bucharest workshop 27th sept 2017 - detectivii apei...

Bucharest Workshop

27th Sept 2017

Why you should NOT use %SIV

Component Analysis of Real Losses

D Pearson

NRW as % of System Input Volume just doesn’t work!

Cor Merks, Mark Shepherd, Marco Fantozzi and Allan Lambert

Can I get in on

the Act

Performance Indicators (PIs) for Water Supply Services

3

• “An individual Performance Indicator should be

unique and collectively appropriate for representing all

the relevant aspects of an utility’s performance in a

true and unbiased way, thus reflecting the managing

activity. Each performance indicator should contribute

to the expression of the level of actual performance

achieved in a certain area and during a given period

of time, allowing for a clear comparison with targeted

objectives and simplifying an otherwise complex

evaluation.”

3rd Edition (2017) of Performance Indicators for Water Supply Systems.

The PI set related to environmental and economic water losses

4

Indicator

Sub-

indicator Definition

WR1 Inefficiency of use of water resources (%) Percentage of water that enters the system and is lost by leakage and overflows

up to the point of customer metering.

Fi46 Non-revenue water by volume (%) Percentage of the system input volume that corresponds to non-revenue water.

Fi47 Non-revenue water by cost (%) Percentage of the system input volume that corresponds to the valuation of

non-revenue water components.

3rd Edition (2017) of Performance Indicators for Water Supply Systems.

The IWA PI set related to operational water losses (1 of 3)

5

Indicator

Sub-


Op23 Water losses per connection (m3/connection/year) Total (apparent and real) losses, expressed in terms of annual volume lost per

service connection. This indicator is adequate for urban distribution systems.

Op24 Water losses per mains length (m3/km/day) Total (apparent and real) losses, expressed in terms of annual volume lost per

mains length. This indicator is adequate for bulk supply and low service

connection distribution systems.

Op25 Apparent losses (%) Percentage of the water provided to the system (system input volume minus exported

water) that corresponds to apparent losses. This indicator is adequate for urban distribution

systems.

• 3rd Edition (2017) of Performance Indicators for Water Supply

Systems.


6

Indicator

Sub-


Op26 Apparent losses per system input volume (%) Percentage of the water entering the system (exported water inclusive) that corresponds to

apparent losses. This indicator is adequate for bulk supply and low service connection

distribution systems.

Op27 Real losses per connection (l/connection/day when system is pressurised) Real losses, expressed in terms of the average daily volume lost per connection. This

indicator is adequate for urban distribution systems.

Op28 Real losses per mains length (l/km/day when system is pressurised) Real losses, expressed in terms of the average daily volume lost per mains length. This

indicator is adequate for bulk supply and low service connection distribution systems.


Systems.


7

Indicator

Sub-


Op29 Infrastructure Leakage Index (-) Ratio between the actual real losses and an estimate of the minimum real losses

that could be technically achieved for the system operating pressure, average

service connection length and service connection density.


Systems.

˗ Current version of the PI set is basically the same as the first one (1st Edition, 2000)

· Precise definitions have received a number of minor changes and refinements

8

˗ % of System Input Volume (SIV) and % of Water

Supplied (WS) are not/never agreed as operational

water losses PIs · Not in the 1st (2000), 2nd (2006) nor 3rd (2017) Edition of Performance Indicators for Water

Supply Systems

· Not in the 4th Edition (2016) of AWWA Manual M36 Water Audits and Loss Control

Programs

˗ Statement PI 2017 Conference in Vienna, Austria on

May 15-17, 2017: ˗ “Everyone knows %s of SIV must not be used for target-setting and/or making technical

comparisons.“

˗ Nevertheless, unfortunately these volumetric % PIs are

being used, and this presentation intends to reinforce

earlier messages that this must be stopped

9

˗ Need for Fit for Purpose PIs is unchanged; EU

Reference document (© EU, 2015)

˗ Strengths and weaknesses of PIs have

become more clear over the last 17 years

˗ An increasing number of national

organisations, countries, water utilities and

leading water professionals have decided on

moving away from volumetric percentage PIs

NASA pictures of the Earth from the Moon; the Earth didn’t change, our view on it changes.

Context matters!

10

% of System

Input Volume

m3/km

mains/day

wsp*

Litres/service

connection/

day wsp*

Litres/conn/day

/metre of

pressure wsp*

Infrastructure

Leakage Index

ILI (incl. UARL)

% of time pressurised? No Yes Yes Yes Yes

water exported? No Yes Yes Yes Yes

length of mains? No Yes No No Yes

number of connections? No No Yes Yes Yes

average pressure? No No No Yes Yes

connections/km mains ? No No No No Yes

length of services ? No No No No Yes

how low could you go? No No No No Yes**

* when system pressurised ** Unavoidable Annual Real Losses UARL

Does the Performance

Indicator make allowance

for:

Performance Indicator for Real (Physical) Losses

11

˗ Volumetric PIs are good for target-setting and

tracking progress

˗ Litres/connection/day/metre of pressure also

allows for differences in pressure

˗ The ILI is designed for technical performance

comparisons between systems

˗ % of System Input Volume just doesn’t work: · strongly influenced by changes and differences in consumption per connection – variables

which may vary substantially from one year to another, not under control of the

undertaking

· does not make allowance for any system-specific key factors (agreed at PI 2017

Conference)

· gives misleading perspective of true performance

Simulated example I (1 of 4)

12

Water Supplied, Authorized Consumption

and Water Loss in Million Gallons per

Day:

˗ Utility has no Water Loss Control

program in place

˗ Steadily rising Water Losses over 13

years; doubled from 3 to 6 MGD

˗ A Major Industry moved to the City so

Authorized Consumption has

increased

Will Jernigan, P.E., Cavanaugh, NAWL Conference 2015 Atlanta

Simulated example I (2 of 4)

13

Water Supplied and Authorized

Consumption in MGD; Water Loss in

MGD and by % of Water Supplied:

˗ Water Loss as a % of Water Supplied

is not an indicator of performance

˗ It seems our performance in volumetric

% drastically improved when the Major

Industry came to the City → misleading

˗ Performance 2014 seems similar as

2001 → misleading

Simulated example II (3 of 4)

14

Water Supplied, Authorized Consumption

and Water Loss in MGD:

˗ Comprehensive Water Loss Control

program in place

˗ Steady reductions in total Water Loss

over 13 years from 6 to 3 MGD

˗ Our Major Industry stays in the City but

changes to Treated Effluent

Will Jernigan, P.E., Cavanaugh, NAWL Conference 2015 Atlanta

Simulated example II (4 of 4)

15

Water Supplied and Authorized

Consumption in MGD; Water Loss in MGD

and by % of Water Supplied:

˗ Water Loss as a % of Water Supplied is

not an indicator of performance

˗ It seems like our performance in

volumetric % today is the same as 13

years ago → misleading

16

Examples of failure to track progress:

˗ CASSA, Romania

˗ Zagreb, Croatia – a water utility

experience

˗ Manila, Philippines – a Miya project

CASSA, Romania (1 of 1)

17

˗ Operating company CASSA

˗ Zones increased 45 to 212

˗ Mains length increased by 125%

˗ Number of connections increased

by 107%

˗ €300M invested in bulk metering,

GIS, hydraulic modelling leakage

detection equipment, replacement

of customer metering, SCADA and

training of operatives

www.leakssuite.com/zero-sum-leakage-pi/

Play the NRW Percentage GameJocul Procentelor

m3/yr

System Input Volume (SIV)

Volum Intrat in Sistem (SIV)

2007

70409


m3/yr



less Metered Consumption

minus Consum

Non Revenue Water (NRW)

Rezulta Pierdere34711

35698

2007

70409


m3/yr %SIV




minus Consum


Rezulta Pierdere34711 49%

2007

70409 100%

35698 51%


m3/yr %SIV m3/yr




minus Consum


Rezulta Pierdere34711 49%

35698 51%

2007 2015

70409 100% 62677


m3/yr %SIV m3/yr




minus Consum


Rezulta Pierdere34711 49% 30276

35698 51% 32401

2007 2015

70409 100% 62677


m3/yr %SIV m3/yr %SIV




minus Consum


Rezulta Pierdere34711 49% 30276 48%

35698 51% 32401 52%

2007 2015

70409 100% 62677 100%


m3/yr %SIV m3/yr %SIV %SIV




minus Consum


Rezulta Pierdere

Perceived change in performance False

Perceptia modificarilor in perfomanta Fals

Always Zero Sum - % change in metered consumption + % change in NRW = zero

-1%

1%

34711 49% 30276 48%

35698 51% 32401 52%

2007 2015 Change in Performance

70409 100% 62677 100% 0%


m3/yr %SIV m3/yr %SIV %SIV m3/yr




minus Consum


Rezulta Pierdere

Perceived change in performance False True

Perceptia modificarilor in perfomanta Fals Adevarat

34711 49% 30276 48% -1% -4435

-7732

35698 51% 32401 52% 1% -3297


70409 100% 62677 100% 0%


m3/yr %SIV m3/yr %SIV %SIV m3/yr %change




minus Consum


Rezulta Pierdere

Perceived change in performance False True True

Perceptia modificarilor in perfomanta Fals Adevarat Adevarat

-13%34711 49% 30276 48% -1% -4435

-7732 -11%

35698 51% 32401 52% 1% -3297 -9%


70409 100% 62677 100% 0%


m3/yr %SIV m3/yr %SIV %SIV m3/yr %change




minus Consum


Rezulta Pierdere

Infrastructure Leakage Index 27 11

-13%34711 49% 30276 48% -1% -4435

-7732 -11%

35698 51% 32401 52% 1% -3297 -9%


70409 100% 62677 100% 0%

Zagreb, Croatia (1 of 1)

18

˗ Utility with high leakage

˗ Introduced district

metering and pressure

management in 2012

˗ In 2013 significant

reductions in annual

volumes were achieved

˗ Good work undertaken in

the field, but not according

to performance judged on

change in % of SIV

System Input Volume Revenu Water Non-Revenu Water Apparent Losses Real Losses

Mm3 Mm3 Mm3 Mm3 Mm3

2012 120,7 49,4 71,3 2,0 69,3

2013 114,1 47,3 66,8 1,9 64,9

Change (2013-2012) -6,6 -2,1 -4,5 -0,1 -4,4

% Change -5,5% -4,3% -6,3% -5,0% -6,3%

System Input Volume Revenu Water Non-Revenu Water Apparent Losses Real Losses

% of SIV % of SIV % of SIV % of SIV % of SIV

2012 100,0% 40,9% 59,1% 1,7% 57,4%

2013 100,0% 41,5% 58,5% 1,7% 56,9%

% Change 0,0% 0,5% -0,5% 0,0% -0,5%

Water Balance Annual Volumes expressed in Million cubic metres (Mm3)

Water Balance Annual Volumes expressed as & of System Input Volume

Year

Year

www.leakssuite.com/zero-sum-leakage-pi/

Always a Zero-Sum calculation: one +X%, the other –X%, or both 0%

Manila, Philippines (1 of 3)

19

˗ Miya partnered with Maynilad Water Services on a

NRW reduction project between 2008 and 2014

˗ At the beginning of the project NRW was 1,580 million

litres/day, three million people could not be connected

and supplied, and millions of others suffered from

intermittent supply, extremely low pressures and poor

water quality

˗ Project goal was to build NRW management capacity,

establish a NRW management system, re-structure

and improve the water distribution network, and reduce

physical and commercial losses to enable supply to

the entire population in the service area


20

˗ Results: · NRW reduced from 1,580 to 650 million litres/day

· Number of customers increased from 700,000 to 1,160,000

· Tremendously improved level of service

· 1,500 DMAs established, 1,500 km of pipelines replaced

· 277,000 leaks detected and repaired

· Maynilad’s net income tripled

˗ Additional revenues during this seven

year period already exceeded the

investments


21

Understanding Leakage



– Causes of Real Losses

– Origins of the BABE Concept

– Reported and Unreported leaks, and Background leakage

– How much do reported bursts contribute to Real Losses?

– Relating leak duration to Utility policies

– Unavoidable Background leakage

– Annual BABE component calculations: basic and detailed

– How low could you go? UARL calculated using BABE concepts

– UARL equations: 1999 original and 2015 EU Reference Doc version

– Performance Indicators and ILI

International Tools for Practical Leakage Management

92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17

EU Reference document Good Practices on Leakage Management WFD CIS WG PoM Main Report and Case Studies

IWA Best Practice International Water Balance and Terminology

IWA recommended Key Performance Indicators (KPIs) for Non Revenue

Water and Real Losses

Influence of reduced burst frequency on annual repair costs, extension of residual infrastructure life and economics of

pressure management

Night flow component analysis using BABE & FAVAD concepts.

Average Zone Point AZP and Night-Day Factor NDF

Economic Intervention Policy for Active Leakage Control based on rate of Rise

of Unreported Leakage

How low could you go? System-specific equations for Unavoidable Annual

Real Losses UARL, and Infrstructure Leakage Index ILI

Uncertainty in Water Balance and Night Flow calculations

Economic Leakage Levels with and without pressure management

Quick predictions for systems with high initial burst frequency Pressure: burst frequency

relationships

Component analysis of Real Losses (Bursts and

Background Estimates, BABE)

A more generally applicable 2-part equation with a non-pressure-dependent component

Economic Leakage Levels using minimum total

cost approach without pressure management)

Pressure: leak flow rate relationships: Fixed

and Variable Area Discharges, FAVAD

Developed by

John May

Leakage Performance Categories (WBI Banding System) for assessment of Real Loss Technical

Management Performance and appropriate actions for improvement

Implementation

IWA Water Loss Specialist Group

Realistic modelling of leakage & intrusion flows through leak openings in pipes, with FAVAD & Leakage Number concepts: theory and practical application

Guidance Notes on Apparent Losses and Water Loss Reduction Planning, with 9 Separate Appendices

IWA 2nd Water Loss Task Force

UK Leakage

Control

InitiativeConcepts

Development and testingKey to colours

IWA 1st Water Loss Task

Force

Component Analysis of Real Losses –

Part 1











Causes of Real Losses

• Real (Physical) Losses: the water lost from all individual

leaks, bursts and overflows on the system infrastructure

• For each individual event, the average flow rate and the

duration determine the volume of Real Losses

• Causes of Real Losses include:

– excess pressure, water hammer

– corrosion and breaks in pipes

– leaks at joints and fittings (valves, hydrants)

– accidental and deliberate damage to hydrants and

standpipes

– leakage and overflows from service reservoirs

Origins of the BABE Concept

• BABE is the acronym for Background and Bursts

Estimates of components of real losses

• BABE is an overview concept of the parameters

and processes which influence real losses

• BABE was developed by Allan Lambert during the

UK National Leakage Control Initiative 1991-94,

using internationally applicable principles, to co-

ordinate the work of the 9 separate research groups

• BABE is not a particular piece of software –

thousands of software programs now use BABE

concepts, in many countries

Something old, something new

• BABE was developed as the first comprehensive

conceptual model of components of real losses, after

extensive reading of international experience

• Parts of the concepts relate to common experience

– for example, not all leakage can be detected

• Other parts are a development of simpler concepts

– for example, splitting average duration of leaks into

different time components

– rationalisation of diverse pressure:leakage

relationships

• BABE uses internationally applicable principles

– but usually benefits from some local customisation

Before BABE

• No serious attempt had been made to model

components of leakage - it was thought to be

too difficult a task

• Local leakage practice was based on ‘rules of

thumb’ which were not transferable

technology, or ‘what we have done up to

now, which seems to work for us’

Components of Annual Real Losses

• There are many different components which go to make up the volume of annual real losses

• If we can separately identify these components, and the parameters which influence them, we can learn to manage and control real losses better

• Avoid empirical relationships and ‘rules of thumb’ which only apply in special situations

• The overall concept, and relationships between parameters must be logical and based on good physical, hydraulic and engineering principles


Part 2











Different Types of Leaks

• Background & Bursts Estimates concept categorises 3 types of leaks: – Reported bursts: notified to Utility by 3rd parties,

generally high flow rate, short duration

– Unreported bursts: moderate flow rates, long run times, found by intervention (active leakage control)

– Background leaks: low flow rates, not detectable by sound if not visible, mainly at joints and fittings

BABE Concepts

Reported and Unreported Bursts, and

Background Leakage

Parameter Reported bursts Unreported burstsBackground

Leakage

Visible? Yes, usually No, usually No

Flow Rates Relatively high ModerateLow

(<250 l/hr?)

Detectable? Yes, usually Yes, usually No

Total run times Usually short

Usually long - depends

on Active Leakage

Control policy

Continuous

Volume lost from a burst or leak = flow rate x run time

So we must consider flow rate x run time for each category

of bursts and leaks to assess overall Real Losses

Typical Flow Rates of Bursts and Leaks

at 50m pressure

• Mains bursts – wide range, to 50 m3/hr or more – strongly influenced by pipe diameter and material

– unreported leaks have lower flows than reported leaks

• Service connections bursts – 0.25 to 2 m3/hr

– slightly influenced by pipe diameter and material

– flows of reported and unreported leaks are similar

• Background leaks on mains and service connections are individually less than around 0.25 m3/hour

• A UK domestic tap runs at around 0.6 m3/hr

Reported Bursts

• Most people think that

mains bursts are the main

source of leakage

• But it usually adds up to no

more than about 5% of

total leakage

Unreported Bursts

• Those that don’t come to

the attention of the

operator from customers

• Not seen, not causing

problems of supply

• That’s why you need a

leakage department

Examples of Mains Burst Flow rates

at 50m Pressure

0

2

4

6

8

10

12

14

16

18

20

0 25 50 75 100 125 150 175 200 225 250

Pipe diameter (mm)

Ave

rage

flo

w r

ate

(litr

es/s

ec)

Ring cracks in Cast-

Iron Mains (Canada)

Managing Leakage Report E,

Reported bursts

Managing Leakage

Report E,

Unreported bursts

Volume of water lost from a single leak

or burst

• Assume that the leak occurs instantaneously and the flow rate does not increase with time: – Volume lost = flow rate (m3/day) x duration (days)

• In BABE concepts, the duration is divided into 3 elements (which depend on Utility’s policies) – awareness: from start of leak to time Utility is aware

there is ‘a problem’

– location: time then taken to locate the leak

– repair: time taken to repair or shut off the leak

Leak Run Time Awareness

A L R

LEAK RUN TIME

FL

OW

RA

TE

RUN TIME

Leak Volume Loss = ( A+ L+R ) Time x Flow Rate

(at a given pressure)

= *Awareness + Location + Repair

* Awareness time is different for Reported & Unreported Leaks

Time components Awareness, Location

and Repair of leaks

Relating Leak Duration to Utility

Policies

• A Utility’s ability to manage real losses volume

therefore depends on its management of the

average duration of different types of leaks

• Average Awareness, Location and Repair times

can all be related to Utility Policies using BABE

concepts

– but many Utilities only record average repair times

Background Leakage

• Background leakage occurs mostly at joints and

fittings

- flow rates (generally <250 l/hr?) are too small to be detected

by sonic means if they do not reach the surface

– so they run continuously

• Every distribution system has background leakage

– and in well-managed systems background real losses are

50% or even more of the total annual real losses

• Background leakage is very sensitive to pressure

Parameters that Influence Unavoidable

Background Leakage

• Length of mains Lm – A proxy for number of joints and fittings

• Number of Service Connections Ns

– from main to property line

• Length of private service connections Lp

– from property line to customer meter

• Average pressure Pav or AZNP (nightflows)

IWA Unavoidable Background Leakage

• IWA values at 50 metres pressure (AQUA, Dec 1999) based

largely on UK (post-mandatory targets) data from DMAs with

well-maintained infrastructure, Germany, Australia, NZ are:

– mains 20 l/km/hour

– service connection, main to property line, 1.25 l/service/hour

– private underground pipes property line to meter (15m average length),

0.5 l/service/hr or 0.033 litres/metre/service/hr

• Use these for international application to calculate

‘best achievable’ background leakage – use N1 = 1.5 to correct for pressures different to 50 metres

Unavoidable Background Leakage (UBL)

• The equation for UBL on mains and service

connections up to the property line, is

UBL (litres/hour) = (20 x Lm + 1.25 x Nc) x (AZNP/50)1.5

where Lm is mains length (km),

Nc is number of service connections, main to property line

AZNP is average zone night pressure (metres)

• If customer meters are located after the property

line, the equation for UBL is

UBL(litres/hour) = (20 x Lm +1.25 x Nc + 33 x Lp) x (AZNP/50)1.5

where Lm is total length of private pipe, property line to meters (km),

UBL as a component of night flow

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

0 10 20 30 40 50 60 70 80 90 100

Average Zone Night Pressure (metres)

UB

L, li

tres

/con

nect

ion/

hour

DC = 20/km

DC = 30/km

DC = 50/km

DC = 80/km

Figure 1 of Cape Town paper (for Zones with meters at property line)


Part 3











A basic annual BABE model

(External Meters)

Simple BABE Components of Real Losses Model Data entry Guidelines Calculated

Name of System For Year 1998/99

Infrastructure Data BABE Calculation of Real Losses

Average pressure, metres 54.0 Background Reported Unreported

Service Connections: Number 57300 Leakage Bursts Bursts Total

Distribution Mains: Length (km) 1220 m3x 10

3/yr m

3x 10

3/yr m

3x 10

3/yr m

3x 10

3/yr

Service Connection density (/km) 47.0 Mains 312 149 31 492

Assumed Infrastructure Condition Factor = 1.30 Service Conns 915 177 20 1112

Totals 1227 326 50 1604

Estimate from 'Benchloss' Annual Water Balance = 1596

Data for Calculation of Losses from Bursts Average Average Average Average Average Burst

Assumed N1 value for bursts events = 1.50 Awareness Location Repair/Shutoff Duration loss/burst FrequencyAnnual Number of Bursts l/sec @ 50m Hours Hours Hours days m3

Mains - Reported 406 at 7.0 8 2 3 0.54 368 33.3 per 100km/yrMains - Unreported 3 at 7.0 36 300 24 15.00 10182 0.2 per 100km/yr

Service Conns - Reported 413 at 2.0 48 2 3 2.21 428 7.21 per 1000 conns/yrService Conns - Unreported 5 at 2.0 168 300 24 20.50 3976 0.09 per 1000 conns/yr

Effect of Bursts triggered by Pressure Instability

If 10% of bursts are triggered by pressure instability, then 37.7 m3x 10

3/yr lost as a result

If variable purchase cost of water is 0.45 $/m3 water lost through triggered bursts costs 16.9 $ x 10

3/year

If average cost of a burst repair is 800 $ then annual cost to repair triggered bursts = 66.2 $ x 103/year

Background Losses Background losses vary with pressure to power N1= 1.50

Assumed background loss = 1.30 times Unavoidable background losses

IWA values, for Infrastructure in Good Condition at 50m average pressure m3/day @50m m3/d at actual m3x 10

3/yr at actual

Service Connections 1.25 l/conn/hr = 2235 2508 915.5

Distribution Mains 20 l/km/hr = 761 854 311.9

Total 2996 3363 1227.3

A New Zealand City

• It is impossible to eliminate all physical (real) losses from distribution systems

• ‘Unavoidable Annual Real Losses’ (UARL) is the lowest technically achievable volume for each individual system, assuming infrastructure is in good condition and well-maintained

• Calculation of UARL depends upon:

– Length of mains

– Number of service connections, and location of customer meter (relative to property line)

– Average operating pressure

http://www.leakssuite.com/concepts/uarl-and-ili/

Annual Real Losses: how low can you

go?






How low could you go?

Unavoidable Annual Real Losses

• The UARL formula uses a BABE component analysis for

systems with well-maintained infrastructure in good

condition

• Initially assumptions are at 50 metres pressure

• Assumes Unavoidable background leakage UBL

• 13 mains bursts/100 km/year (5% unreported)

– reported: 12 m3/hr for 3 days; unreported 6 m3/hr for 50 days

• 3 comms pipe bursts/1000 conns/yr (25% unreported)

– reported: 1.6 m3/hr for 8 days; unreported 1.6 m3/hr for 100 days

• 2 supply pipe bursts/1000 conns/yr (25% unreported)

– reported: 1.6 m3/hr for 9 days; unreported 1.6 m3/hr for 101 days

BABE calculation of Unavoidable

Annual Real Losses at 50m pressure

Infrastructure Component

Unavoidable Background Leakage

Detectable Reported Leaks and Bursts

Detectable Unreported Leaks and Bursts

On Mains

20 litres/km/hr

12.4 bursts/100 km/yr.

at 12 m3/hr for 3 days

= 864 m3/burst

0.6 bursts/100 km/yr.

at 6 m3/hr for 50 days

= 7200 m3/burst

On Service Connections, Main

to Property Line 1.25 litres/conn/hr

2.25/ 1000 conns/yr.

at 1.6 m3/hr for 8 days

= 307 m3/burst

0.75/1000 conns/yr.


= 3840 m3/burst

On Service Conns from Property Line to

meter, if customer meter is not located at the property line

0.50 litres/conn/hr*

1.5/ 1000 conns/yr.*


= 346 m3/burst

0.50/1000 conns/yr*.

at 1.6 m3/hr for 101

days= 3878 m3/burst

Table 1 of paper: same data as 1999 AQUA paper but volume/burst added

* For average length of 15 metres

UARL converted to daily leakage

UARL (litres/day)= (18 x Lm + 0.8 x Nc + 25 x Lp ) x P

Main length km

Average

Pressure

(metres) Number of conns.

Length of underground

private pipes,

property line to meter (km)

How do we take account of key

system-specific factors?

Unavoidable Annual Real Losses ‘per day’ equation (1999)

Unavoidable Annual Real Losses

(UARL)

• Unavoidable Annual Real Losses can be calculated from the 1999 equation:

UARL (litre/day) = (18 x Lm + 0.8 x Nc + 25 x Lp) x P

where: • Lm = mains length (km)

• Nc = number of service connections (main to property line)

• Lp = length of private pipe (km), property line to meter

• P = average pressure (metres)

UARL formula for EU Reference

Document

• Europeans prefer to use total service pipe

lengths main to meter (Lc) and m3/year

– so deduct 4 m/conn x 25 from Nc coefficient = 0.80 –

0.04 x 25 = 0.80 – 0.10 = 0.70

• UARL (litre/day) = (18 x Lm + 0.7 x Nc + 25 x Lc) x P

– then multiply by 365 days and divide by 1000

UARL (m3/yr) = (6.57 x Lm + 0.256 x Nc + 9.13 x Lc) x P – Lm = mains length (km)

– Nc = number of service connections (main to property line)

– Lp = length of private pipe (km), property line to meter

– P = average pressure (metres)

UARL (m3/year) = (6.57 x Lm + 0.256 x Nc + 9.13 x Lc ) x P

Main length Lm

km

Average

Pressure

(metres) Number of conns Nc Length of underground

connections

main to meter Lc km

Unavoidable Annual Real Losses EU Reference Doc.

2015 UARL formula using Lc (main to meter)

UARL in litres/connection/day;

for customer meters at property line

Good KPI for tracking progress in individual systems, but not

for comparing performance between systems

UARL in m3/km mains/day, for customer meters

at property line;

for customer meters at property line

Also OK for tracking progress in individual systems, but not

for comparing performance between systems

Limits of application of UARL formula

Have gradually been extended since the original AQUA Dec 1999 paper:

• Density of connections : lower limit of 20 conns/km removed in 2005

• Average pressure: lower limit around 25 metres, varies with pipe materials

• System size: lower limit reduced from 5000 to 3000, now Lm & Ns formula

Lambert et al, AQUA

Lambert & McKenzie

Liemberger & McKenzie

Lambert Parameter Limits

1999 2001 2005 2009

Minimum 20 20 Removed No lower limit Density of Connections/km Maximum 100 Removed No upper limit

Minimum 20 25 25 See Figure 3 Average Pressures (m) Maximum See Figure 3

System Size Minimum Not stated Ns >5000 Ns> 3000 (150 x Lm + Ns) > 3000

Current Annual

Real Losses

CARL

Pressure Management

Pressure

Management

Active LeakageControl

Speed and

Quality of

Repairs

Speed and quality

of repairs

Active

Leakage

Control

Pipe Materials Management:

selection,installation,

maintenance,renewal,

replacement

Pipeline and

Assets

Management:

Selection,

Installation,

Maintenance,

Renewal,

Replacement

UARL

calculation based on

mains length,

number of services,

customer meter location

and average pressure

Infrastructure

Leakage Index

ILI = CARL/UARL

Potentially Recoverable

Real Losses

Unavoidable

Annual Real

Losses UARL

ILI – What does it mean?

• An ILI of 10 means that Real Losses volume is 10 times the lowest technically achievable real losses for the system at current average pressures

• An ILI close to 1.0 represents world-best performance in managing Real Losses

– desirable where water is in short supply, or expensive, or both

• ILI measures the overall effectiveness of infrastructure management of real losses at current average pressures

• This does not imply that the current pressure management is optimal

– so always quote average pressure alongside ILI

ILIs for 71 Systems in 10 European

High Income Countries (2014)

Comparison of Performance Indicators

% of System

Input Volume

m3/km

mains/day

wsp*

Litres/service

connection/

day wsp*

Litres/conn/day

/metre of

pressure wsp*

Infrastructure

Leakage Index

ILI (incl. UARL)

% of time pressurised? No Yes Yes Yes Yes

water exported? No Yes Yes Yes Yes

length of mains? No Yes No No Yes

number of connections? No No Yes Yes Yes

average pressure? No No No Yes Yes

connections/km mains ? No No No No Yes

length of services ? No No No No Yes

how low could you go? No No No No Yes**

* when system pressurised ** Unavoidable Annual Real Losses UARL

Does the Performance

Indicator make allowance

for:

Performance Indicator for Real (Physical) Losses

ILI on OFWAT Coaxial plot

Source: Lambert, IWEX, 1999

Which Leakage KPI should I use?

Evidence-based conclusions from 16 EU Case Studies

Volume

per year

litres/

service

connection

m3/km

mains

litres/

billed

property

% of System

Input

Volume

% of

Water

Supplied

Infrastructure Leakage

Index, with Pressure

YES,

for large

systems

YES* YES*YES

(UK)NO NO

Only if all justifiable

pressure management

completed

NO NO NO NO NO NO YES

NO NO NO NO NO NOYES, together with

other context factors

* Choose services connection density > 20/km; if not, choose mains; or base choice on country custom and practice

OBJECTIVE

GOOD PRACTICE PERFORMANCE INDICATOR FOR LEAKAGE, FIT FOR PURPOSE

SET TARGETS AND TRACK

PERFORMANCE, FOR AN

INDIVIDUAL SYSTEM

TECHNICAL PERFORMANCE

COMPARISONS OF

DIFFERENT SYSTEMS

DRAW GENERAL

CONCLUSIONS FROM SINGLE

OR MULTIPLE SYSTEMS

Good Practices on Leakage Management – Case Study document._Final.pdf

https://circabc.europa.eu/sd/a/ec13ae01-7800-4114-8d72-98827b509e18/Good Practices on Leakage Management - Case Study document._Final.pdf




National Organisations and Countries

adopting ILI

%SIV ILI

IWA

AWWA

WSAA

Malta

WBTI

Austria

Denmark

Croatia

Italy

Germany

South Korea

2018

EU Reference document

2000 2003 2006 2009 2012 2015

International ILIs

• http://www.leakssuite.com/global-ilis/australian-ilis/

http://www.leakssuite.com/global-ilis/new-zealand-ilis/

• http://www.leakssuite.com/global-ilis/european-ilis-2/

http://www.leakssuite.com/global-ilis/ilis-in-flanders-2014/

• http://www.leakssuite.com/global-ilis/austrian-ilis/

http://www.leakssuite.com/global-ilis/croatian-ilis/

• http://www.leakssuite.com/global-ilis/danish-ilis-2/

http://www.leakssuite.com/netherlands-ilis/

• http://www.leakssuite.com/global-ilis/portuguese-ilis/

http://www.leakssuite.com/global-ilis/north-america-ilis/

• http://www.leakssuite.com/global-ilis/canadian-ilis/

http://www.leakssuite.com/ili-overviews-by-country/

http://www.leakssuite.com/global-ilis/australian-ilis/












http://www.leakssuite.com/global-ilis/european-ilis-2/
















http://www.leakssuite.com/global-ilis/austrian-ilis/










http://www.leakssuite.com/global-ilis/danish-ilis-2/










http://www.leakssuite.com/global-ilis/portuguese-ilis/












http://www.leakssuite.com/global-ilis/canadian-ilis/












Key References

• http://173.254.28.127/~leakssui/wp-

content/uploads/2012/11/1994_Lambert-April-IWEM-1994A.pdf

• http://173.254.28.127/~leakssui/wp-

content/uploads/2012/11/1996_LambertMorrisonBurstsBackground-

CIWEM-1996A.pdf

• http://173.254.28.127/~leakssui/wp-

content/uploads/2012/11/1999_LambertBrown-et-al-AQUA-

1999M.pdf

• http://173.254.28.127/~leakssui/wp-

content/uploads/2012/11/2009_LambertWaterlossCapetown-

2009J.pdf

• EU Good Leakage Practices

• http://www.leakssuite.com/ciwem-pps%e2%80%8f/

• http://www.leakssuite.com/ili-in-uk-update/

http://173.254.28.127/~leakssui/wp-content/uploads/2012/11/1994_Lambert-April-IWEM-1994A.pdf









http://173.254.28.127/~leakssui/wp-content/uploads/2012/11/1996_LambertMorrisonBurstsBackground-CIWEM-1996A.pdf







http://173.254.28.127/~leakssui/wp-content/uploads/2012/11/1999_LambertBrown-et-al-AQUA-1999M.pdf











http://173.254.28.127/~leakssui/wp-content/uploads/2012/11/2009_LambertWaterlossCapetown-2009J.pdf





http://www.leakssuite.com/eu-good-practice-on-leakage-management/

http://www.leakssuite.com/ciwem-pps%e2%80%8f/



http://www.leakssuite.com/ili-in-uk-update/







UARL and Burst Frequency

• The UARL components for frequency and flow rates

of reported and unreported bursts were calculated

for a pressure of 50 metres

• When average pressures greater, or less than 50m

are used in the formula, all the leak flow rates of the

bursts change, but the frequency of bursts does not

• That is because the current international knowledge

of pressure:bursts relationships was not developed

until 10 years after the UARL formula was

published.

Service Connections or Properties?

• The number of service connections which

experience leakage is not necessarily the same as

the number of billed properties.

• In some countries (eg UK) they are generally

approximately equal to each other

• This is not usually the case elsewhere, e.g.

– Singapore has 943,000 metered accounts but only

180,000 service connections (up to master meters)

• Always use number of service connections, not

number of billed properties, when making

comparisons of real losses on distribution systems

Thank You

bucharest workshop 27th sept 2017 - detectivii apei...

Documents