ARTICLE IN PRESS

NDT&E International 43 (2010) 365–374

Contents lists available at ScienceDirect

NDT&E International

0963-86

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/ndteint

Experimental investigation of reflection in guided wave-based inspection forthe characterization of pipeline defects

Xiaojuan Wang a, Peter W. Tse a,n, Chris K. Mechefske b, Meng Hua a

a Smart Engineering Asset Management Laboratory (SEAM), Department of Manufacturing Engineering & Engineering Management, City University of Hong Kong, Tat Chee Avenue,

Kowloon, Hong Kong, Chinab Department of Mechanical Engineering, Queen’s University, Ontario, Canada

a r t i c l e i n f o

Article history:

Received 15 December 2009

Accepted 11 January 2010Available online 22 January 2010

Keywords:

Guided waves

Pipeline inspection

Defect characterization

95/$ - see front matter & 2010 Elsevier Ltd. A

016/j.ndteint.2010.01.002

esponding author. Tel. +852 2788 8431; fax:

ail address: mepwtse@cityu.edu.hk (P.W. Tse

a b s t r a c t

In guided wave-based pipeline inspection, the reflection from a defect usually includes sufficient defect-

relevant information. It has been found that the reflection of guided waves at defect is the joint result of

interference between reflections from both its front and back edges. However, the majority of published

work has only studied and considered the overall resulting reflection signal for the inspection-related

tasks. This paper reports the findings of extensive experimental investigation on the effects of a defect’s

geometric parameters on the two reflection signals from the respective edge of defect. We show that

the two edge reflections present different signal features, which further results in the complexity of

the overall reflection signal from the defect. We accordingly propose a new strategy that considers the

extraction of two edge reflection signals embedded in overall reflection signal and the use of the

identified edge signals so as to enable an accurate and quantitative pipeline defect characterization.

& 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Pipelines represent crucial infrastructure in the oil, gas,chemical and water transport industries. Significant researchhas been conducted on topics related to pipeline inspection tomeet the continuous requirements of civil and industrial users. Toaccurately and efficiently carry out planned maintenance andreplacement operations on pipelines, the ability to characteriseexisting or developing defects during pipeline inspection isimportant for the practical application of non-destructive testing(NDT) techniques. Many of the methods currently available arecapable only of providing a qualitative assessment of pipelinedefects due to various limitations, such as a low resolution. Theuse of ultrasonic-guided waves is a new and advanced techniquein the NDT field [1,2] and, thanks to its continuous developmentin both theory and practice, it has become increasingly attractivefor pipeline inspection.

The advantages of guided waves, which include a high degreeof sensitivity and versatility, offer great potential for the detectionof various defects, but many difficulties in interpreting the wavedata collected for accurate defect characterization remain. Thecomplexities that must be considered when using guided wave-based applications relate to wave excitation and propagation aswell as interaction with the defect. First, multiple modes are often

ll rights reserved.

+852 2194 2289.

).

generated in each guided wave excitation [3,4]. The resultingwave signals present complicated waveforms because of theexistence of the unwanted modes. The important parametersrelated to defect characterization, such as peak amplitude, peaklocation and the arrival time of each mode, are consequentlydifficult to identify. Second, guided waves are inherentlydispersive, which causes the wave packet of the mode to spreadin space and time after it propagates over a certain distance [5].This continuous wave distortion renders many of the availablesignal-processing analysis tools, such as wavelet analysis, insuffi-ciently accurate. Third, the details of the interaction of the guidedwaves with the discontinuities in the pipeline are not sufficientlyelaborate because of the extremely complicated wave properties[6]. This interaction may exhibit unknown features because of theinterference from the various signal components involved. There-fore, to ensure the accuracy of defect characterization, theproblems inherent in the guided wave technique must becarefully considered. Proper control on transduction and excita-tion of guided waves could help to solve the first two aforemen-tioned problems, so the study reported herein primarily takes intoaccount the third that concerns the defect reflection.

The reflection signal that results from the interaction of thepropagating guided waves with the defect in pipeline underexamination, in principle, includes substantial defect-relevantinformation. This information may indicate the presence, location,severity and/or other features of the defect. Better understandingof the details of wave reflection is always helpful to characterisethese features of defect. This paper thus aims to investigate this

ARTICLE IN PRESS

X. Wang et al. / NDT&E International 43 (2010) 365–374366

issue on the basis of existing research achievement throughfurther experiments to discover new underlying rules so thataccurate and quantitative defect characterization could becomeenabled.

radial depth

circumferential extent

axial extent

Fig. 2. Illustration of three-dimensional parameters (axial extent, radial depth and

circumferential extent) of the defect.

2. Characterization of pipeline defects

Defect characterization-related topics have stimulated a greatdeal of interest in guided waves-based pipeline research andapplications [7–9]. The defect that exists in a pipeline can beapproximately represented as a rectangular shape considering theloss of material from pipeline in three dimensions, as shown inFig. 1. The geometric parameters of a defect are, accordingly,described in terms of radial depth and circumferential and axialextent in a simplified manner. It is believed that a defect can becharacterised by identifying and quantifying the effects of itsgeometric parameters on the reflection signal. Many researchershave investigated the related issues through laboratoryexperiments and finite element (FE) simulations. For example,Alleyne et al. [7] and Demma et al. [10], respectively, reported thatthe reflection of incident modes L(0,2) and T(0,1) are close to thelinear functions of the circumferential extent and radial depth of adefect. Bai et al. [11] used an efficient numerical procedure for thethree-dimensional reflection problem to analyze how the lengthand depth of a circumferential defect affect the scatteringcharacteristics in a pipe. They indicated that the change in thereflection coefficient with increasing circumferential crack lengthL is linear to L when the incident modes consider L(0,2) and F(1,3).Cawley et al. [1] also considered the effect of the axial extent andconcluded that the reflection from a defect is a strong function ofthat defect’s axial extent. The findings of these studies provide auseful foundation for further research. Although these previousresearch has identified reflection as important for defectcharacterization, and found that a reflected wave is stronglycorrelated with the geometric parameters of the defect, theproblem of determining the size or severity of pipeline defect hasnot been figured out yet, especially in a quantitative and accuratemanner. The research aiming at developing characterizationtechnique is still ongoing [12].

Demma et al. [10,13] performed substantial parameter study ofthe reflection from crack and notch in pipeline. They reported thatsuch reflection is the joint result from interference between thereflections at the front and back edges of that defect. It is thusdifficult to obtain independent information on the geometricparameters of defect directly from the refection signal, especiallywhen accurate process models are not available. AlthoughDemma et al. indicates the existence of different components inoverall reflection signal, they did not proceed to consider the useof those components for defect characterization, and the majorityof published work is still only concerning or studying the overallresulting reflection signal in the inspection-related tasks.

Given all of the difficulties and limitations that exist in theextant research, the purpose of the work presented herein was topropose a new strategy based on the use of embedded reflection

Fig. 1. Defect can be represented as a rec

components for accurately and quantitatively characterizingpipeline defects in guided wave-based inspection. The respectivereflection from front and back edges of defect will be investigatedthrough extensive experiments so as to identify different features

of edge reflection signals embedded in the reflection signal, and therelationship between edge reflections and the geometric parameters

of defect. Our experimental results were compared and verifiedwith FE simulations each other. The findings will offer insight intomore understanding of the reflection of the guided waves from adefect, and accordingly, the new strategy based on edge reflectionwill be proposed for accurate and quantitative determination ofcharacteristic sizes of pipeline defect.

3. Experimental setup

To identify features of edge reflection signals and reveal therelationships between edge reflections and the defect geometricparameters, experiments were conducted on three same steelpipes that had an external diameter of 34 mm, a wall thickness of4 mm and a length of 2030 mm. We introduced an artificial notchin each pipe as defect, which had specified axial and circumfer-ential extent and radial depth, as shown in Fig. 2. Over the courseof the experiments, the values of these three parameters weregradually increased using a milling machine to simulate defects ofdifferent sizes or degrees of severity.

We adopted the longitudinal L(0,2) mode, as it is easy to exciteand has relatively simple acoustic fields [14]. Frequencies rangingfrom 100 to 240 kHz were chosen because the L(0,2) mode excitedaccordingly is non-dispersive, as can be seen from Fig. 3, whichshows the phase and group velocity dispersion curves of the pipeunder examination. These curves were calculated using theDISPERSE program [15]. The use of the L(0,2) mode and thefrequency range chosen reduced the signal complexity discussedin Section 1 to a certain extent as we considered a single modeand its non-dispersive operating range, thus focusing ourinvestigation on the problem closely relevant to defect reflection.

The experimental setup and instruments used are depictedschematically in Fig. 4. A windowed toneburst consisting of five

tangular shape in three dimensions.

ARTICLE IN PRESS

0.000

1

2

3

4

5

6

7

8

9

F(1,1)

L(0,1)

F(1,2)

L(0,2)

Vph

(m/m

s)

F(1,3)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

Vgr

(m/m

s)

L(0,2)F(1,3)

F(1,2)

L(0,1)

F(1,1)

group velocity dispersion curve phase velocity dispersion curve

Frequency (MHZ)0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

Frequency (MHZ)

Fig. 3. Phase and group velocity dispersion curves of the tested pipe sample.

Signal Generator

Pulser/Receiver

PC

Transducersdefect

pipe

PC-based DAQ

Fig. 4. Schematic representation of experimental setup for pipe inspection.

X. Wang et al. / NDT&E International 43 (2010) 365–374 367

cycles at the chosen frequency was delivered through an arbitrarysignal generator. The transduction system was carefully designedto excite a single L(0,2) mode into the pipe under examination. Aring consisting of a certain number of piezoelectric transducers(PZTs) was bonded at one end of the pipe to generate and receiveguided waves. The PZTs were made of length expander-typepiezoelectric material and distributed axisymmetrically, thusensuring that only the expected longitudinal mode was excitedwhilst the flexural modes were suppressed. This experimentalsystem is similar to that used in other research [1,10]. The defectswere initially machined 1300 mm from the end at which the PZTswere bonded. To ensure that the experimental data would have areliable degree of comparability, each measurement wasperformed with strictly consistent experimental settings, exceptfor the changes in the defect geometric parameters.

FE-based simulation was conducted in parallel to corroborateour experimental results. The commercial FE program ANSYS wasused for the model generation and simulation analysis. Themodelled pipes had the same parameters, including material,diameter, thickness and length, as those used in the experiments.The series of defects were introduced into the pipeline model byremoving certain elements based on the requirements of thedefect position and dimensions set in the experiments. To ensurethat the simulation results would be sufficient in terms ofaccuracy and resolution, we optimised the model parametersand operational setting. The details of this optimisation arereported in a published conference paper [16]. To provide a briefexample here, Fig. 5 shows the modelled pipe with a 3-mm-widecircumferential defect and the applied loading force as excitationsignal.

4. Results and discussion

A series of experiments were performed to investigate theedge reflections of the guided waves created by defects withdifferent geometric parameters. The focus of our analysis here ison the effects of the concerned defect parameter on the reflectionsat front edge and back edge of defect, respectively, especially theeffect of the axial extent, as this parameter was identified as thefirst important factor in the complexity of reflection.

4.1. Relationship between the edge reflections and the axial extent of

a defect

4.1.1. Analysis of experimental results

Tests were performed on a pipe with a notch, the radial depthand circumferential extent of which were kept constant at 21.25%of the pipe wall thickness and 100% of the pipe circumference,respectively. The axial extent was gradually changed from 3 to170 mm. Data were captured after the manufacture of each defect.While manufacturing the defects according to the proposed sizes,the front edge of each that was nearest to the transducers was keptunchanged, whereas the axial extent to the back edge wasincreased, as shown in Fig. 6.

Fig. 7 shows the collection of signals reflected from the defectwith varying axial extents under excitations at centre frequenciesof 145, 175 and 200 kHz, respectively. These three groups of dataexhibited a similar trend in the change of the signal pattern. Thedata collected at the frequency of 175 kHz are used for oursubsequent discussion because the defect responses at thisfrequency were relatively stronger than those at the otherfrequencies.

The parallel FE simulation was designed to corroborate theeffectiveness of the experimental data we had collected. Fig. 8presents a comparison of the reflection signals resulting from thedefects with axial extents of 6, 14, 24 and 42 mm, respectively.The solid lines are the reflection signals obtained from theexperiments, and the dashed lines are the FE simulation results.Excellent agreement was demonstrated between them, thusincreasing the credibility of both.

Fig. 7 shows that when the axial extent is increased, the wavepacket of the reflected signal is extended in time with anincreased number of wave cycles and amplitudes of irregularvariation. For example, some of the signals in this figure representthe superposition effect in amplitude, whereas others represent

ARTICLE IN PRESS

Fig. 5. Example of the meshed pipe model with a defect introduced and loading force applied.

front edge of defect back edge of defect

Fig. 6. Illustration of the front and back edges of a pipe defect.

X. Wang et al. / NDT&E International 43 (2010) 365–374368

its cancellation. It indicates that the signal that results from thedefect reflection involve overlapping signals. These overlappingsignals are the result of the reflections at different edges of thedefect, a conclusion that was validated in our experiments whenthe axial extent was sufficiently long (86 and 170 mm in the pipesamples examined). Two separate signals, caused by the frontedge and back edge of the defect, were clearly observed over time,and are highlighted with a red line in Fig. 7.

It is interesting to note the coincidence of the same wavepattern in the beginning portion of the collected reflection signals.Fig. 9 presents a plot of the correlation coefficients of reflections ofdefect at different axial extents, with the reflection of 170-mm-length defect employed as a reference. When the value of thecorrelation coefficients is 1, the signals at these points have thesame pattern. It shows in Fig. 9 that the longer the axial extent,the greater the coincidence of the signals at axial extents withconsecutively increasing values. This phenomenon indicates thatthe signals generated from the front edge of the defect are thesame for all defects, regardless of the axial extent. Thisassumption is further verified by the signals at the axial extentsof 86 and 170 mm, as these signals represent the two separateparts of the signal indicated above. The signals reflected from thefront edge of the defect at these two axial extents are completelycoincident. It further implies that the signal from the back edge ofthe defect must be complicated as the other part of overallreflection that presents irregular change in Fig. 7.

4.1.2. Front- and back-edge signals embedded in the reflection

As noted in our foregoing discussion, the signal reflected fromthe entire defect mainly comprises two parts: the signal from thereflection process of the incident waves at the front edge of thedefect, and the signal from a more complicated wave interactionprocess that includes the transmission of waves through the frontedge of the defect, reflection from its back edge and the secondtransmission of the reflected wave through the front edge. Here,we refer to these two parts as the ‘front-edge signal’ and the‘back-edge signal’. The former primarily depends on the geo-metric profile of the front edge of the defect, including its depthand circumference. The pattern and energy of the back-edgesignal are related to a greater number of factors, including thegeometric profile of both edges of the defect and its defective area.

The different features of these two edge signals will inevitablyintroduce complexity of resulting reflection and difficulty inevaluating defect directly. It should be noted that the effect ofreverberations following these two main reflections on theresulting signal can be neglected since they are almost attenuatedwith high-order factor in the cases where axial extent of defect isnot sensitively small with respect to wavelength of propagatingwaves.

4.1.3. Separation of overlapping front- and back-edge signals

Since the front- and back-edge signals include the informationwith different features about the characteristics of the defect, theoverlapping of these two signals makes the reflection becomeextremely complex obviously. Therefore, it is useful to extract thetwo and then analyze individually or jointly in defect character-ization. Such an extraction can make the complex reflectionproblem decomposed into several relatively small problemsrelated to edge reflection. The edge signal involves less defectparameters, so it is relatively easier to be analyzed than the wholereflection problem directly. Moreover, two edge signals canprovide more independent and correlated information sourcesof defect. For example, once the extracted signals have beenidentified, the axial extent of the defect, which is directly relatedto the relative distance between two edge reflections, can beeasily and accurately derived based on the wave group velocityand the time shifts of the two signals.

We were able to identify the front-edge signal in ourexperiments by extending the axial extent of the defect (to either86 or 170 mm), which allows the separate signals from the twoedges to be clearly observable over time. Employing this signal asa reference, the back-edge signal embedded in the overallreflection at different axial extents could be extracted simply bysubtracting the reference from each reflection signal collected.The results obtained from this process are presented in Fig. 10,which shows that the cycle numbers in all of the back-edgesignals remained almost constant and that the amplitudeschanged with varying axial extents. Small amount of noiseappears in the rear of each signal due to the reverberationbetween the two edges of defect.

The individual features of the front- and back-edge signals canbe further elucidated by the separation results obtained. Both ofthese signals in the different defect scenarios exhibit nearly thesame patterns in terms of their number of cycles, frequency andmodulation. The amplitude of the front-edge signal remains thesame for all of the scenarios with various axial extents becausethis signal depends primarily on the front edge of the defect, aspreviously discussed. The amplitude of the back-edge signal ismore complex, presenting different variation trends at differentaxial extents. Fig. 11 shows the energies of the reflectioncoefficient spectrum for the back-edge signals at these different

ARTICLE IN PRESS

5.1 5.2 5.3 5.4 5.5 5.6x 10-4

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

Time (s)

Am

plitu

de (v

)

5.1 5.2 5.3 5.4 5.5 5.6x 10-4

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

Time (s)

Am

plitu

de (v

)

5.1 5.2 5.3 5.4 5.5 5.6

x 10-4

-0.25-0.2

-0.15-0.1

-0.050

0.050.1

0.150.2

0.25

Time (s)

Am

plitu

de (v

)

Experimetnal dataFE simulation data

Experimetnal dataFE simulation data

Experimetnal dataFE simulation data

Experimetnal dataFE simulation data

5.1 5.2 5.3 5.4 5.5 5.6 5.7

x10-4

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

Time (s)

Am

plitu

de (v

)

axial extent = 6mm axial extent = 14mm

axial extent = 42mm axial extent = 24mm

Fig. 8. Comparison of defect reflections obtained from experimentation and FE simulation.

Fig. 7. Reflection signals from the defect with varying axial extents, and constant radial depth and circumferential extent under the excitation of 145, 175 and 200 kHz.

X. Wang et al. / NDT&E International 43 (2010) 365–374 369

extents. The Y coordinate is the root-mean-square (RMS) value ofthe reflected back-edge signals, with the RMS value of thecommon front-edge signal taken as the normalisation factor.The X coordinate is the axial extent to wavelength ratio for theincident wave at 175 kHz (31.157 mm). From Fig. 11(a), it can beseen that the reflection coefficients of the back-edge signal at X

from 0 to 1.4 demonstrate a roughly periodical variation due tothe effect of the defect’s front edge on the generation of the back-edge signal. With an increase in the axial extent, the reflectioncoefficients of the back-edge signal exhibit the relatively stablevariation depicted in Fig. 11(b). This is because the front- andback-edge signals are completely separate in time in these

situations, and then the latter primarily depends on thegeometric features of the defect’s back edge. Further, it can alsobe noted that the reflection coefficients of the front-edge signalsare always larger than those of their back-edge counterparts.

We were able to obtain these results easily with our collectedexperimental data as the reference, namely, that the front-edgesignal can be identified when the axial extent of the defect ismade long enough. In practical applications, the extraction of theedge reflection components can be achieved by applying dataanalysis-based method to the collected signal directly. Thetechnique for this problem has been developed by the authors,which will be reported in another forthcoming paper.

ARTICLE IN PRESS

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

0.52 0.54 0.56 0.58 0.6-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

-0.2

0

0.2

0.52 0.54 0.56 0.58 0.6-0.2

0

0.2

front-edge signal

3mm

6mm

8mm

11mm

16mm

14mm

18mm

21mm

24mm

27mm

37mm

29mm

42mm

back-edge signal

86mm

(ms) (ms)

Fig. 10. Extraction of front- and back-edge signals from the experimental reflection signal collected from the defects with varying axial extents, constant radial depths and

constant circumferential extents.

0.18 0.185 0.19 0.195 0.2 0.205 0.210.996

0.9965

0.997

0.9975

0.998

0.9985

0.999

0.9995

1

Time

Cor

rela

tion

coef

ficie

nts

86mm42mm37mm27mm18mm11mm 6mm

Fig. 9. Correlation function curves of defect reflections, with the reflection from a 170 mm length defect employed as a reference.

X. Wang et al. / NDT&E International 43 (2010) 365–374370

4.1.4. Determination of the axial extent of a defect based on the

separated signals

The time shift between the front- and back-edge signalscorresponds to the time that the guided waves take to propagatebetween the front and back edges of the defect, namely, twice itsaxial extent. Thus, once the two edge signals have been identified,the axial extent of the defect can be calculated by the following

equation:

L¼ ðD� vgrÞ=2; ð1Þ

where L is the length of the defect in the axial direction, D isthe time shift between the front- and back-edge signals, and vgr isthe group velocity of the propagating guided waves. An accuratedetermination of this time shift is the first key to calculating the

ARTICLE IN PRESS

0.00.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

Nor

mal

ized

refle

ctio

n co

effic

ient

s at

RM

S

front-edge signal back-edge signal

1.00.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

Nor

mal

ized

refle

ctio

n co

effic

ient

s at

RM

S

front-edge signal back-edge signal

Axial extent of defects (% wavelength of excited wave) Axial extent of defects (% wavelength of excited wave)0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

Fig. 11. Relationship between the axial extent of defect and two reflection components (front-edge signal and back-edge signal).

Table 1Comparison of time shift results obtained by using methods based on actual

measurements, experimental data and simulation data, respectively (fre-

quency=175 kHz, group velocity at 175 kHz=5.35 m/ms).

Axial

extent,

mm

Actual

time shift,

ms

Time shift by

experimental data,

ms

Error,

%

Time shift by

simulation data,

ms

Error,

%

3 1.12 1.10 �1.79 1.5 33.99

6 2.24 2.39 6.70 2.7 20.53

14 5.23 5.53 5.42 5.4 3.25

16 5.98 6.48 7.72 6.3 5.35

21 7.85 8.51 8.41 8.1 3.18

24 8.97 9.10 1.45 9.3 3.68

27 10.09 10.13 0.40 10.2 1.09

29 11.03 11.15 1.09 11.1 0.63

37 13.83 13.99 1.16 14.1 1.80

42 15.71 15.75 0.32 15.9 1.21

86 32.15 31.49 �2.05 32.1 �0.15

Table 2Comparison of axial extent results obtained by using methods based on actual

measurements, experimental data and simulation data, respectively (fre-

quency=175 kHz, group velocity at 175 kHz=5.35 m/ms).

Actual axial

extent , mm

Axial extent by

experimental

data, mm

Error, % Axial extent

by simulation

data, mm

Error,

%

3 2.94 �1.92 4.01 33.75

6 6.39 6.55 7.22 20.38

14 14.79 5.66 14.45 3.18

16 17.33 8.34 16.85 5.33

21 22.76 8.40 21.67 3.18

24 24.34 1.43 24.88 3.66

27 27.10 0.36 27.29 1.06

29 29.82 2.85 29.69 2.39

37 37.42 1.14 37.72 1.94

42 42.13 0.31 42.53 1.27

86 84.24 �2.05 85.87 �0.15

00

5

10

15

20

25

30

35

Tim

e sh

iftin

g be

twee

n th

e re

flect

sfro

m fr

onte

nd a

nd b

acke

nd (u

s)

Axial extent of defects (mm)

Results from experiment dataResults from simulation dataResults from actual measurement

20 40 60 80

Fig. 12. Relationship between the axial extent of defect and the time shift of two

edge signals.

X. Wang et al. / NDT&E International 43 (2010) 365–374 371

axial extent of the defect. As the front- and back-edgesignals present a similar pattern in the time domain, the cross-correlation function (CCF) of the two is computed to estimatetime shift D. In this method, the time shift between the twosignals corresponds to the maximum value of their CCF [17], asshown below.

D¼max1

T

Z T

0xðtÞyðtþtÞdt

� �; ð2Þ

where T is observation time, and x and y are front- and back-edgesignals, respectively.

The estimation results of the time shift (D) and the axial extent(L) of the defect from the actual measurements and the experi-mental and simulation data are presented in Tables 1 and 2, fromwhich it can be seen that the experimental and simulation resultsare very close to the actual measurements. The relationshipbetween the time shift of the two edges of the reflected waves andthe axial extent of the defects based on actual measurements,experiments and simulations are also presented in curve form inFig. 12, which again demonstrates the excellent agreement amongthem. Moreover, it was identified from the correlation andcomparison processes that the phase of the front-edge signal is

the reverse of that of the back-edge signal. These conclusions alsovalidate the acceptability of our extraction operation of the front-and back-edge signals.

ARTICLE IN PRESS

X. Wang et al. / NDT&E International 43 (2010) 365–374372

4.2. Relationship between the edge reflections and the

circumferential extent of a defect

As most pipeline defects are non-axisymmetric in reality, it isalso necessary to investigate the relationship between the edgereflections and circumferential extent of a defect. Accordingly, weconducted experiments employing defects that represented sixdifferent circumferential extent scenarios: 60/360, 120/360, 180/360, 240/360, 300/360, and 360/360 (as a percentage of the fullcircumferential extent). The radial depth of the defects was fixedat 1 mm (25% of the pipe wall thickness), and their axial extentwas held constant at 170 mm. The use of defect with long axialextent was to ensure that the reflections from the front and backedges would be separate over time themselves so that the effect ofthe circumferential extent on the respective edge signal can bedirectly identified.

Our experimental results are shown in Fig. 13. As otherresearchers have also observed, mode conversion occurs atdefects with a non-axisymmetric circumferential extent [18];that is, other modes are generated in addition to mode L(0, 2).Bai [10] indicated that incident mode L(0, 2) is transmitted withmost of the energy in it before and after the reflection. As L(0, 2) isthe primary reflection mode, its amplitude is relatively large andclearly distinguishable. To determine the effects of other modes onthe mode L(0,2) reflection at both edges of a defect, the time shiftbetween the modes L(0, 2) of the two observed edge signals wasidentified employing the CCF method to calculate the axial extentof the defect. The result obtained through this procedure(165.85 mm) is approximately equal to the real value (170 mm),thus verifying that mode L(0,2) was not interfered with othermodes in reflection. The reflection echo of mode L(0,2) is thusdirectly used in analysis.

It can be observed from Fig. 13(a) that the time shift of the twoedge signals remains constant when the axial extent of the defectremains the same. With an increase in the circumferential extent,the amplitudes of both the front- and back-edge signals can beobserved to increase. To present the change in the two reflectionsmore clearly, the curves of reflection coefficients of the edgesignals as a function of the circumferential extent of the defect are

-101

-101

-101

-101

-101

0.52 0.54 0.56 058 0.6 0.62 0.64-101

Nor

mal

ized

refle

ctio

n co

effic

ient

s at

RM

S

front-edge signalL(0, 2)

back-edge signalL(0, 2)

(ms)

Fig. 13. Reflection signals from the defects with various circumferen

plotted and shown in Fig. 13(b). Both reflections are normalisedby the RMS value of the incident signal. It can be seen that thereflection from the front edge has almost the same change rate asthat from the back edge, both of which increase roughly linearlywith respect to the circumferential extent of the defect. Thereflection coefficient of the back-edge signal is always smallerthan that of the front-edge signal in every scenario.

4.3. Relationship between the two edge reflections and the radial

depth of the defect

To investigate the relationship between radial depth and thefront- and back-edge signals of a defect, we performed experi-ments on pipe models whose defects extended to the fullcircumference, had a constant axial extent and varying radialdepths. The constant value of the axial extent chosen for thisseries of testing was 86 mm, at which, as previously discussed, thefront- and back-edge signals are ensured to be separate over timeso that the effect of radial depth on edge reflections can be clearlyrevealed. The radial depths considered were 0.85, 1.64, 2.42 and3.21 mm, which corresponded to 21%, 41%, 61% and 80% of thepipe wall thickness, respectively.

Fig. 14(a) shows the resulting data, and the reflection coefficientsof the front- and back-edge signals as a function of the radial depthof the defect are shown in Fig. 14(b). It can be clearly seen that thefront-edge signal increases monotonically or roughly linearly withrespect to the radial depth. Of particular interest, however, is thatthe back-edge signal fails to exhibit the same monotonic increase.This means that a defect with a back edge of greater depth canproduce a smaller back-edge reflection than will a defect with a backedge of less depth. Such feature of nonmonotonic trend cannot bepresented by the overall reflection signal, which always has a roughlinear relationship with the change of radial depth as reported byother researchers. We surmise that this is because the transmissionenergy of the waves that reach the back edge of defect is diminishedwhen the reflection at the front edge of defect becomes strong to aspecific value. What can be stated with assurance is that the back-edge signal depends on more factors or parameters than does its

00.00

0.02

0.04

0.06

0.08

0.10

0.12front-edge signalback-edge signal

Circumferential extent of defects (% whole circumference)10 20 30 40 50 60 70 80 90 100 110

tial extents, constant axial extents and constant radial depths.

ARTICLE IN PRESS

-101

-101

-101

0.52 0.53 0.54 0.55 0.56 0.57

-101

0.2

0.02

0.04

0.06

0.08

0.10 front-edge signal back-edge signal

Nor

mal

ized

refle

ctio

n co

effic

ient

s in

RM

S

Radial depth of defects (% thickness of pipe)

front-edge signal L(0, 2)

back-edge signal L(0, 2)

(ms)0.3 0.4 0.5 0.6 0.7 0.8 0.9

Fig. 14. Reflection signals from the defects with various radial depths, constant axial extent and constant circumferential extents.

X. Wang et al. / NDT&E International 43 (2010) 365–374 373

front-edge counterpart. Moreover, combined with the findings inSection 4.2, it indicates that the circumference extent and radialdepth reveal different features in their effects on two edgereflections, which provide the possibility of characterizing thesetwo geometric parameters further through proper decompositiontechniques.

4.4. Proposed strategy for the characterization of pipeline defects

Based on the useful observations and findings thus farpresented, we here propose a new strategy for the accurate andquantitative characterization of pipeline defects. This strategyprimarily involves the extraction of two edge signals from singleoverall reflection signal and the use of these identified edgesignals for defect characterization. As discussed, two parts areembedded in a defect’s reflection signal: front-edge signal andback-edge signal. The patterns of the front- and back-edge signalsremain nearly constant for defect scenarios of varying axial andcircumferential extent and radial depth. However, their ampli-tudes demonstrate different features with a continuous increasein a specified defect geometric parameter. The front-edge signal isquite simple, and its amplitude characteristics always change in alinear fashion. The amplitudes of the back-edge signal, in contrast,exhibit periodic, monotonic or nonmonotonic variations withdifferent geometric parameters. Furthermore, it can be concludedthat the axial extent of defect is closely related to the phaseinformation of two edge signals while its circumferential extentand radial depth have important effect on their amplitudes in ajoint manner. Identifying these edge signals once extracted,together with their phase and amplitude features, can providesufficient information to enable accurate and quantitative defectcharacterization.

The axial extent of the defect can be easily determined fromthe time shift between the two identified edge signals, as this shiftin time bears a direct linear relationship with the length of thedefect. Each identified edge signal includes geometric informationon the corresponding edge of the defect, so the radial depth andcircumferential extent of that defect can be determined throughintroducing proper decomposition and analytical techniques onthose edge signals. The results reported in Sections 4.2 and 4.3have shown great feasibility for this purpose. The further

quantitative description of a defect requires a consideration ofthe attenuation and dispersion of the guided waves duringpropagation, which allows the refinement of the obtained results.In addition, it is worthy to note that the issues of multiple-modeand dispersion as aforementioned in Section 1 should be avoidedas much as possible to simply the overall reflection signal firstthrough a good controlling on transduction and excitation so thatthe proposed strategy could be implemented better.

Our proposed strategy decomposes the reflection of guidedwaves at the defect in a very complex form into the relativelysimple reflection problems that occur at its front and back edges.Such process obviously reduces the complexity of the concernedproblem because the destructive effect of the overlap betweentwo edge signals on reflection can be efficiently removed. At thesame time, it provides more information sources about the defect,thus considerably enhancing the reliability and accuracy ofquantitative defect characterization. In order to completelyachieve the objectives based on this proposed strategy, there areseries of follow-up researches that need to be conducted, forexample, the extraction of edge reflections and further decom-position of each reflection component.

5. Conclusion

The study reported herein investigated the reflection of theguided waves at the front and back edges of pipeline defect, andparticular attention is paid in this work to the respective effects ofthe geometric parameters of defect on each edge reflection signal.The results show that the complexity of a defect’s reflection signalresults from the overlap between the reflected edge signals withdifferent features. Extensive experimentation, corroborated by FEsimulation, was performed on artificial defects of various sizesin pipeline for reflection study. A new strategy is accordinglyproposed for the accurate and quantitative characterization ofpipeline defects in using guided wave-based inspection method.

The findings presented in this paper can serve as a foundationfor further work on this topic. The notch-type defect considered inthis paper involves two obvious changes of cross-sectional area atthe boundaries between the defective area and pipe wall, so thefront-edge signal and back-edge signal can be naturally identifiedas primary components of overall reflection signal to be extracted

ARTICLE IN PRESS

X. Wang et al. / NDT&E International 43 (2010) 365–374374

for defect characterization. The proposed strategy can beextended to more complicated defects or real situations since itappears as a simplification method for complex problem based onthe extraction of reflection components. The defective boundariesof any defect will be taken as the features related to size orseverity information of that defect, which generate the strongreflections to overlap in overall reflection signal. The complexlength profile of defect will not have a significant effect onreflection signal since it only involves the tapers extending overmuch small height or distant. That is, the overall reflection signalfrom the defect will mainly include the reflection componentscaused by the boundaries of defective area of concerned defect. Inthe frame of proposed strategy, the destructive effect of axialextent on reflection can be eliminated through the extraction ofboundary reflection components. Although each extracted com-ponent in practical cases will be related to more characteristicparameters of defect besides the circumferential extent and radialextent as discussed in this paper, its complexity is lower certainlycompared to the overall reflection signal. Therefore, it is expectedthat the strategy proposed herein will enable or facilitate theaccurate and quantitative characterization of defect parameters,especially the axial extent of defect. This strategy needs to berefined when the defect is very small since the reverberationsfollowing the first back-edge reflection have to be considered.

In general, the interaction of the guided waves of a real defect,such as corrosion, is a relatively more complicated phenomenondue to propagation in the rough defective area and reflection ofwaves at the irregular edges. Such practical problems requirefurther research to identify their physical mechanisms from atheoretical perspective and to generalise the results of thisresearch to more situations.

Acknowledgements

The work that is described in this paper is fully supported by agrant from City University of Hong Kong (Project no. 7002326).

References

[1] Cawley P, Lowe MJS, Simonette F, Chevalier C, Roosenbrand AG. The variationof the reflection coefficient of extensional guided waves in pipes from defectsas a function of defect depth, axial extent, circumferential extent andfrequency. J Mech Eng Sci 2002;216:1131–43.

[2] Rose JL, Ditric JJ, Pilarski A, Rajana K, Carr FT. A guided wave inspectiontechnique for nuclear steam generator tubing. NDT & E Int 1994;27:307–30.

[3] Li J, Rose JL. Implementing guided wave mode control by use of a phasedtransducer array. IEEE Trans Ultrason Ferroelectr Freq Control 2001;48(3):761–8.

[4] Hou J, Leonard KR, Hinders MK. Automatic multi-mode lamb wave arrivaltime extraction for improved tomographic reconstruction. Inverse Problems2004;20:1874–88.

[5] Wilcox P, Lowe M, Cawley P. The effect of dispersion on long-range inspectionusing ultrasonic guided waves. NDT & E Int 2001;34:1–9.

[6] Alley ND, Cawley P. The interaction of lamb waves with defects. IEEE TransUltrason Ferroelectr Freq Control 1992;39(3):381–97.

[7] Alley ND, Lowe MJS, Cawley P. The reflection of guided waves fromcircumferential notches in pipes. J Appl Mech-T ASME 1998;65:635–41.

[8] Ditric JJ. Utilization of guided elastic waves for the characterization ofcircumferential cracks in hollow cylinders. J Acoust Soc Am 1994;96(6):3769–75.

[9] Bai H, Shah AH, Popplewell N, Datta SK. Scattering of guided waves bycircumferential cracks in steel pipes. J Appl Mech-T ASME 2001;68:619–31.

[10] Demma A, Cawley P, Lowe M. The reflection of the fundamental torsional modefrom cracks and notches in pipes. J Acoust Soc Am 2003;114(2):611–25.

[11] Bai H, Shah AH, Popplewell N, Datta SK. Scattering of guided waves bycircumferential cracks in composite cylinders. Int J Solids Struct2002;39:4583–603.

[12] /http://www3.imperial.ac.uk/pls/portallive/docs/1/55745699.PDFS.[13] Demma A, Cawley P, Lowe M, Roosenbrand A, Pavlakovic B. The reflection of

guided waves from notches in pipes: a guide for interpreting corrosionmeasurements. NDT&E Int 2004;37:167–80.

[14] Alleyne DN, Cawley P. Long range propagation of lamb waves in chemicalplant pipework. Mater Eval 1997;55:504–8.

[15] Pavlakovic B, Lowe M. A general purpose approach to calculating thelongitudinal and flexural modes of multi-layered, embedded, transverselyisotropic cylinders. In: Thompson DO, Chimenti DE, editors. Review ofprogress in quantitative NDE. New York: Plenum Press; 1999. p. 239–46.

[16] Tang S., Tse P., Wang X. The revelation of propagating ultrasonic guidedwaves through simulation when they encountered defects in a pipe. In:Proceedings of the ASME 2009 International Design Engineering TechnicalConferences & Computers and Information in Engineering Conference, 30August–2 September 2009 , San Diego.

[17] Ianniello J. Time delay estimation via cross-correlation in the presence oflarge estimation errors. IEEE Trans Acoust Speech Signal Process1982;30(6):998–1003.

[18] Lowe M, Alleyne D, Cawley P. The mode conversion of a guided wave by apart circumferential notch in a pipe. J Appl Mech-T ASME 1998;65:649–56.