bucharest workshop 27th sept 2017 - detectivii apei...
TRANSCRIPT
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-
indicator Definition
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.
The IWA PI set related to operational water losses (2 of 3)
6
Indicator
Sub-
indicator Definition
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.
• 3rd Edition (2017) of Performance Indicators for Water Supply
Systems.
The IWA PI set related to operational water losses (3 of 3)
7
Indicator
Sub-
indicator Definition
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.
• 3rd Edition (2017) of Performance Indicators for Water Supply
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
Play the NRW Percentage GameJocul Procentelor
m3/yr
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
Rezulta Pierdere34711
35698
2007
70409
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
Rezulta Pierdere34711 49%
2007
70409 100%
35698 51%
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
Rezulta Pierdere34711 49%
35698 51%
2007 2015
70409 100% 62677
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
Rezulta Pierdere34711 49% 30276
35698 51% 32401
2007 2015
70409 100% 62677
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr %SIV
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
Rezulta Pierdere34711 49% 30276 48%
35698 51% 32401 52%
2007 2015
70409 100% 62677 100%
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr %SIV %SIV
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
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%
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr %SIV %SIV m3/yr
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
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
2007 2015 Change in Performance
70409 100% 62677 100% 0%
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr %SIV %SIV m3/yr %change
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
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%
2007 2015 Change in Performance
70409 100% 62677 100% 0%
Play the NRW Percentage GameJocul Procentelor
m3/yr %SIV m3/yr %SIV %SIV m3/yr %change
System Input Volume (SIV)
Volum Intrat in Sistem (SIV)
less Metered Consumption
minus Consum
Non Revenue Water (NRW)
Rezulta Pierdere
Infrastructure Leakage Index 27 11
-13%34711 49% 30276 48% -1% -4435
-7732 -11%
35698 51% 32401 52% 1% -3297 -9%
2007 2015 Change in Performance
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
Manila, Philippines (2 of 3)
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
Manila, Philippines (3 of 3)
21
Understanding Leakage
Component Analysis of Real Losses
Component Analysis of Real Losses
– 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
– 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
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
Component Analysis of Real Losses
Part 2
– 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
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)
Component Analysis of Real Losses
Part 3
– 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
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.
at 1.6 m3/hr for 100 days
= 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.*
at 1.6 m3/hr for 9 days
= 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
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/
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/
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