Data Mining on Leakage Frequencies

Capstone Project Proposal #2

I.     Proposed Topic

Data Mining on Leakage Frequencies

II.     Introduction

Today’s times is crucial for the survival of human beings.  With climate change, global warming and other events that would put the entire living creatures on the brink of extinction, action for preservation is not an option.  More than ever, the clamor for water preservation is indispensable.  Whatever means for conservation would redound to the benefit of the human race.

Water conservation has been the buzzword especially in the water utilities.  They are mandated not only to provide potable water but to help in its preservation as well.  Among the biggest culprit in water loss in the industry are pipe leakages which adds up to the non-revenue water.

Non-revenue water is the difference between the volume of water put into a water distribution system and the volume that is billed to customers (Frauendorfer & Liemberger, 2010).  It is one of the lowest hanging fruits in order to improve the efficiency of water utilities around the world. Especially the water losses in the distribution systems are obvious when analysing data from water utilities. It is more than a decade since IWA presented the ‘best practice’ standard water balance but many water utilities have still no overview of the situation. Around the world we still see that non-revenue water accounts for 25% – 50% of the total water supply (Hvilshøj, 2015).

Based on the Benchmarking activity spearheaded by the Digos Water District (DWD) and currently participated in by 16 out of 54 Category B Water Districts, the industry average for non-revenue water is 1,335,670 cu.m. or equivalent to 23.63% of water produced by the participating water districts.  These Category B Water Districts are Baliwag WD, Bislig City WD, Bocaue WD, Guagua WD, Ilocos Norte WD, Malaybalay City WD, Mariveles WD, Metro Kalibo WD, Metropolitan Tuguegarao WD, Plaridel WD, Silang WD, Sultan Kudarat WD, Surigao Metro WD, Tabaco WD, Valencia WD, and Sorsogon City WD.

There are a number of factors for this water loss – flushing activities, illegal connections and pipe leakages, among others.  Various measures to reduce non-revenue water had been implemented in the water districts but some technology are too expensive especially for the small water districts.  Since the widest area for any water district wherein non-revenue water would probably emanate from are pipe leakages, this would be the focus of this proposal.

To aid the water district in water conservation, its data for pipe leakages and pipe aging will be mined.   The system will then give an analysis on which pipes would be pulled-out for repair or changed with a new one.

III.     Purpose of the Study

The purpose of this project is to develop a data mining and analysis tool which determines leakage frequencies.  The objectives are as follows:

  1. To promote water conservation by reduction of non-revenue water arising from pipe leakages;
  2. To augment the efficiency of pipelines and maintain high standard of water quality; and
  3. To reduce cost of repairs by preempting huge pipeline burst.

IV.     Specific Research Issues/Questions

This project arises from the following issues:

  1. What are the factors that may have caused for burst pipelines?
  2. How prevalent are the pipe leakages in a franchise area of a water district?
  3. Is there a reduction in the non-water revenue after repairs thereof had been made?

These issues will be discussed below.

  1. What are the factors that may have caused for burst pipelines?

There are a lot of factors for pipe leakages or pipe burst if the leak is of magnitude intensity. These factors are generalized as (1) human error and (2) natural wear-and-tear of materials.  Human error, such as the pipe was accidentally hit during a construction on the ground, was very prominent.  These accidents were immediately reported to the water utility thus, reparation is uncomplicated and only a few volumes of water was wasted.  However, detecting a pipe leakage through natural wear-and-tear is difficult to detect.  For one, it is located underground, and depending on the quality of the soil, sometimes it would go unnoticed for a longer period hence, significant volumes of water are wasted.

  1. How prevalent are the pipe leakages in a franchise area of a water district?

In 2015, Digos Water District (DWD) had received 2,582 reports for mainline and service line pipe leakages.  These reports are made because they are detected either by concerned citizens or by the DWD employees themselves.  Unfortunately, there is no data as regards the equivalent volume of water wasted nor the period of how long the leakage has been in existence before its detection.  In addition, length of pipes could not be easily identifiable because old pipes were not numbered for identification purposes.  The challenge would be to assign each pipe with a number for identification so that a history thereof would be set such as the frequency of leaking in the said pipe.

  1. Is there a reduction in the non-water revenue after repairs thereof had been made?

As stated in the previous section, the water district has no data regarding the volume of water wasted in pipe leakages.  Although there had been a gradual decrease in the non-revenue water, it cannot be said that it is directly attributable to the repairs made on the pipes.  It must be taken into consideration that non-revenue water is not solely water loss through flushing and pipeline leakages but it includes illegal connections as well.

V.     Proposed Methodology

  1. Updating of records of pipelines (number, size, date installed, other pertinent data) will be implemented. This will be integrated in the GIS/mapping system of the water district.
  2. A database of pipeline records will be put into place.
  3. A system will show the pipeline data, age, leakage history, and others as well as analysis on leakage frequency will be developed.
  4. Preventive maintenance will be manually made to existing pipelines should the threshold in leakage frequency be obtained.

VI.     List of Readings

Literature on Non-Revenue Water (NRW)

According to the Asian Development Bank (ADB) paper by Frauendorfer & Liemberger, (2010), non-revenue water (NRW) is the difference between the volume of water put into a water distribution system and the volume that is billed to customers.  It comprises of three components as follows:

  • Physical (or real) losses comprise leakage from all parts of the system and overflows at the utility’s reservoirs. They are caused by poor operations and maintenance, the lack of active leakage control, and poor quality of underground assets.
  • Commercial (or apparent) losses are caused by customer meter under-registration, data handling errors, and theft of water in various forms.
  • Unbilled authorized consumption includes water used by the utility for operational purposes, water used for firefighting, and water provided for free to certain consumer groups.

Although it is widely acknowledged that NRW levels in developing countries are often high, actual figures are elusive. Most water utilities do not have adequate monitoring systems for assessing water losses, and many countries lack national reporting systems that collect and consolidate information on water utility performance. The result is that data on NRW is usually not readily available. Even when data is available, it is not always reliable, as some poorly performing utilities are known to practice “window dressing” in an attempt to conceal the extent of their own inefficiency.

The main objective of a water utility is to satisfy customer demand. A high level of NRW has a severe and direct impact on the ability of utilities to meet this objective and therefore has a negative impact on customers. High physical losses often lead to intermittent supply, either because of limited raw water availability or because of water rationing, which may be needed to reduce supply hours (and therefore hours of water leakage) per day.

In addition to substandard service, intermittent supply poses a significant health risk, as contaminated groundwater, or even sewerage, can enter the leaking pipes during supply interruptions and very low pressure periods. The avoidance of this significant public health risk should be reason enough to reduce leakage to enable continuous supply. High leakages also increase flow rates in the pipe network, which can cause unnecessarily high pressure losses that affect customers and often lead to supply interruptions during peak demand hours.

Information on Physical Losses and Leakage Reduction

Nearly 60% of NRW in the developing world is due to physical losses (as opposed to faulty metering and theft), i.e., treated water is lost in the network due to leaks and overflows at storage tanks. Reducing this loss by half will allow the utilities to serve an additional 100 million people with no further withdrawals from natural sources. This loss reduction will also allow the utilities to save the energy associated with treating and pumping roughly 8 billion cubic metres of water (Narayanan, Vasan, Sarangan, & Sivasubramaniam, 2014).

Physical losses can occur along the entire distribution system, from storage reservoirs and the primary network to the smallest service connections. When people think about leakage, they normally think of big and spectacular pipe bursts. These often cause a lot of damage but are insignificant in volume compared to all the other leaks that do not come to the surface.

Normally around 90% of water that is physically lost from leaks cannot be seen on the surface. These leaks might eventually become visible after many years, but until then, large volumes of water are lost every year. Sometimes, undetected leaks can be quite large, such as those that run directly into a sewer or a drain. Therefore, a water utility that does not practice a policy of efficient and intensive active leakage control will always have a high level of leakage, except if the infrastructure is new and/or in excellent condition (Frauendorfer & Liemberger, 2010).

Many water utilities in Asia practice passive leakage control, meaning that they repair only those leaks that are visible. This is clearly not enough since 90% of the leaks are usually not visible on the surface. This means it takes far too long, often many years, until the utility is even aware that there is a leak. Since awareness time largely determines the volume of water lost from a pipe burst, utilities need a strategy to reduce awareness time (Frauendorfer & Liemberger, 2010).

Most water networks are metered only to the extent allowed by the budgets of utilities. In the absence of any metering at all, there is little a utility can do about leak detection except respond to customer complaints or abnormal consumption at head-works (Narayanan et al., 2014).

NRW and the Society

Many utilities that have been successful in addressing NRW have gone beyond technical measures to address community behavior that drives illegal connections and pilferage. This is done with the understanding that water loss is not just an engineering problem but also reflects a sociocultural situation that requires changes in community behavior and attitudes toward water usage.  (Frauendorfer & Liemberger, 2010)

Some water utilities, for instance, took advantage of District Metered Areas (DMAs) as the basis of a decentralized field operations structure as practiced by the Manila Water Company Inc.  In Jamshedpur, India, technical measures have been complemented by efforts to address illegal connections by walk-through surveys and authorizing illegal connections by legitimizing them and adding them to the network.

In Phnom Penh, the public utility was able to reduce NRW by 91% in 15 years through strong commitment and a comprehensive network replacement and physical loss reduction program. On top of that, simple but unique measures were taken to reduce commercial losses. For example, if a meter reader of an area did not, or could not, find an illegal connection, but one of his colleagues did, the colleague received a reward and the meter reader was penalized. (Frauendorfer & Liemberger, 2010)

VII.     Expected Significance of the Study

In spite of the potential benefits, NRW reduction is not a simple matter to implement, and this explains why so many water utilities fail to address this issue effectively. Not only do new technical approaches have to be adopted, but effective arrangements must be established in the managerial and institutional environment—often requiring attention to some fundamental challenges in the utility. (Kingdom, Liemberger, & Marin, n.d.)

The present practice in the Digos Water District as regards NRW reduction is too manual.  For instance, a concerned citizen would call up the office and report a leaky pipes; the DWD personnel would proceed to the area and repair the visible leak.  If only the water district has historical data on all leakages they had repaired corresponding to the estimated volume of water lost in leaks, then it would be easier to detect nearby leaks.  This is because the water district had commenced the implementation of Zone Metering Areas which would more or less show the water that entered a zoned (imaginary) pipe area vis-à-vis the billed consumption for that area.

Water loss is a worldwide issue that demands the highest attention from the water utility sector more so with the Climate Change we have currently experienced.  Although there are various technologies that aid in the detection of leaks, it is too expensive or impractical for usage especially in water utilities in third world countries.  Hence, a system should be developed to data mine the leakage frequencies in a water district to pinpoint possible unnoticed leakage.

VIII.     References

Frauendorfer, R., & Liemberger, R. (2010). The Issues and Challenges of Reducing Non-Revenue Water. Retrieved May 30, 2016, from

Hvilshøj, S. (2015, May 20). Reduction of non-revenue water around the world. Retrieved May 30, 2016, from

Kingdom, B., Liemberger, R., & Marin, P. (n.d.). The Challenge of Reducing Non-Revenue Water (NRW) in Developing Countries. Retrieved June 15, 2016, from

Narayanan, I., Vasan, A., Sarangan, V., & Sivasubramaniam, A. (2014). One Meter to Find Them All: Water Network Leak Localization Using a Single Flow Meter. In Proceedings of the 13th International Symposium on Information Processing in Sensor Networks (pp. 47–58). Piscataway, NJ, USA: IEEE Press. Retrieved from

IX.     GANTT Chart

2 Leakage Frequencies


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s