Detecting Leaks in Water-Distribution Pipes

Leak Detection Methods for Plastic Water Distribution Pipes
Research Project co-funded by:

AWWA Research Foundation and NRC Construction

Leak Finder: NRC's Pipeline Leak Correlator

In many water distribution systems a significant percentage of water is lost while in transit from treatment plants to consumers. According to an inquiry made in 1991 by the International Water Supply Association (IWSA), the amount of lost or "unaccounted for" water is typically in the range of 20 to 30% of production. Unaccounted for water is usually attributed to several causes including leakage, metering errors, and theft – leakage is the major cause. In addition to environmental and economic losses caused by leakage, leaky pipes pose a public health risk as leaks are potential entry points for contaminants if a pressure drop occurs in the system.

Economic pressure, concern over public health risk and simply the need to conserve water motivate water system operators to implement leakage control programs. There are two major steps in any systematic leakage control program: (i) water audits, and (ii) leak detection surveys. Water audits involve detailed accounting of water flow into and out of the distribution system or parts of it. The audits help to identify areas having excessive leakage. Unfortunately, they do not provide information about the location of leaks. To do so, leak detection surveys must be undertaken.

In leak surveys, the water distribution system is checked for leaks by using acoustic equipment which detects the sound or vibration induced by water as it escapes from pipes under pressure. Acoustic equipment include listening devices such listening rods, aquaphones (or sonoscopes), and geophones (or ground microphones). These are used to listen for leak sounds at contact points with the pipe such as fire hydrants and valve. Acoustic equipment also include leak noise correlators. These are modern computer-based instruments that have a simple field setup and work by measuring leak signals (sound or vibration) at two points that bracket a suspected leak. The position of the leak is then determined automatically based on the time shift between the leak signals calculated using the cross-correlation method. Several makes of acoustic leak detection equipment are now commercially available – see below for a list of manufacturers.

PROBLEM STATEMENT

Generally, acoustic leak detection equipment are considered to be satisfactory by most professional operators. This is only the case, however, for metallic pipes. In the case of plastic pipes, the effectiveness of existing equipment is not well established or documented. The equipment was developed mainly with metallic pipes in mind, but the acoustical characteristics of leak signals in plastic and metallic pipes differ significantly. Plastic pipes are "quieter" and do not transmit sound or vibration as efficiently as metallic ones. Problems that are normally encountered with locating leaks with acoustic equipment, e.g., interfering traffic signals, and attenuation of leak signals along pipes, become more detrimental in the case of plastic pipes. Consequently, most operators are skeptical about the effectiveness of acoustic leak detection equipment – a serious problem in view of the increasing use of plastic pipes in water distribution systems worldwide.

RESEARCH OBJECTIVE

The main objective of the research project was to investigate the effectiveness of acoustic leak detection equipment, in particular leak noise correlators, for locating leaks in PVC plastic pipes. Emphasis was placed on evaluating the methods on which the equipment, not on comparing different equipment makes. Objectives of the research also included:

•Survey of leak-detection equipment,
•Characterization of leak signals in plastic pipes,
•Identification of necessary improvements to existing equipment and methods, and
•Evaluation of the potential of technologies from other industries.
RESEARCH APPROACH

The research involved extensive field tests that were carried out under controlled conditions at a specially constructed experimental leak-detection facility at the campus of the National Research Council (NRC) in Ottawa, Canada. Evaluation of commonly used acoustic leak detection equipment was implemented by inviting several experienced leak detection teams from utilities and service companies in Canada and the United States to participate in "blind" leak detection tests. Equipment used by the teams included listening devices and leak noise correlators. The tests involved locating simulated leaks at the experimental site without having prior knowledge about their actual location.

In addition to the blind tests, extensive parametric tests were carried out by the research team. The purpose was to evaluate the effect of several parameters on the accuracy of pin-pointing leaks using the cross-correlation method, and to identify optimum instrumentation and signal processing parameters. The tests were performed using a versatile state-of-the-art vibration measurement and analysis system. Parameters included in the investigation were related to site conditions, instrumentation, and signal processing and analysis. Acoustic characteristics of leak signals were also investigated. These included frequency content, attenuation rate, and variation of propagation velocity with frequency (or dispersion). Leak signals were measured during both winter and summer to evaluate the effect of frozen soil on their acoustic characteristics.

Finally, the potential of locating leaks using alternative non-acoustic technologies was evaluated by inviting experienced users of selected methods to apply them for locating leaks at the NRC site. The potential of the following three methods was evaluated: ground-penetrating radar, thermography, and tracer gas.

EXPERIMENTAL LEAK DETECTION FACILITY

Leak detection tests in this project were carried out at a facility constructed especially for the project in an experimental waterline site at the campus of the National Research Council (NRC) in Ottawa, Canada. The experimental site had an underground PVC test pipe connected to NRC's water distribution network. The pipe is 150 mm (6 in.) in diameter, 200 m (652 ft) in length, and is buried at a depth of 2.4 m (7.87 ft). Soil type at the site is soft silty clay.

The test pipe had several contact or access points where leak sensors could be attached. These included two fire hydrants that are 103 m (338 ft) apart, and six contact points with the pipe in the form of typical 19 mm (¾ in.) copper service connections. Two service connections were located less than 1 m apart across a joint of the test PVC pipe. These were used to measure leak signal attenuation across the joint. In addition to providing contact with the test pipe, service connections were used to simulate interfering noise due to water usage at residential services.

Service connection leaks, a joint leak, and a crack leak were simulated in the test pipe. Each simulated leak can be opened individually and at the desired flow rate by turning appropriate control valves. The area where leaks were created was back-filled with the native clay soil at the site. A manifold consisting of a pressure reducing valve (PRV), a low-flow meter (LFM), a pressure gauge, and a double-check back-flow preventer was installed at the upstream end of the test pipe. Pressure could be set at any level in the range from 139 to 414 kPa (20 to 60 psi). Flow rates ranging from 0.9 to 27 l/min. (0.25 to 7 gpm) could be measured at an accuracy of ± 5 %.

SUMMARY OF FINDINGS

Commercial modern leak noise correlators were generally found to be capable of locating leaks in plastic water distribution pipes. Based on the findings of this study, however, several improvements could be incorporated into existing equipment and field procedures to increase their effectiveness. Improvements for equipment include the revision of automatic mode algorithms, use of higher sensitivity sensors especially in the case of accelerometers, verification of propagation velocities for various pipe types and sizes, procedures to verify proper functioning of sensors, very low-frequency capability of wireless transmission / receiving systems, flexible high and low-pass filter settings (e.g., finer steps and lower limits), optional display of time histories and frequency spectra of leak signals.

On the other hand, improvements of field procedures for locating leaks by correlating leak signals include the use of low-frequency components, on-site measurement of leak signal propagation velocity, verification of proper functioning of sensors, use of hydrophones, and attachment of vibration sensors to pressurized fire hydrants rather than shut-off valves when sufficiently sensitive sensors are available. In the case of the 150 mm (6 in.) PVC test pipe used in this study, the optimum frequency range for correlating leak signals was between 15 and 100 Hz. However, the low-frequency limit may need to be increased or decreased slightly depending on the pipe size and type as well as site conditions.

Finally, initial leak surveys that are normally carried out using listening devices only at access points with distribution pipes may not be effective in detecting leaks due to the high attenuation rate of leak signals in plastic pipes. High resolution surveys using ground microphones may need to be performed instead, but these are time consuming. Thermography and (or) ground-penetrating radar showed promise and could provide efficient tools for initial leak surveys – therefore, it was recommended that their potential be further investigated. The tracer gas method was found to be effective but time-consuming and hence impractical for routine leak locating – however, it could be helpful where other methods fail.

FINAL REPORT

A final report on this project titled "Leak detection methods for plastic water distribution pipes" will be available in Spring 1999 from the AWWA Bookstore [toll-free telephone No. (800)-926-7337]. It is free to AWWA Research Foundation subscribers [telephone No. (303) 347-6121].

RESEARCH TEAM

The project was carried out by the staff of the Urban Infrastructure, Structures, and Acoustics Laboratories of the Institute for Research in Construction at the National Research Council of Canada. Members of the research team were: Osama Hunaidi (project manager and principal investigator) and Wing Chu (co-principal investigator), Alex Wang, Wei Guan, Ted Hoogeveen, Bruce Baldock, and Rock Glazer. Michael Caprara was AWWARF's project manager.

PARTICIPATING ORGANIZATIONS

The cost of the project was shared by the American Water Works Association Research Foundation and the National Research Council of Canada. The Regional Municipality of Ottawa-Carleton and Louisville Water company participated in the project by providing in-kind contributions.

FURTHER INFORMATION

For further information please contact Mr. Michael Caprara, AWWARF, 6666 West Quency Avenue, Denver, CO., 80235-3098, USA. [Tel.: (303) 347-6112, Fax: (303) 730-0851, e-mail: mcaprara@awwarf.com] or Dr. Osama Hunaidi, Institute for Research in Construction, National Research Council of Canada, Ottawa, Canada K1A 0R6 [Tel.: (613) 993-9720, Fax: (613) 952-8102].

TABLES, PHOTOS, AND ILLUSTRATIONS

•Listening devices
•Leak noise correlator
•Schematic of the field setup for the cross-correlation method
•General view of experimental leak detection site
•Contact points with test pipe
•Simulated leaks
•Vibration measurement and analysis system
•Ground-penetrating radar
•Infrared Thermography
•Tracer gas
MANUFACTURERS

•List of manufacturers of acoustic leak detection equipment
PUBLICATIONS

Acoustical characteristics of leak signals in plastic water distribution pipes
Hunaidi, O. Chu, W.T.
Applied Acoustics, 58 (3)
pp. 235-254. 1999-07-01
[Full citation / Référence complète]

Leak detection methods for plastic water distribution pipes
Hunaidi, O. Chu, W.T. Wang, A. Guan, W.
Advancing the Science of Water, Fort Lauderdale Technology Transfer Conference, AWWA Research Foundation (Ft. Lauderdale, Florida, 1999-02-18)
pp. 249-270. 1999-02-18
[Full citation / Référence complète]

Ground-penetrating radar for detection of leaks in buried plastic water distribution pipes
Hunaidi, O. Giamou, P.
Seventh International Conference on Ground Penetrating Radar (GPR'98) (Lawrence, Kansas, 1998-05-27)
pp. 783-786. 1998
[Full citation / Référence complète]

Detecting Leaks in Water-Distribution Pipes
Hunaidi, O.
Construction Technology Update, 40
pp. 6. 2000-10-01
Order / Commander : http://www.nrc- cnrc.gc.ca/virtualstore
[Full citation / Référence complète]

Related Information
DEFAULT VALUE

LeakfinderRT: Pipeline Leak Correlator

Institutes:
•NRC Institute for Research in Construction

Polyethylene is now the preferred material for both new and rehabilitated distribution mains. There has been growing concern, however, surrounding PE pipe joint integrity. This UKWIR project was therefore initiated to quantify the scale of potential leakage problems on existing PE systems and drive improved design and construction methods for installing PE pipe. As electrofusion joints have been found to be significantly more likely to fail than butt fusion and mechanical joints, this is where the focus of the project has been directed. Analysis of failure data and destructive electrofusion joint test records suggest that 20 percent of electrofusion joints will have a life span considerably less than the 50 year design life. Recommendations from the project include: making changes to current standards for PE pipe installation; improved data capture, training and licensing of welders, and testing during installation; and the need to overcome issues associated with coiled pipes.