
email: [email protected]: www.palmer.co.uk
ISBN 0-9538014-0-3
Related Documents
