Leak detection is elementry

The water industry in England and Wales is giving itself a well-earned pat

on the back. With leakage reduced by 37%, enough to supply 12.5M people, things

have come a long way since the ‘dark days’ of the early 1990s.

Add to this Ofwat’s softening view of the need for mandatory leakage targets

for most companies and some might say that perhaps the pat on the back is being

prematurely delivered.

Decade of change

At least one long-standing issue of consensus remains constant – waste is a

bad thing. Whether the waste is water leaking into the ground or money spent

on locating difficult to find leaks, depends on your perspective. Whatever your

point of view, there is no doubt that the science of leakage management has

substantially improved in the last decade. Since privatisation, the creation

of thousands of DMAs has enabled water companies to handle network losses in

manageable chunks. In parallel with this, leakage detection and location technologies

have developed rapidly.

In the mid-1990s a water company facing a large disparity between its reported

leakage and its target might have contemplated region-wide deployment of noise

loggers, devices that register noise anomalies in pipes and thus give an indication

of local leakage. However, such a scatter-gun approach would be considered overkill

today.

So where are we with leakage detection technology? Most water companies no

longer routinely carry out step-testing, the process of systematically closing

distribution valves to localise a leak, thanks, in part, to the targeted rather

than widespread use of noise loggers. They are a powerful tool, but do not do

enough to complete the job. Correlation teams are still required in most companies

to convert a localised leak into a pinpointed leak. Thus there is a three-step

process employed by most water companies. Firstly, an increase in minimum night

flow is detected by a DMA flow data logger, which probably indicates the presence

of a leak. Next, the leak is localised to an area using noise loggers. Last,

a specialist team travels to the area to accurately pinpoint the location of

the leak using a conventional correlator system.

Although this approach works, its efficiency is limited by two factors. The

first of these is that many water companies have struggled to fully justify

the cost of adding telemetry to network data loggers. Using lower cost data

loggers, which have to be manually downloaded, results in a significant time

delay between when the burst occurs and when the data from the relevant logger

is collected and analysed. Secondly, there is an additional delay between the

deployment/collection of noise loggers and the arrival of the correlator team.

This is further exacerbated in some companies because the noise logger activities

and the correlator activities have been outsourced to separate, often competing,

contractors.

One of the equipment suppliers to the leakage industry, Radcom Technologies,

suggests that there is a way to substantially improve the efficiency of the

process. Radcom was one of the companies that pioneered DMA monitoring and network

data logging in the 1980s, and is now keen to extend its role in leakage management

to encompass detection and pinpointing as well.

Simply better

Radcom believes the current method of leak location can be radically improved

and simplified by combining noise logging and leak noise correlation into a

single process.

The system developed for this process, Soundsens, is already in use at many

water companies and is being trialled at several others.

Dr Gurch Chana, technical director and co-founder of Radcom described the system.

“The Soundsens approach is based on the deployment of highly sensitive

and accurate loggers, called ‘pods’. The pods, which are time synchronised,

record sound in short bursts lasting a few seconds. The recording is repeated

multiple times to separate genuine use from suspected leakage.” The duration

of recording, the number of recordings and the dwell time between each recording

is programmable and depends upon the pipe material in question, whether the

test is taking place during the day or at night and whether the pipework being

tested is a trunk or distribution main.

The pods have at their heart an advanced digitally amplified accelerometer,

developed in conjunction with Southampton University, which is capable of detecting

lower levels of sound than can be expected from conventional designs, which

simply correlate sound velocity between two points.

However, the company claims that the key aspect of Soundsens is the way that

the data recorded by the pods is then handled, as Chana continued, “The

data from each pod is transferred to a laptop, where a layout of the pipework

is graphically drawn using either GIS data or as a simple sketch. The layout

enables the software to both link and take into account the spatial relationship

between the pods.”

Once the layout is complete, the software uses proprietary algorithms to amplify

the sound and filter out anomalies. Correlation takes place between each pod

in the array and all of the others. The company says that this process, called

cross-correlation, is unique to the Soundsens system and enables greater certainty

in pin-pointing leaks as it discerns leak noise profiles from other water sounds.

Any leaks are pin-pointed on the pipe layout diagram, tabulated and ranked

in order of probability. The software will also show any correlations that should

be subjected to further investigation. To help with this, the operator can listen

to the sound using the software’s audio feature. The data files can then be

stored for transfer or use at a later date.

The system has already won plaudits. At the IWEX 2001 Innovations Awards, Soundsens

received the highly commended accolade in the measurement and instrumentation

category.

Water companies using Soundsens are reportedly having success with the system

in situations that would normally be difficult for leak noise correlators. In

particular, Soundsens has recently found leaks on trunk mains in the Wessex

Water region and on long length MDPE distribution mains in the Anglian Water

region.

Technology alone cannot answer all of the water industry’s questions about

leakage. However, developments that improve efficiency and save time and money

will help to reduce water loss and go some way to reducing leakage, even if

the remaining questions mean that such a reduction is measured in relative rather

than absolute terms