Aspects of Lamb-Wave Generation

and Transmission
Nicolae Constantin, S tefan Sorohan, Mircea G avan and Viorel Anghel
The University Politehnica of Bucharest, Laboratory for integrity evaluation of composite structures, Splaiul
Independent ei 313, RO-060042 Bucharest, Romania
(Received 11 August 2009; accepted 27 October 2009)
This paper presents numerical and experimental results concerning Lamb-wave actuation and reception in metallic
or layered composite panels and pipes. The numerical simulations were carried out with LS-Dyna, run under
ANSYS platform. The main goals of this approach were to nd better models for actuating Lamb waves and
tracking the wave transmission in order to get efcient in situ structural health monitoring. Prior adequate choice
of the signal frequency using the dispersion curves proved to be benecial for better transmission, in terms of signal
detection, and it extended inspection range. The effects of the transducer-inspected item interface and transducer
shape on the wave transmission were also evaluated. Signal generation with rectangular transducers, batteries of
round piezoceramic transducer (PZT) patches or the use of prisms, which value Snells law, were alternatively
considered for getting guided waves in the inspected item. The experiments showed the performance of various
transducers and devices used on the experimental chain. The Lamb waves were generated by PZT, in various shapes
and congurations, aiming to lower the inspection cost. Experiments always aligned with theoretical results so that
robust models could be selected for further investigations.
1. INTRODUCTION
Composite materials are a widespread class of advanced
materials, with applications in various elds, from everyday
use products to hightech structures. The special demands
concerning reliability of the latter items require sound non
destructive evaluation (NDE) methods for monitoring their
structural health. The Lamb-wave method has been applied in
recent years for inspecting composite panels and shells, and re-
searchers hope to extend a rather short list of available, trusted
global inspection methods. The promising advantages of the
method, mainly the possibility of combining long-range in-
spections with local inspections and the versatility that encour-
ages in situ structural health assessment, are balanced for the
moment by a number of difculties yet to be overcome.
Various aspects concerning the generation, transmission,
and reception of Lamb waves have already been investigated
in a number of works by reputable scientists. The optical
methodshighpower, ultrasonic contactless loudspeakers and
classicalcontact ultrasonic transducershave been used in var-
ious applications in past decades.
13
Piezoceramic transducers
of different types have become dominant in the applications
mentioned in literature in the last 15 years.
47
The generally
low cost and versatility in mounting on the inspected item, in-
cluding within a network or measurement chain, still support
their widespread use. The shape of the inspected structures
created extra difculties, composite pipes proving much more
difcult for any transducer to inspect than composite plates
or shells, due to a much higher diversity of wave modes.
8
Other difcult case studies, like the inspection of sandwich
type structures, repaired structures, or ribstiffened composite
structures, have been carried out adequately.
9, 10
The necessary use of adequate actuation for obtaining cer-
tain modes, such as long-range transmission and sensitive in-
teraction with aws or damages, was another highly valued as-
pect many researchers followed. This issue was generally fol-
lowed with adequate numerical modeling of Lambwave gen-
eration, transmission, and receiving.
913
In recent years, sev-
eral authors performed a more complex analysis, using a mul-
tiphysics approach.
1217
A way to simplify such complex and
timeconsuming simulationshifting to structuralonly mod-
els is presented in Ref.
17
This work attempts to contribute ideas concerning the pos-
sibilities of better simulating generation and transmission of
Lamb waves, considering the shape of the transducer, the in-
terface with the inspected item, and consequent modeling tech-
niques. It also outlines the necessity of having a prior selec-
tion of frequencies in the leastdispersive ranges revealed by
the dispersion curves in order to get higherquality signal gen-
eration and transmission. Our experiments further supported
the attempt to build robust numerical models by evaluating the
Lambwave generation and transmission, produced by various
transducers and devices.
2. MODELING OF LAMB-WAVE
PROPAGATION
2.1. Introductory Theoretical Aspects
Guided waves can usually be actuated using traditional or
modied ultrasonic transducers and a large class of piezo-
electric transducers. Such transducers can range from rather
expensive patches or wafers to cheap circular or rectangular
buzzers. Mounting these transducers on the surface of the in-
spected shell by stiff bonding leads to complex actuation and
phenomena reception, which have to be simulated by a multi-
physics approach.
1217
Such additional difculties are skipped
in the case of a removable transducer, having a weak interface
with the inspected item.
Lambwave generation and propagation can be modeled us-
ing a piezoelectricstructural coupling nite-element analysis.
Piezoelectric modeling consists of the coupling of structural
and electric elds, exploiting the natural material properties of
quartz and ceramics. A voltage difference applied to a piezo-
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010 (pp. 39)
3
Nicolae Constantin, et al.: ASPECTS OF LAMB-WAVE GENERATION AND TRANSMISSION
electric material creates a displacement, and vibrating a piezo-
electric material generates a voltage difference.
The electromechanical constitutive equations describing the
linear material behavior are:
18
{T} = [c]{S} [e]{E}; {D} = [e]
T
{S} + []{E}, (1)
where T is the stress vector, D is the electric ux density vec-
tor, S is the strain vector, E is the electric eld vector, c is the
elasticity matrix, e is the piezoelectric stress matrix and is the
dielectric matrix (evaluated at constant mechanical strain).
Using the variational principle, it is possible to derive a
secondorder, timedependent system of equations that can be
discretized using FEM, and that include the piezoelectric ef-
fect:
19
_
[M] [0]
[0] [0]
_ _
{ u}
{

V }
_
+
_
[C] [0]
[0] [0]
_ _
{ u}
{

V }
_
+
_
[K] [K
z
]
[K
z
]
T
[K
d
]
_ _
{u}
{V }
_
=
_
{F}
{L}
_
, (2)
where the submatrices and vectors used are M-structural mass
matrix, C-structural damping matrix, K-structural stiffness
matrix, K
Z
-piezoelectric stiffness matrix, K
d
-dielectric coef-
cient matrix, F-applied nodal force vector, L-applied nodal
charge vector, u-displacement vector, and V -electric potential
vector. The dot and double dots denote derivatives with respect
to time. To integrate this system, a full transient analysis using
the Newmark method can be performed using the ANSYS 7.0
code.
20
For a structural-only analysis, the equation of motion is ob-
tained from Eq. (2):
[M]{ u} + [C]{ u} + [K]{u} = {F}. (3)
Equation (3) can be again integrated using an implicit inte-
gration method. Moreover, it can also be integrated using the
explicit technique with the LS-Dyna module from the ANSYS
code which, for a short time analysis, is more efcient.
If damping and the applied forces are neglected in a struc-
tural analysis, the equation of motion (3) is transformed in an
equation describing the free vibrations:
[M]{ u} + [K]{ u} = {0}, (4)
where, in general, the geometric stiffness matrix [K

], due to
the initial stress distribution in the structure, is included in the
stiffness matrix of the structure [K]. The mass and stiffness
matrices are symmetric and constant. At this stage, the bound-
ary conditions are considered imposed in Eq. (4). Its solution
is in the form
{U} = {}e
it
, (5)
where {} is the mode shape, independent of time, and is
the circular eigenfrequency.
Including the solution and its second derivative into Eq. (4)
yields
(
2
[M] + [K]){} = {0}. (6)
This generalized eigenvalue problem has n pairs of eigenval-
ues
2
j
and associated eigenvectors {
j
}. Usually, the eigen-
vectors are normalized with respect to the mass matrix, orthog-
onal with respect to the mass and stiffness matrices, and they
are arranged in the ascending order of the eigenvalues.
Figure 1. Lamb mode obtained using modal analysis and adequate boundary
conditions.
Figure 2. Main dispersion curves obtained for the composite plate.
Beside the mode shape, the usual parameter in engineering
practice is the eigenfrequency:
f
j
=

j
2
. (7)
The smallest eigenfrequency is the fundamental one. If the
structure has an inside mechanism or rigid body motion, one
obtains null eigenfrequencies corresponding to these congu-
ration displacements. If the structure has geometric symmetry,
then coincident eigenfrequencies may be obtained.
The mode shapes obtained from the modal analysis, by solv-
ing Eq. (6), may be particular Lamb modes (or guided waves),
and the corresponding frequencies in Eq. (7) are the Lamb
mode frequencies if Eq. (6) compels with the boundary con-
ditions of Lamb modes (Fig. 1), making it possible to obtain
associated dispersion curves. In order to get the Lamb modes,
each pair of the symmetric nodes from the two ends of the
model was coupled. For example, considering the symmet-
ric left and right nodes n
L
and n
R
, the horizontal and vertical
4
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010
Nicolae Constantin, et al.: ASPECTS OF LAMB-WAVE GENERATION AND TRANSMISSION
Figure 3. Lamb modes in the composite plate at 200 kHz, obtained with a
nite-element model under plane strain assumption.
Figure 4. Equivalent PZT-substrate interaction considered in an SOM during
the generation of Lamb waves.
a) b) c)
Figure 5. Interaction forces generated by PZT wafers with different geome-
tries: a) Square 10x10 mm
2
, b) Circular 10 mm
2
, c) Rectangular 20x5 mm
2
.
DOFs are always linked: UX(n
L
) = UX(n
R
) and UY(n
L
) =
UY(n
R
). With these restrictions, the mode shapes obtained
from a modal analysis correspond to Lamb modes.
The existence of damping properties in a polymeric lami-
nate, as well as the inaccuracy of the nine engineering con-
stants required by all commercial codes for advanced modeling
of such materials, may alter the nal results.
2.2. Lamb Waves Propagating in Plates
Numerical and experimental analysis of wave actuation and
transmission was conducted on a glass/epoxy laminate with
2.56 mm-average thickness and consisting of eight roving
fabric reinforced layers. The engineering constants are pre-
sented in Table 1. All layers have parallel warp and ll
yarns oriented in the reference/principal directions 1 and 2.
Actuation frequencies were chosen after obtaining the disper-
sion curves (Fig. 2) in the warp direction, using Eq. (eq.6) and
boundary conditions according to Fig. 1, by modal analysis.
21
Robust FE modeling of Lambwave transmission has to ob-
serve some meshing rules: uniform mesh, with at least eight
elements on a wave length.
20, 22
As displacement distribution
Table 1. Engineering constants of the composite plate.
Young modulus Poisson ratio Shear modulus Mass density
[GPa] (minor) [GPa] [kg/m
3
]
E
1
= 13.94
21
= 0.32 G
12
= 4 1530
E
2
= 13.94
32
= 0.18 G
23
= 2.5 1530
E
3
= 7.79
31
= 0.18 G
13
= 2.5 1530
(a)
(b)
(c)
Figure 6. Total displacement eld USUM [m] at t = 40 s: (a) for the square
PZT (scale factor = 50,000), (b) for the circular PZT (scale factor = 50,000),
and (c) for the rectangular PZT (scale factor = 20,000).
(a)
(b)
Figure 7. Total displacement eld USUM [m] at t = 40 s: (a) for the normal
beam (scale factor = 2,000) and (b) for the angle beam (scale factor = 5,000).
along wave propagation is not constant or linear, even for fun-
damental A0 and S0 modes (Fig. 3), the use of plate or shell
type nite elements has to be restricted to low wave frequency,
unless higherorder plate elements are available (but with an
additional computingtime penalty). Accordingly, 2D or 3D
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010
5
Nicolae Constantin, et al.: ASPECTS OF LAMB-WAVE GENERATION AND TRANSMISSION
(a)
(b)
Figure 8. Main dispersion curves obtained for the composite tube.
(a)
(b)
Figure 9. Total displacement eld [m] at t = 80 s for excitation with (a)
rectangular PZT and (b) square PZT, both stiff-bonded; scale factor = 100,000.
(a)
(b)
Figure 10. L(0,1) and L(0,2) modes, excited with four square 8x8 mm
2
PZTs,
with in-phase 50 kHz signals (left) and in-phase 125 kHz signals (right); scale
factor = 100,000.
solid elements are recommended, with appropriate meshing,
both on the plate thickness and in the propagation direction
(Fig. 3).
In order to simplify the FE models, the multiphysics models
were reduced to structuralonly models (SOMs), in which the
whole PZTsubstrate interaction is simulated by the resultant
forces applied to the PZT contour on the substrate (Fig. 4).
17
The vertical force, coming from outofplane interaction, was
neglected in this analysis, as such simplication was found not
to change the nal results.
17
Consequently, Lambwave ac-
tuation and propagation analysis could be done with SOM on
only onequarter of a doubly symmetric virtual plate, having
the actuating transducer in its centre.
The actuation function for any implied quantity (voltage,
pressure, or force) was generated through a Hanning window:
p(t) =
_
1
2
P
0
_
1 cos
_

0
t
n
0
__
cos(
0
t); if t
n
0
f
0
0; if t >
n
0
f
0
(8)
where
0
= 2f
0
, f
0
is the central frequency, n
0
is the number
of periods, and P
0
the maximum amplitude.
The modeled quadrant was meshed using hexahedral nite
elements (SOLID164 from ANSYS LS-Dyna) with an average
size of 0.75 mm. More than 70,000 nite elements and 90,000
nodes were needed in all. The actuators were not modeled,
only the interactions between them and the composite plate
being considered. Three geometries for PZTs have been con-
sidered in SOMs for distributing the interaction forces (Fig. 5).
For a wave actuation with a maximum force on the contour
P
0
= 200 N/m and n
0
=1.5, with f
0
=100 kHz (non-dispersive
frequency range, see Fig. 2b) actuation parameters, the eld
6
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010
Nicolae Constantin, et al.: ASPECTS OF LAMB-WAVE GENERATION AND TRANSMISSION
(a)
(b)
Figure 11. L(0,2) mode, excited with four square 8.5x8.5 mm
2
PZTs, with
in-phase 125 kHz signals, at t = 100 s (left) and at t = 400 s (right).
of total displacements obtained in each case after 40 sthe
end of the analysis intervalis presented in Fig. 6. The time
integration step was about 0.16 s, which leads to a total anal-
ysis time of less than 10 minutes on a standard PC. The wave
propagation was tracked until end reections appeared.
Removable PZTs, having non-adhesive contact with the in-
spected plate, produced a different wave actuation, which is
presented in the next section in two variants. Figure 7a presents
the wave created on the previously studied plate by the constant
pressure corresponding to a normal beam excitation, modu-
lated in time by the same Hanning windowed burst signal, with
P
0
= 1 MPa, n
0
= 1.5, and f
0
= 100 kHz. For the same sym-
metry reasons, only a quarter of the whole structure was mod-
eled. Figure 7b shows the wave created in the same plate by
an anglebeam excitation, obtained with a prism. The angle
and the material of the prism were chosen so that, according to
Snells law, an A0 wave is excited at the same 100 kHz cen-
tral frequency in the non-dispersive wave propagation range.
This time, due to simple symmetry, half of the virtual structure
was modeled, using over 145,000 nite elements and 185,000
nodes. In this latter analysis, the reference time was considered
the moment when the longitudinal wave transmitted through
the prism excited the A0 Lamb wave in the plate. In both mod-
els, the total displacement elds in the plate were obtained after
the same standard 40 s time interval.
2.3. Lamb Waves Propagating in Pipes
Analysis of Lambwave propagation in pipes or tubes is
considerably more complex due to the higher number of
modes: longitudinal, exural, torsional.
23, 24
In this section,
a composite pipe with a 169 mm outer diameter and 4.6 mm
wall thickness, made of a quasiorthotropic glass/epoxy com-
posite, is analyzed. The engineering constants are presented
in Table 2, where 1, 2, and 3 are the loop, longitudinal and
radial direction, respectively. The dispersion curves were ob-
tained using the numerical approach nalized by Eq. (6); the
ones corresponding to longitudinal modes is shown in Fig. 8.
Considering the symmetry of the virtual pipe, with a solidly
bonded transducer in the middle of a 500 mm segment, only
1/8 of it was modeled. Even for this reduced model, 135,000
hexahedral elements with around 1 mm edge were needed, de-
ned by over 170,000 nodes. For a Hanning windowed signal,
with P
0
= 200 N/m, n
0
= 3.5, and f
0
= 50 kHz (nondispersive
frequency range again; see Fig. 8a), excited by a rectangular 20
x 6 mm transducer, the total displacement eld obtained after
80 s with an SOM is presented in Fig. 9a. The same signal,
applied through a square 10 x 10 mm PZT, is producing, like
in composite plates, isodirectional propagation (Fig. 9b).
Obtaining a directional wave is always very attractive for in-
spection purposes, since such a wave is guided across the item
with less interference with reected signals. Signal directiv-
ity obtained with rectangular PZTs led to the idea of replac-
ing them with a battery of singlesquare or round transduc-
ers much cheaper and recommended for experiments and eld
inspectionsby modeling the propagation of signals generated
in phase or with various delays or time gaps. So far, this type
of simulation was made on plates.
10, 25
Results on compos-
ite pipes have shown that the wave type and its propagation is
consistently dependent on the frequency of the generated sig-
nal (Fig. 10), also relating to the dispersion curves in terms of
dissipative propensity function of frequency. Because of the
short distance allowed by the 3D FE models that are possible
to run on the available computers, a clear directionality could
not be found. For that reason, an SOM using shell elements
was used next, requiring less computational resources. This
time, 1/4 of a 2 m long pipe was modeled, and the results are
presented in Fig. 11.
Directionality of L(0,2) mode, good at the generating point,
degraded quite rapidly, and the small size difference of the
PZTs in the two models could not signicantly inuence the
results. It is a solid enough reason to keep, for the mo-
ment, the much more expensive inspection technology that
uses PZT rings applied on the pipe for generating L and/or T
waves, which proved to be efcient in the inspection of metal-
lic pipes.
26
3. EXPERIMENTAL RESULTS
Propagation of Lamb waves was experimentally evaluated
on a 600 x 350 x 2.56 mm
3
composite plate, made of the same
material considered in the numerical model and having a mid-
plane delamination induced with a 100 mm nonadhesive lm
in the center. Various PZTs were mounted on it in the solid or
removable bond variant (Fig. 12).
The experiment was carried out with rectangular PZTs,
solidly bonded to the composite plate 400 mm from each
other, at equal distances from the plate edges. The 50 x 30
x 0.2 mm
3
patches were made of PIC 255 ceramic mate-
rial, with a maximal force of 52.5 N for 200 V input and a
Youngs modulus ranging from 19 to 22.3 GPa. The numeri-
cal simulation was performed with a 2D SOM. Actuations al-
Table 2. Engineering constants of the composite pipe.
Youngs modulus Poissons ratio Shear modulus Mass density
[GPa] (minor) [GPa] [kg/m
3
]
E
1
= 25.3
21
= 0.1581 G
12
= 3 1800
E
2
= 10
32
= 0.4 G
23
= 3 1800
E
3
= 10
31
= 0.1976 G
13
= 3 1800
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010
7
Nicolae Constantin, et al.: ASPECTS OF LAMB-WAVE GENERATION AND TRANSMISSION
Figure 12. Experimental setup for Lamb-wave transmission in a composite
plate.
Figure 13. Wave propagation tracked in the experimental and numerical
model of the composite plate.
ways used a Hanningwindowed signal with P
0
= 50 V, n
0
=3,
and f
0
=100 kHz. Time of ight and dispersion curves cross
checking proved that the received signals were S0 waves. The
simulated signal shows higher amplitude for the overlapped,
reected waves coming from both ends (Fig. 13). The mea-
sured signal is slightly delayed and appears to be contaminated
by noise caused by the waves reected from the holes in the
composite pipe (neglected in SOM) and, possibly, also caused
by noise due to external electromagnetic interference.
The experiments on a physical 6inch composite pipe
(Fig. 14), with geometry and material properties quite close
to those considered in the numerical model (see Table 2), led
to the results shown in Fig. 15. The signal, produced with
4 PZT buzzers having dimensions (10 mm in diameter) and
layup close to those considered in the numerical model (see
Fig. 10), followed a generatrix on which no damage occurred
and another on which a 20 J energy, lowvelocity impact event
was produced. That area, about 25 mm in diameter, is so
far the damagesize threshold possible to identify through this
method. Signal attenuation, resulting from the damaged area,
was observed on both models. The extra signal delay obtained
in the experiment shows that numerical simulation requires
further accuracy, especially on the side of material modeling.
4. DISCUSSION AND CONCLUSIONS
The theoretical and experimental modeling of Lambwave
actuation and transmission examined the capabilities of PZT
patches to adequately implement the Lambwave method. Par-
Figure 14. Inspected composite pipe.
a) b)
Figure 15. Reception of L(0,2) mode on the undamaged and damaged path in
(a) the numerical model and (b) the experimental model.
ticularly, the numerical models made evident two parameters
that clearly inuence wave generation and transmission across
the inspected item: (i) the interface between the transducer and
the item surface and (ii) the shape of the PZT patch. The inter-
face governed the generation mechanisms and, thus, the wave
modes transmitted in the item. The strong interface, obtained
for the solidly bonded PZTs, has to imply a multiphysics ap-
proach for a robust simulation. There are still aspects on which
future research must focus, such as the characteristics of the
adhesive existing in the interface and the inuence of the thick-
ness of the solidly bonded PZT patch.
Directionality of rectangular transducers was proved in both
model types, but it has to be further explored and evaluated
in terms of range. Using batteries of common, cheaper iso-
transmitting transducers proved to be a good solution for get-
ting guided waves, both in plate and pipe scanning.
Once again, outlined here is the usefulness of dispersion
curves for a priori frequency selection of the generated sig-
nal, which can help numerical simulation and experiments in
relation to achieving efciency in terms of range and accuracy.
Moreover, in connection with time of ight, the dispersion
curves make easier mode-recognition possible.
As a whole, this paper posits that numerical simulation com-
bined with experiments are indispensable for achieving better
Lambwave transmissionhighly needed for accurate inspec-
tion results in reliable and efcient SHM. Adequate choice of
transducers, transducer mounting and/or networking, actuation
signal shape, and frequencywindowing are of the utmost im-
portance and should be further explored to obtain longerrange
8
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010
Nicolae Constantin, et al.: ASPECTS OF LAMB-WAVE GENERATION AND TRANSMISSION
and widerscanning capabilities, thus qualifying this method
as a reliable globalinspection tool.
ACKNOWLEDGMENTS
The research work was nancially supported by Sure2Grip
CRAFT project, operated in the FP6 program under COOP
CT2004513266 contract.
REFERENCES
1
Royer, D., Noroy, M. H. and Fink, M. Optical gener-
ation and detection of elastic waves in solids, Journal
de Physique IV, Colloque C7, supplement au Journal de
Physique III, 4, C7673C7684, (1994).
2
Ke, W., Castaigns, M. and Bacon, C. 3D Finite Element
simulations of an air coupled ultrasonic NDT system, in
Ultrasonic Wave Propagation in Non Homogenous Media,
Eds. Alain L eger and Marc Deschamps, Springer Proceed-
ings in Physics, N. Y., 195206, (2009).
3
Jia, X., Auld, B. and Quentin, G. Normalmode theory
for analyzing the Lamb waves generated by a wedge trans-
ducer, Proceedings of the World Congress on Ultrasonics,
Berlin, 317321, (1995).
4
Diaz Valdes, S. H. and Soutis, C. Health monitoring of
composite using Lamb waves generated by piezoelectric
devices, Plastics, Rubber & Composites, 29(9) 475481,
(2000).
5
Cawley, P. and Alleyne, D. The use of Lamb waves for
the long-range inspection of large structures, Ultrasonics,
34(2), 287290, (1996).
6
Zhongquing, S., Lin, Y. and Ye, L. Guided Lamb waves
for identication of damage in composite structures: A re-
view, Journal of Sound and Vibration, 295 (35), 753780,
(2006).
7
Kessler, S. S., Spearing, S. M. and Soutis, C. Damage de-
tection in composite materials using Lamb waves, Smart
Mat. Structures, 11(2), 269278, (2002).
8
Rose, J. L. Ultrasonic Waves in Solid Media, Cambridge
Univ. Press, N.Y., (1999).
9
Diamanti, K., Soutis, C. and Hodgkinson, J. M. Non-
destructive inspection of sandwich and repaired compos-
ite laminated structures, Composites Science & Technology,
65(13), 20592067, (2005).
10
Diamanti, K., Soutis, C. and Hodgkinson, J. M. Piezoelec-
tric transducer arrangement for the inspection of large com-
posite structures, Composites A, 38(4), 11211130, (2007).
11
Giurgiutiu, V. and Zagrai, A. N. Characterization of piezo-
electric wafer sensors, Journal of Intelligent Materials Sys-
tems and Structures, 11(12), 959975, (2000).
12
Nieuwenhuis, J. H., Neumann, J., Greve, D. W. and Oppen-
heim, I. J. Generation and detection of guided waves us-
ing PZT wafer transducers, IEEE Trans. Ultrasonics, Fer-
roelectrics, and Frequency Control, 52(11), 21032111,
(2005).
13
Sorohan, S t., Constantin, N., Anghel, V. and G avan, M.
Finite element analysis of generation and detection of
Lamb waves using piezoelectric transducers, Mathematics
in Industry, 11, SpringerVerlag Berlin Heidelberg, Eds.
Gabriela Ciuprina and Daniel Ioan, 8996, (2007).
14
Moulin, E., Assaad, J., Delebarre, C., Grondel, S., Bal-
ageas, D. Modeling of integrated Lamb waves generation
systems using a coupled nite element-normal modes ex-
pansion method, Ultrasonics, 38(18), 522526, (2000).
15
Wang, X., Lu, Y. and Tang, J. Damage detection using
piezoelectric transducers and the Lamb wave approach:
I. System analysis, Smart Materials and Structures, 17,
025033, (2008).
16
Von Ende, S., Lammering, R. Modeling and simulation of
Lamb wave generation with piezoelectric plates, Mechan-
ics of Advanced Materials and Structures, 16(3), 188197,
(2009).
17
Constantin, N., Sorohan, S t., Anghel, V. and G avan, M.
Finite element analysis on the mechanism of Lamb wave
generation with piezoelectric transducers, Proceedings of
the International Semiconductor Conference (CAS 2006),
Sinaia, Romania, 2, 403406, (2006).
18
Giurgiut iu, V. and Lyshevski, S. E. Micromechatronics:
Modeling, Analysis, and Design with Matlab, CRC Press,
(2004).
19
Allik, H. and Hughes, J. R. Finite Element for Piezoelectric
Vibration, International Journal for Numerical Methods in
Engineering, 2, 151157, (1970).
20
ANSYS, Inc. Theory. Reference and User Guides (2000).
21
Sorohan, S t., Constantin, N., Anghel, V. and G avan, M. Nu-
merical extraction of dispersion curves for Lamb wave in-
spections on complex structures, Proceedings of the Inter-
national Congress on Ultrasonics, Vienna, 9497, (2007).
22
Sorohan, S t., Constantin, N., Petre, C., Anghel, V. and
G avan, M. Numerical simulation of Lamb waves in plate
and pipe-like structures, Proceedings of the International
Conference on Structural Analysis of Advanced Materials,
Bucharest, 185192, (2005).
23
Nelson, R. B., Dong, S. B., Karla, R. D. Vibration and
waves in laminated orthotropic circular cylinders, Journal
of Sound and Vibration, 18(3), 429444, (1971).
24
Graff, K. F. Wave Motion in Elastic Solids, Dover Publica-
tions Inc., N. Y., (1991).
25
Constantin, N., Sorohan, S t., G avan, M., Anghel, V. and
Raetchi, V. Evaluation of wave directivity obtained by
transducer arrays, Proceedings of the International Semi-
conductor Conference, 2, Sinaia, 347350, (2008).
26
Cawley, P. and Davies, J. Long-range guided wave
inspectionCurrent capabilities and future directions, Pro-
ceedings of the International Congress on UltrasonicsICU
2009, Santiago, (2009).
International Journal of Acoustics and Vibration, Vol. 15, No. 1, 2010
9