Journal of Mechanical Engineering and Sciences (JMES)

ISSN (Print): 2289-4659; e-ISSN: 2231-8380; Volume 5, pp. 602-610, December 2013
© Universiti Malaysia Pahang, Malaysia
DOI: http://dx.doi.org/10.15282/jmes.5.2013.6.0057

WAVE PROPAGATION SCATTERING DUE TO DEFECTS ON THIN

Z. M. Hafizi1, 2, J. Epaarachchi1 and K. T. Lau1

ABSTRACT

Engineering structures which are based on laminated composites have a high probability
of unexpected damage developing during service. The damaged formations must be
monitored from the beginning before they head towards structural failure which could
result in substantial damage. This leads to the necessity of a structural health monitoring
(SHM) system to be installed during the construction of laminated composite structures.
However, an understanding of damage area detection and damage characteristics is
essential before a SHM system can be integrated into the structures. This article presents
the effects of wave propagation through an existing area of damage on composite plates.
Theoretically, a propagating wave that starts from any source will vary when crossing
an area with damage. This study shows that high frequency wave propagation (kHz
range) reacts differently when passing through the damaged area, compared with low
frequency wave propagation. The results of the study will lead to an effective damage
detection method, utilizing the available vibration source, especially for the condition
monitoring of thin laminated composite structures.

Keywords: SHM, wave propagation, wavelet analysis, Gl/epoxy composites.

INTRODUCTION

Composite material is one of the catalysts for the growth of modern structures, and
especially the development of smart structures. For instance, the high ratio of material
strength compared to its weight means that it has become one of the main choices in
aircraft construction. However, the structure of the composite material is exposed to the
risk of the formation and propagation of internal damage, which is hard to predict.
Failure of the structure can start from various causes, either during the manufacturing
process (e.g. voids) or when the structure is being used (e.g. impact or fatigue).This is a
strong reason why an effective SHM system needs to be installed in every composite
structure. One of the concepts in SHM, so-called wave-propagation-based SHM, is
becoming popular recently. The idea is based on the propagation of acoustic waves. In
general, this method is usually referred to as guided waves, ultrasonic guided waves, or
Lamb waves. (Croxford, Wilcox, Drinkwater, & Konstantinidis, 2007) have claimed
that guided acoustic waves may be the only detection method that combines an
acceptable level of damage detection sensitivity with a significant propagation range.
Moreover, waveform analysis of the guided waves can provide more detailed
information on the location and nature of smaller defects (Mal, Ricci, Banerjee, & Shih,
2005). A thorough literature study has been done by (Diamanti & Soutis, 2010) on the

602
 Wave Propagation Scattering due to Defects on Thin Composite Plates

use of Lamb waves in aircraft composite structure and they concluded that this
technique can lead to an active SHM for laminated composite structures, utilizing
piezoelectric wafers embedded into layered composite structures. Composite materials
display a wide variety of failure mechanisms as a result of their complex structure and
manufacturing processes, which include fiber failure, matrix cracking, buckling and
delamination (Ibrahim, Sapuan, & Faieza, 2012; Orifici, Herszberg, & Thomson, 2008;
Umar, Zainudin, & Sapuan, 2012). Damage can develop and propagate very slowly
from inside the composite layers (e.g. matrix cracks, delamination and matrix-fiber
debonding) (Adebisi, Maleque, & Rahman, 2011; Shan, Ghazali, & Idris, 2013). As a
result, it will affect the material properties of the composites, such as their strength and
stiffness. The formation of damage such as matrix cracking can sometimes be seen with
the naked eye; however, damage such as delamination is a silent killer, in that it is
almost impossible to detect from the surface of the structure. There are various methods,
highlighted by the researchers, to identify the damage in composite structures, and
particularly for thin laminated composite plates such as the fiber Bragg gratings,
ultrasonic, acousto-ultrasonic, x-ray imaging, and acoustic emission(AE) methods
(Bhaskar & Sharief, 2012; Kahandawa, Hafizi, Epaarachchi, & Lau, 2012; Lam, Lau,
Ling, Su, & Tam, 2009; Muravin, Muravin, & Lezvinsky, 2010; Popovics, 2009).
However, the passive fault detection system was chosen as the focus in this study, as it
can be used online, is more practical, requires less equipment and is a relatively cheap
system. One of the main challenges for a passive system is to create a classifying
technique which can evaluate the condition of the examined structures. In other words,
passive monitoring must be combined with good signal analysis in order to produce a
robust and reliable system. AE techniques may suit the need. Although AE usage is
highly established for metallic materials, there is a huge challenge in detecting good AE
signals that can be correlated with any damage formation and propagation, especially
where composite materials are involved.
This present study highlights a different approach/concept which may be a more
practical alternative in real application. The idea is to manipulate the available vibration
source in order to identify the existence of damage in thin composite plates. This
concept can also be considered as a form of passive monitoring. We consider a
composite structure that is constantly exposed to constant vibration; in any frequency
range, this will produce wave propagation in the structure. For thin plates, this wave
propagation is called the Lamb waves. When the wave propagates through areas with
damage, such as matrix cracking or delamination, this wave will change its form and
some characteristics, as it is affected by the existence of the damaged area. This article
will reveal the results of several experiments that have been carried out on a thin
composite sample in which a small hole has been made to indicate the presence of the
damaged area.

EXPERIMENTATION

265 mm × 97 mm × 4.4 mm of Gl/epoxy resin laminates with a stacking sequence of

603
 Hafizi et al. / Journal of Mechanical Engineering and Sciences 5(2013) 602-610

Hole

Figure 1. Gl/epoxy resin laminates for the experiments.

High Frequency Case

Figure 2 shows the experimental setup. Two piezoelectric sensors (numbers‘1’ and ‘2’,
as shown in Figure 2) were coupled to the surface of the plate. The sensors were
individually connected to two PAC AE Node Systems (data acquisition from Physical
Acoustic Corporation) for waveform acquisition and were synchronized with the help of
AE Win software. The sampling rate for acquisition was set to 1Mega sample per second
and the threshold was set to 45 dB.

Clamped Composite laminates Clamped

A B
1 2

1 2
C

Figure 2. Setup for high frequency source’ test.

Point B indicates the hole’s location. A pencil lead break (PLB) test was done at
point A. A PLB test was chosen as it can excite high frequency wave propagation, at
approximately 30 kHz. The signal propagation due to the lead breaking will be acquired
by both piezoelectric sensors. For this case, sensor ‘2’ captured the signal which was
propagating across the artificially damaged area (hole). Finally, for comparison, the lead
breaking was also done at point C, as this area has no damage in between the sensors.

Low Frequency Case

The test arrangement was as shown in Figure 3. Both piezoelectric sensors were
connected directly to a digital storage oscilloscope and the sampling rate was set to 100
kHz. An impact with a hammer at point F was the method used to excite the low
frequency wave propagation on the thin plates. The wave propagated and traveled

604
 Wave Propagation Scattering due to Defects on Thin Composite Plates

across point B and was then captured by sensor ‘1’. At the same time, sensor ‘2’also
detected this signal, but it was not propagating through any damaged area before
reaching the sensor. The impacts were repeated at points E and F, where no wave
propagation was expected to cross the damaged area.

B
1 D
E
2 F

Figure 3. Arrangement for low frequency source test.

Figure 4. Specimen prepared before the low frequency impact by hammer.

RESULTS AND DISCUSSION

When a Lamb wave propagates past the damaged area, it will experience a wave
scattering. The wave scattering effect due to defects in composite materials has been
discussed theoretically by a number of researchers. The wave scattering varies
depending on the propagating wave frequency range. Based on several experiments that
have been done, the propagation of a high frequency wave experienced a very clear
wave scattering, compared with the wave propagation of the low frequency range.

High Frequency Case

Figures 5 and 6 show the waveform and its respective fast Fourier transform (FFT) of
Lamb wave signals due to the PLB at points A and C. It is difficult to evaluate the
difference between the signals obtained from the two sensors, except for an obvious
reduction in the signal amplitude and signal energy, which is caused by the effects of
attenuation. The same is true when looking at the results of comparisons of their FFT
analysis. Therefore, further analysis is needed to see more clearly the effects of this
wave scattering. However, the two major modes of wave propagation can still be

605
 Hafizi et al. / Journal of Mechanical Engineering and Sciences 5(2013) 602-610

observed, and they are always associated with Lamb wave propagation: the flexural and
extensional modes.

Figure 5. Response due to PLB at point A: (a) and (c) are from sensor ‘1’, while (b) and
(d) are from sensor ‘2’.

Figure 6.Response due to PLB at point D: (a) and (c) are from sensor ‘1’, while
(b) and (d) are from sensor ‘2’.

Meanwhile Figure 7 shows the result after the continuous wavelet transform
(CWT) is performed on the signals. The CWT, one of the time-frequency analyses, can
provide extra information from any given time domain signal (Hamstad,

606
 Wave Propagation Scattering due to Defects on Thin Composite Plates

O‘GALLAGHER, & Gary, 2002; Jingpin, Bin, & Cunfu, 2008; Zohari, Epaarachchi,
& Lau, 2013). The CWT of a function, as defined by (Chui, 1992), can be expressed as


 t  
WT f s,   f t  * 
1

s   s 
dt (.1)

where and the superscript * indicates the complex conjugate. The term is the
basic wavelet. The parameter in Equation 1 stands for the scale of the basic wavelet
and is related to the signal frequency. Meanwhile, the parameter stands for the shift
or position of the basic wavelet and can be related to the time of the signal. Plotting the
wavelet transform magnitude on the axis gives the time-frequency view of a
signal.

Figure 7. Wavelet analysis of the (a) signal from sensor ‘1’ due to PLB at A; (b) signal
from sensor ‘2’ due to PLB at A; (c) signal from sensor ‘1’ due to PLB at C; and (d)
signal from sensor ‘2’ due to PLB at C.

There are many basic wavelets available and the appropriate choice will give a
better result. In this study, a Morlet wavelet, which is identical to the Gabor wavelet
(Simonovski & Boltežar, 2003) and has a similar shape to an impulse (Lin, 2001) was
used. It can be defined as (Lin, 2001; Simonovski & Boltežar, 2003)

(2)

The scale s can be related with the frequency by this relation:

(3)

607
 Hafizi et al. / Journal of Mechanical Engineering and Sciences 5(2013) 602-610

where the coefficient can be written as the wavelet center frequency, and depends
on the sampling frequency and the selected minimum scale, as explained by
(Simonovski & Boltežar, 2003). Now we can clearly observe the effect of changes in
wave propagation which are caused by the presence of the damaged area. The circles in
Figure 7 (a, b) indicate the changes of the signal waveform in terms of time-frequency
analysis due to the existence of the artificial damage. Meanwhile, the results in Figure 7
(c, d) show no significant variation.
For high-frequency wave propagation (kHz), the whole wave will traverse the
existing artificial damage area (hole). This is because the wavelength is very small
compared to the size of the hole. As a consequence, this leads to the overall impression
of the waves undergoing a scattering effect as shown in Figure 7. However, this
outcome should not be confused with the wave dispersion and attenuation effect that
always occurs when a Lamb wave propagates in a thin plate.

(a) Impact at point D

(b)Impact at point E

(c) Impact at point F

Figure 8.Effect of the low frequency wave traveling across the perforated region
(indicated by red circle).Two lines in each figure indicate the two waveforms captured
by the two different sensors (sensor ‘1’ and sensor ‘2’).

Low Frequency Case

In the meantime, for low-frequency wave propagation, the wave packet did not entirely
traverse the hole at point B. Part of the propagated wave seems to change a little
compared with the wave which propagated without passing the perforated region. After
a few tests, it can be observed that 80 to 90% of the low frequency wave which travels
past the artificial damage area will show the changes described in the result in Figure
8.Figure 8 (a) shows that, if the hammer impact is done at point D (see Figure 3), there
is a significant difference between the waveforms captured by sensor ‘1’ and sensor ‘2’,

608
 Wave Propagation Scattering due to Defects on Thin Composite Plates

although the waveform actually came from the same source. This situation did not occur
if the impact was at another location (E and F). This is because none of the source will
propagate across point B (the hole).

CONCLUSION

This study has successfully reported on the effects of wave propagation when
traversinga defective area. In the case of a real composite structure, propagation may
originate from various sources such as engine vibration and rotation of ball-bearings.
This investigation is still in the early stages but is already very useful in developing an
effective monitoring system for composite structures.

ACKNOWLEDGEMENT

The authors would like to thank the Faculty of Engineering and Surveying and the
Centre of Excellence in Engineered Fibre Composites, University of Southern
Queensland, for providing laboratory facilities and financial support. They would also
like to thank the Universiti Malaysia Pahang together with the Ministry of Higher
Education, Malaysia for scholarship support.

REFERENCES

Adebisi, A. A., Maleque, M. A., & Rahman, M. M. (2011). Metal matrix composite
brake rotor: Historical development and product life cycle analysis.
International Journal of Automotive and Mechanical Engineering, 4, 471-480.
Bhaskar, H. B., & Sharief, A. (2012). Effect of solutionizing on dry sliding wear of
al2024-beryl metal matrix composite. Journal of Mechanical Engineering and
Sciences, 3, 281-290.
Chui, C. K. (1992). An introduction to wavelets (Vol. 1): Academic press.
Croxford, A., Wilcox, P., Drinkwater, B., & Konstantinidis, G. (2007). Strategies for
guided-wave structural health monitoring. Proceedings of the Royal Society A:
Mathematical, Physical and Engineering Science, 463(2087), 2961-2981.
Diamanti, K., & Soutis, C. (2010). Structural health monitoring techniques for aircraft
composite structures. Progress in Aerospace Sciences, 46(8), 342-352.
Hamstad, M., O‘GALLAGHER, A., & Gary, J. (2002). A wavelet transform applied to
acoustic emission signals: Part 2: Source location. Journal of Acoustic Emission,
20, 62-82.
Ibrahim, M. S., Sapuan, S. M., & Faieza, A. A. (2012). Mechanical and thermal
properties of composites from unsaturated polyester filled with oil palm ash.
Journal of Mechanical Engineering and Sciences, 2, 133-147.
Jingpin, J., Bin, W., & Cunfu, H. (2008). Acoustic emission source location methods
using mode and frequency analysis. Structural Control and Health Monitoring,
15(4), 642-651.
Kahandawa, G. C., Hafizi, Z. M., Epaarachchi, J., & Lau, K. T. (2012, Dec.9–12).
Detecting delamination in a composite structure using an embedded fbg – ae
hybrid system. Paper presented at the 7th Australasian Congress on Applied
Mechanics (ACAM 7), Adelaide, Australia.

609
 Hafizi et al. / Journal of Mechanical Engineering and Sciences 5(2013) 602-610

Lam, P.-M., Lau, K.-T., Ling, H.-Y., Su, Z., & Tam, H.-Y. (2009). Acousto-ultrasonic
sensing for delaminated gfrp composites using an embedded fbg sensor. Optics
and lasers in engineering, 47(10), 1049-1055.
Lin, J. (2001). Feature extraction of machine sound using wavelet and its application in
fault diagnosis. NDT & E International, 34(1), 25-30.
Mal, A., Ricci, F., Banerjee, S., & Shih, F. (2005). A conceptual structural health
monitoring system based on vibration and wave propagation. Structural Health
Monitoring, 4(3), 283-293.
Muravin, B., Muravin, G., & Lezvinsky, L. (2010). The fundamentals of structural
health monitoring by the acoustic emission method. Paper presented at the
Proceedings of the 20th International Acoustic Emission Symposium,
November.
Orifici, A., Herszberg, I., & Thomson, R. (2008). Review of methodologies for
composite material modelling incorporating failure. Composite Structures,
86(1), 194-210.
Popovics, J. S. (2009). Recent developments in ndt and shm in the united states. Non-
Destructive Testing in Civil Engineering, NDTCE, 9.
Shan, C. W., Ghazali, M. I., & Idris, M. I. (2013). Improved vibration characteristic of
flexible polyurethane foam via composite formation. International Journal of
Automotive and Mechanical Engineering, 7, 1032-1042.
Simonovski, I., & Boltežar, M. (2003). The norms and variances of the gabor, morlet
and general harmonic wavelet functions. Journal of Sound and Vibration,
264(3), 545-557.
Umar, A. H., Zainudin, E. S., & Sapuan, S. M. (2012). Effect of accelerated weathering
on tensile properties of kenaf reinforced high-density polyethylene composites.
Journal of Mechanical Engineering and Sciences, 2, 198-205.
Zohari, M. H., Epaarachchi, J. A., & Lau, K. (2013). Modal acoustic emission
investigation for progressive failure monitoring in thin composite plates under
tensile test. Key Engineering Materials, 558, 65-75.

610