UKWIR “BIG QUESTIONS” - ACHIEVING ZERO LEAKAGE BY 2050
PROPOSED RESEARCH PROGRAMME
The “Zero Leakage” programme, which is part of UKWIR’s “Big Questions” programme, is being developed in two stages. In Stage 1, universities and consultants were engaged to investigate the current state of knowledge in each of five subject areas, namely:
- Basic mechanisms of leakage
- Leak detection and location
- Leak repair
- Leak-free new networks
- Water accounting
The objective of this stage was to determine the following for each of the five areas:
What research and development has been done
- What research and development is currently in progress
- What are the gaps, and therefore what new research and development needs to be done to facilitate the path towards zero leakage.
This stage was completed in 2016. Following this work, a research roadmap was developed, and 15 proposed projects were identified. Of these, 10 would be suitable for development by universities as academic projects in cooperation with UKWIR and the UK water companies. It is thought that all of these would be suitable for co-funding by UKWIR and other funding bodies, and many would be suitable candidates for academic programmes such as STREAM or WISE.
All of the projects are listed in the table below. Projects 1, 2, 13 and 15 are expected to be carried out by consultants. Projects 3 to 12 inclusive are expected to be university-led as 3 or 4-year PH.D or Eng.D projects. Project 14 would be carried out by volunteering water companies working with their contractors, and coordinated by the Water UK Leakage network. Further details are given in the following sections and in Annex 1.
Modern technology has made it possible to collect much greater quantities of data, and at higher resolution. Leakage analysis methods and leak detection technology have both made many advances in recent years, but data collection and manipulation processes have hardly changed. The basic principle of measuring minimum night flow into a DMA, and then subtracting estimates of household and non-household night use to give leakage, remains unchanged in the past 30 years.
However the recent growth in smart networks, and particularly the use of smart meters for revenue purposes, could offer many new opportunities for better leakage management. It is essential that these opportunities and benefits are identified now, so that water companies can take them into account when making their choices of which smart technologies to invest in.
The aims of this project will be to :
- Investigate and document the smart network technologies which are now available, and which water companies are currently using them, throughout the world.
- Gather information and data from those companies which are using smart network technologies, specifically with reference to how these technologies are being used to support leakage management.
- Identify the opportunities for better leakage management from the following, amongst others :
- More data; different data; more frequent data; better quality data, etc
- More permanent monitors with greater processing power
- Rapid integration with other data sets – e.g. AMR data from customer meters
- Real time analysis and diagnosis
- Live modelling
- Advanced pressure management and transient control
- Real time transmission of outputs to field staff
- Describe the advantages and disadvantages for leakage management of the different types of smart technologies with different levels of data discrimination.
- Make recommendations for the most useful technologies for leakage management.
Project 2 - Understanding the balance between customer use, supply pipe leakage, plumbing losses and meter under-registration
It is not possible to measure leakage from a water distribution network directly. The traditional approach to leakage analysis, which has remained unchanged for at least 35 years, is to measure the total net flow entering a district metering area (DMA) at night, subtract estimates of customer use at night in both household and non-household properties, and assume that the remainder is leakage. However the real situation is much more complex.
The flow entering a DMA at night comprises leakage from the water company’s network, leakage from customer supply pipes, plumbing losses within the property, and genuine use of water by customers. Under-registration or non-registration of low flows introduces a significant degree of uncertainty into the data provided by the customer’s meter. Household night use allowances are derived from company consumption monitors, which also suffer from under-registration. Also, all of these processes suffer from the fact that, from data analysis alone, it is impossible to differentiate between supply pipe leakage, plumbing losses and continuous use. There is a growing body of evidence that the proportion of DMA night flow which is accounted for by components other than leakage, particularly plumbing losses, may be significantly greater than previously thought.
This project will use recently developed flow measurement techniques to investigate these factors across a representative sample of household properties within several water companies. The data obtained will be used to provide greatly improved estimates of :
- Plumbing losses, which are part of consumption
- Intermittent night use, also part of consumption
- Water running into storage at night, also part of consumption
- Background leakage on underground supply pipes, which is part of the total leakage KPI
Other factors which will be investigated include occupancy rates, seasonal variations, unaccounted for consumption, flows at low rates into storage / WC cisterns, and meter under registration, and the impact of all of these on the resulting leakage estimates.
A significant amount of research has already been done on the processes by which buried pipes of various materials degrade over time while they are in service. This includes investigations of the impacts of various environmental factors such as soil type, bedding material, pressures, external loading (e.g. traffic) and water quality. However this work has been directed almost exclusively towards the prediction of catastrophic failures. A significant amount of research has already been done on the processes by which buried pipes of various materials degrade over time while they are in service. This includes investigations of the impacts of various environmental factors such as soil type, bedding material, pressures, external loading (e.g. traffic) and water quality. However this work has been directed almost exclusively towards the prediction of catastrophic failures.
Very little work has been done to investigate how these various degradation mechanisms are manifested as leakage. However, pipes can deteriorate to a certain extent without leaking. These projects will examine the consequences of such degradation in relation to potential leakage mechanisms, both in terms of time to first leak (and its variability) and the subsequent evolution of leakage rates. They will identify the role of material type, inherent defects and the internal and external environmental factors listed above on the initiation and growth of degradation processes that can lead to through-wall leakage without resulting in pipe failure, i.e. the processes that are important in creating and sustaining leakage before a final burst occurs.
Variations in temperature are considered to influence pipe failure (e.g. cold weather events) although the precise mechanism responsible for increased leakage and burst rates remains unclear. In particular, the relative contribution to increased winter leakage is not well understood.
As these mechanisms are likely to be very different for different pipe materials, it is envisaged that this work will be developed as three separate but related projects as follows:
- Project 3 – cast iron
- Project 4 – asbestos cement
- Project 5 – service pipe materials (probably lead, galvanised iron and copper).
Through-wall leakage on modern polyethylene and PVC pipes is thought to be very rare, and normally only caused by third party damage, so these materials will not be considered.
In pipe networks, joints and other fittings must provide the mechanical continuity to allow water to flow under pressure without leakage. Joints must also enable the pipes themselves to resist variations in both the service and environmental loadings over time. However; there is evidence that the materials used in joint systems between pipes are prone to degradation processes that can be independent of the pipe material they are joining, e.g. the lead run joints used to connect cast iron pipes and the natural rubber sealants used in asbestos cement pipe systems. Any degradation of the joint compromises its ability to accommodate movement. This creates the opportunity for leakage at the joint (through ground movement and thermal effects) and increases the risk of pipe leakage and failure.
Whilst the coupling of different metals is known to cause potential issues, the interactions that can occur between the various materials within joints and other fittings is only poorly understood. There remains a need for more research into development of leaks in joints as a result of both degradation processes and the thermo-mechanical effects during their time in service.
These projects are similar to Projects 3, 4 and 5 described above, but relate to the influence of joint deterioration on leakage rather than deterioration of the pipe itself. A key difference is the fact that the basic mechanisms of degradation of pipe joints and jointing materials, and the impact on this of the various internal and external environmental factors, are currently less well understood than those for the pipes themselves. So these projects will begin by reviewing the available literature (and other evidence) of historic failure of joints and connections in small diameter pipes across the range of different exposure conditions. They will then seek to clarify how the condition and integrity of pipe joints degrade in service, and what factors control this process. They will determine the primary causes that initiate defects in different joints systems and how any degradation induced changes go on to create leakage under normal service exposures.
The work will compare the likely performance of old and new examples of a given joint type to gather evidence to establish whether joint leakage is the result of fundamental material issues or stem from manufacturing or installation problems. This will require both on site and laboratory-based approaches.
As these mechanisms will be different for the jointing systems used with different pipe materials, it is envisaged that this work will be developed as two separate but related projects as follows:
- Project 6 – cast iron jointing systems
- Project 7 – asbestos cement jointing systems
It is known that leakage on polyethylene pipe systems occurs almost entirely at joints and fittings. However this has already been investigated, and will be covered by one of the non-academic projects (Project 15 – Leak-free new networks).
There is growing evidence that some of the leakage occurring on distribution networks may be partly or wholly caused by the way the network is operated. The relationship between leakage and pressure is reasonably well understood, but the impact of normal diurnal pressure variations has not been studied. Some work has recently been done to understand the impact of pressure transients on pipe bursts, but the impact on leakage has not been investigated. It is known that large transient pressure spikes can cause pipe failures, but there is evidence that repeated transient pressure waves of lower amplitude can also cause degradation of pipes and joints by fatigue mechanisms. This effect may result in leakage long before it causes pipe failure. Operations of valves and other fittings by water company staff can cause pressure transients, as can the activities of industrial customers, particularly in the filling of site tanks. Routine activities of third party operators may also lead to leakage or bursts.
This project would build on existing research to establish the prevalence of these effects in typical water distribution networks. It would also investigate, in the laboratory and in the field, the potential for all of these mechanisms to damage the pipes and joints in a way that can cause leaks or influence leakage levels. It is hoped that this could lead to recommendations to water companies for minimisation of the impact on leakage, e.g. pressure management schemes or network calming.
Detection of leaks on plastic pipes using acoustic methods is difficult, due to the rapid attenuation of the leak noise signals with distance from the leak. There has been a tendency within academia and the industry to try to compensate for this by the development of ever more sophisticated hardware and more complex processing techniques. However there is evidence that it may be more effective to address the fundamental issues of acoustic wave propagation in pipes to reduce the attenuation losses. Recent work suggests that a 20dB improvement in the leak noise attenuation at 100Hz could increase the distance at which a leak can be detected in plastic pipes by 50 metres.
This project will investigate several options for enhancing the transmission of leak noise on plastic pipes, including:
(i) Reduction of the radiation of leak noise into the surrounding soil, by partial decoupling of the pipe from the soil.
(ii) Optimisation of sensor locations and configurations, to take advantage of the natural resonances within a pipe network. This could significantly increase the bandwidth of the measured leak noise, which would improve the effectiveness of leak noise correlators
(iii) Reduction of acoustic coupling losses at discontinuities, i.e. at pipe joints.
(iv) Making use of both flexural waves and axial waves, i.e. detection and correlation in three dimensions instead of one.
Leak detection methods may be classified into two categories:
- Methods based on the detection of the transient signal that occurs during the burst process when the leak first occurs.
- Methods based on the detection of the continuous steady state signal that occurs after the burst process.
The first method requires permanent deployment of sensing instrumentation, whereas the second method uses temporary deployment of instruments (e.g. correlators).
The signal processing methods developed for these two types of signal are very different. For the detection of transient signals, processing methods such as the inverse transient method, cepstral methods which aim to separate the reflection due to the leak from reflections due to other parts of the pipe system, and wavelets have also been used to give a time-frequency representation of measured pressure signals. Signal processing methods developed for the detection of the steady state signal are usually variants of the basic correlation method. Leak noise correlators have the major advantage that the correlation process removes uncorrelated noise from the two sensors used in the system.
This project will provide a direct comparison of the spectral content of the pressure signals in the fluid and vibration signals in the pipe wall due to a transient signal from a bursting leak with the steady state signal that occurs following the burst. A comparison will also be made of the of the peak transient signal amplitude for the two types of signal.
As a result, the project will provide:
- Guidelines on the relative effectiveness of transient and steady state detection methods.
- Information on the relative benefit of permanently instrumenting a pipe in an attempt to observe the leak as it first occurs, compared with instrumenting the pipe following a suspected leak.
Optical fibres are capable of sensing or measuring a number of different parameters at close spacing over long distances, potentially many kilometres from a single installation. Optical fibres laid along a water main could detect and locate noise and vibration, caused either by the fracture of a pipe at the moment it happens, or by the escape of water from a pipe after the fault has occurred. It could also measure temperature, indicating either the presence of moisture in a dry soil or the change in soil water temperature when water escapes from a pipe. Similarly it could detect strain on the pipe wall, indicating a break or an imminent break.
This project will investigate the practicality and the usefulness of such techniques in the context of leakage management. It will consider two situations:
(i) Installation of optical fibres with the laying of a new pipe.
(ii) retrofitting of optical fibres to existing pipes.
Leakage management techniques using leak noise loggers or correlators both involve the making of detailed recordings of leak noise, which potentially contains a large amount of information about the leak. However current techniques only use this information to determine whether a leak is present or not. Recent work at Sheffield University has shown, using high quality laboratory-based data and sophisticated signal processing techniques including feature recognition software, that information such as the nature, shape, size and flow rate of a leak can be extracted from the recorded leak noise signal. However further work is required to refine and extend this work, and to demonstrate the practicality of the technique in real field situations.
This project would have three main tasks:
(i) Use of the existing laboratory data, and also new data generated as part of this work, to optimise the feature extraction from the signal, including the trialling of modern neural network techniques, to relate the features of the leak noise signals to the parameters of the leak.
(ii) Acquisition of additional data from a controlled buried infrastructure facility, to provide full size representative tests on a range of leak types and materials. Use of this data for further training of the leak characterisation system.
(iii) Work in conjunction with a UK water company to acquire data from real leaks in the field. This would entail the recording of leak noise before and after repair, together with the capture of detailed information about the leak at the time of excavation.
Distribution Maintenance staff at water companies know well that many leaks and bursts, once excavated, prove to be at the location of a previous repair, and occur as a result of a failure of the old repair. However it is not known how much data is collected on this, and there is no quantitative evidence of the magnitude or significance of this problem at national level. Nor has there been any study of the reasons for the failures, i.e. whether they are due to deterioration of the clamp or other repair materials over time, or whether they are caused by faulty workmanship at the time of the initial repair.
Many companies do record the types of failure within their records of mains and service bursts. However these descriptions are often very brief (e.g. “pin-hole”), and the fact that the failure was at a previous repair may not be recorded. This project will initially assess the availability of suitable data, in collaboration with participating water companies. The national mains failure database may be a valuable source of data.
The objectives of the project will be:
(i) To assess the proportions of bursts on mains of different materials which occur as a result of the failure of a previous repair, and thereby to determine whether this is a significant problem.
(ii) Where bursts have occurred at the site of a previous repair, to determine where possible whether these have resulted from deterioration of the materials used for the repair over time, or by faulty workmanship at the time of the original repair.
(iii) To provide guidance for maintenance operatives carrying out repairs, on how to minimise the likelihood of future failures
Water pipes made of medium density polyethylene (mdpe), which has been by far the dominant material used for new networks in the UK since the 1980s, do not degrade with time in the ground. Therefore it should be possible to lay mdpe networks which do not leak. However, UKWIR report 10/WM/08/43, “Leakage on Polyethylene Pipe Systems”, published in 2011, identified the fact that many of the polyethylene pipe networks laid in the UK do exhibit significant levels of leakage, including some networks that have been laid relatively recently.
If zero leakage is to be achieved, then it will be essential that newly-installed networks are leak-free, both at the time they are installed and for many decades thereafter. The Phase 1 work in this area (Lot 4) concluded that the tools, techniques, materials and fittings required to lay leak-free new networks are already readily available in the UK, and that no new research or development is required. However the 2011 report concluded that the optimum tools and techniques are often not used, in an attempt to reduce unit costs. Therefore the challenge is to ensure that water companies apply appropriate controls on:
- Selection of materials and fittings
- Quality control on site
- Standards and testing
- Training of staff
The objectives of this project will be:
(i) To review the analysis and conclusions from the 2011 report
(ii) To collect additional leakage data on polyethylene networks laid since 2010, to assess whether the leakage levels on mdpe networks are still at the levels described in the 2011 report.
(iii) Try to relate the leakage levels measured on recently-laid new networks to the materials and methods used during installation.
(iv) To draft a specification for the materials, techniques and methods that would be required to ensure as far as possible, that new mdpe networks are leak-free, both at the time they are installed and into the future. This will include consideration of the following among other issues :
- Use of better fittings and couplings for polyethylene pipe.
- Better workmanship and quality control on site.
- Better jointing techniques
- Service pipe layouts and meter position on new developments.
- Pressure-testing of the whole system including services
- Different contract styles, for example including long-term performance guarantees.
(v) To demonstrate the whole-life cost case for higher specification new networks, in terms of reduced maintenance costs.
Project 15 - Establishment of some high specification new DMAs, to demonstrate that zero leakage can be achieved
This project will follow on from Project 14 above, and ideally should run in parallel with the later stages of Project 14. Once the specification required to ensure that new networks will be leak-free has been drafted and agreed, selected companies will be asked to volunteer to construct a network for a new development in accordance with this specification. Companies will be asked to set the new development up as a new and separate DMA, so that leakage levels can be monitored within the area of the new network. This new DMA may be smaller than the company’s normal DMA size.
Companies will then carry out intensive leakage monitoring of these new DMAs, including fast logging to establish the best possible allowances for household night use. If leakage is found, companies will be expected to remedy this, and to continue this intensive process for a year following completion of construction.
The resulting leakage levels, and the measures taken to achieve them, will be reported by the companies through the Leakage Network. The aim will be to demonstrate whether or not zero leakage can indeed be achieved and sustained by using the specification produced in Project 14.
UKWIR Programme Lead for Leakage
Annex 1 - ZERO LEAKAGE BIG QUESTION - PROPOSED RESEARCH PROGRAMME
|Project No||Area||Project||Project Duration||Researchers|
|1||Water accounting||Use of smart meters and smart networks for leakage management||1 year||Consultants|
|2||Water accounting||Understanding the balance between customer use, supply pipe leakage, plumbing losses and meter under-registration.||2 years||Consultants|
|3||Basic mechanisms||Understanding of how deterioration of pipes evolves into leakage - cast iron||4 years||Eng D student|
|4||Basic mechanisms||Understanding of how deterioration of pipes evolves into leakage - asbestos cement||4 years||Eng D student|
|5||Basic mechanism||Understanding of how deterioration of pipes evolves into leakage - service pipes||4 years||Eng D student|
|6||Basic mechanisms||Degradation of pipe joints, and the impact on leakage - iron pipes||4 years||Eng D student|
|7||Basic mechanism||Degradation of pipe joints, and the impact on leakage - asbestos cement pipes||4 years||Eng D student|
|8||Basic mechanisms||How does network operation affect leakage?
Impact of diurnal pressure variations, transients/fatigue, third party operators, water quality, etc.
|4 years||Eng D student|
|9||Leak detection||Leak detection Enhancing the transmission of leak noise on plastic pipes||3 years||Eng D student|
|10||Leak detection||Leak detection Comparison of transient and steady state leak detection methods 2 years Post-doctoral researcher||2 years||Post-doctoral researcher|
|11||Leak detection||Leak detection Use of fibre optics in water pipes to detect leakage||3 years||Post-doctoral researcher or Eng D student|
|12||Leak detection||Leak detection Analysis of leak noise to determine characteristics of the leak before excavation.||3 years||Eng D student|
|13||Leak repairs||Leak repairs Incidence and causes of repeat bursts at old repairs||1 year||Consultants|
|14||Leak-free new networks||Leak-free new networks What do we need to do differently to lay leak-free networks?||1 year||Consultants|
|15||Leak-free new networks||Leak-free new networks Establishment of some high specification new DMAs, to demonstrate that zero leakage can be achieved||3 years||Companies with the Leakage network|