ACOUSTIC EMISSIONEarly Uses of Acoustic Emission:-The first acoustic emission used by an artisan may well have been in making pottery (the oldest variety of hardfired pottery dates back to 6,500 BC) . In order to assess the quality of their products, potters t raditionally relied on the audible cracking sounds of clay vessels cooling in the kiln. These acoustic emissions were accurate indications that the ceramics were defective and did indeed structurally fail

Acoustic Emission in Metalworking:-(Tin Cry)

It is reasonable to assume that the first observation of acoustic emission in metals was tin cry, the audible emission produced by mechanical twinning of pure tin during plastic deformation. This phenomenon could occur only after man learned to smelt pure tin, since tin is found in nature only in the oxide form. It has been established that smelting (of copper) began in Asia Minor as early as 3,700 BC. · The deliberate use of arsenic and then tin as alloying additions to copper heralde1 the beginning of the Bronze Age somewhere between the fourth and third millennium BC. The oldest piece of pure 'tinfound to date is a bangle excavated at Thermi in Lesbos. The tin has been dated between 2,650 and 2,550 BC. It is 41 mm (1.6 in.) in diameter and . consists of two strands of pure tin , on wrapped around the other and hammered flat at the end. During the manufacture af this bangle, the craftsman could have heard considerable tin cryEarly Documented Observations of Tin Cry:-The first documented observations of acoustic emission may have been made by the eighth century Arabian alchemist Jabir ibn Hayyan (also known as Geber). His book Summa Perfectionis Magisterii (The Sum of Perfection or The Sum of Perfect Magistery) was published in English translation in 1678; the Latin edition was published in Berne in 1545. In it he writes that Jupiter (tin) gives off a "harsh sound" or "crashing noise." He also describes Mars (iron) as "sounding much" during forging. This sounding of iron was most likely produced by the fonnation of martensite during cooling….Since the time of the alchemists, audible emissions have become known and recognized properties of cadmium and zinc as well as tin. Tin cry is commonly found in books on chemistry published in the last half of the nineteenth century. For example, Worthington Hooker in 1882 describes "the cry of tin" as owing to the friction on minute crystals of the metal against each other."DETECTING AND RECORDING ACOUSTIC EMISSION:-The transition from the incidental obsezvation of audible tin cry to the deliberate study of acoustic emission phenomena consisted of three separate and unrelated experiments in which instrumentation was used to detect, amplify and record acoustic emission events occurring in the test specimens. The first experiment instrumented specifically to detect acoustic emission was conducted in Germany and the results were published in 1936 by Friedrich Forster and Erich Scheil. They recorded the "Gerausche" (noises) caused by the formation of martensite in 29 percent nickel steel

In the United States, Warren P. Mason, H.J. McSkimin and W. Shockley performed and published the second instrumented acoustic emission experiment in 1948. At the suggestion of Shockley, experiments were directed toward obsezvation of moving dislocations in pure tin specime

by means of the stress waves they generated. The experiment's instrumentation was capable of measuring displacements of about 10-7 mm

occurring in times of 10-6 seconds. The third instrumented experiment was performed in England by D.J. Millard in 1950 during research for his Ph.D. thesis at the University of Bristol. He conducted twinning experiments on single crystal wires of cadmium. Twinning was detected using a Rochelle salt transducerKaiser's Study of Acoustic Emission Sources:-The early obsezvations of audible sounds and the three instrumented experiments were not directed at a study of the acoustic emission phenomenon itself, nor did the researchers carry on any further investigations in acoustic emission. The genesis of today's technology in acoustic e mission was the work of Joseph Kaiser at the Technische Hochschule Miinchen in Germany.In 1950 Kaiser published his Doktor-Ingenieur dissertation where he reported the first comprehensive investigation into the phenomena of acoustic emission. Kaiser used tensile tests of conventional engineering materials to determine: (1) what noises are generated from within the specimen; (2) the acoustic processes involved;

(3) the frequency levels found; and(4) the relation between the stress-strain curve and the frequencies noted for the various stresses to which the specimens we re subjected

In 1965, however, Robinson used more sensitive equipment to show that acoustic emission occurred at much lower load levels than had been reported earlier, and hence, could be used to monitor earlier microcracking (such as that involved in the growth of bond cracks in the interfacial region between cement and aggregate). In 1970, Wells built a still more sensitive apparatus, with which he could monitor acoustic emissions in the frequency range from about 2 to 20 kHz. However, he was unable to obtain truly reproducible records for the various specimen types that he tested, probably due to the difficulties in eliminating external noise from the testing machine. Also in 1970, Green reported a much more extensive series of tests, recording acoustic emission frequencies up to 100 kHz. Green was the first to show clearly that acoustic emissions from concrete are related to failure processes within the material; using source location techniques, he was also able to determine the locations of defects. It was this work that indicated that acoustic emissions could be used as an early warning of failure. Green also noted the Kaiser effect, which suggested to him that acoustic emission techniques could be used to indicate the previous maximum stress to which the concrete had been subjected. As we will see below, however, a true Kaiser effect appears not to exist for concrete.

Nevertheless, even after this pioneering work, progress in applying acoustic emission techniques remains slow. An extensive review by Diederichs et al. (et al means: and others), covers the literature on acoustic emissions from concrete up to 1983. However, as late as 1976, Malhotra noted that there was little published data in this area, and that “acoustic emission methods are in their infancy.” Even in January, 1988, a thorough computer-aided search of the literature found only some 90 papers dealing with acoustic emissions from concrete over about the previous 10 years; while this is almost certainly not a complete list, it does indicate that there is much work to be carried out before acoustic emission monitoring becomes a common technique for testing concrete. Indeed, there are still no standard test methods which have even been suggested for this purpose

His most significant discovery was the irreversibility phenomenon which now bears his name, the Kaiser effect. He also proposed a distinction between burst and continuous emission. Kaiser concluded that the occurrence of acoustic emission arises from frictional rubbing of grains against each other in the polycrystalline materials he tested and also from intergranular fracture.Kaiser continued his research at the Institut fiir Metallurgie und Metallkunde der Technischen Hochschule Mlinchen until his death in March 1958. His work provided the momentum for continued activities at the Institut by several of his coworkers, including Heinz Borchers and Hans Maria Tensi, and also furnished the impetusfor further research elsewhere in the world.

What is AEAcoustic emission is the technical term for the noise emitted by materials and structures when they are subjected to stress. Types of stresses can be (1) mechanical, (2) thermal or (3) chemical. This emission is caused by the rapid release of energy within a material due to events such as crack initiation and growth, crack opening and closure, dislocation movement, twinning, and phase transformation in monolithic materials and fiber breakage and fiber- matrix debonding in composites The subsequent extension occurring under an applied stress generates transient elastic waves which propagate through the solid to the surface where they can be detected by one or more sensors. The sensor is a transducer that converts the mechanical wave into an electrical signal (piezoelectric) . In this way information about the existence and location (triangulation by multi-transducers) of possible sources is obtained. Acoustic emission may be described as the "sound" emanating from regions of localized deformation within a material Until about 1973, acoustic emission technology was primarily employed in the non-destructive testing of such structures as pipelines, heat exchangers, storage tanks, pressure vessels, and coolant circuits of nuclear reactor plants. However, this technique was soon applied to the detection of defects in rotating equipment bearingsAcoustic Emission:-Acoustic Emission (AE) refers to generation of transient elastic waves during rapid release of energy from localized sources within a material The source of these emissions in metals is closely associated with the dislocation movement accompanying plastic deformation and with the initiation and extension of cracks in a structure under stress. Other sources of AE are: melting, phase transformation, thermal stresses, cool down cracking and stress build up, twinning, fiber breakage and fiber- matrix debonding in composites.

Acoustic fundamentals:-Airborne sound Basic physicsThe human ear perceives sounds chiefly through the medium of the sur- rounding air. A sound source sets the air vibrating, causing a cycle of com- pression and expansion. Superimpo- sed over normal air pressure, these oscillations propagate in the form of waves. Upon reaching the human ear, these sound waves cause our eardrums to vibrate, thus triggering the process of hearing The human ear perceives sounds chiefly through the medium of the sur- rounding air. A sound source sets the air vibrating, causing a cycle of com- pression and expansion. Superimpo- sed over normal air pressure, these oscillations propagate in the form of waves. Upon reaching the human ear, these sound waves cause our eardrums to vibrate, thus triggering the process of hearing.

Fig. 1 shows a compression/expansioncurve which is „higher“ than

F Fig. 2, represents a louder sound. On the other hand, in Fig. 2 the airborne sound pressure vibrationSound field parameters(Sound velocity+Sound pressure+ Sound power)These air vibrations can be measured and physically analyzed in terms of their key variables, referred to as „sound field parameters“. Some of these parameters are described be- low.Sound velocity:-The sound velocity „c“ is the speed at which sound waves travel - about 333 m/s under normal conditions.Sound pressure:-The term „sound pressure“ refers to the alternate compression and expansion of air caused by a sound source. These pressure variations are measured in µbar (microbars).Sound power:-Sound power is a theoretical quantity which cannot be measured. It is calculated and expressed in watts (W). To illustrate the difference between sound pressure and sound power, let us consider the example of a trumpet player. What we hear coming out of the instrument are sound pressure waves that trigger the process of hearing via our eardrums. What we don’t hear is the amount of work done by the play- er to produce the sound, i.e. the „power“ input made by blowing into the mouthpiece. This power is necessary to generate the sound waves (reduced according to the trumpet’s efficiency); it is referred to as sound power or acoustic power.

As we move away from the trumpet player, his music appears to fade, i.e. decrease in loudness. In a room with a strong echo the instrument will sound differently than in a room de- corated with heavy drapery and car- pets. Thus, the sound pressure per- ceived by our ear is dependent on di- stance and space. But regardless of what we hear (i.e. of distance and space conditions), the trumpet player must expend the same amount of energy. In other words, sound power is not dependent on distance and space. This is what makes this para- meter so valuable. As an objective quantity that cannot be influenced, it constitues an excellent starting point for all acoustic calculations.

AE TECHNIQUE:-The AE technique (AET) is based on the detection and conversion of high frequency elastic waves emanating from the source to electrical signals. This is accomplished by directly coupling piezoelectric transducers on the surface of the structure under test and loading the structure. The output of the piezoelectric sensors (during stimulus) is amplified through a low-noise preamplifier, filtered to remove any extraneous noise and further processed by suitable electronics. AET can non-destructively predict early failure of structures. Further, a whole structure can be monitored from a few locations and while the structure is in operation. AET is widely used in industries for detection of faults or leakage in pressure vessels, tanks, and piping systems and also for on-line monitoring welding and corrosion.

NDTNondestructive testing (NDT) is the process of inspecting, testing, or evaluating materials, components or assemblies for discontinuities, or differences in characteristics without destroying the serviceability of the part or system. In other words, when the inspection or test is completed the part can still be used.

In contrast to NDT, other tests are destructive in nature and are therefore done on a limited number of samples ("lot sampling"), rather than on the materials, components or assemblies actually being put into service.

These destructive tests are often used to determine the physical properties of materials such as impact resistance, ductility, yield and ultimate tensile strength, fracture toughness and fatigue strength, but discontinuities and differences in material characteristics are more effectively found by NDT.

Today modern nondestructive tests are used in manufacturing, fabrication and in-service inspections to ensure product integrity and reliability, to control manufacturing processes, lower production costs and to maintain a uniform quality level. During construction, NDT is used to ensure the quality of materials and joining processes during the fabrication and erection phases, and in-service NDT inspections are used to ensure that the products in use continue to have the integrity necessary to ensure their usefulness and the safety of the public.

It should be noted that while the medical field uses many of the same processes, the term "nondestructive testing" is generally not used to describe medical applications.

NDT Test Methods

Test method names often refer to the type of penetrating medium or the equipment used to perform that test. Current NDT methods are: Acoustic Emission Testing (AE), Electromagnetic Testing (ET), Guided Wave Testing (GW), Ground Penetrating Radar (GPR), Laser Testing Methods (LM), Leak Testing (LT), Magnetic Flux Leakage (MFL), Microwave Testing, Liquid Penetrant Testing (PT), Magnetic Particle Testing (MT), Neutron Radiographic Testing (NR), Radiographic Testing (RT), Thermal/Infrared Testing (IR), Ultrasonic Testing (UT), Vibration Analysis (VA) and Visual Testing (VT).

The six most frequently used test methods are MT, PT, RT, UT, ET and VT. Each of these test methods will be described here, followed by the other, less often used test methods.

EXPLAIN SOME OF THEM:-

Radiographic Testing (RT):-

industrial radiography involves exposing a test object to penetrating radiation so that the radiation passes through the object being inspected and a recording medium placed against the opposite side of that object.  For thinner or less dense materials such as aluminum, electrically generated x-radiation (X-rays) are commonly used, and for thicker or denser materials, gamma radiation is generally used.

Gamma radiation is given off by decaying radioactive materials, with the two most commonly used sources of gamma radiation being Iridium-192 (Ir-192) and Cobalt-60 (Co-60).   IR-192 is generally used for steel up to 2-1/2 - 3 inches, depending on the Curie strength of the source, and Co-60 is usually used for thicker materials due to its greater penetrating ability. 

The recording media can be industrial x-ray film or one of several types of digital radiation detectors.  With both, the radiation passing through the test object exposes the media, causing an end effect of having darker areas where more radiation has passed through the part and lighter areas where less radiation has penetrated.  If there is a void or defect in the part, more radiation passes through, causing a darker image on the film or detector, as shown in Figure 8.

RT Techniques

Film RadiographyFilm radiography uses a film made up of a thin transparent plastic coated with a fine layer of silver bromide on one or both sides of the plastic.  When exposed to radiation these crystals undergo a reaction that allows them, when developed, to convert to black metallic silver.  That silver is then "fixed" to the plastic during the developing process, and when dried, becomes a finished radiographic film.

To be a usable film, the area of interest (weld area, etc.) on the film must be within a certain density (darkness) range and must show enough contrast and sensitivity so that discontinuities of interest can be seen.  These items are a function of the strength of the radiation, the distance of the source from the film and the thickness of the part being inspected.  If any of these parameters are not met, another exposure ("shot") must be made for that area of the part.

Magnetic Particle Testing (MT):-

Magnetic Particle Testing uses one or more magnetic fields to locate surface and near-surface discontinuities in ferromagnetic materials.  The magnetic field can be applied with a permanent magnet or an electromagnet.  When using an electromagnet, the field is present only when the current is being applied.  When the magnetic field encounters a discontinuity transverse to the direction of the magnetic field, the flux lines produce a magnetic flux leakage field of their own as shown in Figure 1.  Because magnetic flux lines don't travel well in air, when very fine colored ferromagnetic particles ("magnetic particles") are applied to the surface of the part the particles will be drawn into the discontinuity, reducing the air gap and producing a visible indication on the surface of the part.  The magnetic particles may be a dry powder or suspended in a liquid solution,

and they may be colored with a visible dye or a fluorescent dye that fluoresces under an ultraviolet ("black") light.

YOKEMost field inspections are performed using a Yoke, as shown at the right.  As shown in Figure 2(a), an electric coil is wrapped around a central core, and when the current is applied, a magnetic field is generated that extends from the core down through the articulated legs into the part.  This is known as longitudinal magnetization because the magnetic flux lines run from one leg to the other. 

When the legs are placed on a ferromagnetic part and the yoke is energized, a magnetic field is introduced into the part as shown in (b).  Because the flux lines do run from one leg to the other, discontinuities oriented perpendicular to a line drawn between the legs can be found.  To ensure no indications are missed, the yoke is used once in the position shown then used again with the yoke turned 90o so no indications are missed.  Because all of the electric current is contained in the yoke and only the magnetic field penetrates the part, this type of application is known as indirect induction.

Prods:-

Prod units use direct induction, where the current runs through the part and a circular magnetic field is generated around the legs as shown in Figure 3.  Because the magnetic field between the prods is travelling perpendicular to a line drawn between the prods, indications oriented parallel to a line drawn between the prods can be found.  As with the yoke, two inspections are done, the second with the prods oriented 90o to the first application.

Coils:-

Electric coils are used to generate a longitudinal magnetic field.  When energized, the current creates a magnetic field around the wires making up the coil so that the resulting flux lines are oriented through the coil as shown at the right.  Because of the longitudinal field, indications in parts placed in a coil are oriented transverse to the longitudinal field.

Heads:-

Most horizontal wet bath machines ("bench units") have both a coil and a set of heads through which electric current can be passed, generating a magnetic field.  Most use fluorescent magnetic particles in a liquid solution, hence the name "wet bath."   A typical bench unit is shown at the right.  When testing a part between the heads, the part is placed between the heads, the moveable head is moved up so that the part being tested is held tightly between the heads, the part is wetted down with the bath solution containing the magnetic particles and the current is applied while the particle are flowing over the part.  Since the current flow is from head to head and the magnetic field is oriented 90o to the current, indications oriented parallel to a line between the heads will be visible.  This type of inspection is commonly called a "head shot.

Liquid Penetrant Testing (PT):-The basic principle of liquid penetrant testing is that when a very low viscosity (highly fluid) liquid (the penetrant) is applied to the surface of a part, it will penetrate into fissures and voids open to the surface. Once the excess penetrant is removed, the penetrant trapped in those voids will flow back out, creating an indication. Penetrant testing can be performed on magnetic and non-magnetic materials, but does not work well on porous materials. Penetrants may be "visible", meaning they can be seen in ambient light, or fluorescent, requiring the use of a "black" light. The visible dye penetrant process is shown in Figure 7. When performing a PT inspection, it is imperative that the surface being tested is clean and free of any foreign materials or liquids that might block the penetrant from entering voids or fissures open to the surface of the part. After applying the penetrant, it is permitted to sit on the surface for a specified period of time (the "penetrant dwell time"), then the part is carefully cleaned to remove excess penetrant from the surface. When removing the penetrant, the operator must be careful not to remove any penetrant that has flowed into voids. A light coating of developer is then be applied to the surface and given time ("developer dwell time") to allow the penetrant from any voids or fissures to seep up into the developer, creating a visible indication. Following the prescribed developer dwell time, the part is inspected visually, with the aid of a black light for fluorescent penetrants. Most developers are fine-grained, white talcum-like powders that provide a color contrast to the penetrant being used.

PT Techniques:-

Solvent RemovableSolvent Removable penetrants are those penetrants that require a solvent other than water to remove the excess penetrant.  These penetrants are usually visible in nature, commonly dyed a bright red color that will contrast well against a white developer.  The penetrant is usually sprayed or brushed onto the part, then after the penetrant dwell time has expired, the part is cleaned with a cloth dampened with penetrant cleaner after which the developer is applied.  Following the developer dwell time the part is examined to detect any penetrant bleed-out showing through the developer.

Water-washableWater-washable penetrants have an emulsifier included in the penetrant that allows the penetrant to be removed using a water spray.  They are most often applied by dipping the part in a penetrant tank, but the penetrant may be applied to large parts by spraying or brushing.  Once the part is fully covered with penetrant, the part is placed on a drain board for the penetrant dwell time, then taken to a rinse station where it is washed with a course water spray to remove the excess penetrant.  Once the excess penetrant has been removed, the part may be placed in a warm air dryer or in front of a gentle fan until the water has been removed.  The part can then be placed in a dry developer tank and coated with developer, or allowed to sit for the remaining dwell time then inspected.

Post-emulsifiablePost-emulsifiable penetrants are penetrants that do not have an emulsifier included in its chemical make-up like water-washable penetrants.  Post-emulsifiable penetrants are applied in a similar manner, but prior to the water-washing step, emulsifier is applied to the surface for a prescribed period of time (emulsifier dwell) to remove the excess penetrant.  When the emulsifier dwell time has elapsed, the part is subjected to the same water wash and developing process used for water-washable penetrants.  Emulsifiers can be lipophilic (oil-based) or hydrophilic (water-based).

Difference between AET and other non-destructive testing (NDT):-The difference between AET and other non-destructive testing (NDT) techniques is that AET detects activities inside materials, while other techniques attempt to examine the internal structures of materials by sending and receiving some form of energyAcoustic Emission Nondestructive Testing:-Acoustic emission examination is a rapidly maturing nondestructive testing method with demonstrated capabilities for monitoring structural integrity, detecting leaks and incipient failures in mechanical equipment, and for characterizing materials behavior. The first documented application of acoustic emission to an engineering structure was published in 1964 and allof the available industrial application experience has been accumulated in the comparatively short time since then.Comparison with Other Techniques:-Acoustic emission differs from most other nondestructive methods in two significant respects. First, the energy that is detected is released from within the test object rather than being supplied by the nondestructive method, as in ultrasonics or radiography. Second, the acoustic emission method is capable of detecting the dynamic processes associated with the degradation of structural integrity. Crack growth and plastic deformation are major sources of acoustic emission. Latent discontinuities that enlarge under load and are active sources of acoustic emission by virtue of their size, location or orientation are also the most likely to be significant in terms of structural integrity.

Usually, certain areas within a structural system will develop local instabilities long before the structure fails. These instabilities result in minute dynamic movements such as plastic deformation, slip or crack initiation and propagation. Although the stresses in a metal part may be well below the elastic design limit; the region near a crack tip may undergo plastic deformation as a result of high local stresses. In this situation, thepropagating discontinuity acts as a source of stress waves and becomes an active acoustic emission source .

Charlie Chong/ Fion Zhang

Acoustic emission examination is non-directional. Most acoustic emission sources appear to function as point source emitters that radiate energy in spherical wavefronts. Often, a sensor located anywhere in the vicinity of an acoustic emission source can detect the resulting acoustic emission.

This is in contrast to other methods of nondestructive testing, which depend on prior knowledge of the probable location and orientation of a discontinuity- in order to direct a beam of energy through the structure on a path that willproperly intersect the area of interest.

Advantages of Acoustic Emission Tests over other nondestructive testing methods:-1.- Acoustic emission is a dynamic inspection method in that it provides a response to discontinuity growth under an imposed structural stress; Static discontinuities will not generate acoustic emission signals.

2. -Acoustic emission can detect and evaluate the significance of discontinuities throughout an entire structure during a single test.

3. Since only limited access is required, discontinuities may be detected thatare inaccessible to the more traditional nondestructive methods.

4. - Vessels and other pressure systems can often be requalified during an in- service inspection that requires little or no downtime.

5.-The acoustic emission method may be used to prevent catastrophic failure of systems with unknown discontinuities, and to limit the maximum pressure during containment system tests

Types of AETAcoustic emissions are broadly classified into two major types namely:- continuous type (associated with lattice dislocation) burst type. (twinning, micro yielding, development of crackThe waveform of continuous type AE signal is similar to Gaussian random noise, but the amplitude varies with acoustic emission activity. In metals and alloys, this form of emission is considered to be associated with the motion of dislocations. Burst type emissions are short duration pulses and are associated with discrete release of high amplitude strain energy. In metals, the burst type emissions are generated by twinning, micro yielding, development of cracks.

What is Normal (Gaussian) distribution:-In probability theory, the normal (or Gaussian) distribution is a very common continuous probability distribution. Normal distributions are important in statistics and are often used in the natural and social sciences to represent real-valued random variables whose distributions are not known he normal distribution is remarkably useful because of the central limit theorem. In its most general form, under mild conditions, it states that averages of random variables independently drawn from independent distributions are normally distributed. Physical quantities that are expected to be the sum of many independent processes (such as measurement errors) often have distributions that are nearly normal.[3] Moreover, many results and methods (such as propagation of uncertainty and least squares parameter fitting) can be derived analytically in explicit form when the relevant variables are normally distributed

Charlie Chong/ Fion Zhang

The normal distribution is sometimes informally called the bell curve. However, many other distributions are bell-shaped (such as Cauchy's, Student's, and logistic). The terms Gaussian function and Gaussian bell curve are also ambiguous because they sometimes refer to multiples of the normal distribution that cannot be directly interpreted in terms of probabilities

Schematic diagram of a basic four-channel acoustic

The Main Elements Of A Modern Acoustic Emission Detection System:-

Charlie Chong/ Fion Zhang

Application of Acoustic Emission Tests:-

1- mechanical property testing and characterization;

2- . pre-service proof testing;

3- in-service (requalification) testing;

4- on-line monitoring;

5- in-process weld monitoring;

6- mechanical signature analysis;

7- leak detection and location; and

Successful Applications:-

1.-Periodic or continuous monitoring of pressure vessels and other pressure containment systems to detect and locate active discontinuities.

2.-Detection of incipient fatigue failures in aerospace and other engineering structures.

3.- Monitoring materiais behavior tests to characterize various failuremechanisms.

4.-Monitoring fusion or resistance weldments during welding or during the cooling period.

5.-Monitoring acoustic emission response during stress corrosion cracking and hydrogen embrittlement susceptibility tests.

1

8. geological applications

Acoustic Methods for Locating Leaks inMunicipal Water Pipe Networks

The recovery of water loss from leaks in transmission and distribution pipes can provide a solution, at least partially, to water shortages caused by insufficient water resources and / or limited water treatment capacity. This paper introduces a new, low-cost and easy-to-use system that will help water utilities to dramatically improve the effectiveness of locating leaks in all types of pipes, including traditionally difficult plastic pipes. The system has promising potential for all water utilities, including small and medium-sized ones and utilities in developing countries.

Introdution

Water utilities in many parts of the world are facing growing challenges in their attempts to meet the demand for drinking water. For example, in the United States more than 36 states expect to experience water shortages over the next 10 years (EPA, 2003). Another major example is China where more than 400 of its 600 large and medium sized cities suffer from water shortages, with at least 100 cities, including Beijing, seriously threatened (Tai, 2004). Several factors are contributing to this situation. Climate change, manifested by extended periods of drought, is adversely impacting water resources. Population growth caused by migration to large urban centers and temperate regions is exerting increasing pressure on existing water supplies. The problem is compounded if water treatment infrastructure is operating near capacity and funds required for expansion are scarce. Limited treatment capacity can create drinking water shortages even in water-rich regions. In the face of these challenges, the recovery of water loss from leaks in transmission and distribution pipes can provide a solution, at least partially, to shortages caused by insufficient water resources and / or limited treatment capacity.

Water transmission and distribution networks deteriorate naturally with time and subsequently loose their initial water tightness. Causes of the deterioration include corrosive environments, soil movement, poor construction standards, fluctuation of water pressure, and excessive traffic loads and vibration. Water is lost due to leakage in different components of the networks that include transmission pipes, distribution pipes, service connection pipes, joints, valves, fire hydrants, and storage tanks and reservoirs. In addition to the physical losses due to leakage, many networks suffer from so called apparent losses. These are caused by under- registration of customer meters, accounting errors, and unauthorized water use.

The amount of water loss is typically between 20 to 30% of production, with leakage being the main component (Cheong, 1991). For some distribution systems, the

loss can be in excess

2

of 50% (AWWA, 1987). In addition to their help in meeting water demand, detection and repair of pipe leaks help to minimize water quality breaches that may result from the entry of contaminants via leaks. They also help to reduce the high cost of energy wasted on the treatment and pumping of leaking water (Colombo & Karney, 2002). The energy-wasting aspect of leakage is important as significant savings can be realized. Energy to supply water is the second largest cost after labour for water systems in developed countries, and the cost may easily consume 50% of a municipality’s budget in the developing world (James et al., 2002).

Acoustic equipment is commonly used to locate leaks in municipal water pipes. These include noise loggers, simple listening devices, and leak noise correlators. The economic viability and leak detection effectiveness of noise loggers is questionable. Also, they are not suitable for pinpointing leaks. The effectiveness of listening devices such as ground microphones for pinpointing leaks depends greatly on the experience of the user and the process is time- consuming. Leak noise correlators are more efficient, yield more accurate results and depend less on user experience than do listening devices. Existing correlators, however, require extensive training and can be unreliable for quiet leaks in cast and ductile iron pipes and for most leaks in plastic and large-diameter pipes. Correlators are also expensive and remain beyond the means of many water utilities and leak detection service companies.

A new leak noise correlation system is introduced in this paper. The system, called

LeakfinderRTTM, incorporates several new developments, most importantly an enhanced correlation method that dramatically improves the effectiveness of locating leaks in all types of pipes, including traditionally difficult plastic pipes (see Hunaidi et al. (2000) regarding plastic pipe difficulties). The system uses personal computers as a platform, which eliminates the need for a major component of the usual hardware of correlators, dramatically reducing the system’s cost. Also, the use of

Microsoft WindowsTM makes LeakfinderRT very easy to use, as most potential users are already familiar with Windows.

Features, capabilities, and performance examples of the new LeakfinderRT leak noise correlation system are presented. The paper also includes a brief overview of leakage management, a review of acoustic and other leak detection techniques, a discussion of the factors influencing the effectiveness of acoustic techniques and best practices for pinpointing leaks in plastic pipes.

Overview of Leakage Management

Management of leakage comprises four main components: (i) quantifying the total water loss, (ii) monitoring of leakage, (iii) locating and repairing leaks, and (iv) pipe pressure management.

Quantifying the total amount of lost water is achieved by conducting a system-wide water audit, known internationally as a water balance. Guidelines for conducting water audits have been published by the American Water Works Association (AWWA, 1999) and by the International Water Association (IWA, 2000).

3

A move to harmonise the AWWA and IWA guidelines is underway (WLCC, 2003). Like financial audits that account for all the debits and credits of a business, water audits account for all water flowing into and out of a utility’s water delivery system. An audit can be performed over an arbitrary period of time, but normally it is computed annually over a period of 12 months. Audits provide a valuable overall picture about various components of consumption and loss, which is necessary for assessing a utility’s efficiency regarding water delivery, finances, and maintenance operations. Also, water audits are necessary for planning other leakage management practices.

4

Monitoring of leakage involves dividing the distribution system into well-defined areas that each can be supplied through a single pipe where a flow meter capable of measuring low flow rates is installed. These areas are known as district meter areas or DMAs. The boundaries of DMAs can sometimes occur naturally but generally they have to be created by the closing of appropriate valves. Guidelines for setting up, maintenance and leakage monitoring of DMAs have been published by a research consortium of water companies in the United Kingdom (UKWIR, 1999). The size of a typical DMA can be between 500 and 3000 properties. Leakage in DMAs is monitored by measuring the minimum night flow rate monthly or quarterly or on continual basis if flow meters are connected telemetrically to a SCADA system. Leakage is suspected if the minimum night flow rate is greater than a previously measured level or if it exceeds a certain threshold. The latter is determined as the sum of the flow rate of water used by all night-time commercial and industrial users in the district, flow rate of water used by all residential properties based on average night flow rate per property, and unavoidable leakage rate. DMAs make it possible to quickly and efficiently identify areas of the pipe network that suffer from excessive leakage, which are then targeted for leakage detection and localization operations. Analysis of minimum night flow rates can also be used to refine (or check) the accuracy of water audits.

The exact positions of leaks are commonly pinpointed by using ground microphones and leak noise correlators and possibly by using non-acoustic methods such as thermography, ground- penetrating radar, and tracer gas (Hunaidi et al., 2000). Pinpointing of leaks can be time consuming and therefore leak detection surveys are normally undertaken prior to pinpointing to narrow down the area of the leak to a pipe section(s). Step testing can identify pipe sections with leaks in a DMA. This involves the monitoring of the district meter’s flow rate while successively closing valves within the DMA, starting with the valve that is farthest away from the meter. A significant reduction in the flow rate is an indication of leakage in the last shut- off section. Step testing has to be performed at night and can be time consuming and dangerous. In recent years, its use has dwindled in favour of acoustic surveys using noise loggers, acoustic listening tools, or leak noise correlators. Acoustic and non-acoustic leak detection and pinpointing equipment are described in the next section.

Pipe pressure affects leakage in a number of ways (TWGWW, 1980) and a substantial reduction in leakage can be realised by pressure management. The lower the pressure, the lower the pipe break frequency. Also, pressure transients can fracture pipes and damage their joints. Frequent pressure fluctuations may cause fatigue failure in pipes, especially plastic ones. Most importantly, the higher the pressure, the higher the leakage rate. Theoretically, the flow rate of a fluid through an opening is proportional to the square root of the pressure differential across the opening, provided that the dimensions of the opening remain fixed. However, pipe leak openings may enlarge with pressure. Therefore, much greater reductions in leakage can be realised than predicted by the square root relationship, especially for small leaks from joints and fittings in most pipe types and large leaks in plastic pipes (Lambert,2001). A linear relationship between pipe pressure and leakage level is widely used by leakage management practitioners.

5

Leak Detection Equipment

Acoustic Equipment

Listening devices. These devices include listening rods and ground microphones and may be either mechanical or electronic. They use sensitive mechanisms or materials such as piezoelectric elements to sense leak-induced sound or vibration. Modern electronic listening

6

devices incorporate signal amplifiers and noise filters to make the leak signal stand out. Leak inspectors conducting leak surveys work their way around the pipe network systematically and use listening rods at appropriate pipe fittings to detect the characteristic hissing sound created by leaking water. The leak detection effectiveness of listening surveys depends on the size of leaks, ambient noise from road traffic and water draw, and the degree of detail of the survey. General surveys, performed by listening at only convenient fittings such as fire hydrants and / or valves, mainly detect large leaks. On the other hand, detailed surveys conducted by listening at all pipes fittings, including curb-stops (or stop-taps), can detect small leaks. Ground microphones are used to pinpoint leaks by listening for leak noise at the ground surface directly above pipes at small intervals. This process is time consuming and its success depends on the experience of the user.

Noise loggers. These are compact units composed of a vibration sensor (or hydrophone) and a programmable data logger. They are used to leak survey large areas but they are not suitable for pinpointing leaks. Loggers are deployed in groups of 6 or more at adjacent pipe fittings, e.g., fire hydrants and valves 200 to 500 m apart, and left there overnight. The units are normally programmed to collect pipe noise data between 2 and 4 AM. The loggers are collected the next day and the stored data are downloaded to a personal computer before the loggers are deployed at the next location. The logged data are analysed statistically, e.g., frequency analysis of leak noise levels, to detect the presence of leaks. Recent models of acoustic noise loggers can be deployed permanently – leak noise is processed using onboard electronics and the stored result is transmitted wirelessly to a roaming receiver. The economic viability and leak detection effectiveness of temporarily or permanently deployed noise loggers is questionable. van der Klejj and Stephenson (2002) found that both permanently and temporarily deployed loggers are not an economical alternative to skilled and well-equipped leak inspectors. For network-wide coverage, permanent loggers had a minimum payback period of 25 years. When the loggers were used in temporary mode, i.e., moved from one survey area to the next, they were three times less efficient than acoustic surveys. van der Klejj and Stephenson (2002) also report that the number of leaks found by noise loggers and by general listening surveys were similar; however, the loggers failed to detect approximately40% of leaks found in detailed listening surveys. Acoustic loggers can be advantageous over listening surveys in instances where the latter cannot be undertaken during daytime due to high ambient noise.

Leak noise correlators. These are portable microprocessor-based devices that can be used in either leak survey or pinpointing modes. They are based on the cross-correlation method, which involves the measurement of leak noise (either sound or vibration) at two locations on a pipe section. Measured noise is transmitted wirelessly to the correlator, which then determines the position of the leak based on the time shift of the maximum correlation of the two leak signals, propagation velocity of leak noise, and the distance between sensing points. The distance between sensors can be read from distribution system maps when the correlator is used in survey mode but it should be measured onsite accurately when it is used in pinpointing mode. Propagation velocities for various pipe types and sizes are programmed in most correlators, but they should be measured onsite to improve pinpointing accuracy,

7

especially for non-metallic pipes. Leak noise correlators are more efficient, yield more accurate results and are less dependent on user experience than listening devices. However, existing equipment requires extensive training and can be unreliable for quiet leaks in cast and ductile iron pipes and for most leaks in plastic and large diameter pipes. Correlators are also expensive and remain beyond the means of many water utilities and leak detection service companies. The new correlator presented in this paper overcomes these shortcomings.

8

Non-acoustic Equipment

Tracer gas technique. In this technique, a non-toxic, water-insoluble and lighter-than-air gas, such as helium or hydrogen, is injected into an isolated section of a water pipe. The gas escapes at a leak opening and then, being lighter than air, permeates to the surface through the soil and pavement. The leak is located by scanning the ground surface directly above the pipe with a highly sensitive gas detector.

Thermography. The principle behind the use of thermography for leak detection is that water leaking from an underground pipe changes the thermal characteristics of the adjacent soil, for example, making it a more effective heat sink than the surrounding dry soil. Thermal anomalies above pipes are detected with a ground or air-deployed infrared camera.

Ground-penetrating radar. Radar can be used to locate leaks in buried water pipes either by detecting voids in the soil created by leaking water as it circulates near the pipe, or by detecting sections of pipe which appear deeper than they truly are because of the increase in the dielectric constant of water-saturated adjacent soil. Ground-penetrating radar waves are partially reflected back to the ground surface when they encounter an anomaly in dielectric properties, for example, a void or pipe. An image of the size and shape of the object is formed by radar time-traces obtained by scanning the ground surface. The time lag between transmitted and reflected radar waves determines the depth of the reflecting object.

Factors Influencing the Effectiveness of Acoustic Equipment

The effectiveness of acoustic leak-detection equipment depends on several factors including pipe size, type, and depth; soil type and water table level; leak type and size; pipe pressure; interfering noise; and sensitivity and frequency response of the equipment.

Pipe material and diameter have a significant effect on the attenuation of leak signals in the pipe. For example, leak signals travel farthest in metal pipes and are attenuated greatly in plastic ones (see Hunaidi & Chu (1999) for acoustical characteristics of plastic pipe leaks). The larger the diameter of the pipe, the greater the attenuation, and the harder it is to detect the leak. Pipe material and diameter also affect the predominant frequencies of leak signals the larger the diameter and the less rigid the pipe material, the lower the predominant frequencies. This effect makes leak signals susceptible to interference from low-frequency vibrations, for example, from pumps and road traffic.

Soil type and the water table level influence the strength of leak signals at the ground surface significantly. Leak sounds are more audible on sandy soils than on clayey ones, and on an asphalt or concrete surface than on grass. Leak signals are muffled if the pipe is below the water table level.

Acoustical characteristics of leak sounds vary with leak type and size. Splits and corrosion pits in pipe walls may induce stronger leak signals and higher frequencies than leaking joints and valves. Generally, the larger the leak, the louder the leak signal, but this may not be true for very large leaks. The higher the pipe pressure, the louder the leak signals.

The more sensitive and quieter the leak sensors, and the higher the signal-to-noise ratio of the signal conditioning and recording equipment, the smaller the leaks that can be detected.

9

Modern acoustic equipment incorporates signal-conditioning components such as filters and amplifiers to make leak signals stand out. Filters remove interfering noise occurring outside

10

the predominant frequency range of leak signals. Amplifiers improve the signal-to-noise ratio and make weak leak signals audible. If the frequency response of the equipment does not extend to sufficiently low frequencies, it can miss leaks in plastic and large-diameter pipes.

The LeakfinderRT System

Overview

LeakfinderRT is a new system for locating leaks in municipal water distribution and transmission pipes using the cross-correlation method. This leak noise correlation system is fully realized in software for personal computers (PCs) running under Microsoft Windows XP, 2000, or 9x. The system utilizes the PC’s soundcard and other multimedia components to record and playback leak signals. It also uses the PC’s central processing unit (CPU) to perform the cross-correlation operation and associated signal conditioning. Modern PCs incorporate fast CPUs and high-resolution soundcards and hence offer several advantages over existing commercial hardware implementation of the cross-correlation method. Hardware components of the LeakfinderRT system are shown in Figure 1 and are composed of leak sensors, wireless signal transmission system, and a PC. The menu-driven user interface of the system’s software is shown in Figure 2. The software can be installed on either a notebook PC or a desktop one having a soundcard with a stereo line-in port.

The use of PCs eliminates the need for a major component of the usual hardware of leak noise correlators. Most potential users of LeakfinderRT already have PCs, which reduces the cost of the system. The use of Microsoft Windows makes the process of using the fully menu-driven system a simple and intuitive one as most potential users are already familiar with Windows.

LeakfinderRT incorporates several new developments, most importantly an enhanced correlation function. For narrow-band leak noise, this new function dramatically improves the definition of correlation peaks. This is important for plastic pipes, multiple-leak situations, and in settings where leak sensors have to be closely spaced. Also, the enhanced correlation function is more effective than the traditional correlation function for small leaks and for situations of high background noise.

Computer records and correlates leak noise Wireless receiver picks up

broadcasted leak noise Wireless transmitters broadcast leak noise sensed by leak sensors

Hydrophones (optional)sense leak-induced sound

11

Accelerometers sense leak-induced vibration

Figure 1 Hardware components of LeakfinderRT

12

Figure 2 Main interface and input windows of LeakfinderRT

Correlation of Leak Noise

Locating leaks in water distribution and transmission pipes is a classical application of the cross-correlation method. Two things make this possible. First, the propagation velocity of leak sounds in water pipes is nearly constant over the dominant frequency range of leak sounds. Second, water-filled pipes transmit leak signals for long distances. Therefore, the shape of leak signals does not change significantly as they travel away from the leak, which is a pre-requisite for a successful correlation.

A typical field set-up of the correlation technique is shown in Figure 3. The correlation function of leak noise signals measured at the two points that bracket the location of a suspected leak provides information about the time delay (or lag) between the two signals. The time delay between the two leak signals is the result of one measurement point being closer to the leak location than the other. If the two measurement points are symmetrically positioned about the leak location, leak signals will arrive simultaneously at the two points and the time delay will be zero. On the other hand, if the leak location is exactly at the position of one of the two measurement points (or equivalently it is not between the two points), the time shift will be equal to the distance between the measurement points divided by the propagation velocity of leak noise in the pipe.

The correlation magnitude of two leak noise signals is the summation of their product as a function of time shift. In simple terms, the correlation value at time shift is computed by first shifting one of the signals by relative to the other signal. Then the two signals are multiplied, point-by-point, and the products are summed. The correlation function will display a peak at the time shift, which corresponds to the actual delay between the two leak noise signals (this is the time at which the two signals overlap).

LeakfinderRT determines the time delay max corresponding to the peak of the cross- correlation automatically. In reference to Figure 3, the time delay between the two leak noise signals is related to the location of the leak relative to measurement points by

max L 2 L1

cwhere L1 and L2 are the positions of the leak relative to sensors 1 and 2, respectively, and c is the propagation velocity of the leak sound in the pipe. By substituting L2 = D – L1 in the above equation, the position of the leak relative to point 1 is found as

13

L D c max

1 2

where D is the distance between the sensors, either measured on site or read off system maps. The propagation velocity can be specified if it was measured onsite or it can be calculated theoretically by LeakfinderRT based on input for pipe material type and diameter.

In the presence of material and geometric discontinuities, e.g., joints of dissimilar pipes and sharp bends, the cross-correlation method may lead to false results. This is because leak signals are partially or completely reflected at the discontinuities. Reflected leak signals create spurious peaks in the cross-correlation function which can be mistakenly interpreted as real leaks. For example, the presence of an out-of-bracket reflector at distance LR from the closest leak sensor may create a spurious cross-correlation peak at distance LR from the actual position of the leak. An in-bracket reflector may create a spurious peak at the location of the reflector itself.

If there is more than one leak between sensor positions 1 and 2, the cross-correlation function will have a peak corresponding to each leak. However, if the leaks are closely spaced, the peaks will overlap and in turn distort the corresponding time delay. The peak width depends on the bandwidth of the leak noise; the wider the frequency bandwidth of leak signals, the narrower the cross-correlation peak. The frequency bandwidth of leaks in metal pipes is much wider than that of leaks in plastic ones. For metal pipes, it may be possible to resolve leaks that are 6 m apart; for plastic pipes it may not be possible to accurately resolve leaks that are less than 20 m apart.

LeakfinderRT uses an enhanced cross-correlation function which is calculated indirectly in the frequency domain using the inverse Fourier transform of the cross-spectral density function instead of the usual shift-and-multiply method in the time domain. For narrow-band leak signals, the enhanced correlation function provides improved resolution, i.e., better definition of peaks, in comparison with the traditional correlation function. This is helpful for plastic pipes, for situations of multiple leaks, and for settings where leak sensors are closely spaced. Also, the enhanced correlation function is more effective than the traditional function for small leaks or for situations of high background noise. Performance examples are presented further on to demonstrate these advantages for leaks in both plastic and metal pipes. The enhanced correlation function does not require the usual filtering of leak signals for removing interfering noise. This is a major advantage over the traditional correlation method because it eliminates the uncertainty involved in selecting filter cutoff frequencies.

Measurement of Leak Noise

Leak signals are measured using either vibration sensors or hydrophones. Accelerometers, which sense the acceleration of vibration induced by leak signals in the pipe wall or fittings, are normally used to measure leak signals in metal pipes. These sensors are attached to the pipe directly, if accessible; if not they can be attached to fire hydrants or underground valves. A magnetic base is fitted to the base of these sensors and hence they are easy to install on metallic surfaces. Accelerometers can also be used to measure leaks in plastic pipes but they are only effective for large leaks.

LeakfinderRT’s special low-frequency vibration sensors and hydrophones are significantly more effective than accelerometers. Hydrophones, which are underwater microphones that sense pressure perturbations (sound) induced by leak noise in the water core of the pipe, are installed at fire hydrants or air release valves by using special fittings. Installation of

14

Computer

Receiver

Pipe

SuspectedLeak

Figure 3 Schematic illustration of the field setup of the cross-correlation technique for locating a suspected leak

hydrophones is more difficult and more time-consuming than that of accelerometers. However, they are more effective for measuring leak signals in difficult situations, e.g., in the case of weak signals from small leaks, at sites with high background vibration levels, and for large sensor-to-sensor distances. Also, hydrophones are more effective than accelerometers for low-frequency leak signals, which are found in plastic pipes and large-diameter ones (plastic or otherwise).

LeakfinderRT’s low-frequency vibration sensors are almost as effective as hydrophones for plastic pipes but they are much easier to install because they are attached to the pipe directly or to fully pressurized fire hydrants. For example, the smallest service connection leaks in a152-mm PVC pipe, under 60 psi pressure, correlated over a distance of about 100 m that could be successfully located using low-frequency vibration sensors and hydrophones were1.7 and 0.85 liters per minute (see Figures 4c and 4d), respectively. In comparison, the smallest leak that could be located with accelerometers under the same conditions was 20 liters per minute. A joint leak having a flow rate of 6 and 3.25 liters per minute in the same pipe, under 60 and 20 psi pressure, respectively, was also successfully located with both low- frequency vibration sensors and hydrophones about 100 m apart (see Figure 4a and 4b for correlation of vibration sensor-measured signals).

The shorter the sensor-to-sensor spacing the better. For plastic pipes, it may not be possible to locate leaks with spacing greater than 100 m with the low-frequency vibration sensors. For larger distances, the use of hydrophones is recommended. For metal pipes, sensor-to-sensor spacing can be as large as 500 m but a maximum spacing of 200 m is recommended.

Signals from leak sensors are transmitted wirelessly to the PC for processing by using radio frequency transmitters and receivers. The received signals are fed into the line-in port of the PC’s soundcard. LeakfinderRT’s wireless system transmits leak signals almost without distortion – for input levels at 200 micro volts, its flat frequency response starts to roll down and its phase starts to become nonlinear at about 10 Hz (-6 dB point is at 5 Hz). It also has a low noise floor; e.g., it is capable of transmitting leak signals as small as 0.5 micro volts.

15

Recording of Leak Noise

LeakfinderRT records leak sounds by using the soundcard of the PC. The soundcard is set automatically to stereo mode (i.e., dual channel input) and to 16-bit resolution for maximum accuracy. Signals from the wireless receiver’s line-out should be fed into the STEREO line-in port of the PC’s soundcard. The signals are then converted by the soundcard to digital samples at the rate of 11025 Hz (i.e., samples per second).

By default, leak sounds are recorded and correlated by LeakfinderRT for an indefinite duration. The cross-correlation results are displayed on screen and continuously updated in real time while leak signals are being recorded. This default mode is convenient for situations in which leak signals are weak and may need to be correlated for a long time to average out interfering noise. The correlation operation can be terminated at any time. Normally, in the presence of a leak, a duration of 30 to 60 seconds is sufficient to obtain a definite peak in the cross-correlation function, but a much longer duration may be needed for small leaks and / or noisy environments. Recorded signals can be saved to the PC’s hard drive. Leak signals are saved in standard PCM sound file format and can be later converted to ASCII format for exporting to other applications, if needed.

Before recording and correlating leak signals, the recording volume must be adjusted to utilise as much as possible of the soundcard’s voltage range, without overloading it. This helps to achieve a high signal-to-noise ratio. LeakfinderRT has a signal preview function to help check the level or volume of leak sounds (signals are not saved to disk during previewing). The level of leak noise signals is displayed via two level meters (see Figure 2). The recording volume must be adjusted so that the signal level does not fall in the red range of the level meters, just like recording music on a home stereo audio system.

At the start of the correlation operation the user is prompted to input only the following three parameters:

– pipe type,– pipe diameter, and– sensor-to-sensor spacing.

The software sets all other correlation and signal processing parameters automatically. The user can opt to input pipe wall thickness (in addition to the diameter), propagation velocity if known (instead of diameter), or none of the preceding (pipe type only). For multi-type pipes, properties can be input for different pipe sections.

Playback of Leak Noise

A file containing leak sounds saved by LeakfinderRT on the PC’s hard disk can be played back. The software does not alter the raw data in a saved leak sounds file and hence the data can be played back and re-correlated repeatedly. Also, the complete time history of saved leak signals can be displayed graphically on screen and printed.

The speed at which leak signals are played back can be increased arbitrarily. This is helpful when playing back noise signals of leaks in plastic and large-diameter pipes. In these pipes, leak signals are dominated by low-frequency components, mainly in the infrasound range, and thus cannot be heard by an unaided human ear. Speeding up the playback of low-frequency signals shifts their frequency components to a higher range at which the sensitivity of human hearing is high enough to discern the leak sound.

16

Propagation Velocity of Leak Noise

LeakfinderRT calculates the propagation velocity of leak noise using a theoretical formula based on input for pipe type, diameter and wall thickness. If the wall thickness is not input, the software estimates it based on pipe type and diameter. If the diameter is also unknown, the propagation velocity is calculated as the average of velocities corresponding to several diameters. It should be emphasised that velocity values calculated theoretically are approximate and are provided only for preliminary leak location. Leak positions based on these approximate values can be inaccurate, especially for non-metallic pipes. For improved accuracy, it is recommended that the propagation velocity be measured onsite using a known in-bracket or out-of-bracket simulated leak, for example, by drawing water at a fire hydrant or service connection. An in-bracket leak produces signals that are more similar than those produced by an out-of-bracket one and is preferred. LeakfinderRT has a built-in velocity calculator that facilitates this operation.

Output

Frequency spectra of leak signals, coherence, and correlation functions are output graphically by LeakfinderRT. Frequency spectra provide information about the frequency content of leak signals. The coherence function provides a measure of the relationship between recorded leak signals, i.e., whether they were induced by the same source or not. The closer the coherence function is to 1, the more related the signals. Small coherence values indicate high noise-to- signal ratio (e.g., in the case of leaks creating weak noise). The correlation function provides information about the time delay between leak noise signals, which in turn is used to calculate the leak location.

Any of the above functions can be displayed individually in the main interface window where they are updated in real time while the recording and correlation of leak signals is in progress. They can also be displayed together in a separate window at the end of the recording and correlation process. The graphs have a zoom button and a live cursor. For the correlation function, the software re-calculates the leak position corresponding to any time shift of any peak or point in the correlation function. The output also includes the time shift corresponding to the peak of the correlation function, and the leak position. The output can be sent to a Windows default printer, copied to a Windows clipboard for pasting in other applications, or exported to an Excel spreadsheet.

Performance Examples

LeakfinderRT was verified extensively using pure tones, random signals, and field tests on in- service water distribution pipes. Performance examples of LeakfinderRT’s enhanced correlation method versus the traditional method are shown in Figures 4 and 5 for plastic and metal pipes, respectively. These results demonstrate the following advantages and capabilities of the LeakfinderRT system utilizing the enhanced correlation method:

– Figure 4a: Improved peak definition for narrow-band leak noise in PVC pipes.– Figure 4b: Superior sensitivity for detecting small leaks in PVC pipes under a

very low pressure of 20 psi (~15 m).– Figures 4c and 4d: Smallest detectable PVC pipe leaks using LeakfinderRT’s low

frequency vibration sensors (1.7 liters per minute) and using hydrophones (0.85 liters per minute).

– Figure 5a: Improved peak definition for resolving multiple leaks.– Figure 5b and 5c: Effectiveness for locating small leaks in metal pipes.– Figure 5c and 5d: Effectiveness for situations of high background noise.

17

Conclusions

A new system for locating pipe leaks based on the cross-correlation method is introduced. The system incorporates several new developments, most importantly an enhanced correlation method. For narrow-band leak noise, this new function dramatically improves the definition of correlation peaks. This is important for plastic pipes, multiple-leak situations, and in settings where leak sensors have to be closely spaced. Also, the enhanced correlation function is more effective than the traditional one for small leaks or for situations of high background noise. The enhanced function does not require the usual filtering of leak signals for removing interfering noise. This is a major advantage because it eliminates the uncertainty involved in selecting filter cutoff frequencies.

Another major advantage of the new system is the use of low-frequency vibration sensors instead of the inconvenient use of hydrophones to locate leaks in traditionally difficult plastic pipes. The effectiveness of these low-frequency vibration sensors is demonstrated for locating small service leaks and joint leaks in PVC pipes, even when under low pipe pressures.

In addition to the effectiveness of the new system, it is very easy to use and low in cost. This is a result of using personal computers as a platform to both record and analyse leak signals, which eliminates a major component of the usual hardware of leak noise correlators. Consequently, the new system has promising potential for all water utilities, including small and medium-sized ones and utilities in developing countries.

References

AWWA (1999). Water Audits and Leak Detection. Manual of Water Supply Practices M36, American Water

Works Association, Denver, CO.AWWA (1987). Leaks in Water Distribution Systems – A Technical/Economic Overview. American Water

Works Association, Denver, CO.Cheong, L.C. (1991). Unaccounted-for Water and the Economics of Leak Detection. Proc. International Water

Supply Congress and Exhibition, Copenhagen, published in Water Supply, Volume 9, pp. IR 1-1 to 1-6. Colombo, A.F., and Karney, B.W. (2002). Energy and Costs of Leaky Pipes: Toward a Comprehensive Picture.

Journal of Water Resources Planning and Management, Volume 128, No. 6, pp. 441-450.EPA (2003). EPA Will Help Consumers Locate and Purchase Water-Efficient Products. Newsroom,

Environmental Protection Agency, U.S.A., posted at “www.epa.gov/newsroom/headline_090503.htm”.

Hunaidi, O., Chu, W., Wang., A., and Guan, W. (2000). Detecting Leaks in Plastic Water Distribution Pipes.

Journal AWWA, Volume 92, No. 2, pp. 82-94, posted at: “irc.nrc.gc.ca/fulltext/nrcc43058.pdf”.Hunaidi, O., and Chu, W. (1999). Acoustical Characteristics of Leak Signals in Plastic Water Distribution Pipes.

Applied Acoustics, Volume 53, pp. 235-254, posted at: “irc.nrc.gc.ca/fulltext/nrcc42673.pdf”.IWA (2000). Manual of Best Practice: Performance Indicators for Water Supply Services. International Water

Association, London, U.K.James, K., Godlove, C.E., and Campbel, S.L. (2002). Watergy: Taking Advantage of Untapped Energy and

Water Efficiency Opportunities in Municipal Water Systems. Alliance to Save Energy, Washington, DC, posted at: “www.watergy.org”.

Lambert, A. (2001). What Do We Know About Pressure / Leakage Relationships in Distribution Systems? Proc.

IWA Specialised Conference: System Approach to Leakage Control and Water Distribution SystemsManagement, Brno, Czech Republic, 16-18 May 2001, pp. 89-96.

Tai, Y. (2004). China Faces Serious Water Shortages. The Washington Times, 23 March 2004, posted at

“www.washtimes.com/upi-breaking/20040322-121420-7382r.htm”.

18

TWGWW (1980). Leakage Control Policy and Practice. Report No. 26 by Technical Working Group on Waste of Water, National Water Council, Department of the Environment, U.K.

UKWIR (1999). Manual of DMA Practice. Published by UK Water Industry Research Limited, London. U.K. van der Klejj, F.C., and Stephenson, M.J. (2002). Acoustic Logging – The Bristol Water Experience. Proc. IWA

Specialised Conference: Leakage Management – A Practical Approach, International Water Association,20-22 November 2002, Lemesos, Cyprus, posted at “www.leakage2002.com”.

WLCC (2003). Applying Worldwide BMPs in Water Loss Control (Report by AWWA Water Loss Control

Committee). Journal AWWA, Volume 95, No. 8, pp. 65-79.

Enhanced correlation Traditional correlation

19

(a) 6 liters / minute joint leak in 152-mm PVC pipe (sensor type: low-frequency vibration sensor, sensor spacing: 102.6 m, pipe pressure: 70 psi, 02.010703)

Enhanced correlation Traditional correlation

(b) 3.25 liters / minute joint leak in 6-inch PVC pipe (sensor type: low-frequency vibration sensor, sensor spacing: 102.6 m, pipe pressure: 20 psi, 03.010703)

Enhanced correlation Traditional correlation

(c) 1.7 liters / minute service connection leak in 152-mm PVC pipe (sensor type: low-frequency vibration sensor, sensor spacing: 102.6 m, pipe pressure: 60 psi, 06.200703)

Enhanced correlation Traditional correlation

(d) 0.85 l liters minute service connection leak in 152-mm PVC pipe (sensor type: hydrophone, sensor spacing: 102.6 m, pipe pressure: 60 psi, 11.270603)

Figure 4 Performance of LeakfinderRT’s enhanced correlation method versus traditional method for PVC pipe leaks

Enhanced correlation Traditional correlation

20

(a) Multiple leaks in 100-year old, 100-mm cast iron pipe (sensor type: accelerometer, sensor spacing: 199 m, pipe pressure: unknown, 01.040301)

Enhanced correlation Traditional correlation

(b) 1 liter / minute fire hydrant leak in152-mm ductile iron pipe (sensor type: accelerometer, sensor spacing: 128.5 m, pipe pressure: 60 psi, 14.200703)

Enhanced correlation Traditional correlation

(c) 2 liters / minute fire hydrant leak in152-mm ductile iron pipe (sensor type: accelerometer, sensor spacing: 134.5 m, pipe pressure: 60 psi, 17.090303)

Enhanced correlationTraditional correlation

(d) 25 liters / minute fire hydrant leak in152-mm ductile iron pipe (sensor type: accelerometer, sensor spacing: 128.5 m, pipe pressure: 60 psi, 15.200703)

Figure 5 Performance of LeakfinderRT’s enhanced correlation method versus traditional method for metal pipe leaks