Ultrasonics 49 (2009) 202–205

Contents lists available at ScienceDirect

Ultrasonics
journal homepage: www.elsevier.com/locate/ultras

Experimental study of the A0 and S0 Lamb waves interaction

with symmetrical notches
Farouk Benmeddour *, Sébastien Grondel, Jamal Assaad, Emmanuel Moulin
IEMN, UMR/CNRS 8520, OAE Department, Université de Valenciennes et du Hainaut Cambrésis, Le Mont Houy 59313 Valenciennes Cedex 9, France

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this work is to study the fundamental Lamb modes interaction with defects in isotropic plates.
Received 22 July 2007 For these experimental investigations, symmetrical notches with various depths milled in aluminum
Accepted 7 August 2008 plates are considered. Moreover, the incident Lamb wave of a specific mode is generated by means of
Available online 15 August 2008
two identical thin piezoceramic transducers placed at the opposite sides of the plate. The waves scattered
by the notch are recorded with conventional transducers located on the plate surface in front and behind
Keywords: the defect. The selection of the A0 or the S0 modes is obtained by exciting the transducers with anti-
Lamb waves
phased or in-phased signals, respectively. Furthermore, a calibration process is investigated to correct
NDT
Notches
errors caused by the presence of the receiver between the emitters and the defects. The power reflection
and transmission coefficients are then obtained and the power balance is verified. Finally, these measure-
ments are compared successfully with those obtained by a numerical method using the finite-element
modeling described in a previous work.
Ó 2008 Elsevier B.V. All rights reserved.

1. Introduction Generally, discontinuities occur reflections and mode conver-

Techniques based on guided ultrasonic Lamb waves have
proved their capability for the health monitoring of plate-like
structures. The principal advantage is that these waves can carry-
ing out energy over long distances and detect internal defects by
interrogating the entire thickness of the structure. However, some
difficulties are confronted in the interpretation of the received
experimental signals because of the multimode and dispersive nat-
ure of Lamb modes.
In the past, a great deal of work has been investigated concern-
ing the interaction of Lamb waves with damage and discontinuities
[1]. The excitation and reception of a specific Lamb mode is per-
formed generally by using the coincidence principle as mentioned
in the literature [2]. However, this technique is not adapted for
integrated health monitoring systems. To solve this problem, a
few studies propose to work with a comb transducer at the emis-
sion to select and isolate a Lamb mode in an elastic isotropic plate
[3]. In the same way, Grondel et al. [4] have used thin piezoceramic
transducers to generate and receive a selected Lamb mode in com-
posite structures. These thin transducers can be bounded or inte-
grated in the structure [5] to ensure a permanent and a real time
health monitoring.
* Corresponding author. Fax: +33 327511189.
E-mail address: Farouk_Benmeddour@yahoo.fr (F. Benmeddour).
sions. The separation of the Lamb waves packets is not easy and
therefore requires appropriate signal processing techniques. For
example, El Youbi et al. [6] and Rguiti et al. [7] have used multi-ele-
ment and multi-electrode thin piezoceramic transducers, respec-
tively, to be able to apply the two-dimensional Fourier transform
(2D-FT) to received signals in order to isolate and study the sensi-
tivity of each Lamb mode to the encountered holes.
The aim of this work is the development of a simple experimen-
tal device to study the interaction of the fundamental Lamb modes
with Symmetrical Notches (SN). The work deals essentially with a
calibration process applied to the received signals in order to filter
and correct wave packet’s amplitudes. A selected fundamental
Lamb mode is launched with two thin piezoceramic transducers
placed on the opposite sides at the plate edge. The A0 and S0 modes
are selected and generated by exciting these transducers with anti-
and in-phased electrical signals, respectively. The studied fre-
quency of the fundamental Lamb modes is taken under the cut-
off frequency of the A1 and S1 modes. Section 2 presents the exper-
imental set-up, the measurements procedure and the calibration
process. Section 3 deals with the experimental results of the inter-
action of a fundamental Lamb mode with symmetrical notches.
The power reflection and transmission coefficients are computed
to characterize the behavior of the fundamental Lamb modes with
symmetrical notches. Finally, a comparison between the experi-
mental and numerical results obtained in a previous work [1] is
given.

0041-624X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.ultras.2008.08.002
 F. Benmeddour et al. / Ultrasonics 49 (2009) 202–205 203

2. Experimental set-up
2.1. General description
Fig. 1 illustrates the instrumentation used in the experimental
investigation to study the fundamental Lamb modes interaction
with symmetrical notches. A pulse generator (HM 8035) is used
to trigger simultaneously the oscilloscope (Le Croy type LT344)
and the two arbitrary function generators (HP 33120A) which
can deliver a tone burst modified by a Hanning window function
to the emitter. The emitter is composed of two identical piezoce-
ramic transducers (PZT-27) placed at the opposite sides of the left
plate edge. These transducers are 1 mm thick to be thin, 50 mm
long to be unidirectional [8] and 6 mm large to be close to the free
mechanical lateral resonance-frequency of 200 kHz.
The application of anti-phased electrical signals to the two piez-
oceramic transducers generates the A0 mode, whereas the in-
phased excitation generates the S0 mode. Two conventional trans-
ducers (Panametrics) are chosen in reception for their easy han-
dling and are placed on the plate surface before and after the
notch at the desired locations. These transducers are more sensi-
tive to normal displacements than tangential ones. Local honey
and gel coupling are employed for the emitters and the receivers,
respectively. All the signals from the sensors are recorded using
the digital oscilloscope and transferred via the GPIB bus to a com-
puter for signal processing.
The six aluminum plates used in the experimental investiga-
tions have the following dimensions: 6 mm thick, 300 mm wide
and 500 mm long. Moreover, all these plates except one contain
50 mm wide notches, milled across the full plate width normal
to the transmission path. The tested plates and notches have thick-
nesses equal to 2d and 2dp, respectively, where d is the half thick-
ness of the plate and p takes values from 0 to 1 with a constant
increment. The notches depths are equal to 0.5, 1, 1.5, 2 and
2.5 mm which corresponds to the ratios p = 5/6, 4/6, 3/6, 2/6, and
1/6, respectively. The desired notches width and depths are ob-
tained with a conventional milling machine.
and verify the Shannon sampling theorem. The acquired signals are
filtered by an implemented averaging method in the oscilloscope.
In fact, 50 signals are acquired in few seconds and averaged in or-
der to get a high signal to noise ratio (SNR). Errors between these
signals are less than 0.5%. Then, the result is transmitted to the
computer for recording and signal processing. Furthermore, in or-
der to verify the repeatability process and the gel coupling effect,
the receiver is removed and replaced at the same location 20 times.
Hence, an error bar can be computed for each experience. Fig. 2b
shows the average of the 20 electrical signals measured with the
transducer (C) at a distance of l1 = 185 mm (Fig. 2a) from the left
edge of the healthy plate when the A0 mode is launched whereas
Fig. 2c illustrates the case of the S0 mode. In each case, the selection
of one Lamb mode is successful, which validates the experimental
setup. Since the S0 mode has a higher group velocity than the A0
mode, its reflection on the right edge of the plate (re0 ) is visible.
2.3. Calibration process
Experimental measurements are carried out to study the recei-
ver presence effect on the transmitted wave packets. Therefore, an
identical conventional receiver (C0 ) is placed between the emitter
and the conventional effective receiver (C) as depicted on Fig. 3a.
Hence, Figs. 3b and c depict measured signals when the A0 and S0
modes are launched, respectively. It can be shown that the trans-
mitted signal measured with the effective receiver (C) is disturbed.
(a) without receiver C’
C
(b) A launched (c) S launched
0 0
0.4 0.4
in (A0)
Amplitude (V)

0.2 0.2 in (S )
0 re’ (S )
0
2.2. Measurement procedure 0 0

The arbitrary function generators are programmed to excite the −0.2 −0.2
two thin transducers with in-phased and anti-phased signals. −0.4 −0.4
These signals correspond to 5 sinusoidal cycles at a frequency of 0 100 200 0 100 200
Time (µs) Time (µs)
200 kHz windowed by a Hanning function. Then, the oscilloscope
is programmed to acquire 2000 temporal points at a sampling fre- Fig. 2. Experimental configuration for measurements in a healthy plate (a) and
quency of 10 MHz which corresponds to 0.1 ls between each point measured signals when the A0 (b) and S0 (c) modes are launched.

(a) with receiver C’

C’ C
Computer

Pulse
generator Oscilloscope
200.000,00 kHz

Function (b) A0 launched (c) S launched

200.000,00 kHz 0
generators 0.4 0.4
Amplitude (V)

Fixation in (A0)
Receivers 0.2 0.2 in (S0)
device re’ (S )
C 0
C 2dp 2 0 0
Honey 1
l1 l2 2d=6 mm
−0.2 −0.2
SN −0.4 −0.4
Emitter 0 100 200 0 100 200
Transducers Aluminium plate Time (µs) Time (µs)

Fig. 1. Experimental device to study the interaction of Lamb modes with symmet- Fig. 3. Experimental configuration for the calibration process (a) and measured
rical notches. signals when the A0 (b) and S0 (c) modes are launched.
204 F. Benmeddour et al. / Ultrasonics 49 (2009) 202–205

In fact, the presence of the conventional transducer (C0 ) along the
path of the propagation can be considered as a surface perturba-
tion because of its dimensions (39  17  15 mm). Hence, Fig. 3b
show that the amplitude of the A0 mode (Fig. 2b) decreases to
about 40%. Furthermore, its shape is modified by the multiple
reflections from the edges of the receiver. Contrarily to this, the
S0 mode (Fig. 3c) decreases only to about 5% (Fig. 2c).
In practice, when a fundamental Lamb mode is launched from
the left edge of the plate, the incident wave packet encounter the
conventional receiver C1 placed at l1. Then, the disturbed transmit-
ted wave packet is reflected by the left edge of the notch and re-
ceived by C1. Hence, the measured reflection wave packet must
be corrected by multiplying its amplitude by 40% when the A0
mode is launched whereas the amplitude of the reflection wave
packet is multiplied by only 5% when the S0 mode is launched.
For the transmitted wave packet after the notch, the first recei-
ver C1 is taken off. Hence, the transmitted amplitude does not need
any correction. However, in order to take into account implicitly
the dispersion, the diffraction and the attenuation effects, the inci-
dent wave packet is acquired on a healthy plate with C2 (Fig. 4a) at
the same location l2 (see Eq. (2)) as the transmitted wave packet in
a plate containing a SN (Fig. 4b).
The power reflection and transmission coefficients are com-
puted at the central working frequency which corresponds to the
maximum of the Lamb wave packet envelope as
 2
e 
 S nðreÞ ðl1 ; dÞ
RExp
1;n ¼  ;
eS nðinÞ ðl1 ; dÞ
and
 2
e 
 S nðtrÞ ðl2 ; dÞ 
T Exp
2;n ¼   ;
eS nðinÞ ðl2 ; dÞ
where the exponent Exp denotes experimental measurement, S the
received signal, l1 and l2 are the measurement distances from the
left edge of the plate. The subscriptions in, re and tr denote incident,
reflected and transmitted wave, respectively, and n is the Lamb
mode (A0 or S0). Subscriptions 1 and 2 denote that the measurement
is performed before and after the notch, respectively. The tilde sym-
bol (e) indicates that the corresponding value is determined at the
central frequency by taking the maximum of the wave packet
envelope.
3. Experimental results
Figs. 5a and b display the electric signals measured before and
after the notch, respectively, when the A0 mode is launched for a
plate containing a 2.5 mm (p = 1/6) deep SN. In the Fig. 5a, the first
and the second wave packets correspond to the incident (in) and
the first reflection (re1) from the notch of the A0 mode. Further-
more, multiple reflections from the edges of the notch and the
plate are observed. However, Fig. 5b displays the second transmis-
sion (tr2) after the SN of the A0 mode.
It is important to precise that by means of the flight time and
the shape observation of each wave packet, there is no significant
mode conversion from the incident A0 to the S0 mode and from the
incident S0 to the A0 mode.
Since the reflection and transmission Lamb modes wave packets
are identified, quantified and corrected, power reflection and
transmission coefficients can be computed.
The power balance is determined as described in a previous
work. Hence, the incident mode generates [1] a first reflection
(re1) and a first transmission (tr1) waves from the first edge of
the SN. Then, the first transmitted wave (tr1) generates a second
reflection (re2) and a second transmission (tr2) waves from the sec-
ond edge of the notch. Consequently, the power balance can be ex-
pressed as
R1;n þ R2;n þ T 2;n ¼ 1; ð3Þ
where R1 is the first power reflection, R2 the second power reflec-
ð1Þ
tion and T2 the second power transmission coefficients.
Figs. 6a and b show power reflection and transmission coeffi-
cients comparisons between numerical [1] and experimental re-
sults of the interaction of the A0 and S0 modes, respectively, with
ð2Þ different SN depths. Several experimental measurements (20) are
performed to determine the displayed error bars as mentioned in
the previous section. Good agreement is found for both A0 and S0
cases. Therefore, by analogy to numerical results, the power bal-
ance is verified. Moreover, Figs. 6a and b show that the power
reflection coefficient of the A0 mode is greater than of the S0 mode.
a b
0.4 0.4
in (A )
0
Amplitude (V)

0.2 re1 (A0) 0.2

The above equations are valuable only for a selected fundamen- tr (A )
2 0
tal Lamb mode in non-dispersive frequency range taken under the 0 0
cut-off frequencies of the A1 and S1 modes and when the incident,
the reflected and the transmitted signals are acquired on the sur- −0.2 −0.2
face of the same plate thickness (2d). However, when the launched
wave packet is composed of the A0 and S0 Lamb modes, they must −0.4 −0.4
0 100 200 0 100 200
be separated before the computation of the power coefficients by Time (µs) Time (µs)
applying arithmetic operations (addition and subtraction) to the
acquired signals at the same location on the opposite sides of the Fig. 5. Electric signals measured in a plate when the A0 mode interacts with a
plate. 2.5 mm (p = 1/6) deep SN at l1 = 185 mm (a) and l2 = 315 mm (b).

a C
2 a b
1 T (A ) 1
l2 2 0 T2 (S0)
R and T

0.5 0.5
b C
2
R1 (A0) R (S )
1 0
l2 0 0
1/6 2/6 3/6 4/6 5/6 1/6 2/6 3/6 4/6 5/6
p p

Fig. 4. Experimental configuration of the calibration process for the transmission Fig. 6. Comparison between numerical and experimental results when the
region of a: healthy plate (a) and plate containing a notch (b). launched Lamb mode A0 (a) or S0 (b) interacts with a SN.
 F. Benmeddour et al. / Ultrasonics 49 (2009) 202–205
Hence, the A0 mode is more sensitive to this kind of discontinuities
than the S0 mode.
4. Conclusion
In this work, an experimental device is built to launch an indi-
vidual A0 or S0 mode and to study their sensitivity to symmetrical
notches. A calibration process is performed to get a high SNR and to
correct signals amplitude due to the presence of the conventional
receiver. Different symmetrical notches depths are investigated.
Then, the power reflection and transmission coefficients are com-
puted with an acceptable error. The technique allows a faster anal-
ysis and a direct comparison between the numerical and the
experimental results. At the end, the experimental measurements
are confronted with success to the numerical predictions [1]. It is
shown that on one hand, the power balance is verified and on
the other hand, the A0 mode is more sensitive to the symmetrical
notches than the S0 mode by the consideration of the power
reflection coefficients.
205
References
[1] F. Benmeddour, S. Grondel, J. Assaad, E. Moulin, Study of the Fundamental Lamb Modes
Interaction with Symmetrical Notches, NDT & E International 41 (1) (2008) 1–9.
[2] D.N. Alleyne, P. Cawley, A two-dimensional Fourier transform method for the
measurement of propagating multimode signals, Journal of the Acoustical
Society of America 89 (3) (1991) 1159–1168.
[3] J.L. Rose, S.P. Pelts, M.J. Quarry, A comb transducer for mode control in guided
wave NDE, IEEE Ultrasonics Symposium (1997) 1033–1036.
[4] S. Grondel, C. Paget, C. Delebarre, J. Assaad, K. Levin, Design of optimal
configuration for generating A0 Lamb mode in a composite plate using
piezoceramic transducers, Journal of the Acoustical Society of America 112 (1)
(2002) 84–90.
[5] E. Moulin, J. Assaad, C. Delebarre, Modeling of Lamb waves generated by
integrated transducers in composite plates using a coupled finite element-
normal modes expansion method, Journal of the Acoustical Society of America
107 (1) (2000) 87–94.
[6] F. El-Youbi, S. Grondel, J. Assaad, Signal processing for damage detection using
two different array transducers, Ultrasonics 42 (2004) 803–806.
[7] M. Rguiti, S. Grondel, F. El-Youbi, C. Courtois, M. Lippert, A. Leriche, Optimized
piezoelectric sensor for a specific application: detection of Lamb waves, Sensors
and Actuators A 126 (2006) 362–368.
[8] E. Moulin, N. Bourasseau, J. Assaad, C. Delebarre, Directivity of integrated
piezoelectric Lamb wave sources, IEEE Ultrasonics Symposium 107 (2001)
1081–1084.