See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/228706704

An evaluation of acoustic reflectometry for

leakage and blockage detection

ARTICLE in ARCHIVE PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART C JOURNAL OF

MECHANICAL ENGINEERING SCIENCE 1989-1996 (VOLS 203-210) · JUNE 2008
Impact Factor: 0.56 · DOI: 10.1243/09544062JMES873

CITATIONS READS

12 90

7 AUTHORS, INCLUDING:

Barry Lennox Xuejing Wang

The University of Manchester Harbin Institute of Technology
102 PUBLICATIONS 867 CITATIONS 8 PUBLICATIONS 33 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Barry Lennox

Retrieved on: 05 January 2016
 Detection of Leakage in a Pipeline using Acoustic Waves

Kassandra M. Papadopoulou, David Glover, Barry Lennox, David MacKay*,

M. Nouar Shamout, Antony R. Taylor and John J. Turner

University of Manchester, Manchester, UK

*
BP Exploration, UK

Abstract

This paper describes the results obtained from an experimental study into the use of acoustic

signals to detect and locate leakages and blockages in pipelines and pipeline networks. The

paper discusses the physical basis of the technique, describes the experimental investigation

and presents a series of results. The conclusion is that this is an effective method for leakage

and blockage detection in simple pipelines and that it can be extended to monitor the integrity

of more complex pipeline networks. A further advantage is that the method appears to be

equally effective when applied to steel or plastic pipes of varying inside diameters. Hence the

technique appears to be effective in both small and large-scale pipelines and pipeline

networks and lends itself to the remote monitoring system.

1 Introduction

Pipeline systems are an essential and ubiquitous feature of modern society, being employed

for the delivery of utilities such as drinking water, petrochemicals and other fluid substances.

The prevention of leakage has never been of greater importance because of the economic and

environmental consequences of even a small loss but, despite this, disastrous events involving

pipeline leakages are reported with monotonous frequency. In the United States, the Office of

Pipeline Safety records incidents involving pipeline leakages centrally and, for 2004 alone,

reported the loss of 3.5 million gallons of hazardous material, five deaths and US$76 million

worth of property damage (http://ops.dot.gov). The inadequacy of existing methods for

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
1
detecting leakages in pipelines is nicely illustrated by the recent pipeline disaster in the Kenai

National Wildlife Refuge in Alaska. There, the release of 200,000 gallons of contaminated

water and crude oil into an area of great environmental sensitivity went undetected until its

accidental discovery by a recreational snowmobiler [1]. In the most recent example, that of

the newly constructed oil pipeline from Baku through Georgia and then to the Turkish port of

Ceyhan in the Mediterranean, it is reported that horsemen will patrol daily in Georgia and

Azerbaijan to counteract terrorist activities [2].

Monitoring the health of a pipeline or, more particularly, detecting leakage, blockage and

corrosion, is not only important in the transportation of hazardous materials but also in the

monitoring of water distribution systems. The problems posed in the distribution of fresh

water are highlighted by recent studies indicating that thirty-six states in the USA are likely to

suffer water shortages in the next ten years [3]. Similarly, in China, two-thirds of the largest

towns and cities currently have water supply problems [4]. An inquiry by the International

Water Supply Association in 1991 concluded that leakage accounted for between 20 and 30

percent of total water usage [5]. More recent surveys indicate that this figure has not

improved: for example, a study in Saudi Arabia estimated that approximately one-third of the

available water is lost through leakage [6]. The only sensible way to alleviate this problem is

to minimise water losses from pipeline distribution systems.

Many methods for detecting leakages in liquid and gas pipelines have been proposed. These

include, but are not limited to, volume balance, inverse transient analysis, acoustic detection

methods, thermographic, radar and tracer gas techniques. The simplest of these approaches is

probably the compensated volume balance method, which involves measurement and

comparison of the flow rate into and out of a pipeline. If the difference between these

measurements exceeds a threshold value then leakage is assumed. Large leakages can be

detected within one to two hours and leakages as small as 0.5% of the nominal pipeline flow

rate can be detected within three to six hours using this approach [7]. Unfortunately, the

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
2
method requires precise measurements of the flow rate and these, at best, will only be

recorded at the inlet and exit of the pipeline. Thus, the exact location of the leakage cannot be

determined: the literature suggests, furthermore, that the method is insensitive for small

leakages [8].

An alternative to volume balancing is to construct a real time model (RTM) of the pipeline.

The RTM then provides an estimate of what the pressure and flow rate should be at various

points in the pipeline or network. Any deviation from these estimates then suggests that

leakage has occurred. Unfortunately, the simulation model must be very sophisticated and a

detailed knowledge of the structure of the pipeline or pipeline network is necessary. Such

knowledge is often unavailable, particularly in old pipeline systems. An alternative to the

application of RTM models, which are developed from a theoretical analysis of the flow

characteristics within the pipeline, is to use data-based (or data-fusion) techniques to construct

the model. In this approach, measurement data collected from the pipeline is used to generate

a predictive model, such as a non-linear state observer, which describes the relationships

between the various sensors located around the pipeline. Discrepancies between the actual

sensor measurements and those estimated by this model provide an indication of possible

leakages in the pipeline [8]. Despite the relative maturity of this field, its practical

implementation is severely restricted by the limited amount of instrumentation typically

available along pipelines and pipeline networks.

More recently, acoustic devices have been used as a method for detecting leakages within

pipelines. The principal acoustic devices are referred to as listening devices and noise

correlators. Listening devices detect leakages by sensing the acoustic energy generated by the

escaping jet of fluid. Devices such as microphones can be fixed relative to the pipework, or

the detector may be moved, as is the case with ‘pigs’, which travel down the inside of the

pipe. In contrast, noise correlators are portable computer-based systems that measure the

leakage-induced noise at two locations along the pipeline. The correlation between these two

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
3
time varying signals yields the location of the leakage. The disadvantages of most if not all of

these acoustic approaches is that the equipment often requires a highly skilled operator and,

under normal conditions, the techniques perform poorly because background noise from

process equipment and traffic interferes with the accuracy of the devices [9].

In this paper a novel approach to detecting leakage and blockage in single pipelines and

pipeline networks is described and demonstrated. The approach is based on the observation

that the propagation of acoustic waves in a fluid medium is very sensitive to any discontinuity

in its properties. Discontinuities will occur whenever there is a change in the static pressure

level of the fluid, or a variation in the cross-sectional area of the pipe, such as may occur at a

flange, T-piece, orifice plate, valve, a deposition of wax, or a pipe wall failure leading to

leakage. In physical terms, this behaviour simply indicates that very small changes in the

acoustic impedance can be linked to the physical and geometrical properties of the acoustic

domain. This principle was exploited in [10] to identify the location of small leaks in a wind

instrument by analysing the propagation of acoustic signals within the instrument.

In some respects, there are similarities between the acoustic wave technique proposed in this

paper and those methods that have been developed in the last decade to exploit the transient

flow characteristics of fluids within pipelines. These methods, which include inverse transient

analysis [11], fluid transient damping [12] and frequency response analysis [13], monitor the

properties of the fluid in the pipeline when it is subjected to a step or impulsive pressure

change. This pressure change is typically introduced by the rapid opening and closing of a

valve and the leakage detection mechanism involves a comparison of measurements recorded

by the pipeline instrumentation with what might be expected from a theoretical analysis of the

pipeline layout and fluid properties. Despite some success, this technique requires there to be

valves located at appropriate positions in the pipeline and for these valves to be manipulated

when required.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
4
The following section of this paper describes the theory of the proposed acoustic pipeline

monitoring system. This is followed in section 3 with a series of results that illustrate the

ability of the technique to detect and locate leakages and blockage in steel and plastic pipe of

varying lengths and diameters. Finally a list of conclusions is provided and details of on-

going and future work plans presented.

2 Theory

2.1 Background

The basis for the leakage detection method described in this paper is the observation that the

propagation of acoustic waves in a fluid medium is very sensitive to any discontinuity in the

properties of the fluid.

To illustrate this, figure 1 shows an expansion in the diameter of a pipeline containing a

stagnant fluid. If an acoustic wave is injected from the left into the pipe then it will be

partially reflected at the interface, producing reflected and transmitted acoustic components.

For weak plane waves of the type considered here, the waves propagate at the local speed of

sound, which will vary depending up on the local conditions of the fluid. Reflective waves

will occur wherever there is a change in the cross-sectional area of the pipe [10]. In industrial

pipeline systems this will occur wherever there is a valve, ‘T’ piece or blockage, for example.

Further to this, it was reported in [10] that any leakage within a pipe would act like a change

in the cross-sectional area and hence a fraction of the incident acoustic energy would be

reflected. This was demonstrated this principle, to a limited extent, in [10] by using reflected

signals to identify small leaks in a musical wind instrument.

This paper describes an acoustic leakage detection method that exploits the fact that any

leakage present within it will reflect acoustic waves. A detailed discussion of the theory

behind the proposed approach is available in [14]. This work is not reproduced here as the

purpose of this paper is to focus on the application of the proposed technique.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
5
 e d E nergy
mitt
Trans
y
o ta l Energ
T
nergy
ef l e cted E
R

Figure 1: Acoustic transmission at pipe joint

3 Experimental Results

In the experimental results presented in this paper, the signal generated by an acoustic pulse

generator was passed through an amplifier to drive a loudspeaker, which transmitted the

signal into a pipe. A microphone was then used to measure the transmission and reflection of

this wave through the pipe. Prior to there being any leakage in the pipe, the wave will be

reflected from every discontinuity in the pipeline, for example where there is a valve or

orifice plate. If similar acoustic waves are transmitted in to the pipe then the measurement

made by the microphone should remain unchanged. If however, the signal recorded by the

microphone does differ then this will indicate that there is a new discontinuity in the pipe,

which may have been caused by a leakage or blockage.

The ideal frequency component of the acoustic signal injected in to the pipe is the subject of

current research. However, in this work a short period square pulse has been used. This pulse is

translated through the speaker and microphone such that the wave recorded by the microphone

is shown in figure 2.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
6
Figure 2: Acoustic signal recorded in the pipe

3.1 Effect of Leakage

To illustrate the basic application of the proposed technique a length of PVCu pipeline, with

an internal diameter of 150mm was used. The basic configuration of the pipeline is illustrated

in figure 3.

Figure 3: Experimental configuration with leakage

In this pipeline system, a loudspeaker is connected to one end of a straight, open-ended pipe

of length 39.84m. A microphone mounted in a tapping in the pipe wall was located 6.11m

from the loudspeaker. An acoustic wave was injected from the loudspeaker and the reflection

response of the pipe was measured using the microphone. The measurement from the

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
7
microphone is recorded for a sufficient length of time for the acoustic wave reflected at the

end of the pipe to return to the microphone. The microphone was sampled at a rate of 50 kHz.

This measurement is referred to here as the reference signature of the pipeline. A leakage was

then introduced into the pipeline. This leakage was simulated by drilling a circular hole of

diameter 25mm in the pipe wall. The leakage was located 22.91m from the loud speaker.

With the leakage present, a second acoustic wave with the same amplitude and frequency

content used earlier was injected in to the pipeline. The measurement made by the

microphone with the leakage present, referred to as the leakage signature, was then recorded

and compared with the original reference signature. Figure 4 shows a time trace of the two

signatures measured by the microphone.

Figure 4: Comparison of reference and leakage signatures

The acoustic waves were transmitted into the pipeline at time 0s and as figure 4 shows the

two traces are very similar. The first peak which occurs in the traces at approximately 0.018s

represents the passage of the acoustic wave across the microphone. Note that the speed of

sound in air is approximately 340m/s. Hence it takes approximately 0.018s for the acoustic

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
8
wave to travel the 6.11m between the loudspeaker and microphone. The subsequent peaks

recorded in the traces are the reflected signals produced whenever the acoustic wave

encounters a discontinuity within the pipeline. The peaks labelled A, B, C, D and E result

from the following features in the pipeline:

A. This peak is caused, in part, by the acoustic wave being reflected from the first T-

section along the pipeline. A second component of this peak is a secondary reflection.

This secondary reflection is caused by the original signal being reflected first off the

microphone. This reflected signal is then reflected off the loud speaker and this

secondary reflection is then recorded by the microphone. These two waves will have

travelled 18.33m and 19.55m respectively.

B. This peak is the result of the acoustic wave being reflected between the microphone

and loud speaker twice. In other words the acoustic wave travels speaker →

microphone → speaker → microphone → speaker → microphone. This is a total

distance of approximately 30.55m.

C. This peak is the microphone measurement of the signal reflected from the second T-

section. This wave has travelled 34.39m.

D. This peak is a result of the acoustic wave being reflected at the bend in the pipeline.

There is a change in the internal cross-sectional area of the pipeline at the bend and

this change produces a reflection. This wave has travelled 53.89m.

E. This peak is caused by the acoustic wave being reflected from the end of the pipeline.

The distance travelled by the wave is 73.57m.

Careful examination of figure 4 shows that the reference and leakage traces are virtually

identical until approximately 0.12s where the two traces begin to differ slightly. This

difference is highlighted in the top right plot in figure 4 which shows an expanded section of

the trace. This difference in the traces is a result of the leakage that has been introduced in to

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
9
the pipeline. To highlight this difference further, figure 5 shows a trace which is the

difference between the reference and leakage traces, this is referred to as the difference trace.

Note that the x-axis has now been converted to a distance measurement. This is the distance

from the loudspeaker and is obtained by multiplying the time value in figure 4 by the speed of

sound. This figure shows that the first significant difference in the observed trace, which is

caused by the leakage, occurs after 39.68m. After this point the two signatures are clearly

different. This result means that in travelling from the speaker to the leakage and then back to

the microphone, the acoustic wave has travelled a distance of 39.68m. Given that the distance

between the speaker and microphone is 6.11m then the distance between the microphone and

leakage is calculated to be:

39.68 − 6.11
d leakage = = 16.79m
2

The actual distance between the leakage and the microphone is 16.80m. Hence the proposed

technique is able to detect the existence of leakage and locate this with some accuracy.

In this example, the speed of sound has been assumed to be constant at 340m/s. This value

will change slightly as the operating conditions vary. Section 3.4 explains how the system can

be configured to compensate for this.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
10
 0.7

0.5

0.3

0.1
Voltage (V)

-0.1

-0.3

-0.5

-0.7

-0.9
0 20 40 60 80 100
Distance (m)

Figure 5: Difference Trace

3.2 Effect of Leakage Size

Experiments were performed to determine the effect that the size of the leakage had on the

acoustic response of the pipeline. In these experiments the equipment described in section 3.1

was used again. The only difference in the arrangement was that on this occasion the location

of the leakage was moved to a point 11.55m from the speaker, or 5.44m from the microphone.

In these experiments, a series of acoustic signatures were recorded when a leakage of varying

size was introduced into the pipeline. The leakages were of diameter 5mm, 10mm, 15mm,

22mm and 29mm. Figure 6 shows the difference signatures recorded for each of these leakage

sizes. This figure illustrates that as the size of the leakage increases then the amplitude of the

difference signature also increases. The difference signature can therefore not only be used to

detect and locate the presence of leakage, but can also be used to identify the relative size of

the leakage. Note that the difference signal does not change significantly until approximately

0.05s. Assuming the speed of sound to be equal to 340m/s then the acoustic wave that has

been reflected by the leakage has travelled 17.0m before it reaches the microphone. This

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
11
 ⎛ 17.0 − 6.11 ⎞
corresponds to a leakage located 5.45m ⎜ ⎟ away from the microphone. This
⎝ 2 ⎠

compares extremely well with the actual distance of 5.44m.

Figure 6: Difference trace for leakages of varying size

3.3 Application to Steel Pipe

In this experiment the ability of the technique to detect and locate the presence of leakage in a

steel pipe was investigated. The pipeline used was 40m in length and had an inside diameter

of 305mm. The first stage in this experiment involved the measurement of the speed of sound

within the pipeline. In the earlier experiments, the speed of sound was assumed to be equal to

340m/s. However, it is known that this figure may vary as the operating conditions and the

pipeline material change. For improved accuracy, it is therefore necessary to determine the

speed of sound before the acoustic technique is applied.

The speed of sound was determined by measuring the time it took for an acoustic wave to

travel between two microphones situated along the pipeline. In this experiment, the distance

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
12
between the microphones was 14.19m. For this test an acoustic wave was introduced in to the

pipeline whilst the two analogue signals produced by the two microphones were sampled at a

frequency of 50kHz. The results from this test indicated that the speed of sound within the

pipeline was 341.1m/s. In contrast, the standard gas laws suggested that the speed of sound

under the observed conditions of 18OC and prevailing atmospheric pressure should be

341.94m/s. These results, therefore, indicate that the calculated speed of sound was in error

by approximately 0.24%. Although, in this test it was relatively easy to calculate the speed of

sound in the pipeline using the standard gas laws, in many operating pipelines the exact

composition of the pipeline material and fluid contents may be unknown. The experimental

approach applied here offers an accurate method to determine the average value of this speed

along the pipe length between the two sensors.

Following the identification of the speed of sound, the method employed to detect the

presence of leakage in section 3.1 and 3.2 was again applied. Figure 7 shows a typical result

obtained from this equipment. In this test, a leakage hole of diameter 20mm was located

16.3m away from the microphone. The difference trace presented in figure 7 indicates that the

leakage is located 16.9m away from the microphone, representing an error of 0.6m in the

predicted distance.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
13
 0.6

0.4
Voltage (V)

0.2

-0.2

-0.4

-0.6
0 20 40 60 80 100

Distance (m)

Figure 7: Difference trace with steel pipe

The results presented in sections 3.1, 3.2 and 3.3 confirm that the method can be applied for

detecting and locating leakages in both steel and PVCu pipeline. In the next section the same

acoustic technique is shown to be capable of detecting and locating partial blockages within

more complicated pipeline network systems.

3.4 Experimental Results for Blockage

In this experiment a small moulded piece of dental plaster was inserted in to a straight length

of PVCu pipeline of 150mm internal diameter. The plaster insert created a partial blockage

that corresponded to approximately 20% of the total pipeline cross-sectional area. The

difference between the acoustic traces recorded before and after the insertion of the partial

blockage is presented in figure 8. As with the leakage examples, the location of the blockage

can be clearly identified as the point on the trace at which the signal begins to deviate

significantly from zero. This point corresponds to a total distance of approximately 20.86m.

This implies that the distance from the speaker to the blockage and back to the microphone is

20.86m. With the microphone located 6.11m away from the loud speaker, this indicates that

the distance between the microphone and blockage is approximately 7.4m ⎛ 20.86 − 6.11 ⎞ .
⎜ ⎟
The
⎝ 2 ⎠

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
14
actual distance between the partial blockage and the microphone was 7.30m, and hence the

error is 0.1m.

0.7

0.2
Voltage (V)

-0.3

-0.8

-1.3
0 20 40 60 80 100
Distance (m)

Figure 8: Difference trace for partial blockage

3.6 Results for pipeline networks

In this section, the applicability of the technique to the more difficult but practically important

case of pipeline networks is discussed. Once again, the working principle of the method is to

inject an acoustic impulsive wave into the pipeline. The reflection response of the network is

then detected by one or more microphones. A series of experiments were carried out to assess

the ability of this technique to detect and locate leakage within relatively complex networks.

As with the previous experiments, these tests were carried out under static conditions with the

pipe network filled with air. The configuration for the pipeline network is provided in figure

9. The pipeline material for these tests was PVCu of 44 mm internal diameter.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
15
 C

Leak 1
mic2

mic1

Figure 9: A schematic drawing of the pipe network

The network used in this experiment contained two independent loudspeakers and two

microphones, labelled mic1 and mic2. An acoustic wave was injected into the pipeline

network using the two loudspeakers in turn and the resulting acoustic trace was recorded from

each of the microphones. The network contained a leakage at the point marked leak 1, which

was a distance of 1.90m from mic1. The size of the hole was approximately 10mm in

diameter. The reason for using two loudspeakers and microphones is so that the system is able

to identify the precise location of the leakage, as will now be illustrated.

The first step in the experiment was to record the reference trace of the pipe network. This

was achieved by injecting an acoustic pulse in each loudspeaker in turn and recording the

signal measured by mic1 and mic2. The leakage was then introduced in to the network and the

signals measured by mic1 and mic2 were again recorded. The difference trace recorded by

mic1 is provided in figure 10. This signal shows that there is no difference between the two

signals until approximately 0.025 seconds. From knowledge of the speed of sound and the

distance between the speaker and microphone, it is possible to identify the location of the

leakage as being approximately 1.92m from mic1. This means that referring back to figure 9,

the signal recorded from mic1 has identified the leakage as being at either point A or B, both

these locations being a distance of 1.92m from mic1. By analysing the signal recorded by

mic2 it is then possible to identify the distance between the leakage and that microphone. In

this experiment this indicated a leakage at point A or C. Hence the location of the leakage is

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
16
isolated to point A, as this is the only location in the network to be consistent with the results

from mic1 and mic2. Thus the experiment has identified the location of the leakage to within

2cm of its actual location.

1
0.8
0.6
0.4
0.2
Voltage (V)

0
-0.2
-0.4
-0.6
-0.8
-1
0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

Time (s)

Figure 10: Difference trace for Pipeline Network

4 Conclusions

The primary conclusions from the work which has been completed so far are as follows:

• The method has been shown to work well in both steel and PVCu pipelines whereas more

traditional acoustic techniques become unreliable in plastic pipelines (Hunaidi et al, 2000).

This is an important issue since many gas and oil pipelines now have an internal plastic

liner.

• The approach also provides an indication of the relative size of any leakage since the

proportion of the incident acoustic energy which is reflected depends on the change in

impedance (and this is a function of leakage area). Tests indicate that the proposed

approach is capable of identifying leakages as small as 1% of the pipeline diameter.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
17
• Initial results indicate that the approach is unaffected by ambient noise conditions.

Moreover, in situations where background noise does exist, the frequency content of the

acoustic wave can be adjusted to reduce the interference.

• The approach has been demonstrated to work in pipelines filled with stagnant and flowing

fluid, and experiments using both liquid and gas have been performed. It is important to

observe, here, that traditional acoustic techniques cannot usually be applied with stagnant

fluid conditions since there must be a leakage flow to generate the noise signal. The

advantage of the present method is that it is applicable under stagnant pipeline conditions,

with or without a leakage flow: for this reason, it can be used to detect wall failures (with

the potential for leakage) in pipelines where there may only be an occasional fluid flow.

• A particularly strong advantage of the method is that it enables blockage to be detected as

readily as leakage. As has been demonstrated, a full or partial blockage produces a change

in the cross-sectional area of the pipe and, hence, a change in the acoustic impedance. This

too will be detected by the method.

• The use of multiple acoustic sources, multiple sensors and directionally sensitive processing

techniques enables leakage and blockage to be detected and located in complex pipeline

networks. No prior knowledge of the pipeline layout is required, unlike many other leakage

detection methods, and the method remains one which is particularly suitable for remote

operation.

• Linking the source and the sensors into a SCADA system enables the health of the pipeline

to be monitored continuously and remotely. Leakage or blockage may then be detected and

located almost in real time.

Whilst there appear to be clear benefits in using the proposed approach for monitoring the

health of industrial pipelines, it is recognised that the research completed so far is limited both

as to physical understanding and experimental validation in realistic pipeline configurations.

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
18
These issues are currently being addressed so that the technique can be applied with greater

confidence in real operational pipeline systems.

Acknowledgements

The authors would like to thank BP Exploration for providing funding for this work. One of

the authors (MNS) was a postgraduate student and received support from the Syrian

Government during his period in Manchester.

References

[1] Audubon, 2002, Toxic Tundra. Auban, Washington.

[2] Page, J., 2005, ‘Black gold brings hope of return to the glory days of a century ago’,

The Times, 25th May, pp.38-39.

[3] Hunaidi, O, Wang, A, Bracken, M, Gambino, T and Ficke C, 2004, ‘Acoustic methods

for locating leaks in municipal water pipe networks, International Water Demand

Management Conference, Jordan

[4] Tai, Y., 2004, ‘China faces serious water shortages’, The Washington Times, 23rd

March

[5] Cheong, LC, 1991, ‘Unaccounted-for water and the economics of leak detection’,

Water Supply, 9(1), pp.1-6.

[6] Arab News, 2004, ‘Kingdom needs massive investment for water security’, June 3rd.

[7] Ellul, I., 1989, ‘Pipeline leak detection’, The Chemical Engineer, 461, pp.39-44.

[8] Billman and Isermann, 1984, ‘Leak detection methods for pipelines’, IFAC World

Congress, Hungary.

[9] Hunaidi, O, Chu, W, Wang, A and Guan, W, 2000, ‘Detecting leaks in plastic pipes’,

Journals of American Water Works Association, 92 (2), pp.82-94.

[10] Sharp, D. B., 1996, ‘Acoustic pulse reflectometry for the measurement of musical wind

instruments’, Ph.D. thesis, University of Edinburgh

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
19
[11] Liggett, JA and Chen, L, 1994, ‘Inverse transient analysis in pipe networks’, Journal of

Hydraulic Engineering, 120 (8), pp.934-955.

[12] Wang, X, Lambert, MF, Simpson, AR, Liggett, JA amd Vitkovsky, JP, 2002, ‘Leak

detection in pipelines using the damping of fluid transients’, Journal of Hydraulic

Engineering, 128 (7), pp.697-711.

[13] Lee, PJ, Vitkovsky, JP, Lambert, MF, Simpson, AR and Liggett, 2005, ‘Leak location

using the pattern of the frequency response diagram in pipelines: a numerical study’,

Journal of Sound and Vibration, 284, pp.1051-1073.

[14] Shamout, M.N., Lennox, B and Turner, J.T., 2005, ‘A novel acoustic technique for the

remote detection of leakage in pipelines and pipeline networks: Basic principles’,

submitted to the proceedings of the IMechE

Papadopoulou, K, Wang, X, Lennox, B, Turner, J, (2008), ‘An evaluation of acoustic reflectometry for leakage and
blockage detection’, IMechE Part C, Mechanical Engineering Science, 222 (6), 959-966
20