Original Article

Structural Health Monitoring

2017, Vol. 16(2) 142152

Characterization of delamination-type The Author(s) 2016

Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
damages in composite laminates DOI: 10.1177/1475921716666411
journals.sagepub.com/home/shm
using guided wave visualization and
air-coupled ultrasound

Rabi S Panda, Prabhu Rajagopal

and Krishnan Balasubramaniam

Abstract
This article reports on the characterization of delamination damages in composite laminates using wave visualization
method. A combination of plate-guided ultrasound and air-coupled ultrasonics is used to locate and visualize delamina-
tions. The study focuses on the physics of Lamb wave propagation and interaction with delaminations at various
through-thickness locations and positions. Three-dimensional finite element simulations are used to study, in detail, the
changes in wave features such as mode velocity, wavelength and wave refraction in the delamination region. These wave
features provide information on the location, position and orientation of the delamination. These studies are validated
by experimental measurements. The influence of position of source and delamination on wave refraction in the delami-
nation region is examined. This method also correlates the results obtained from experiments and finite element simula-
tions to theoretical dispersion curves in order to distinctly determine the delamination location.

Keywords
Lamb waves, delamination, wave visualization, air-coupled ultrasound, non-destructive testing, structural health
monitoring

Introduction over a long distance with less attenuation. The entire

thickness of the laminate can be interrogated by various
The usage of fibre-reinforced plastic (FRP) composite Lamb modes,3 allowing detection of surface and sub-
materials in aerospace, military and civilian structural surface damages. Hence, these waves can be used for
applications has increased significantly due to attrac- NDE as well as structural health monitoring (SHM) of
tive mechanical properties such as high specific strength composite structures.1,2 Damage detection in composite
and modulus. Arising due to faulty manufacturing or laminates utilizing Lamb waves requires understanding
operational conditions such as impact or fatigue load- the physical principles of their propagation and interac-
ing and temperature, delamination is a common dam- tion with damages. Over the last decade, studies on the
age type in composites that reduces the strength, interaction of Lamb waves with delamination-type
stiffness and integrity of the structure. Delaminations defects have been reported by many researchers in com-
are sub-surface damages and hence difficult to be posite laminates.59
detected and characterized, especially in the inaccessi-
ble/hidden regions of complex composite structures.
Thus, composite materials require stringent non- Centre for Nondestructive Evaluation (CNDE), Department of
destructive testing/non-destructive evaluation (NDT/ Mechanical Engineering, Indian Institute of Technology Madras, Chennai,
NDE) techniques to ensure the overall integrity and India
reliability of the structures. Ultrasonic guided wave14
methods are promising for rapid inspection of large Corresponding author:
Rabi S Panda, Centre for Nondestructive Evaluation (CNDE),
areas of structure from a single transducer position. Department of Mechanical Engineering, Indian Institute of Technology
Lamb waves are ultrasonic waves guided by geome- Madras, 312 Machine Design Section, Chennai 600036, Tamil Nadu, India.
trical boundaries of plate-like structures and can travel Email: rabisankarpanda@gmail.com
Panda et al. 143

Of the various transduction methods available, air- along 0 direction in the eight-layered quasi-isotropic
coupled ultrasonic testing (AUT) technique is emerging GFRP composite laminate of 2.64 mm thickness. Here,
as a promising method for non-contact inspection of we must clarify that, here and in the rest of this article,
composites.7,1012 The advantages of AUT include along 0 direction means the guided waves travel
rapid automated scanning, excitation of a pure Lamb along the fibre direction of 0 ply in the composite lami-
wave mode11,13 and inspection of irregular and inacces- nate. Dispersion curves are plotted using DISPERSE22
sible regions. Non-contact inspection using Lamb software by considering the material properties and
modes to visualize damage in plate-like structures has orientation of each individual ply as input data accord-
been studied widely.1419 ing to the layup of composite laminate. Low-frequency
This article investigates the physics of Lamb wave excitation of 100 kHz was considered in the study, to
propagation and interaction with delamination-type minimize the number of modes and wave attenuation in
defects in a quasi-isotropic composite laminate using a the received signal. This frequency range is also ideal
non-contact wave visualization approach (the method for air-coupled ultrasonic inspection of composites. At
is highlighted in previous work by the authors; see such low frequencies, only the fundamental symmetric,
Panda et al.20 and Sasanka et al.21). The studies are car- S0, and anti-symmetric, A0, modes exist (see Figure
ried out in the time domain, in view of simplicity and 1(a)). The S0 mode mainly consists of in-plane particle
ease of practical implementation. This approach pro- displacements, whereas the A0 mode has out-of-plane
vides a series of images of propagating waves, which particle displacements. On account of out-of-plane par-
spans the scanned surface of the test piece. Wave fea- ticle displacement, the A0 mode can be measured conve-
tures such as refraction to interaction with defects and niently with air-coupled ultrasonic transducer.
changes to wave velocity and wavelength due to disper- Moreover, the A0 mode has a shorter wavelength per-
sion are recorded and exploited for detection of defects. mitting study of smaller features. The A0 mode was
The physics of wave refraction is shown to be the basis excited and received by appropriate positioning11 of an
for delamination location in thickness direction of the air-coupled ultrasonic transducer (illustrated in Figure
laminate. Three-dimensional (3D) finite element (FE) 2) based on Snells law. From the incident angle disper-
simulations validated by experiments are used to study sion curve shown in Figure 1(b), the receiving angle of
low-frequency A0 mode interaction with square-shaped A0 mode at 100 kHz frequency was deduced to be
delaminations in glass-fibre reinforced plastic (GFRP) approximately 21.
plate-like structures. The size of the delamination is
chosen appropriately to capture changes to wave field Guided wave visualization approach
when traversing the delamination regions. Qualitative
results are followed by quantitative studies of the influ- The presence of delamination causes a composite lami-
ence of various defect parameters on the visualized nate to divide into two sub-laminates of reduced thick-
wave field. nesses. As can be seen from Figure 1(a), in the low
The article is organized as follows. A brief back- frequency-thickness regime, the A0 mode is dispersive.
ground to the method used and problem studied is As the thickness decreases, the A0 mode velocity
given in section Background. The details of the experi- decreases, thereby decreasing the wavelength. If this
ments and 3D FE simulations carried out using air- change in wavelength is measured, we can locate and
coupled ultrasonic transducers are presented in section visualize the defect region from outside. More details
Methods. Section Results and discussion first pre- on the practical implementation of this approach in
sents results for experimental verification of FE simula- our work are given in the Results and discussion
tion approach for the case of delamination located at a section.
corner of the composite plate, with a discussion of phy-
sics of wave propagation and interaction with delami- Problem studied
nations. Later on, the section presents results for
delaminations located at the centre for the plate. The The article analyses GFRP composite laminates with
article concludes with directions for further work in engineered defects at various through-thickness loca-
section Summary and conclusions. tions (lamina interfaces) of the laminate to study Lamb
wave propagation, detection and visualization of
defects using the air-coupled ultrasonic technique (illu-
Background strated in Figure 2). The A0 mode is excited in the lami-
nate. Square-shaped delaminations, located at a corner
Lamb wave mode selection and excitation of the laminates (illustrated in Figure 4(a)), are consid-
Figure 1(a) and (b) shows the phase velocity and inci- ered for the study. The position and through-thickness
dent angle dispersion curves of Lamb waves travelling location of the defect are determined from the
144 Structural Health Monitoring 16(2)

Figure 1. Plot showing (a) phase velocity dispersion curve and (b) incident angle dispersion curve.

interaction of the wave mode with the defect. First, we

consider delaminations located at a corner of the plate:
for this case, numerical (FE) results are validated with
experiments since air-gap-type delaminations are more
convenient to be created artificially. Based on thus vali-
dated FE models, we then study the case of air-gap
delaminations located at the centre of the plate, where
experimental samples are more difficult to create artifi-
cially. These numerical models are used to study the
effect of defect position on wave scattering features.
Details of physics of the problem are explained in the
Results and discussion section.

Methods
Procedure for experiments
Experiments were performed on eight-layered quasi-
Figure 2. Schematic diagram of experimental setup. isotropic composite laminates of layup [0/ + 45/45/90]S
Panda et al. 145

Table 1. Material properties of lamina. bonding quality of the PZT crystal are also important.
Since in this study the interest is more in understanding
Exx (GPa) Eyy (GPa) nxz nyz Gxz (GPa) r (kg/m3) the interaction of guided waves with the delamination,
44.68 6.90 0.28 0.355 2.54 1990 we did not optimize the crystal dimensions. A 100-kHz
central frequency air-coupled transducer (The Ultran
Group, USA, http://www.ultrangroup.com) of 25 mm
with delaminations at lamina interfaces in through- diameter was used for measuring the out-of-plane dis-
thickness locations. This type of ply layup is very placement. The receiving angle of the transducer, based
commonly used in aerospace structures. The material on the phase velocity of the A0 mode, was calculated
properties of the composite laminates, as in Ramadas using Snells law. The vertical position and angle of
et al.,9 are presented in Table 1. Thicknesses of each receiving transducer were adjusted manually. The air-
lamina and laminate were 0.33 and 2.64 mm, respectively. coupled transducer was approximately 15 mm away
The size of each laminate was 300 mm 3 200 mm with a from the laminate surface, corresponding to the best
square-shaped delamination of size 60 mm 3 60 mm at signal amplitude of the received wave mode.
one corner. A schematic of the laminate is shown in The scanning was conducted for a specific area of
Figure 4(a). The presence of the delamination was con- 140 mm 3 120 mm of the laminate, including the dela-
firmed using conventional ultrasonic air-coupled through- mination region with a resolution of 0.5 mm 3 0.5 mm
transmission C-scan. by the moving air-coupled transducer. Lamb waves
The experiments were carried out in a custom-built were excited and propagated in the composite laminate.
C-scan-like setup, as shown in Figure 2. The experi- The excitation signal was a 100-kHz narrow-band five-
mental setup consists of a custom-built scanner, a pul- cycle sinusoidal tone burst pulse modulated by a
serreceiver, a 100-MHz A/D card, a motion controller Hanning window. The Lamb waves propagate in the
and a personal computer. A data acquisition and signal laminate and leak to the surrounding air medium due
processing program was used to filter and average the to out-of-plane displacement of surface particles. These
captured data to improve the signal-to-noise ratio of leaky Lamb wave23 signals were collected at each data
the digitized signal. The pitchcatch arrangement was point, shown in Figure 3(a) as A-scans by a built-in
employed to generate and receive A0 mode Lamb wave. data acquisition system. At a particular time instant,
A commercially available, adhesively bonded circular the displacement field was obtained from the collected
PZT crystal (Sparkler Ceramics Pvt. Ltd, http:// A-scans as shown in Figure 3(b).
www.sparklerceramics.com) of 12 mm diameter, The time response at each measurement point was
10 mm thickness and 100 kHz central frequency was averaged 16 times to improve the signal-to-noise ratio.
used to excite the A0 mode in the composite laminate. A 30- to 500-kHz band-pass filter was applied to limit
Here, we note that A0 mode generation is a thickness- the bandwidth of the output signal. The raw data con-
mode (thickness-dependent) phenomenon for the PZT tain time signals from each of the scanned points. The
crystal. Additionally, the impedance matching and the data were then exported to the MATLAB software

Figure 3. (a) A-scan receiving points and (b) displacement field at a particular time instant.
146 Structural Health Monitoring 16(2)

program and processed to get the out-of-plane displa- mode wavelength to provide a converged solution for
cement field across the scanned surface of the specimen elastic wave propagation.26 The incremental time step
at a particular instant in time. The resultant images used was 0.01 ms. A 100-kHz narrow-band five-cycle
from MATLAB were plotted as a two-dimensional sinusoidal tone burst pulse modulated by a Hanning win-
(2D) colour image. The wave propagation in the lami- dow was used as the excitation signal. The A0 mode was
nate and the interaction with the delaminated region excited by applying the out-of-plane (along the z-direc-
can be visualized from these images. tion) nodal displacement to the surface FE nodes at the
edge of the laminate to avoid edge reflections as shown in
Figure 4(a). The delamination region in the laminate was
Procedure for 3D FE simulations modelled by disconnecting the surfaces of adjacent layers
The 3D numerical simulations were carried out using imitating an air-gap. The region without delamination
commercially available FE package24 on similar sam- was connected to the adjacent layers, thus ensuring conti-
ples used in experiments to visualize the Lamb wave nuity of displacements in the numerical model. Without
propagation and interaction with delamination. A sche- loss of generality, the delamination can exist at any
matic diagram of the configuration used in FE simula- through-thickness location between a pair of ply inter-
tions is shown in Figure 4(a) and the ply layups in faces. The size of delamination considered was of 60 mm
Figure 4(b). The dimensions of the composite laminates 3 60 mm. Out-of-plane displacement was captured at the
chosen were 150 mm 3 120 mm 3 2.64 mm, to avoid receiver location. Simulations were repeated by changing
the early reflections from model boundaries and to the delamination location along through-thickness of the
minimize the total run time. laminate between a pair of lamina interfaces.
The lamina properties are the same as those used in
experiments, and for simplicity, viscoelastic properties
are not considered (see Table 1). In the numerical mod- Results and discussion
els, each lamina was modelled individually and proper-
ties were assigned depending on the layup. Each lamina
Qualitative measurements
was modelled using eight-noded 3D solid brick ele- The results obtained from the experiment are presented
ments (C3D8R) having three degrees-of-freedom at in Figure 5. The images (see Figure 5(a) and (b)) show
each node. The incremental time step size and mesh ele- the A0 mode wave propagation and interaction with
ment size of the FE simulations are of importance for delamination located between third and fourth layers
the stability and the accuracy of the solution. The incre- of the laminate at various time instants.
mental time step size and element size were selected as Several characteristic changes such as the reduction
stated, for example, in Ramdhas et al.25 The size of the in amplitude, change in wavelength, speed of propaga-
elements was 0.165 mm in the thickness direction and tion and deflection of wavefronts of the selected A0
0.25 mm in both length and width directions. The FE mode are observed at the delamination region. Since
mesh ensures that at least 40 elements exist per A0 the delamination was created with air-gap between the

Figure 4. (a) Composite laminates with defect orientation and position and (b) ply layups.
Panda et al. 147

Figure 5. Experimental results showing the A0 mode propagation and interaction with delamination between third and fourth
laminas at different times: (a) 114 ms and (b) 220 ms.

layers, there was an impedance mismatch between the thickness locations for a particular propagating angle,
defective and healthy region. The impedance of air is up are shown in Figure 6. In this study, the propagating
much lesser than that of the laminate material. When angle is defined as the angle between the top horizontal
the Lamb wave interacts with the delamination region, edge of the laminate and the line joining the point of
some of the energy is reflected back, some mode con- loading/excitation to the intersection point of a wave-
verted9 into other modes and the remaining is trans- front with the vertical defect contour. The propagating
mitted into it. Therefore, wave amplitude is reduced in angle was calculated for a certain time instant (see
the delamination region as a result of reflection and Figure 6(b)) and was fixed at 78. This changes with
mode conversion. Figure 5(b) does not show reflected time, and accordingly, the wave deflection ud can be cal-
energy because we received only the forward travelling culated. These results, especially Figure 6(b), are in
A0 mode, since we kept the air-coupled receiver at a good agreement with those generated by the experimen-
particular angle, say +u degrees with the normal (see tal wave visualization method, presented in Figure 5(b).
Figure 2). To receive the backward travelling wave It was observed from the experiments and simula-
modes, that is, the reflection from the defect bound- tions that there is a strong correlation between the
aries, we would have to keep the air-coupled receiver at wave deviation and the location of delamination along
u degrees with the normal (see Kazys et al.,13 Figure the thickness of the laminate. The wave deviates from
1(b), page e820). The mode-converted signal is also not its original direction of propagation when it interacts
shown in Figure 5(b), as in the present setup the receiv- with the defect. Application of Snells law shows that a
ing probe was set to pick only the A0 mode. The wave- reflected wave and a refracted wave are expected in all
length change is affected by the velocity change. This is cases, with the exception of normal incidence. The
because the laminate is getting divided into two sub- refracted wave bends towards or away from the normal
laminates by the defect. The wavelength reduces in the depending on the velocity differences in healthy and
thinner sub-laminate and it gradually increases as the defective regions of the laminate. As it is known that
sub-laminate thickness increases. The deflection of the mode velocities vary with the position of defects, this
wavefront occurs due to the mismatch in impedance phenomenon causes wavefronts to bend to different
between the healthy and defective region. In a healthy extents for defects in different depths. This suggests
region, wave enters and leaves without leaving any that studying the wave path deviations can provide
traces of energy, while multiple reflections may occur information on the location of defective regions.
in the defective zone due to the energy trapped15 in it.
In FE simulations, the delaminations were created Quantitative measurements
between a pair of lamina interfaces. The results Studies were conducted to check whether the wave field
obtained from FE simulations when A0 mode wave measurements can provide a quantitative guidance on
interacts with delaminations at various through- the position and orientation of defects.
148 Structural Health Monitoring 16(2)

Figure 6. Contours of displacement magnitude as obtained from FE simulations at 96 ms for A0 mode interaction with defects
between different layers: (a) no defect, (b) one and two, (c) two and three and (d) three and four.

Angle of wave deflection ud with through-thickness location of study the value of v1 and u1 are fixed, equation (1) shows
delamination. The angle of deviation ud which is the that u2 }1=v2 , that is, the wave angle in top sub-laminate
wave deflected in the defective region for a particular varies inversely with the wave velocity. As can be seen in
propagating angle up was calculated for various delami- phase velocity dispersion curve, Figure 1(a), the velocity
nation locations along the laminate thickness. An of A0 wave mode increases with laminate thickness. The
example of the calculation of the wave deflection angle top sub-laminate thickness increases with the location of
into the delamination is shown in Figure 6(b). Figure 7 delamination along the thickness direction, thereby
shows the calculated values from FE simulations and increasing the wave velocity, and hence, the angle of
average experimental values. We observe very good deflection/refraction angle decreases. The location of the
agreement between experimental results and FE predic- delamination in thickness direction of the laminate can
tions. This result shows that obtaining the wave devia- be identified by measuring this refraction angle.
tion angle can provide a practical quantitative parameter
to locate the delamination in the thickness direction.
The basis for the trend of the angle of deviation can Wave refraction. As in Rajagopal and Lowe,26 the wave
be explained based on Snells law refraction was studied in terms of the frequency domain
ratio of the beam compensated refracted and incident
v1 sin u1 = v2 sin u2 1 signals, given as follows
where v1 and v2 are the mode velocities in the healthy p
laminate and top sub-laminate, respectively, and u1 and R(v): dR
Refraction ratio = p 2
u2 are the corresponding wave angles. Since in our I(v): dI
Panda et al. 149

Figure 8. Variation of refraction ratio with through-thickness

Figure 7. Variation of angle of wave deflection with delamination location.
delamination location in thickness direction.

equation (4) that Refraction ratio } v2 , that is, the

where R(v) and I(v) are the frequency spectra of the wave refraction directly varies with the wave velocity in
resultant displacement obtained from the refracted and the top sub-laminate. Figure 8 also shows values
incident signals, respectively; dR is the distance from obtained for refraction ratio based on calculations in
the refraction point to the point where the refraction is equation (4). The velocities for each sub-laminate were
monitored; and dI is the distance from the source to the obtained from DISPERSE22 for these calculations. We
point where the incident signal is monitored. (See, for find good agreement with FE and experimental results.
instance, Figure 4(a) for an illustration.) Figure 8 As the top sub-laminate thickness increases with the
shows average experimentally measured and FE pre- location of delamination in successive laminas along
dicted values of the refraction ratio for the monitored the thickness of the laminate changes, the velocity and
location, plotted with the through-thickness location of
wavelength of the wave mode increase. (See the phase
delamination.
velocity dispersion curve, Figure 1(a).) Due to an
The above trend of refraction ratio again follows
increase in velocity in the top sub-laminate and trans-
based on the dispersive property of the A0 mode. It
mission of more energy to the defect region, the value
was observed that the value of refraction ratio increases
of refraction ratio increases along the laminate thick-
along the laminate thickness. Based on acoustic impe-
ness. By finding this wave refraction value, we can pre-
dance, we can write, assuming plane wave behavior in
dict the location of the delamination along the
the far-field,
thickness of the laminate.
2Z2
Refraction ratio = 3
Z1 + Z2 Velocity curves from dispersion. The presence of a delami-
where Z1 and Z2 are the acoustic impedances of the nation between a pair of laminas causes a laminate to
healthy laminate and top sub-laminate, respectively. divide into two sub-laminates. Behaviour of wave
Z = rv, where r is the density of the material of the modes in sub-laminates can be understood from the
laminate and v is the wave velocity. Let v1 and v2 be dispersion properties of concerned mode. In an eight-
the mode velocities in the healthy laminate and top layered laminate, the defect could be present after first,
sub-laminate, respectively. The density of the material second or seventh layers. In these cases, dispersion
of the laminate and sub-laminate is the same, that is, properties of all layered combinations must be consid-
r = r1 = r2 . Therefore, from equation (3) ered. In composites, dispersion properties change as
propagating angle changes. The data were collected by
2v2 plotting dispersion curves for various combinations of
Refraction ratio = 4 layers, such as single-layered, two-layered and up to
v1 + v2
eight-layered laminates (as in Table 2) at different pro-
When the top sub-laminate thickness is small, the pagation angles.
wave velocity (v2 ) is very low compared to the velocity Using the data collected from disperse plots, the
in healthy region (v1 ) and can be neglected in the normalized velocity curves were generated for funda-
denominator of equation (4). So it is clear from mental A0 and S0 modes. The maximum velocity
150 Structural Health Monitoring 16(2)

Table 2. Layer combinations.

Layers Layup Thickness (mm)

First [0] 0.33

First and second [0/+45] 0.66
First to third [0/+45/45] 0.99
First to fourth [0/+45/45/90] 1.32
First to fifth [0/+45/45/90/90] 1.65
First to sixth [0/+45/45/90/90/45] 1.98
First to seventh [0/+45/45/90/90/45/+45] 2.31
First to eighth [0/+45/45/90/90/45/+45/0] 2.64

Figure 9. Normalized velocity curves for different layered combinations.

(Vmax = 4767 m/s) among all velocities in different Change in wave features with defect position and propagating
directions was taken to normalize them. A reference angle. The influences of the propagating angle on the
semicircle of unit radius was drawn to compare mode angle of wave deflection in the defect region were also
velocities with normalized velocity (Vn) at any angle. studied with the help of FE models. The delamination
Figure 9 shows the normalized velocity curves for some was modelled at the centre of the laminate as illustrated
of the layer combinations (sub-laminates), as given in in Figure 10. The angle of wave deflection ud at the defect
Table 2. It was observed from the velocity curves that region was calculated for different propagating angles up
when the location of the delamination changes gradu- and through-thickness defect locations. The variation in
ally from the first layer to eighth layer, the velocity and angle of deflection with propagating angles and through-
wavelength for a particular mode increase. thickness defect locations are given in Table 3.
The studies mentioned so far in this article are These studies demonstrate that the leaky guided
focused on a single propagating angle up and a particu- wave scanning approach can provide effective visuali-
lar position of the defect. For better understanding of zation of defects in composite laminates, along with a
this phenomenon, the study is extended to various pro- guidance as to the location (and also possibly, the
pagating angles and different defect positions. orientation) of defects along the thickness.
Panda et al. 151

Table 3. Deflection values for different propagating angles.

Defect location (between layers) Angle of propagation, up (deg) Angle of deflection, ud (deg)
up1 up2 up3 ud1 ud2 ud3

First and second 75 62 47 52 48 34

Second and third 75 62 47 39 36 23
Third and fourth 75 62 47 32 29 18
Fourth and fifth 75 62 47 22 20 14
Fifth and sixth 75 62 47 17 16 09
Sixth and seventh 75 62 47 13 13 05
Seventh and eighth 75 62 47 07 06 04

Lamb wave such as amplitude reduction, change in

wavelength, speed of propagation and deflection of the
wavefronts were observed in the defective region. It
was also observed from the FE simulations that the
wave refraction angle in the defective region varies with
position of source and delamination (Figures 6(b) and
10), as given in Table 3. The correlation of the results
obtained from experiments and FE simulations to theo-
retical dispersion curves validates the technique pre-
sented here.
The results of these studies provide improved physi-
cal insight into the wave interaction at delaminations at
various depths across laminate thickness. The wave
visualization approach using air-coupled ultrasound
was found to be a useful tool for analysing the wave
mode interactions with damages in composites. Guided
Figure 10. Displacement contour showing deflection of A0 Lamb wave modes can be used to locate and find the
mode in defective zone (FE simulation). depth of a delamination by measuring the mode wave-
length change and wave refraction angle in the delamina-
tion region. Non-contact inspection of composite
laminates using air-coupled ultrasonic technique has
Summary and conclusion advantages over laser-based and PZT filmbased SHM.
This article studied the A0 mode Lamb wave propaga- Laser-based sensing is not convenient for composite mate-
tion and interaction with delamination in a [0/45/45/ rials because of the low reflectivity, low sensitivity, high
90]S quasi-isotropic composite laminate at the low- noise, high cost and single-point measurements. Similarly,
frequency regime using visualization approach. The PZT filmbased Lamb wave sensing is intrusive and not
acoustic wave field visualization images were generated accepted by the industry in many applications. The air-
with air-coupled ultrasonic technique using leaky coupled receiver has better efficiency for measuring the
Lamb waves. The investigation was carried out with A0 mode because it integrates normal displacements over
3D FE simulations and validated by experiments. FE a surface corresponding to that of its field of view.27 The
simulations and experiments were conducted on quasi- findings are important to further advance this damage
isotropic laminates containing delaminations at various detection technique to complex structures.
lamina interfaces across the laminate thickness. The
results in both cases are in good agreement as shown in Acknowledgment
Figures 5 and 6. Also, quantitative parameters such as The authors thank Mr DVS Sasanka for his assistance during
wavefront deflection and refraction ratio were mea- experiments and helpful discussions.
sured at the delamination region to know the effective-
ness of the selected method. The variability in the Declaration of Conflicting Interests
results obtained is due to the orientation of the receiv- The author(s) declared no potential conflicts of interest with
ing transducer, material attenuation and instrumental respect to the research, authorship and/or publication of this
errors. Several characteristic changes of A0 mode article.
152 Structural Health Monitoring 16(2)

Funding 13. Kazys R, Demcenko A, Zukauskas E, et al. Air-coupled

The author(s) disclosed receipt of the following financial sup- ultrasonic investigation of multi-layered composite mate-
port for the research, authorship, and/or publication of this rials. Ultrasonics 2006; 44: e819e822.
article: The authors gratefully acknowledge support from the 14. Michaels TE, Michaels JE and Ruzzene M. Frequency
NPMASS program managed by The Aeronautical wavenumber domain analysis of guided wavefields. Ultra-
Development Agency (ADA) Bangalore, India. sonics 2011; 51(4): 452466.
15. Park B, An YK and Sohn H. Visualization of hidden
delamination and debonding in composites through non-
References contact laser ultrasonic scanning. Compos Sci Technol
1. Janardhan PM and Balasubramaniam K. Lamb- 2014; 100: 1018.
wave-based structural health monitoring technique for 16. Ramadas C, Padiyar J, Balasubramaniam K, et al.
inaccessible regions in complex composite structures. Lamb wave based ultrasonic imaging of interface dela-
Struct Control Hlth 2014; 21(5): 817832. mination in a composite T-joint. NDT&E Int 2011;
2. Raghavan A and Cesnik CES. Review of guided-wave 44(6): 523530.
structural health monitoring. Shock Vib Digest 2007; 17. Tian Z, Yu L and Leckey C. Delamination detection and
39(2): 91114. quantification on laminated composite structures with
3. Rose JL. Ultrasonic waves in solid media. Cambridge: Lamb waves and wavenumber analysis. J Intel Mat Syst
Cambridge University Press, 1999. Str 2015; 26(13): 17231738.
4. Su Z, Ye L and Lu Y. Guided Lamb waves for identifica- 18. Tian Z, Leckey C and Seebo J. Guided wave imaging for
tion of damage in composite structures: a review. J Sound detection and evaluation of impact-induced delamination
Vib 2006; 295(35): 753780. in composites. Smart Mater Struct 2015; 24: 105019.
5. Guo N and Cawley P. The interaction of Lamb waves 19. Yashiro S, Takatsubo J and Toyama N. An NDT tech-
with delaminations in composite laminates. J Acoust Soc nique for composite structures using visualized Lamb-wave
Am 1993; 94(4): 22402246. propagation. Compos Sci Technol 2007; 67: 32023208.
6. Kessler SS, Spearing SM and Soutis C. Damage detection 20. Panda RS, Rajagopal P and Balasubramaniam K. An
in composite materials using Lamb wave methods. Smart approach for defect visualization and identification in
Mater Struct 2002; 11(2): 269278. composite plate structures using air-coupled guided ultra-
7. Liu Z, Yu H, He C, et al. Delamination detection in com- sound. In: Proceedings of review of progress in quantitative
posite beams using pure Lamb mode generated by air- nondestructive evaluation (ed DE Chimenti and LJ Bond),
coupled ultrasonic transducer. J Intel Mat Syst Str 2014; Boise, ID, 2025 July 2014, vol. 1650, pp. 12991306.
25(5): 541550. Melville, NY: AIP.
8. Padiyar MJ and Balasubramaniam K. Quantitative char- 21. Sasanka DVS, Madhavan V, Padiyar MJ, et al. Visualiza-
acterization of interface delamination in composite T- tion studies of Lamb wave propagation and interactions
joint using couplant-free Lamb wave methods. J Reinf with anomalies in composite laminates using air-coupled
Plast Comp 2016; 35(4): 345361. ultrasonics. In: Proceedings of review of progress in quanti-
9. Ramadas C, Janardhan PM, Balasubramaniam K, et al. tative nondestructive evaluation (ed DE Chimenti, LJ
Delamination size detection using time of flight of anti- Bond and DO Thompson), Baltimore, MD, 2126 July
symmetric (Ao) and mode converted Ao mode of guided 2013, vol. 1581, pp. 10911097. Melville, NY: AIP.
Lamb waves. J Intel Mat Syst Str 2010; 21(8): 817825. 22. DISPERSE users manual, version 2.0.11, January 2016.
10. Bhardwaj MC. Phenomenally high transduction air/gas 23. Chimenti DE and Martin RW. Nondestructive evalua-
transducers for practical non-contact ultrasonic applica- tion of composite laminates by leaky Lamb waves. Ultra-
tions. In: Proceedings of review of progress in quantitative sonics 1991; 29(1): 1321.
nondestructive evaluation (ed DO Thompson and DE Chi- 24. ABAQUS analysis users manual, version 6.12, January
menti), Chicago, IL, 2225 July 2008, vol. 1096, pp. 2016.
920927. Melville, NY: AIP. 25. Ramdhas A, Pattanayak RK, Balasubramaniam K, et al.
11. Castaings M and Cawley P. The generation, propagation, Symmetric low-frequency feature-guided ultrasonic waves
and detection of Lamb waves in plates using air-coupled in thin plates with transverse bends. Ultrasonics 2015; 56:
ultrasonic transducers. J Acoust Soc Am 1996; 100(5): 232242.
30703077. 26. Rajagopal P and Lowe MJS. Short range scattering of
12. Michaels JE and Michaels TE. On the measurement of the fundamental shear horizontal guided wave mode nor-
guided wavefields via air-coupled ultrasonic transducers. mally incident at a through-thickness crack in an isotro-
In: Proceedings of review of progress in quantitative non- pic plate. J Acoust Soc Am 2007; 122(3): 15271538.
destructive evaluation (ed DE Chimenti and LJ Bond), 27. Castaings M, Singh D and Viot P. Sizing of impact dam-
Boise, ID, 2025 July 2014, vol. 1650, pp. 792798. Mel- ages in composite materials using ultrasonic guided
ville, NY: AIP. waves. NDT&E Int 2012; 46: 2231.