Some misconceptions about ILI in the UK
Infrastructure Leakage Index ILI, first published in 1999, was developed by an IWA Task Force as a key performance indicator designed for comparisons of technical leakage management performance, and is now widely used internationally. A 2015 CIWEM Policy Position Statement (PPS) on Leakage supports the use of leakage performance indicators such as ILI which function across companies and geographical locations. Although there is currently no regulatory requirement for UK Utilities to assess ILIs, some UK ILIs are available since 2004, but some misconceptions still exist which may be constraining the wider use of ILI in the UK.
In the UK, Company annual average leakage targets are set on the basis of an evaluation of Sustainable Economic Leakage Level (SELL) in Ml/day, using various methodologies. Each Company will have its own SELL, depending upon a range of factors including the benefits of advanced pressure management, and other investments in addition to economic active leakage control..
Recommendations of the EU Reference document ‘Good Practices on Leakage Management’, summarised in the Table below, show that the KPIs used in the UK (m3/km/day, litres/property/day) are not fit for the purpose of technical performance comparisons of leakage management in different systems and countries. In contrast, ILI is designed for that purpose.
Since 2003 most of the UK Companies (around 25 in number, including Northern Ireland and Scotland) have contributed leakage performance data to a Water UK Leakage Network group. This data, which includes some of the basic parameters used to calculate ILI, is not published or available to researchers or the media, so the Leakage Network is the only group able to calculate ILI on a consistent national basis for the UK.
However, all members of the Leakage Network have, for several years, been able (if they wish) to calculate their ILIs using a free LEAKSSuite customised software UKIWAPICalcs. For 2011/12, nine UK Companies did so, and these ILIs are included in the European ILIs data set on an anonymous basis. This international comparison using ILI suggests that although economic leakage levels of UK water utilities are not the lowest in Europe and internationally, they lie within a leading group of technically advanced countries.
In September 2014, the Leakage Network prepared a Paper ‘ILI – the UK Position’ which was circulated to UK Water Companies but not made publicly available for comment. Unfortunately, the paper contains several significant errors and mis-assumptions, outlined below with comments.
1 ‘ILI – the UK Position’ accepts that ILI can be an appropriate measure to compare the general range of UK leakage levels with those in other countries, but states that such comparisons can be misleading because
a) UK quoted leakage levels include leakage on customer supply pipes, whereas in many other countries leakage is only calculated up to the meter.
b) Number of properties per connection may be much higher in some countries. Comment: this is precisely why the UARL formula uses number of service connections up to the meter, not number of properties as in UK traditional practice.
Comment on 1(a): The three infrastructure components of the UARL equation specifically allow for both meter location and density of connections. In the UK, distribution losses ILI and Total Losses ILI (including customer underground supply pipe leakage) are both calculated in the UKIWAPICalcs free software, and comparison of these figures provides a useful quality control feature for assumptions relating to assessment of supply pipe losses.
2 ‘ILI – the UK Position’ states that in the UK, ILI is not considered suitable for assessment of the appropriate level of leakage in an area, or for target setting, which should be based on an evaluation of SELL.
Comment: As with ILI (see previous Table), assessments of SELL are only fit for purpose for setting targets if all justifiable pressure management has been completed, or has been built into SELL calculations. Expressing SELL (in Ml/d) as an ILI is a useful cross-check particularly for Water Resources Zones, and can identify systematic errors in bulk metering between Zones.
3 Other material used in ‘ILI – the UK Position’ to justify statements in (1) and (2) above is shown below. Each in its own way demonstrates a lack of understanding of the UARL formula, and of pressure-dependent component analysis of leakage.
a) the parameters ‘number of service connections’, ‘length of communication pipes’ and ‘length of supply pipes’ are not easily quantified
Comment: number of communication pipes (not length) is required for the UARL formula; systematic estimates of supply pipe length are acceptable. By common consent most annual leakage occurs on services, so Companies should have this data to reliably assess their SELL.
b) It is quite possible that a reduction in pressure could result in reduction in leakage but an increase in the ILI.
Comment: This possibility may exist for changes in flow rates of existing leaks in systems with a high % of rigid pipes (mains and services), a high burst frequency, and high ILI. However, when reductions in annual burst numbers are included in the calculation it is a most unlikely scenario at the range of ILIs in the UK (1 to 3). Improved predictions of changes in ILI allowing for leak flow rates and bursts can now be made for individual situations. Where flexible pipe materials predominate, ILIs will generally reduce to some extent after pressure management.
c) The calculation of UARL is based on only 3 parameters – number of connections, total length of network (including mains, communication pipes and supply pipes), average operating pressure.
Comment: the UARL formula is based on 4 variables: 3 infrastructure variables Lm, Nc, Lp and average pressure P (metres) , see UARL and ILI.
UARL (litres/day) = (18 x Lm + 0.8 x Nc + 25 x Lp) x P
UARL (m3/year) = (6.57 x Lm + 0.256 x Nc + 9.13 x Lp) x P
The coefficients for the 3 infrastructure variables are calculated using a leakage component analysis (Background and Bursts) model with 24 additional parameters for frequencies, flow rates and durations of different type of leaks (background, reported, unreported), using auditable values and assumptions in AQUA Dec 1999
d) The ILI is used widely internationally but it is not used within the UK Water Industry except where international comparisons are required.
Comment: Only a minority of the Companies in the UK Leakage Network seem willing to calculate and state ILIs for international comparisons, and then only on an anonymous basis. However, increasingly some progressive Companies in UK are calculating ILIs for Water Supply Zones, and Snapshot ILIs for DMAs, to supplement traditional UK approaches with the strategic techniques outlined in the EU Reference Document. This approach has recently been used in the UK to reveal data errors in the calculation of the zonal water balances and SELL, and also to highlight zones and DMAs with a relatively high level of loss, which require special attention.
In summary, it seems as though the UK Leakage Network concluded some years ago that as the ILI was not an economic KPI, it was therefore of little value for the UK. The Network did not seek to encourage more diagnostic comparisons of SELLs with technical leakage management performance in other countries. The paper ‘ILI – the UK Position’ continues to promote that view, using some assertions which do not stand up to scrutiny. However, some UK practitioners and Utilities are more willing to blend the best features of traditional UK practices with the international advances supported by the EU Reference Document ‘Good Practices on Leakage Management’ and the CIWEM Policy Position Statement.
UK Utilities and regulators may be interested to attend a UK Workshop on ILI and Snapshot ILI and their applications, to be presented by Allan Lambert and Stuart Trow, in which any uncertainties and appropriate/inappropriate assumptions can be openly discussed. Contact ILMSS Ltd using the Contact Us Webpage for further information.