DOC

Leakage Economics

By Lester Stone,2014-12-23 20:13
11 views 0
Leakage Economics

    Leakage Economics: Plugging the Knowledge Gap

Background

    All water companies in England and Wales, have recently been required to submit to Ofwat their Final Business Plans for the AMP4 period 2005/06 2009/10. A

    requirement of these plans was that they should include a statement of their proposed supply ~ demand arrangements over the AMP4 period. There is an expectation by Ofwat and the Environment Agency that supply ~ demand submissions should demonstrate that optimal strategies have been chosen from the wide mix of feasible supply, demand management and leakage management options. Optimal in this sense is taken to mean economically optimal and therefore one of the outputs from this ‘Least Cost Plan’ (LCP) will be the leakage levels to be targeted in each year of the plan

    period.

    Target leakage levels, at least for the time being, will continue to be based on economic principles, although this is a topic currently being debated by the EA and others. In considering the optimal mix of water resource and other options, OFWAT therefore expect companies to have presented reviews of their Economic Levels of Leakage (ELL) and are required to demonstrate that their ELL assessments are in line with ‘best practice’.

    At the present time, industry best practice is taken to mean that which is set out in the OFWAT / DEFRA / EA report on ‘Leakage Target Setting for Water Companies in England and Wales’ (also known as the Tripartite Report), which was published as a consultation document in March 2002. This has now been published as a final version in the light of comments from water companies, consultants and other stakeholders. Whilst providing a valuable framework for ELL calculations, thankfully the guidelines are not prescriptive, and it is recognised that there is no single ‘approved’ method.

    Indeed a number of methodologies exist across the industry for determining ELL, all of which may be equally valid. However, what is paramount is that they should reflect the relationships that exist between leakage and cost within companies, and be based on local company data rather than national defaults. Two other important aspects of the Tripartite Report that OFWAT will no doubt be looking for evidence for, are that ELLs have:

    ; been calculated at zonal level (rather than just for the company as a whole) ; included consideration of the Environmental and Social costs (and benefits) of

    leakage and leakage control

Back to Basics

    There are two approaches set out in the Tripartite Report for calculating ELL :

    ; For short-run target setting and interim reviews of ELL, a method is set out whereby

    target leakage levels are determined by balancing leakage control costs against the

    marginal cost of water

    ; For long-run target setting and comprehensive reviews of ELL, a full LCP approach

    is recommended, which identifies the least cost mix of leakage policy, supply and

    demand management options over a plan period.

    The main difference between the two is that short-run ELLs take no account of the savings which could accrue through the deferment of capital, which might otherwise be

    1

    required for the development of new schemes to meet deficiencies in supply. It is therefore the latter which will have been included in the company Business Plan submissions. However, short run ELLs can be used as an input to the LCP, and except in zones where there are supply ~ demand issues, long run and short-run ELLs will not be dissimilar.

    The key aspect of any ELL review is the establishment of the relationship between cost and leakage. The generally held view is that the variable costs of leak detection (and possibly the variable costs of repairs) increase as leakage levels decrease (Figure 1).

    1.80

     L = (L1+L2+NRR)/21.60

     Detection1.40 Detected Repair

    1.20

     1.00

     0.80

     0.60

    UC (? per l/p/h per property)

     0.40

     0.20

     0.00L0.002.004.006.008.0010.0012.0014.0016.0018.0020.00

     (l/p/h)

    Figure 1 Leakage Cost Relationship

    This may be described mathematically by a number of different functional forms of which the following power-law function is typical:

     bLUnit Cost of Leakage Reduction (UC) = a .

    In this case, UC is the cost of reducing leakage by one unit of leakage per property (? per l/p/h per property) at a given leakage, . a and b are coefficients established L

    from curve fitting, using actual cost and leakage data. in this case is the average L

    leakage level for which the unit cost applies.

    Leakage cost relationships may also be implicitly modelled within conceptual models of leakage which include parameters such as break out rates, leak flow rates, intervention frequencies etc. BABE is a well known example of such a model.

    Apart from the functional form of the equation there are two critical parameters which define the position and shape of the curve:

1) the leakage level at which unit costs are infinite

    2) the natural rate of rise in leakage

    2

Background Leakage or Policy Minimum?

    The leakage cost curve is asymptotic to a level of leakage at which unit costs are infinitesimally high. This hypothetical level is often referred to as the Background Level of Leakage (BLL), made up from the sum of many small undetectable leaks in a network. The Tripartite Report introduced an additional term ‘Policy Minimum Leakage’ (PML), defined as “the lowest level of leakage which can be achieved through intensive active leakage control using conventional methods, current technology and reasonable effort“. However this has a slightly different meaning to the common definition of the BLL, which refers to an unattainable level of leakage. This is not quite the same therefore as the lowest level achieved, and which might be expected to be always higher than the BLL.

    Nevertheless, lowest levels achieved from samples of DMAs may be used as the basis for estimating BLL values. In previous studies undertaken by the author, the lowest leakage levels following intensive leak detection and repair activity have been 1successfully fitted to the ‘Managing Leakage’ equation which relates BLL to night

    pressures in the network, length of main, number of service connections and infrastructure condition.

    All of the exit levels from leakage interventions will not necessarily give good indications of background levels since this depends on the intensity of the leak detection effort, which has been undertaken during the intervention periods. As a result, two approaches to this have been taken:

    1) to use a very low percentile value of the exit levels over the period of the

    interventions

    2) to use the absolute minimum levels achieved over the period of the interventions

    Needless to say, there is continuing debate in the industry regarding the term ‘Background Leakage’ and because of its influence on leakage targets, there is

    increasing pressure by OFWAT on water companies to provide robust estimates based on an analysis of local data, and not national default values.

Natural Rate of Rise in Leakage

    The total natural rate of rise (NRR) in leakage may be thought of as the continuing increase in leakage that would occur in the absence of any leak repairs. It is made up

    of two components the breakout of new leaks in the network plus the growth (increase in volume) of existing leaks. Of this total NRR, a proportion will comprise visible (customer reported) leaks, which will be duly repaired. It is the remaining portion which is normally used in leakage economic studies, and defines the leakage which must be overcome through active leakage control (or other means) before any observable reduction in leakage over a period (normally a year) is observed. It is therefore used to define the effort / expenditure levels that would be needed to hold leakage at a given level.

     1

    BLL (l/hr) = PCF . ( e . L + f . N ) Where: PCF = Pressure Correction Factor, calculated using the power law function: 1.118PCF = (AZNP/50) L =length of mains (m) N = number of connections e and f are coefficients to be estimated through regression analysis

    3

Two methods have been employed by the author for estimating NRR:

    1) a ‘bottom up’ approach based on the increases in DMA night flow data between

    leakage intervention periods, and;

    2) a ‘top down’ approach based on area or zone leakage levels taken over a year,

    burst numbers and average leak flow rates (Figure 2).

    Neither method is perfect. Method 1 for example will require a large enough sample of DMA data to encapsulate all the variability in network characteristics across a company. Furthermore, if only a small number of ‘between intervention periods’ is used, this may not overcome seasonal variability in the data. There also needs to be separate identification and treatment of visible (customer reported) leaks. Method 2 will generally overcome the network variability and seasonal variability issues, but depends on reliable leak numbers (by type) and robust values for leak flow rates for different types of leak. Nevertheless, both methods have been successfully used in tandem in studies for a number of clients, and have produced reasonably consistent results. Typical values are in the range 2.5 to 3.5 l/p/h per year. Clearly there is a relationship between NRR and infrastructure condition, seasonal and other factors, and further work is needed within the industry to improve our understanding of the mechanics of NRR, to improve the robustness of leakage economic calculations. In addition, it should be possible to further apportion NRR into the different assets in the network; mains, communication pipes, customer supply pipes - so that the economic analysis can potentially be applied to these different parts of the distribution system.

     NRRt

     Lstart Volume

     of leak

     repairs

     Lend

     Start ofEnd of yearyear

    Figure 2 NRR Calculation (Method 2)

Economic Levels of Leakage

    The unit cost relationship illustrated in Figure 1 may also be described in Total Cost terms as follows:

     ddTC = c . [(L + NRR BLL) (L BLL)] 12

    Where TC is the total annual cost (? per year per property) of reducing the average leakage in year 1 (L) to an average leakage in year 2 (L), c and d are coefficients, 12

    and can be derived from the values of a and b in the unit cost equation.

    4

    It is this form of the leakage cost relationship that we have often used in the calculation of ELLs and short-run budget setting for clients. The advantage of this form of the equation is that it can be used to assess the annual expenditure necessary to:

; maintain leakage at a given level i.e. where L1 = L2, or

     L2. ; move from one level of leakage to another, L1

    Therefore, by selecting a glide path of leakage from the current level to some notional target in a future year (and then maintained thereafter), the future annual cost stream of leakage control can be forecast.

    For the given leakage profile, water supply costs are then calculated using the marginal cost of water and appropriate forecasts of demand (consumption plus leakage) in each year.

    Present Value costs of leakage control, plus water supply costs, are then calculated for a range of notional targets to give the familiar trade-off curves between leakage control and supply costs, from which the ELL can be deduced (Figure 3).

    PV of

    Total

    Total CostBackgroundCostLeakageCost of

    Water

    Detection + Repair

    Costs

    LeakageELL

    Figure 3 ELL Calculation

Environmental and Social Costs

    As mentioned previously, company Least Cost Plans are expected to have included a sensitivity analysis of the effects of environmental and social costs on leakage policy, supply and demand management options. In the context of recent short-run ELL studies undertaken by RPS the effects of including environmental and social costs 2have been analysed using the Environment Agency best practice guidelines. All best

    practice approaches tend to recommend the use of local data wherever possible. Although the EA guideline provides a step-by-step methodology for determining E&S costs, the local information that would be necessary to apply it objectively, is not in our experience, readily available within companies. In our view, considerable future

     2 Benefits Assessment Guidance and Valuations for Water Resource and Water Quality Schemes’, RPA, 2003

    5

    investment will be required on data collection and development of the methodology if the robustness of E&S costing is to become comparable with other aspects of leakage economic analysis. The fallback position within the guideline in the absence of local data is to apply the ‘benefits transfer’ approach. Here, the cost and benefit assessments are taken from known study sites (often geographically remote from the study area), appropriately adjusted to account for different technical and environmental characteristics, and then applied to the sites in question.

    Our studies so far tend to have focussed on two principal categories of impact:

    ; Traffic Disruption: assessments at zone level of the traffic disruption costs

    associated with incremental changes in the level of mains repairs for the range of

    road classifications.

; Environmental impacts resulting from changes in river abstraction:

    assessments at resource zone level of the direct benefits, which would accrue from

    changes in river flow at abstraction sites assuming a 1:1 relationship between

    leakage and river flow.

    The resulting values are then applied as potential additions to the marginal cost of water and therefore provide a mechanism for generating lower ELLs. As might be expected, calculated values have varied considerably between zones ranging from 1 or 2 ? per Ml, to values in the same order of magnitude as the marginal cost of water. For some zones therefore, the ELL can be significantly reduced if E&S costs are included.

Target Setting

    In the past it was usual for ELL values to be evaluated for the company as a whole, but new regulatory requirements mean that they now need to be quantified at resource zone level. ELLs therefore represent targets to be used by companies in strategic water resources planning and supply management. However, it is the Leakage Control Manager who is responsible for delivering these targets, by translating them to a lower level such as District Meter Areas. Evaluation of ELLs at DMA level is not currently regarded as Best Practice. Prioritisation of detection resources to DMAs, such that ELLs at zone level (and subsequently at company level) are achieved in the most cost efficient way, is therefore the challenge faced by Leakage Control Managers. RPS have recently completed a study for UKWIR which included an evaluation of the criteria which may be used for this prioritisation.

    Whatever approach is adopted to set leakage targets in the future, there is an expectation by the industry that leakage targets will continue to be driven down. The inclusion of increasingly robust environmental and social costs and benefits may well add some justification to lower economic leakage targets. However, current approaches to Environmental and Social costing can not be regarded as robust, and their inclusion is likely to add to the overall uncertainty surrounding company economic leakage levels. OFWAT acknowledge that further work is required in this area.

“Find and Fix” or Asset Renewal?

    Leakage targets over recent times have been achieved largely by active ‘find and fix’ as the most cost-effective policy for most water companies in England and Wales. With a few notable exceptions, most water companies are now operating at, or indeed below, their economic levels of leakage. Therefore the future challenge for most

    6

    companies is to overcome natural rates of rise in leakage. However, this is against a background of pressure from the industry regulators for further reductions in leakage over time.

    The new AMP4 business plans for the most part appear to have concluded that an increase in the level of mains renewals will be the ‘best’ strategy to help achieve

    leakage objectives, whilst at the same time stemming the tide of increasing asset performance failure. However, there are clear dangers in a wholesale policy shift away from ‘find and fix’.

    It goes without saying that mains renewal will have some impact on levels of company-side leakage, the leakage cost relationship and therefore the ELL. Since both the background level of leakage and the natural rate of rise are both inherently dependent (at least in part) on the state of the infrastructure, the effects on the leakage cost relationships of asset renewal or rehabilitation could be modelled through changes in BLL and NRR. Background levels of leakage for example can be expected to reduce (albeit temporarily) in those parts of the network which are renewed. Currently running detectable leaks will be similarly removed, and natural rates of rise can also be expected to become flatter through reduced break out rates. However, the degree of change and the timescales over which the original levels will eventually be reached are currently not predictable with any degree of certainty.

    Without further industry investment in research and development into these relationships, there is clearly a risk that past achievements in leakage reduction will be nullified if companies relax current ‘find and fix’ policies in favour of mains renewal.

Where do we go from here?

    Given the varying levels of uncertainty associated with the different components of the leakage cost relationships, and other aspects of the ELL calculation (particularly the new requirement to include the effects of E&S costs), it is perhaps surprising that published company leakage targets do not come with government health warnings. There is much to be done by our industry in improving data availability and accuracy and refining our understanding of some of the fundamental mechanisms on which leakage economic analysis relies so that decisions on possible policy shifts can be taken with confidence.

    7

Report this document

For any questions or suggestions please email
cust-service@docsford.com