sensors

Article
Mode Conversion Behavior of Guided Wave in a Pipe
Inspection System Based on a Long Waveguide
Feiran Sun 1 , Zhenguo Sun 1,2, *, Qiang Chen 1,2 , Riichi Murayama 3 and Hideo Nishino 4
1 Department of Mechanical Engineering, Tsinghua University, Beijing 100084, China;
sfr13@mails.tsinghua.edu.cn (F.S.); chenq@tsinghua.edu.cn (Q.C.)
2 Yangtze Delta Region Institute of Tsinghua University, Jiaxing 314006, China
3 Department of Intelligent Mechanical Engineering, Fukuoka Institute of Technology,
Fukuoka 811-0295, Japan; murayama@fit.ac.jp
4 Graduate School of Advanced Technology and Science, Tokushima University, Minami-Josanjima,
Tokushima 770-8506, Japan; hidero.nishino@tokushima-u.ac.jp
* Correspondence: sunzhg@tsinghua.edu.cn; Tel.: +86-10-6277-3860

Academic Editors: Dipen N. Sinha and Cristian Pantea

Received: 22 August 2016; Accepted: 27 September 2016; Published: 19 October 2016

Abstract: To make clear the mode conversion behavior of S0-mode lamb wave and SH0-plate wave
converting to the longitudinal mode guided wave and torsional mode guided wave in a pipe,
respectively, the experiments were performed based on a previous built pipe inspection system.
The pipe was wound with an L-shaped plate or a T-shaped plate as the waveguide, and the S0-wave
and SH0-wave were excited separately in the waveguide. To carry out the objective, a meander-line
coil electromagnetic acoustic transducer (EMAT) for S0-wave and a periodic permanent magnet
(PPM) EMAT for SH0-wave were developed and optimized. Then, several comparison experiments
were conducted to compare the efficiency of mode conversion. Experimental results showed that
the T(0,1) mode, L(0,1) mode, and L(0,2) mode guided waves can be successfully detected when
converted from the S0-wave or SH0-wave with different shaped waveguides. It can also be inferred
that the S0-wave has a better ability to convert to the T(0,1) mode, while the SH0-wave is easier to
convert to the L(0,1) mode and L(0,2) mode, and the L-shaped waveguide has a better efficiency than
T-shaped waveguide.

Keywords: mode conversion; waveguide; guided wave; EMAT; pipe inspection

1. Introduction
Online monitoring of pipes at high temperature for wall thickness and defects is a key to many
industrial non-destructive evaluation (NDE) demands, such as the power generation and the nuclear
industry. It is quite cost effective to avoid plant shutdowns during monitoring of its structures; whereas
“high-temperature monitoring” allows us to conduct the plant structure monitoring without shutting
down the plant [1].
In recent years, the ultrasonic guided wave testing method has been widely applied for pipe
inspections as it carries major advantages such as low attenuation, long distance propagation, and high
detection efficiency [2–6]. There are two techniques which are commonly employed for generating
guided waves: the piezoelectric transducer and electromagnetic acoustic transducer (EMAT) [2].
However, the piezoelectric ultrasonic testing requires a good sonic contact with the test piece,
which restricts its inspection ability for certain applications, especially under high-temperature
conditions [7,8]. EMAT has the ability to generate and detect ultrasonic waves without making a
contact with the test object which makes it suitable to inspect high-temperature pipes with a large

Sensors 2016, 16, 1737; doi:10.3390/s16101737 www.mdpi.com/journal/sensors

 Sensors 2016, 16, 1737 2 of 14

Sensors 2016, 16, 1737 2 of 14

lift-off. Even though the EMAT still requires a cooling system which causes many issues [9–11],
in addition,
Sensors the
2016, 16, lift-off effect also influences the results [12–15].
1737 2 of 14
Therefore, a sensor system for high-temperature pipe inspection, which can transmit and receive
Therefore, a sensor system for high-temperature pipe inspection, which can transmit and receive
ultrasonic guideda sensor
Therefore,
waves using afor long waveguide, has been developed by Murayama [16]. As shown
ultrasonic guided wavessystem using a long high-temperature
waveguide, haspipe beeninspection,
developedwhich can transmit
by Murayama [16].andAs receive
shown
ininFigure
ultrasonic 1,
Figure 1,
the
guidedtested pipe
wavespipe
the tested
was
usingwas
wound
a long with
waveguide,
wound
a long
with a long
plate
has been as the
platedeveloped
waveguide,
by Murayama
as the waveguide,
which
which
could
[16].could keep
As shown
keep
the
ultrasonic
in Figure
the sensor
ultrasonic awayaway
1, thesensor
tested from the high-temperature
pipe from
was wound with a longpipe.
the high-temperature plate The S0-mode
as the
pipe. The lamb
waveguide,
S0-mode wave
which
lamb wavewas
could excited
was keep in the
the
excited
plate-waveguide
ultrasonic
in as the
sensor away from
the plate-waveguide source,
the
as the and it has
high-temperature
source, been
and it has been proved
pipe.proved that
The S0-mode the S0-wave
that thelamb wavecan
S0-wave can convert
wasconvert
excited to to
thethe
in the
longitudinal
plate-waveguide
longitudinal mode
modeas guided
guided wave
wave in
the source, the pipe
inand
the pipe
it has with
been
with theproved
the T-shaped
T-shaped thatwaveguide
the S0-wave
waveguide and
and tocanto
thethe torsional
convert
torsional to
mode mode
the
guided
guided wave
longitudinal
wavemodewith the L-shaped
guided
with the wave in
L-shaped waveguide
the pipe with
waveguide [16],
[16],the as shown
as T-shaped in Figure
shown in waveguide 2. However,
and to thethe
Figure 2. However, the longitudinal
torsional mode
longitudinal
and
and torsional
guided wavemode
torsional with guided
mode the waves
L-shaped
guided have different
waveswaveguide
have different detection
[16],detection abilities
as shownabilities
in Figure for
for 2. different
However,
different types
types of defects
theoflongitudinal
defects duedue
tototheir
and different
torsional
their different oscillation
mode patterns,
guided waves have
oscillation thus both
different
thus of them
both ofdetection are always
them areabilities
alwaysfor employed
differentfor
employed for
types a comprehensive
of defects due
a comprehensive
pipe
pipe evaluation
to their different
evaluation [6,16–18].
oscillation
[6,16–18]. Therefore,
patterns, in
Therefore, inthis
thus system,
both
this of them
system, ififthe
thetransformation
are always employed
transformation ofoffor
testing wave
wavefrom
a comprehensive
testing fromthe
pipe evaluation [6,16–18]. Therefore, in this system, if the transformation
longitudinal mode to the torsional mode is required, the waveguide will be converted from the T-shaped
the longitudinal mode to the torsional mode is required, the waveguide of testing
will be wave
converted from the
from
longitudinal
tothe
theT-shaped
L-shaped mode thetoL-shaped
tofixed the
on torsional
fixedmode is required,
on the
the high-temperature thewhich
waveguide
high-temperature
pipe, will
pipe,
cannot bebeeasily
which converted
cannot from
be
executed the
easily T-shaped
executed
in practice.
to practice.
in the L-shaped fixed on the high-temperature pipe, which cannot be easily executed in practice.

Figure
Figure 1.
Figure 1. Pipe
1. Pipe inspection
Pipe inspection systemusing
inspection system
system usingaaalong
using longwaveguide
long waveguide
waveguide [16].
[16].
[16].

(a)
(a) (b)(b)
Figure 2. Pipeinspection
Figure inspection system with
with different shaped waveguides based on S0-mode lamb wave (a)
Figure2.2.Pipe
Pipe inspectionsystem
system withdifferent
differentshaped
shapedwaveguides
waveguides based
based on
on S0-mode
S0-mode lamb
lamb wave
wave (a)
longitudinal
longitudinal mode
mode with
with T-shaped
T-shaped waveguide;
waveguide; (b) torsional
(b)(b)
torsionalmode with L-shaped waveguide [16].
(a) longitudinal mode with T-shaped waveguide; torsionalmode
modewith
withL-shaped
L-shaped waveguide [16].
waveguide [16].

As a consequence, based on the already established pipe inspection system, a method to change
As a consequence, based on the already established pipe inspection system, a method to change
As a consequence,
the mode basedin
of the testing wave onthe
thepipe
already established
by replacing thepipe inspection
S0-wave EMATsystem, a methodEMAT
with SH0-wave to change
was
the
themode of
mode of the testing wave in the pipe by replacing the S0-wave EMAT with SH0-wave EMAT waswas
proposed in the
thistesting wave
research, whichin the
waspipe
muchby easier
replacing
thanthe S0-wavethe
removing EMAT
fixedwith SH0-wave
waveguide. AnEMAT
SH0-wave
proposed
proposed ininthis
thisresearch,
research, which
which was much
was mucheasier
easierthan
thanremoving
removingthethe fixed
fixed waveguide.
waveguide. AnAn SH0-wave
SH0-wave
periodic permanent magnet (PPM) EMAT and an S0-wave meander-line coil EMAT were developed,
periodic
periodicpermanent magnet (PPM)
(PPM) EMAT andan anS0-wave
S0-wave meander-linecoil coil EMAT were developed,
then the permanent
experimentsmagnetwere conducted EMAT and
involving the mode meander-line
conversion from SH0 EMAT were
mode developed,
and S0 mode
then
thenthe experiments
thewaves
experiments were conducted
were conducted involving the mode conversion from SH0 mode and S0 mode
guided to the longitudinal andinvolving
torsional the
mode mode conversion
guided waves, from SH0 mode
respectively. In and S0 mode
addition, the
guided
mode conversion behavior for each mode of guided wave was studied by calculating and comparingthe
guided waves
waves to
tothe
the longitudinal
longitudinal and
and torsional
torsional mode
mode guided
guided waves,
waves, respectively.
respectively. In In addition,
addition, the
mode
mode conversion
conversion
the mode behavior
behavior
conversion for
for each
efficiency, frommode
mode
SH0of of
andguided
guided wavewas
S0 to wave wasstudied
longitudinalstudied byby calculating
calculating
and torsional, andand comparing
comparing
respectively.
the
themode
modeconversion
conversionefficiency, fromSH0
efficiency, from SH0and andS0S0totolongitudinal
longitudinalandand torsional,
torsional, respectively.
respectively.
2. Method
2. Method
In order to test a pipe with longitudinal and torsional mode guided waves, based on the previous
In pipe
built orderinspection
to test a pipe with longitudinal
system, a method to and torsional
change modeofguided
the mode waves,
the testing based
wave in on
thethe previous
pipe by
built pipe inspection
changing the EMAT on system, a method
the waveguide to change
without the mode
removing of the
the fixed testing was
waveguide wave in the pipe
proposed. That by
changing the EMAT
is, the T-shaped on the waveguide
waveguide and L-shapedwithout removing
waveguide wouldthenot
fixed waveguide
be switched was
with proposed.
each other, andThat
is,the
theS0-mode
T-shaped EMAT needed and
waveguide to beL-shaped
changed waveguide
to the other would
mode guided
not be wave sensor
switched to each
with complete theand
other,
theswitch between
S0-mode EMAT the longitudinal
needed to bemode and torsional
changed mode,
to the other which
mode was much
guided waveeasier to implement
sensor to completeinthe
Sensors 2016, 16, 1737 3 of 14

Sensors 2016, 16, 1737

an industrial NDE. 3 of 14
Thus, the key was to find another mode guided wave in the plate-waveguide
which could separately convert to the longitudinal and torsional mode guided waves using the
opposite
2. Methodwaveguide of the S0-mode.

In order
2.1. Guided to test
Wave a pipe
into a Pipewith longitudinal and torsional mode guided waves, based on the previous
built pipe inspection system, a method to change the mode of the testing wave in the pipe by changing
The axially symmetric L(0,2) and T(0,1) modes are the most widely-used modes for pipelines [19].
the EMAT on the waveguide without removing the fixed waveguide was proposed. That is, the
The oscillation patterns of the L(0,2) and T(0,1) mode guided waves are shown in Figure 3, in which
T-shaped waveguide and L-shaped waveguide would not be switched with each other, and the
the oscillation
Sensors direction of L(0,2) is parallel to its propagation direction, but for T(0,1) mode,
2016, 16, 1737 3 ofthe
14
S0-mode EMAT needed to be changed to the other mode guided wave sensor to complete the switch
oscillation direction is perpendicular to the propagation [16].
between the longitudinal mode and torsional mode, which was much easier to implement in an
an industrial NDE. Thus, the key was to find another mode guided wave in the plate-waveguide
industrial NDE. Thus, the key was to find another mode guided wave in the plate-waveguide which
which could separately convert to the longitudinal and torsional mode guided waves using the
could separately convert to the longitudinal and torsional mode guided waves using the opposite
opposite waveguide of the S0-mode.
waveguide of the S0-mode.
2.1.
2.1. Guided
Guided Wave
Wave into
into aa Pipe
Pipe
The
The axially
axially symmetric
symmetric L(0,2)
L(0,2)and
andT(0,1)
T(0,1)modes
modesarearethe
themost
mostwidely-used
widely-usedmodes
modesfor forpipelines
pipelines[19].
[19].
The oscillation patterns of the L(0,2) and T(0,1) mode guided waves are shown in
The oscillation patterns of the L(0,2) and T(0,1) mode guided waves are shown in Figure 3, in Figure 3, which
in which
the
the oscillation
oscillation direction
direction of L(0,2)
of L(0,2) is parallel
is parallel to its propagation
to its propagation direction,
direction, but for
but for T(0,1) T(0,1)
mode, themode, the
oscillation
oscillation
direction isdirection is perpendicular
perpendicular to the propagation
to the propagation [16]. [16].

Figure 3. Oscillation patterns of the L(0,2) and T(0,1) mode guided waves in pipes [16].

Then, the dispersion curves of the pipe and the plate-waveguide were built using a numerical
method with MATLAB based on an analysis of the dispersion equations [20,21], as shown in Figures 4
and 5. The relevant parameters of the pipe and waveguide can be obtained from Table 1.

Table 1. The parameters of the pipe and waveguide.

Parameter Pipe Waveguide (Plate)

Inner radius (width of plate)/mm 18.195 10
Figure 3. Oscillation
Figure patterns
3. Oscillation of
patterns
Thickness/mm the
of (a)
the L(0,2)
L(0,2) and
and (b) T(0,1)
T(0,1)
3.22 modemode guided
guided waves
waves in in pipes
pipes
0.6 [16].[16].

Velocity of shear wave/(m/s) 3200 3200

Then, the
the dispersion
Then,Velocity
dispersion curves
curves of
of the
the pipe
pipe and
and the
the plate-waveguide
plate-waveguide were
were built
built using
using aa numerical
numerical
method with MATLABof longitudinal
based on anwave/(m/s)
analysis of the 5940 equations [20,21],5200
dispersion as shown in Figures in
4
method with MATLAB based on an analysis of the dispersion equations [20,21], as shown
and 5. The relevant parameters of the pipe and waveguide can be obtained from Table
Figures 4 and 5. The relevant parameters of the pipe and waveguide can be obtained from Table 1. 1.

Table 1. The parameters of the pipe and waveguide.

Parameter Pipe Waveguide (Plate)

Inner radius (width of plate)/mm 18.195 10
Thickness/mm 3.22 0.6
Velocity of shear wave/(m/s) 3200 3200
Velocity of longitudinal wave/(m/s) 5940 5200

(a) (b)
Figure 4.4.TheThe
Figure dispersion curvescurves
dispersion of group
ofvelocity
group invelocity
the pipe:in
(a)the
longitudinal
pipe: (a)mode and (b) torsional
longitudinal modemode.
and
(b) torsional mode.
Sensors 2016, 16, 1737 4 of 14
Sensors 2016, 16, 1737 4 of 14

(a) (b)
Figure 5.
Figure 5. The
The dispersion
dispersion curves
curves of
of group
group velocity
velocity in
in the
the waveguide:
waveguide: (a) lamb wave
(a) lamb wave and
and (b)
(b) shear
shear
horizontaL(SH) wave.
horizontaL(SH) wave.

If the T(0,1) mode is one of the objective guided waves, the exciting frequency should be below
Table 1. The parameters of the pipe and waveguide.
the cut-off frequency of 500 kHz to eliminate the other interference torsional modes in Figure 4b.
Based on Figure 4a, the L(0,1) mode guided wave may exist
Parameter Pipe during Waveguide
the test even if we try to avoid it
(Plate)
through the design of the EMAT, for it has the same frequency range as the L(0,2) mode. However,
Inner radius (width of plate)/mm 18.195 10
at the frequency of around 100 kHz, the group velocity of
Thickness/mm L(0,2) is much higher
3.22 0.6 than that of L(0,1) so
that we can remove the L(0,1)
Velocity partwave/(m/s)
of shear by reception signal selection
3200 [22]. Therefore,
3200 the exciting frequency
Velocity
was chosen as 100 kHzofand
longitudinal wave/(m/s)
the wavelengths 5940could be computed
of S0 and SH0 5200 as 50 mm and 32 mm,
respectively, from Figure 5, which can help in the design of the EMAT.
If the T(0,1) mode is one of the objective guided waves, the exciting frequency should be below
2.2. Method and Assumption
the cut-off frequency of 500 kHz to eliminate the other interference torsional modes in Figure 4b.
BasedBased
on Figure 4a, the L(0,1)
on Figure 2a,b, mode
it can guided
be noted wave
that the may exist during
oscillation theremains
direction test evenunchanged
if we try towhen
avoidthe
it
through
S0-modethe design
guided of the
wave EMAT,
travels fromforthe
it has the sameto
waveguide frequency
the pipe range as the
no matter L(0,2)
which mode.waveguide
shaped However, at is
the frequency of around 100 kHz, the group velocity of L(0,2) is much higher
used. The reason for generation of different modes of guided waves in the pipe is the differentthan that of L(0,1) so
that we can remove
connection patternstheofL(0,1)
the part by reception
different shapedsignal selection that
waveguides, [22]. is
Therefore, the exciting
the T-shaped and frequency
L-shaped
was chosen asThe
waveguides. 100 kHz and the
T-shaped wavelengths
waveguide of S0 andtoSH0
is parallel thecould
axis ofbe computed
the pipe so as that
50 mmtheand 32 mm,
consistent
respectively, from Figure 5, which can help in the design of the EMAT.
oscillation direction is parallel to the axis and the propagation direction is always along the axis, thus
forming the L(0,2) mode. This is the same for the L-shaped waveguide and T(0,1) mode.
2.2. Method and Assumption
Consequently, based on this phenomenon, it can be assumed that a guided wave whose
oscillation direction
Based on Figure is perpendicular
2a,b, to its
it can be noted thatpropagation in adirection
the oscillation plate can convert
remains to the T(0,1)
unchanged when mode
the
guided wave
S0-mode guidedusing the travels
wave T-shaped waveguide
from or L(0,2)
the waveguide to mode
the pipeusing the L-shaped
no matter which waveguide. That is,
shaped waveguide
theused.
is shearThe
horizontaL(SH) wave is aofpolarized
reason for generation different guided
modes of wave in-plane
guided waves with respect
in the pipetois the
the reference
different
interface [23].
connection Thereofare
patterns themultiple
different symmetric and anti-symmetric
shaped waveguides, SH waveand
that is the T-shaped modes. Thiswaveguides.
L-shaped study used
onlyT-shaped
The the lowest order SH is
waveguide mode, SH0,
parallel towhich
the axisis of
a symmetric
the pipe sodispersionless modeoscillation
that the consistent [24] just like the S0
direction
mode.
is Based
parallel on this
to the axisassumption, experiments
and the propagation were performed
direction involving
is always along the mode
the axis, conversions
thus forming from
the L(0,2)
S0-mode
mode. and
This is SH0-mode
the same fortothe
T-mode and waveguide
L-shaped L-mode, respectively,
and T(0,1) as shown in Figure 6.
mode.
Consequently, based on this phenomenon, it can be assumed that a guided wave whose oscillation
direction is perpendicular to its propagation in a plate can convert to the T(0,1) mode guided wave
using the T-shaped waveguide or L(0,2) mode using the L-shaped waveguide. That is, the shear
horizontaL(SH) wave is a polarized guided wave in-plane with respect to the reference interface [23].
There are multiple symmetric and anti-symmetric SH wave modes. This study used only the lowest
order SH mode, SH0, which is a symmetric dispersionless mode [24] just like the S0 mode. Based on
this assumption, experiments were performed involving the mode conversions from S0-mode and
SH0-mode to T-mode and L-mode, respectively, as shown in Figure 6.
 2016, 16, 1737
Sensors 2016, 1737 5 of 14
Sensors 2016, 16, 1737 5 of 14

Figure 6. The mode conversion experiments to prove the assumption.

Themode
Figure6.6.The
Figure modeconversion
conversionexperiments
experimentstotoprove
provethe
theassumption.
assumption.

3. Experimental System
3.3.Experimental
ExperimentalSystem
System
As shown in Figure 7, the experimental system consists of a function generator and a power
As
Asshown
shownin inFigure
Figure7,7,thetheexperimental
experimentalsystem
systemconsists
consistsofofaafunction
functiongenerator
generatorand andaapower
power
amplifier, which generates the 100 kHz frequency and six volt amplitude electric power with four
amplifier,
amplifier,which
whichgenerates
generatesthe the100
100kHzkHzfrequency
frequencyand andsixsixvolt
voltamplitude
amplitudeelectric
electricpower
powerwithwithfour
four
cycles of burst shape, and is amplified 20 times, then input to an electromagnetic induced coil, a pre-
cycles of burst shape, and is amplified 20 times, then input to an electromagnetic induced
cycles of burst shape, and is amplified 20 times, then input to an electromagnetic induced coil, a coil, a pre-
amplifier and a frequency filter, which amplifies the received signal by 50 dB, and selects the signal
amplifier and and
pre-amplifier a frequency filter,
a frequency which
filter, amplifies
which thethe
amplifies received
receivedsignal
signalbyby
5050
dB,
dB,and
andselects
selectsthe
thesignal
signal
between 50 kHz and 200 kHz, and an oscilloscope and central processing unit (CPU), which collects
between 50 kHz and 200 kHz, and an oscilloscope and central processing unit (CPU),
between 50 kHz and 200 kHz, and an oscilloscope and central processing unit (CPU), which collects which collects
and evaluates the received signal. The two waveguides of 1 m long, 10 mm wide and 0.6 mm thick
and
andevaluates
evaluatesthethereceived
receivedsignal.
signal.The Thetwo
twowaveguides
waveguidesofof11mmlong, long,1010mm
mmwidewideand
and0.60.6mm
mmthick
thick
plate were separately connected around the pipe, and the transmit EMAT and receive EMAT were
plate
platewere
wereseparately
separatelyconnected
connectedaround
aroundthethepipe,
pipe,and
andthethetransmit
transmitEMATEMATand andreceive
receiveEMAT
EMATwere were
placed on the terminal of the waveguides. Figure 8 shows the part of the experimental system which
placed on the terminal of the waveguides. Figure 8 shows the part of the experimental
placed on the terminal of the waveguides. Figure 8 shows the part of the experimental system which system which
consists of the EMAT and the waveguide winding around the test pipe.
consists
consistsofofthe
theEMAT
EMATand andthe
thewaveguide
waveguide winding
winding around
around thethe test
test pipe.
pipe.

Figure
Figure 7. Block diagram
7. Block diagram of
of experimental
experimental system.
system.
Figure 7. Block diagram of experimental system.
Sensors 2016, 16, 1737 6 of 14

Sensors 2016, 16, 1737 6 of 14

Sensors 2016, 16, 1737 6 of 14
Sensors 2016, 16, 1737 6 of 14

Figure 8. The photo of the pipe inspection system.

Figure 8. The photo of the pipe inspection system.

4. Development and Optimization of EMATs
Figure 8. The photo of the pipe inspection system.
4. Development and Optimization of EMATs
Figure 8. The photo of the pipe inspection system.
4.1.
4. Meander-Line
Development Coil
and EMAT for S0 Mode
Optimization Wave
of EMATs
4.1. Meander-Line Coil EMAT for S0 Mode Wave
4. Development and Optimization of EMATs
4.1. The meander-line
Meander-Line coil EMAT
Coil EMAT
The meander-line S0has
forEMAT
coil Modebeen
has widely
Wave
been used
widely used to excite
to excite a lambawavelamb in wave in a plate-waveguide
a plate-waveguide [25]. [25].
In this4.1.
study, permanent
In this study, Coil
Meander-Line magnets
permanent
EMAT (PM)
magnets
for S0 (PM)were
Mode usedtoto
were used
Wave provide
provide a magnetic
a static static magnetic
field. The PMfield.
was The
10 PM was 10
The meander-line
mm in10 length, coil
10 mm EMAT
in widthhas
andbeen
4 mm widely used and
in thickness, to excite a lambfield
the magnetic wave in a at
density plate-waveguide
the surfaceat the [25].
mm in length,The mm in width and 4 mm in thickness, and the magnetic field density surface
wasmeander-line
In this study, permanent
326.4 coil
mT. As shown EMAT
magnets has
(PM)
in Figure been
9, widely
were
the used
used
meander-line toprovide
tocoil excite a lamb
was made wave
a static inwire
a plate-waveguide
magnetic
of copper withfield. The [25].
a diameter PM was
was 326.4 mT.
In this Asmm,
ofstudy, shown
permanent in magnets
Figure 9, thethe
(PM) meander-line
were coilequal
was made theofwavelength
copper wire with 10
a diameter
10 mm 0.315
in length, 10 mmand in
the pitch between
width and 4 mm in used
meander to provide
lines
thickness, and athe
was static
to magnetic
half
magnetic field. The
field of25PM
density at was
mm the surface
of 0.315
mmmm, and 10
inaccording
length, the
to mm
the pitch between
in width
drive and 4the
principle. mmmeander
in thickness, linesandwas equal tofield
the magnetic halfdensity
the wavelength
at the surfaceof25 mm
was 326.4 mT. As shown in Figure 9, the meander-line coil was made of copper wire with a diameter
was 326.4
according to the mT. As shown
drive in Figure 9, the meander-line coil was made of copper wire with a diameter
principle.
of 0.315 mm, and the pitch between the meander lines was equal to half the wavelength of 25 mm
of 0.315 mm, and the pitch between the meander lines was equal to half the wavelength of25 mm
according to the
according to drive principle.
the drive principle.

Figure 9. The structure of meander-line coil EMAT for S0-mode guided wave.

In order to excite a pure S0-mode lamb wave in the waveguide, the meander-line coil EMAT
was optimized. First of all, to produce the best static magnetic field, the configuration of the
permanent magnets above the coil was tested in the free waveguide whose dimensions were presented
in Table 1.Figure
Figure 9. The structure
The transmit EMAT of of meander-line
andmeander-line
receive EMATcoil EMAT
were for S0-mode
separately guided
placed wave.side
onguided
each of the
Figure 9.9. The
The structure
structure of meander-line coilcoil EMAT
EMAT forfor S0-mode
S0-mode guided wave.
wave.
waveguide, and there were three reasonable configurations of the PMs to be tested as shown in Figure
10.order
In The densities
to exciteofathe static
pure magneticlamb
S0-mode field wave
were measured in these three
in the waveguide, theconfigurations
meander-lineusing a
coil EMAT
In order
gauss
tometer,
excitealso
ashown
pureinS0-mode
Figure 10.
lamb wave in the waveguide, the meander-line coil EMAT
In
was optimized. First of all, to produce the best static magnetic field, the configuration EMAT
order to excite a pure S0-mode lamb wave in the waveguide, the meander-line coil of the was
was optimized.
permanent
First of all,the
tocoil
produce the best static magnetic field, the configuration of the
optimized. Firstmagnets above
of all, to produce wasbest
the tested in the
static free waveguide
magnetic whose
field, the dimensions were
configuration presented
of the permanent
permanent
magnets magnets
in Table
above 1. the above
Thecoil wasthe
transmit coilin
EMAT
tested wasthetested
and receive inEMAT
the free
free waveguide waveguide
were
whoseseparately whose dimensions
placedwere
dimensions on each sidewere
presented presented
ofinthe
Table 1.
inThe
Table
transmit EMAT and receive EMAT were separately placed on each side of the waveguide, and the
1. The
waveguide, transmit
and there EMAT
were three and receive
reasonable EMAT
configurations were
of theseparately
PMs to be placed
tested as on
shown each
in side
Figure of
there10.
waveguide, Theand
were densities
three there of the static
were
reasonable threemagnetic
reasonable
configurations fieldconfigurations
of thewere
PMsmeasured in the
of
to be tested these
as PMsthree
shown configurations
to in
beFigure
tested10. using ain Figure
as shown
The densities
10. gauss meter, also shown in Figure 10.
of The densities
the static of the
magnetic static
field
(a) weremagnetic
measured field
in were three
these
(b)
measured in these three
configurations using
(c)
configurations
a gauss meter, using a
also
gauss meter, also
shown in Figure 10. shown in Figure 10.
Figure 10. Three configurations of permanent magnets (PMs) of EMAT: (a) two magnets; (b) four
magnets; and (c) eight magnets.

(a) (b) (c)

Figure 10. Three configurations of permanent magnets (PMs) of EMAT: (a) two magnets; (b) four
magnets;
(a)and (c) eight magnets. (b) (c)
Figure
Figure 10.
10. Three
Three configurations of permanent
configurations of permanentmagnets
magnets(PMs)
(PMs)ofofEMAT:
EMAT:(a)(a)
twotwo magnets;
magnets; (b) (b)
fourfour
magnets; and (c) eight magnets.
magnets; and (c) eight magnets.
Sensors 2016, 16, 1737 7 of 14

Sensors 2016, 16, 1737 7 of 14

The amplitude and signal to noise rate (SNR) value of the S0-wave were computed from the
The signal,
received amplitude
then,and signal
these threetoconfigurations
noise rate (SNR)
werevalue of the to
compared S0-wave were
determine thecomputed fromasthe
best one, just in
received signal, then, these three configurations were compared to determine the best one,
Figure 11. It indicated that the amplitude and SNR could reach the highest values when the 4-PM just as in
Figure 11. It indicated that the amplitude and SNR
configuration was selected from these three candidates.could reach the highest values when the 4-PM
configuration was selected from these three candidates.

The amplitude
Figure 11. The amplitude and
and signal
signal to
to noise
noise rate
rate (SNR)
(SNR) value
value of
of S0
S0wave
wavewith
withdifferent
different configurations
configurations
of PMs shown in Figure 10.

Next, the
Next, the number
number of of turns
turns ofof the
the meander-line coil was
meander-line coil was optimized.
optimized. The
The test
test was
was conducted
conducted while
while
changing the number of turns from four to eleven, and the capacitance, inductance,
changing the number of turns from four to eleven, and the capacitance, inductance, resistance, resistance, and
and
impedance of
impedance each coil
of each coil were
were measured
measured using
using an
an LCR
LCR (inductance,
(inductance, capacitance,
capacitance, resistance)
resistance) meter
meter asas
shown in Table 2. Figure 12 shows the relationship of how the amplitude and
shown in Table 2. Figure 12 shows the relationship of how the amplitude and SNR of the S0 waveSNR of the S0 wave
varied while
varied while the
the number
number of of turns
turns increased.
increased. ItIt also
also shows
shows that
that the amplitude increased
the amplitude increased with
with the
the
increasing number
increasing numberof ofturns,
turns,andandthetheSNRSNR obtained
obtained thethe maximum
maximum value
value whenwhen
the the number
number of turns
of turns was
was equal to
equal to ten. ten.

Table 2. The
Table 2. The capacitance, inductance, resistance,
capacitance, inductance, resistance, and
and impedance
impedance of
of the
the meander-line
meander-line coil.
coil.

Number of Capacitance/ Inductance/ Resistance/ Impedance/

Number of Turns Capacitance/nF Inductance/µH Resistance/mΩ Impedance/Ω
Turns nF H m 
4 822.77 2.7921 289.45 1.7815
54 822.77
652.10 2.7921
3.5233 289.45
318.75 1.7815
2.3468
65 462.17
652.10 4.9716
3.5233 386.84
318.75 3.3027
2.3468
7 371.51 6.1839 439.51 4.1030
6 462.17 4.9716 386.84 3.3027
8 280.44 8.1950 506.31 5.4305
97 371.51
238.35 6.1839
9.6395 439.51
533.51 4.1030
6.3814
108 200.12
280.44 11.481
8.1950 626.32
506.31 7.6005
5.4305
11 176.87 12.989 656.48 8.5942
9 238.35 9.6395 533.51 6.3814
10 200.12 11.481 626.32 7.6005
11 176.87 12.989 656.48 8.5942
Sensors 2016, 16, 1737 8 of 14
Sensors 2016,
Sensors 2016, 16,
16, 1737
1737 88 of
of 14
14

Figure 12.
Figure 12. The
The amplitude and SNR
amplitude and SNR of
of S0
S0 wave
wave with
with aa different
different number
number of turns of
of turns of meander-line
meander-line coil.
coil.

As aaaresult,
As
As result,
result, an
anan EMAT
EMAT
EMAT with
withwith aa 10-turn
10-turn
a 10-turn meander-line
meander-line
meander-line coil
coil andcoil
4-PMand
and 4-PM configuration
4-PM configuration
configuration has been
has been
has been developed
developed
developed
as shown inasasFigure
shown
shown in
9,in Figure
Figure
which has9,the
9, which
which has
besthas theconditions
the
drive best drive
best driveaccording
conditionstoaccording
conditions according toand
to
Figures 11 Figures
Figures 11 and
12. 11 and 12.
12.

4.2. PPM
4.2. PPM EMAT
EMAT for
for SH0-Mode
SH0-Mode Wave
Wave
To generate
To generate an an SH0-mode
SH0-mode guided
guided wave,
wave, the
the periodic
periodic permanent
permanent magnet
magnet (PPM) EMAT [26–28]
(PPM) EMAT [26–28]
was employed
was employed in in this
inthis study.
thisstudy. The
study.The PPM
PPM
The PPM EMAT
EMATEMAT consisted
consisted
consistedof aaof
of racetrack-shaped
racetrack-shaped coil with
coil
a racetrack-shaped with aa magnet
magnet
coil with array
array
a magnet
on top
on topon
array (two
(two columns
top columns
(two columns and several
and several rows rows
rows
and several of alternating
of alternating polarity
polarity
of alternating magnets)
magnets)
polarity as shown
as
magnets) shown in Figure
in
as shown Figure 13. The
13.
in Figure The
13.
racetrack
racetrack
The coilcoil
coil
racetrack hadhad
had aa 55 amm
mm width
width
5 mm widthwith
with
with eight
eight turns
turns
eight ofof
of
turns 0.2
0.2 mm
0.2mm diameter
mmdiameter wire.
diameterwire. The
wire.The permanent
Thepermanent
permanent magnet
magnet
was as
was as long
long as
as half
half the
half the wavelength
thewavelength
wavelengthof ofofthe
the SH0
theSH0
SH0 wave,
wave,
wave, and
andandthethe
the magnetic
magnetic field
field
magnetic density
density
field at the
at
density the surface
at surface was
was
the surface
281.1281.1
281.1
was mT. mT.
mT.

(a)
(a)

(b)
(b)
Figure 13.
Figure 13. (a) The
The layout of
of PPM EMAT;
EMAT; (b) The
The outlook of
of PPM EMAT
EMAT for
for SH0
SH0 wave.
wave.
Figure 13. (a)
(a) The layout
layout of PPM
PPM EMAT; (b)
(b) The outlook
outlook of PPM
PPM EMAT for SH0 wave.

Similar to
Similar to the
the optimization
optimization tests
tests for
for the
the meander-line
meander-line coil
coil EMAT,
EMAT, the
the number
number ofof PPM
PPM rows
rows waswas
optimized to
optimized to obtain
obtain aa good
good SNR
SNR signal.
signal. The
The test
test was
was conducted
conducted while
while changing
changing the
the number
number of
of PPM
PPM
rows from
rows from two
two toto eight,
eight, and
and accordingly,
accordingly, the
the length
length of
of the
the racetrack
racetrack coil
coil had
had to
to be
be changed
changed to
to fit
fit the
the
Sensors 2016, 16, 1737 9 of 14

Sensors 2016, 16, 1737 9 of 14

Similar to the optimization tests for the meander-line coil EMAT, the number of PPM rows was
length of the
optimized to PPM array,
obtain andSNR
a good the capacitance,
signal. The testinductance, resistance,
was conducted and
while impedance
changing the of each coil
number were
of PPM
rows from two to eight, and accordingly, the length of the racetrack coil had to be changed to fit and
measured using an LCR meter as shown in Table 3. Figure 14 shows the result of the amplitude the
SNR ofofthe
length theSH0
PPMwave while
array, andthe
thenumber of PPM
capacitance, rows wasresistance,
inductance, set from two
andtoimpedance
eight. Andofiteach
is easy
coiltowere
find
that when the
measured usingnumber
an LCRof PPM
meterrows is equalintoTable
as shown five, the PPM EMAT
3. Figure can generate
14 shows the resultthe
ofbest
the SNR SH0 wave.
amplitude and
SNR of the SH0 wave while the number of PPM rows was set from two to eight. And it is easy to
Table 3. The capacitance, inductance, resistance, and impedance of the racetrack coil.
find that when the number of PPM rows is equal to five, the PPM EMAT can generate the best SNR
SH0 wave.
Number of the
Capacitance/ nF Inductance/  H Resistance/ m Impedance/ 
PPMTable
Rows3. The capacitance, inductance, resistance, and impedance of the racetrack coil.
2 1305.3 1.9406 494.69 1.3159
Number of
3 the PPM Rows 859.75
Capacitance/nF Inductance/µH
2.9478 Resistance/mΩ
730.01 Impedance/Ω
1.9891
4 2 734.08 1305.3 3.4506
1.9406 831.21
494.69 2.3216
1.3159
5 3 584.11 859.75 2.9478
4.3366 730.01
955.75 1.9891
2.8874
4 734.08 3.4506 831.21 2.3216
6 5 503.73 584.11 5.0169
4.3366 1156.0
955.75 3.3643
2.8874
7 6 429.12 503.73 5.0169
5.9027 1156.0
1271.4 3.3643
3.9403
7 429.12 5.9027 1271.4 3.9403
8 370.45 6.8380 1445.7 4.5329
8 370.45 6.8380 1445.7 4.5329

different number
Figure 14. The amplitude and SNR of SH0 wave with different number of
of PM
PM rows.
rows.

5. Experiment
5. Experimentto
toDetect
DetectaaGuided
Guided Wave
Wave into
into the
the Pipe
Pipe Using
Using the
the S0
S0 and
and SH0
SH0 Waves
Waves
In order
In order toto select
select the
the appropriate
appropriate parameters
parameters for for the
the device,
device, such
such as
as the
the function
function generator
generator and
and
amplifier, as
amplifier, as well
well as
as to
to compute
compute the the mode
mode conversion
conversion efficiency
efficiency later,
later, we
we conducted
conducted an an experiment
experiment
for the
for the S0
S0wave
waveand andSH0SH0wave
wavein in
thethe
freefree
waveguide
waveguidebefore traveling
before to thetopipe.
traveling The transmit
the pipe. EMAT
The transmit
and receive EMAT were separately placed on each side of the waveguide so
EMAT and receive EMAT were separately placed on each side of the waveguide so that it could that it could receive ideal
signals ideal
receive of the S0 wave
signals and
of the S0 SH0
wavewave
and SH0through
waveparameter selectionselection
through parameter as shown as in Figure
shown 15. From
in Figure 15.
Figure 15a,b, the velocities of the S0 wave and SH0 wave were separately computed
From Figure 15a,b, the velocities of the S0 wave and SH0 wave were separately computed to be aboutto be about 5000 m/s
and 3200
5000 m/s m/s, respectively,
and 3200 that agreedthat
m/s, respectively, well with the
agreed wellgroup velocity
with the groupcalculated from the dispersion
velocity calculated from the
curve. Thecurve.
dispersion amplitude of the firstofreceived
The amplitude signal ofsignal
the first received the S0of and SH0
the S0 andwasSH0measured as 12.10
was measured V and
as 12.10 V
13.75 V, respectively.
and 13.75 V, respectively.
Sensors 2016, 16,
Sensors 2016, 16, 1737
1737 10
10 of
of 14
14
Sensors 2016, 16, 1737 10 of 14

(a)
(a) (b)
(b)
Figure 15.
Figure 15. The
15. The received
The received signal
received signal in
signal in the
in the free
the free waveguide:
free waveguide: (a)
waveguide: (a) S0-wave;
(a) S0-wave; (b)
S0-wave; (b) SH0-wave.
(b) SH0-wave.
SH0-wave.
Figure

The
The experiments
experiments were were then
then processed
processed in in the
the mode
mode conversion
conversion fromfrom S0 S0 and
and SH0SH0 to to the
the
The
longitudinalexperiments were then processed in the mode conversion from S0 and SH0 to the longitudinal
longitudinal mode and torsional mode guided waves, respectively, using the opposite waveguide to
mode and torsional mode guided waves, respectively, using the opposite waveguide to
mode
confirm and torsional mode guided waves, respectively, using the opposite waveguide to confirm the
confirm the the previous
previous assumption.
assumption. The The distance
distance between
between the the two
two waveguides
waveguides was was setset at
at 0.5
0.5 m,
m, 11 m,
m,
previous
and assumption. The distance between the16 two waveguides wasexample
set at 0.5 m, 1 m, and 1.5 m to
and 1.5
1.5 mm to
to verify
verify the
the received
received results.
results. Figures
Figures 16 and and 17
17 present
present anan example of of the
the received
received signals
signals
verify
convertedthe received results. Figures 16 and 17 present an example of the received signals converted
converted from from the
the S0
S0 wave
wave andand SH0
SH0 wave,
wave, respectively,
respectively, when when thethe distance
distance between
between two two waveguides
waveguides
from
is the S0 wave and SH0 wave, respectively, when the distance between two waveguides is 0.5 m
is 0.5
0.5 mm and
and the
the length
length of of the
the waveguide
waveguide is is 11 m.
m. The
The velocities
velocities ofof the
the L(0,2)
L(0,2) mode,
mode, L(0,1)
L(0,1) mode,
mode, andand
and
T(0,1)the length
mode areof the waveguide
about 5200 m/s, is 1m/s,
2300 m. The and velocities
3200 m/s, of the L(0,2) mode,
respectively, from L(0,1)
the mode, dispersion
developed and T(0,1)
T(0,1) mode are about 5200 m/s, 2300 m/s, and 3200 m/s, respectively, from the developed dispersion
mode
curves. are about 5200 m/s, 2300 m/s, and 3200 m/s, respectively, from the developed dispersion
curves. WeWe could
could then
then compute
compute the the transmit
transmit time time for
for each
each mode
mode ofof guided
guided wave
wave and
and speculate
speculate which
which
curves.
waveform We could then compute the transmit time for each mode of guided wave and speculate which
waveform was was the
the objective
objective mode
mode as as labeled
labeled in in the
the Figures
Figures 16 16 and
and 17.
17. This
This was
was the
the same
same with
with the
the
waveform
signals wasthe
the objective mode as labeled in thewasFigures 16 and 17. This was the samesignals,
with the
signals when the distance between the waveguides was 1 m and 1.5 m. From the received signals, it
when distance between the waveguides 1 m and 1.5 m. From the received it
signals
can when the distance between the waveguides was 1 m and convert
1.5 m. From the received signals,
can bebe proven
proven that
that the
the SH0-mode
SH0-mode guided
guided wavewave can can successfully
successfully convert to to the
the longitudinal
longitudinal mode mode
it
andcan be proven that the SH0-mode guided wave can successfullyofconvert to the longitudinal mode
and torsional
torsional mode
mode guided
guided waves
waves with
with the
the opposite
opposite waveguide
waveguide of the the S0-wave,
S0-wave, andand thethe method
method of of
and torsional
shifting mode guided wavesmode
with thethe opposite waveguide of the can
S0-wave, and the method of
shifting the
the EMATs
EMATs to to change
change the
the mode of of the guided
guided wave
wave in in the
the pipe
pipe can be be implemented.
implemented.
shifting the EMATs to change the mode of the guided wave in the pipe can be implemented.

(a)
(a) (b)
(b)
Figure
Figure 16.
16. The received signal converted from
from S0 wave when the
the distance between the
the waveguides is
Figure 16.TheThe
received signalsignal
received converted
convertedS0from
wave S0
whenwave distance
when the between
distance waveguides
between theis
0.5
0.5 m:
m: (a)
(a) Longitudinal
Longitudinal wave
wave using
using T-shaped
T-shaped waveguide;
waveguide; (b)
(b) Torsional
Torsional wave
wave using
using L-shaped
L-shaped waveguide.
waveguide.
waveguides is 0.5 m: (a) Longitudinal wave using T-shaped waveguide; (b) Torsional wave using
L-shaped waveguide.
 Sensors2016,
Sensors 16,1737
2016,16, 1737 1111
of of
1414
Sensors 2016, 16, 1737 11 of 14

(a) (b)
(a) (b)
Figure
Figure17.17.
The The
received signalsignal
received converted from SH0
converted wave
from SH0when the when
wave distance between
the thebetween
distance waveguides
the is
Figure
0.5 m: 17. Longitudinal
(a) The received wave
signalusing
converted fromwaveguide;
L-shaped SH0 wave(b)
when the distance
Torsional wave between
using the waveguides
T-shaped waveguide. is
waveguides is 0.5 m: (a) Longitudinal wave using L-shaped waveguide; (b) Torsional wave using
0.5 m: (a) Longitudinal wave using L-shaped waveguide; (b) Torsional wave using T-shaped waveguide.
T-shaped waveguide.
Furthermore, the amplitude of each mode of guided wave could be measured from the received
Furthermore,
signal, then the mode the amplitude
conversionofefficiency
each mode of guided
could wave could
be calculated, which beismeasured
defined as from
the the
ratioreceived
of the
signal, Furthermore, the amplitude of each mode of guided wave could be measured from the received
amplitude of the guided wave in the pipe to the amplitude of the first received signal in theoffree
then the mode conversion efficiency could be calculated, which is defined as the ratio the
signal, then the mode conversion efficiency could be calculated, which is defined as the ratio of
amplitude
waveguide. of the
The guided
average wave in
amplitude the pipe to
of pipe the
the to
guidedamplitude
wave forof the first
thefirst
threereceived signal
distances in the free
the amplitude of the guided wave in the the amplitude of the received signalbetween
in the free the
waveguide.
waveguides The average amplitude of the guided wave for the three distances between the
waveguide. was used. The
The average results of
amplitude ofthe
the mode
guidedconversion
wave for thetothree
the different
distancesmode guided
between waves in the
the waveguides
waveguides
pipe was used. The results of the mode conversion to the different mode guided waves in the
was from
used. the
TheS0-wave
results ofand SH0-wave
the mode are shown
conversion to theindifferent
Figure mode
18. It guided
can be waves
observedin thethatpipe
thefrom
mode
pipe from the
conversion S0-waveS0-wave
from and SH0-wave are shown in Figurewave
18. Ithas canthebe best
observed that the andmode
the S0-wave and the
SH0-wave aretoshown the T(0,1) mode
in Figure 18.guided
It can be observed that the efficiency,
mode conversion the
conversion
efficiency from the S0-wave to the T(0,1) mode guided wave has the best efficiency, and the
from the of the S0-wave
S0-wave to the converting
T(0,1) modetoguidedthe L(0,2)
wave mode
has has
the the
bestlowest efficiency.
efficiency, and theHowever,
efficiencyit ofis quite
the
efficiency
the of
opposite
S0-wave
the
forS0-wave converting
the efficiency
converting result
to the L(0,2) mode
to the
convertedL(0,2)from
mode
has the lowest thehas
SH0thewave,
efficiency.
lowest in efficiency.
However, other However,
it iswords, SH0itwave
theopposite
quite the
is for
quite
is
the
mostopposite for the
likely to convert
the efficiency efficiency result
to the L(0,2)
result converted from converted
mode rather
the SH0 from the SH0
thanin other
wave, T(0,1)wave,
mode.the SH0 wave is most likely to is
words, in other words, the SH0 wave
most likely
convert toto
theconvert to therather
L(0,2) mode L(0,2)than
mode therather
T(0,1)than
mode.the T(0,1) mode.

(a) (b)
(a) of mode conversion to different mode guided waves in
Figure 18. The efficiency (b)the pipe (a) converted
from S0
Figure and
18.18.
The (b) converted from SH0.
efficiency
Figure The efficiencyofofmode
modeconversion
conversion to
to different modeguided
different mode guidedwaves
wavesinin
thethe pipe
pipe (a)(a) converted
converted
from S0 and (b) converted from
from S0 and (b) converted from SH0.SH0.
It is also necessary to compare the mode conversion behaviors between the S0-wave and SH0-
wave It when
is alsothey converttotocompare
necessary the samethe
mode
mode guided wave in
conversion the pipe between
behaviors shown intheFigure 19. Therefore,
S0-wave and SH0-
it can be concluded that the S0-wave has a better efficiency to convert to
wave when they convert to the same mode guided wave in the pipe shown in Figure 19. the T(0,1) mode guided wave
Therefore,
itthan theconcluded
can be SH0-wave,that
while
thethe SH0-wave
S0-wave has is more likely
a better to convert
efficiency to thetoL(0,1)
to convert and L(0,2)
the T(0,1) modemode guided
guided wave
waves than the S0-wave, which is a very important characteristic of the mode conversion
than the SH0-wave, while the SH0-wave is more likely to convert to the L(0,1) and L(0,2) mode guided behavior.
waves than the S0-wave, which is a very important characteristic of the mode conversion behavior.
 Sensors 2016, 16, 1737 12 of 14

It is also necessary to compare the mode conversion behaviors between the S0-wave and SH0-wave
when they convert to the same mode guided wave in the pipe shown in Figure 19. Therefore, it can be
concluded that the S0-wave has a better efficiency to convert to the T(0,1) mode guided wave than the
SH0-wave, while the SH0-wave is more likely to convert to the L(0,1) and L(0,2) mode guided waves
Sensors
than2016, 16, 1737 which is a very important characteristic of the mode conversion behavior.
the S0-wave, 12 of 14

(a) (b)

(c)

Figure
Figure19.
19.The
Theefficiency
efficiencycomparison betweenS0
comparison between S0wave
waveandandSH0
SH0 wave
wave when
when they
they convert
convert to the
to the samesame
mode
modeguided
guidedwave
wavein inthe
the pipe:
pipe: (a) T(0,1)
T(0,1) mode;
mode;(b)
(b)L(0,1)
L(0,1)mode;
mode;(c)(c) L(0,2)
L(0,2) mode.
mode.

InInaddition,
addition,the
the average efficiencyofof
average efficiency thethe different
different mode
mode conversions
conversions couldcould be calculated
be calculated when
when the
the same waveguide was used. For example, the average efficiency of mode conversions
same waveguide was used. For example, the average efficiency of mode conversions including S0 to including
S0T(0,1),
to T(0,1),
SH0 to SH0 to L(0,1),
L(0,1), and SH0 and SH0 to
to L(0,2) canL(0,2) can beasregarded
be regarded as the
the efficiency for efficiency forwaveguide,
the L-shaped the L-shaped
waveguide, and also the average efficiency for the T-shaped waveguide. The comparison
and also the average efficiency for the T-shaped waveguide. The comparison between the efficiency between
the
of efficiency of the twois waveguides
the two waveguides shown in Figure is shown in indicates
20, which Figure 20, which
that indicateswaveguide
the L-shaped that the L-shaped
had a
waveguide had a better efficiency than the T-shaped waveguide no matter which mode guided in
better efficiency than the T-shaped waveguide no matter which mode guided wave was excited wave
theexcited
was waveguide.
in the waveguide.
Sensors 2016, 16, 1737 13 of 14
Sensors 2016, 16, 1737 13 of 14

Figure 20.The
Figure20. Theefficiency
efficiencycomparison
comparisonbetween
betweentwo
twowaveguides.
waveguides.

6.6. Conclusions
Conclusions
InInthis
thisstudy,
study,based
basedononthe
theprevious
previouspipe pipeinspection
inspectionsystem,
system,an anSH0-wave
SH0-wavePPM PPMEMAT
EMATand andanan
S0-wave meander-line coil EMAT were developed and optimized based on
S0-wave meander-line coil EMAT were developed and optimized based on the dispersion curves of the dispersion curves of
the
theguided
guidedwaves.
waves.The Theexperiments
experimentswere weresuccessfully
successfullyperformed
performedon onthe
themode
modeconversion
conversionfrom
fromthe
the
SH0-mode
SH0-modetotothe thelongitudinal
longitudinalmodes
modeswith withthetheL-shaped
L-shapedwaveguide
waveguideand andtotothe
thetorsional
torsionalmode
modewith
with
the
theT-shaped
T-shapedwaveguide.
waveguide.The Theresults
resultsshowed
showedthat thatthe
themode
modeswitch
switchbetween
betweenthe thelongitudinal
longitudinaland
and
torsional
torsionalcan canbebeaccomplished
accomplishedby bythetheshift
shiftbetween
betweenthe theS0-EMAT
S0-EMATand andSH0-EMAT
SH0-EMATwithoutwithoutshifting
shifting
the
thefixed
fixedwaveguide,
waveguide,whichwhichwill
willprovide
providesignificant
significantconvenience
convenienceininthe theprocess
processofofpipe
pipeevaluations.
evaluations.
Moreover,
Moreover,the themode
modeconversion
conversionbehaviors
behaviorswere werestudied
studiedbybycalculating
calculatingand andcomparing
comparingthe theamplitude
amplitude
efficiency.
efficiency.ItItcan
canbebeinferred
inferredthat
thatthe
theS0S0wave
wavehashasa abetter
betterability
abilitytotoconvert
converttotothe
theT(0,1)
T(0,1)mode,
mode,while
while
the
theSH0
SH0wavewavehashasa abetter
betterability
abilitytotoconvert
converttotothe theL(0,1)
L(0,1)mode
modeand andL(0,2)
L(0,2)mode,
mode,andandthe
theL-shaped
L-shaped
waveguide
waveguidehas hasbetter
betterefficiency
efficiencythan
thanthetheT-shaped
T-shaped waveguide.
waveguide.

Author Contributions:Z.S.,
AuthorContributions: Z.S.,H.N.
H.N.and
andR.M.
R.M.proposed
proposedthe theidea;
idea;F.S.
F.S.and
andR.M.
R.M.conceived
conceivedand
anddesigned
designedthe the
experiments; F.S. performed the experiments; F.S. also contributed to writing and editing the manuscript. Z.S., Q.C.
experiments; F.S. performed the experiments; F.S. also contributed to writing and editing the manuscript. Z.S.,
and R.M. guided the research and were charged with the critical revisions of the manuscript; all of the authors
Q.C. andthe
analyzed R.M. guided the research and were charged with the critical revisions of the manuscript; all of the
results.
authors analyzed the results.
Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of Interest: The authors declare no conflict of interest.
References
References
1. Burrows, S.E.; Fan, Y.; Dixon, S. High temperature thickness measurements of stainless steel and low carbon
1. Burrows,
steel S.E.; Fan, Y.; Dixon,
using electromagnetic S. High
acoustic temperature
transducers. NDT Ethickness
Int. 2014,measurements of stainless steel and low
68, 73–77. [CrossRef]
2. carbon
Liu, steel
Z.; Hu, Y.;using
Fan, J.electromagnetic
Longitudinal mode acoustic transducers.patch
magnetostrictive NDTtransducer
E Int. 2014,array
68, 73–77.
employing a multi-splitting
2. Liu, Z.; coil
meander Hu,for Y.;pipe
Fan,inspection.
J. Longitudinal
NDT E mode magnetostrictive
Int. 2016, patch transducer array employing a multi-
79, 30–37. [CrossRef]
3. Duan, W.; Kirby, R. A numerical model for the scattering of elastic30–37.
splitting meander coil for pipe inspection. NDT E Int. 2016, 79, waves from a non-axisymmetric defect in
3. a Duan, W.; Kirby,
pipe. Finite Elem. R. A numerical
Anal. Des. 2015,model for the
100, 28–40. scattering of elastic waves from a non-axisymmetric defect
[CrossRef]
4. in a pipe.
Willey, C.L.;Finite Elem. Anal.
Simonetti, Des. 2015,
F.; Nagy, 100, 28–40.
P.B. Guided wave tomography of pipes with high-order helical modes.
4. Willey,
NDT C.L.;
E Int. 2014,Simonetti,
65, 8–21. F.; Nagy, P.B. Guided wave tomography of pipes with high-order helical modes.
[CrossRef]
5. NDT P.B.;
Nagy, E Int.Simonetti,
2014, 65, 8–21.
F.; Instanes, G. Corrosion and erosion monitoring in plates and pipes using constant
5. Nagy,velocity
group P.B.; Simonetti,
Lamb wave F.; Instanes,
inspection.G. Corrosion
Ultrasonicsand
2014,erosion monitoring
54, 1832–1841. in plates[PubMed]
[CrossRef] and pipes using constant
6. group E.;
Leinov, velocity
Lowe,Lamb wave
M.J.S.; inspection.
Cawley, Ultrasonics
P. Ultrasonic 2014, of
isolation 54,buried
1832–1841.
pipes. J. Sound Vib. 2016, 363, 225–239.
6. Leinov, E.; Lowe, M.J. S.; Cawley, P. Ultrasonic isolation of buried pipes. J. Sound Vib. 2016, 363, 225–239.
[CrossRef]
7.7. Cegla,F.B.;
Cegla, F.B.;Jarvis,
Jarvis, A.J.C.;
A.J. C.; Davies,
Davies, J.O.
J.O.High-temperature
High-temperatureultrasonic
ultrasoniccrack monitoring
crack using
monitoring SH waves.
using NDT
SH waves.
E Int.
NDT E 2011, 44, 669–679.
Int. 2011, 44, 669–679. [CrossRef]
8. Cegla, F.B.; Cawley, P.; Allin, J. High-temperature (>500 °C) wall thickness monitoring using dry-coupled
ultrasonic waveguide transducers. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2011, 58, 156–167.
Sensors 2016, 16, 1737 14 of 14

8. Cegla, F.B.; Cawley, P.; Allin, J. High-temperature (>500 ◦ C) wall thickness monitoring using dry-coupled
ultrasonic waveguide transducers. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2011, 58, 156–167. [CrossRef]
[PubMed]
9. Murayama, R.; Kobayashi, M.; Matsumoto, K. Ultrasonic Inspection System Using a Long Waveguide with
an Acoustic Horn for High-Temperature Structure. J. Sens. Technol. 2014, 4, 177–185. [CrossRef]
10. Hernandez-Valle, F.; Dixon, S. Preliminary tests to design an EMAT with pulsed electromagnet for high
temperature. Am. Inst. Phys. 2009, 1096, 936–941.
11. Hernandez-Valle, F.; Dixon, S. Initial tests for designing a high temperature EMAT with pulsed electromagnet.
NDT E Int. 2010, 43, 171–175. [CrossRef]
12. Tian, G.Y.; Sophian, A. Reduction of lift-off effects for pulsed eddy current NDT. NDT E Int. 2005, 38, 319–324.
[CrossRef]
13. Huang, S.; Zhao, W.; Zhang, Y. Study on the lift-off effect of EMAT. Sens. Actuators A Phys. 2009, 153, 218–221.
[CrossRef]
14. Yu, Y.; Yan, Y.; Wang, F. An approach to reduce lift-off noise in pulsed eddy current nondestructive technology.
NDT E Int. 2014, 63, 1–6. [CrossRef]
15. Morrison, J.P.; Dixon, S.; Potter, M.D.G. Lift-off compensation for improved accuracy in ultrasonic lamb
wave velocity measurements using electromagnetic acoustic transducers (EMATs). Ultrasonics 2006, 44,
e1401–e1404. [CrossRef] [PubMed]
16. Murayama, R.; Matsumoto, K.; Ushitani, K. Pipe Inspection System by Guide Wave Using a Long Distance
Waveguide. Mod. Mech. Eng. 2015, 5, 139–149. [CrossRef]
17. Lee, H.; Park, H.W.; Sohn, H. Pipe Defect Visualization and Quantification Using Longitudinal Ultrasonic
Modes. Int. J. Struct. Stab. Dyn. 2014, 14, 1440008. [CrossRef]
18. Murayama, R.; Weng, J.; Kobayashi, M. Pipe inspection system of a pipe by three-modes guide wave using
polarized-transverse wave EMATs. Proc. SPIE 2014, 9302, 93022T. [CrossRef]
19. Xu, J.; Wu, X.; Kong, D. A Guided Wave Sensor Based on the Inverse Magnetostrictive Effect for
Distinguishing Symmetric from Asymmetric Features in Pipes. Sensors 2015, 15, 5151–5162. [CrossRef]
[PubMed]
20. Zheng, X.M.; Zhao, Y.Z.; Shi, Y.W. Calculation for lamb wave dispersion curves. Nondestr. Test. 2003, 25,
66–68.
21. Xu, X.N.; Yu, Z.J.; Zhu, L.Q. A graphical method to solve a dispersion equation of Lamb wave. J. Electron.
Meas. Instrum. 2012, 11, 9. [CrossRef]
22. Lowe, M.J.S.; Alleyne, D.N.; Cawley, P. Defect detection in pipes using guided waves. Ultrasonics 1998, 36,
147–154. [CrossRef]
23. Murayama, R.; Makiyama, S.; Kodama, M. Development of an ultrasonic inspection robot using an
electromagnetic acoustic transducer for a Lamb wave and an SH-plate wave. Ultrasonics 2004, 42, 825–829.
[CrossRef] [PubMed]
24. Petcher, P.A.; Burrows, S.E.; Dixon, S. Shear horizontaL(SH) ultrasound wave propagation around smooth
corners. Ultrasonics 2014, 54, 997–1004. [CrossRef] [PubMed]
25. Murayama, R.; Mizutani, K. Conventional electromagnetic acoustic transducer development for optimum
Lamb wave modes. Ultrasonics 2002, 40, 491–495. [CrossRef]
26. Andruschak, N.; Saletes, I.; Filleter, T. An NDT guided wave technique for the identification of corrosion
defects at support locations. NDT E Int. 2015, 75, 72–79. [CrossRef]
27. Petcher, P.A.; Dixon, S. Weld defect detection using PPM EMAT generated shear horizontal ultrasound.
NDT E Int. 2015, 74, 58–65. [CrossRef]
28. Hill, S.; Dixon, S. Frequency dependent directivity of periodic permanent magnet electromagnetic acoustic
transducers. NDT E Int. 2014, 62, 137–143. [CrossRef]

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).