Page 1
CFRP Composites with Embedded PZT Transducers for
Nonlinear Ultrasonic Inspection of Space Structures
Christos Andreades and Francesco Ciampa*
Department of Mechanical Engineering, University of Bath, Bath, BA2 7AY, UK
*Corresponding Author: [email protected]
Abstract
Spacecraft structures are made of carbon fibre reinforced plastic (CFRP) composites
due to their high strength-to-weight ratio. However, material damage such as micro-
cracks and delamination are likely to occur during spacecraft fabrication, assembly or
on-orbit due to hypervelocity debris impacts. In the latter case, satellite components are
visually inspected during time-consuming and risky astronauts’ extravehicular
activities. Hence, there is a need for real-time monitoring of cracks in spacecraft
composites, especially for future manned missions. The integration of piezoelectric lead
zirconate titanate (PZT) transducers in CFRP composites is a possible solution for the
development of “smart” structures capable of (i) providing in-situ ultrasonic monitoring
of damage, and (ii) preventing the direct exposure of PZTs to the harsh outer space. In a
previous study, the use of a woven E-glass fibre fabric layer between the PZT and the
CFRP plies was proposed as a suitable technique for electrical insulation of embedded
PZTs with no effect on the interlaminar properties of the composite. Nonlinear
ultrasonic experiments on artificially delaminated CFRP plates revealed that the damage
sensitivity based on the second harmonic generation was nearly two times higher than
with conventionally surface-bonded PZTs. In this study, nonlinear ultrasonic
experiments on CFRP test samples with both artificial (in-plane delamination) and real
impact damage proved the capability of the proposed embedded PZTs to detect multiple
defects of various dimensions. The ultrasonic response of damaged specimens was
studied against that of a pristine one, and damage detection was achieved based on the
generation of second harmonics at specific input signal frequencies. In addition, by
scanning the material response with a laser Doppler vibrometer it was verified that for
each of the chosen driving frequencies, the area on the sample’s surface at which the
out-of-plane vibrational velocity was higher matched the position of the associated
damage. Based on the results of this study, the novel sensor embedding technique has
the potential to be used for in-service monitoring of composite spacecraft components
and other critical engineering structures.
Mor
e in
fo a
bout
this
art
icle
: ht
tp://
ww
w.n
dt.n
et/?
id=
2331
4
Page 2
2
1 Introduction
Carbon fibre reinforced plastic (CFRP) composites are extensively used in
spacecraft structures because of the high strength-to-weight ratios, corrosion/fatigue
resistance and design flexibility they offer, especially when compared to metals [1].
Moreover, when CFRP composites are used as the inner wall material in dual-wall
shield systems on spacecrafts the ballistic performance is improved relative to all-
aluminium dual-wall systems of the same weight, and they can also be repaired easily
after impacts using adhesively bonded patches [2, 3]. Spacecrafts in the lower earth
orbit are susceptible to micro-cracks and delamination caused by hypervelocity impacts
from micrometeoroids and pieces of orbital debris (MMOD) [4, 5]. Currently, satellite
components are inspected only visually through long and dangerous extravehicular
activities (EVAs) carried out by trained crew [6]. Therefore, monitoring of cracks and
delamination in composite components of crewed spacecrafts is necessary.
Over the past years, various non-destructive testing (NDT) techniques have been
developed for the detection of material damages. Some examples include techniques
based on linear ultrasonic wave propagation [7, 8], acoustic emission [9-11] and X-ray
scanning [12, 13]. Although there is an ongoing research towards the automation of
these techniques, currently they are labour-intensive, time-consuming and expensive
[14]. Nonlinear elastic wave spectroscopy (NEWS) techniques such as those based on
higher-harmonic generation [15-18], time reversal [19-21] and wave modulation [22-24]
are also very popular due to their higher sensitivity over linear ultrasonic methods to
detect damage at early stages of formation (e.g. micro-cracks, delamination and voids).
NEWS techniques can also be implemented for the inspection of large structures using
only few transducers for the propagation of guided ultrasonic waves [25]. Piezoelectric
zirconate titanate (PZT) is the most common type of transducers because they are
capable of converting changes in strain, pressure, force and acceleration into electrical
signals, based on the piezoelectric effect [26]. Also, they offer fast sensing and
actuation response, high stiffness and resistance to high temperatures [26, 27].
In previous studies, NEWS techniques were successfully applied for the detection of
damage in composite materials, mainly using surface-bonded PZTs [14, 28, 29]. In the
case of monitoring impacts on spacecraft components, external PZTs can be
permanently damaged even if the impact energy is not enough to damage the monitored
components [30]. For this reason, there is a growing interest in the development of
composites with integrated PZTs capable of providing real-time ultrasonic monitoring
of damage, and preventing the direct exposure of PZTs to the space debris clouds. It is
believed that damage detection transducers integrated into the spacecraft shielding can
help operations determine safe encounter distances from the threat [30]. For example, if
shield integrity was confirmed good, more risky near approaches could be planned with
higher science return [30]. In the past, researchers inserted PZTs between the plies of
both CFRP and glass fibre reinforced plastic (GFRP) composites for structural health
monitoring applications. [32-35]. However, the main challenge in the manufacture of
smart CFRP composites is the need for insulation of embedded PZTs from the
electrically conductive carbon fibres. In the majority of previous studies, insulation was
achieved by interlaying polyimide (Kapton) films between the PZT and the CFRP plies
[33-35]. It is known though that polymeric films such as Kapton and Teflon are
commonly placed in composites to constrain the adhesion between layers during the
Page 3
3
curing process, and hence simulate artificial delamination [36]. Therefore the presence
of such films can affect the structural integrity of the composite.
A previous study of the authors proposed an alternative insulation technique where a
woven E-glass fibre fabric layer was interlaced between the conductive surface of the
PZT and the CFRP plies, without affecting the interlaminar properties of the composite
[37]. By conducting NEWS experiments on CFRP laminates with artificial delamination
it was shown that the damage sensitivity of the proposed configuration of embedded
PZTs, based on the second harmonic (A2) generation was around two times higher than
the sensitivity of the same PZTs that were bonded onto the composite surface.
In this study, additional NEWS experiments were performed on a CFRP plate with
two artificial damages and a plate with two impact damages. The aim was to assess the
capability of the proposed type of embedded PZTs to detect multiple damages of
different size on the same plate, based on the generation of A2 harmonics due to damage
excitation. In particular, by propagating elastic waves through the material at selected
frequencies was expected to force the debonded layers at the damage location to either
oscillate (“clapping” motion) or move relative to each other (“rubbing” motion) leading
to the generation of nonlinear elastic effects that would be detected as higher harmonics
(even and odd multiples) of the input signal frequency [38]. The main difference
between the two types of damages is that the artificial damages cause debonding at a
single interface (in-plane delamination) which is a common type of manufacturing
defects whereas the impacts can cause delamination, fibre breakage and matrix cracking
at multiple interfaces (through-thickness damage). Moreover, the out-of-plane
vibrational velocity of the material surface was measured at the locations of damages
using a laser Doppler vibrometer (LV) to examine whether the detection of A2
harmonics at specific input signal frequencies occurred indeed due to excitation of the
debonded layers.
2 Experimentation
2.1 Laminate Manufacturing
Three 140 x 180 mm laminates were manufactured for this study using
unidirectional carbon/epoxy prepregs (T800/M21) in [90°/0°/90°/0°/90°/0°]s lay-up
giving a thickness of around 3.5 mm. The plates were cured in an autoclave for 180
minutes at a pressure of 0.7 MPa and a temperature of 150 �C with a ramp rate of 3
�C/min. As illustrated in Figure 1, the laminates included two PZTs for the
transmission and reception of elastic waves through the material. A woven glass-fibre
fabric layer (10 x 10 mm) was also interlaced between the top surface of each PZT and
the CFRP ply, for electrical insulation from the carbon fibres. The first laminate
included two double-layered patches of different size that were made from Fluorinated
Ethylene Propylene (FEP) release film. Specifically, each patch consisted of two square
layers of FEP film (12µm thick) stacked one on top of the other. The double FEP
patches were used to generate controlled artificial in-plane delamination. The PZTs and
the FEP patches were embedded between layers 8 and 9 from the bottom. To minimise
internal material distortion, the thin wires from to the anode and the cathode of the
PZTs were directed outside the top surface of the plate through small slits on the CFRP
plies, in the fibre direction of each ply (i.e. no fibre cutting was involved). The wires
were connected to 50Ω straight Bayonet Neill-Concelman (BNC) plugs through low
Page 4
4
noise cables (RG174/U). The other two laminates were identical to the first one but
without the
Figure 1: Dimensions of the CFRP plates used in the NEWS experiments.
two double FEP patches. One was kept in pristine condition (control plate) whereas the
other one was impacted at its centre with a hemispherical indentor of 20 mm diameter
and used for the detection of actual damages. It must be noted that two damages of
different size were created by applying two levels of impact energy (low and high
level). Detection of the small impact damage was experimentally demonstrated before
creating the big impact damage. In this paper, the first laminate is referred as artificially
damaged (AD) laminate, the second one as undamaged (UD) laminate and the last one
as impact damaged (ID) laminate.
2.2 Damage Evaluation
Prior to performing the NEWS experiments, the AD- and ID-plates were subject to
stepped linear C-scanning to evaluate the size of internal damages. That was achieved
using a phased array system (National Instruments NI PXIe-1062Q) with a 128-element
probe. The C-scan was performed in steps of 12 elements and the damages were
assessed based on the signal amplitude. As depicted in Figure 2 and Figure 3,
delamination was detected at the locations of the double FEP patches and the impact
damages. In AD-plate, the size of the two artificial damages was very similar to the size
Page 5
5
of the double FEP patches. In ID-plate, the small and big impact damages were
approximately 18 mm and 32 mm in diameter.
Figure 2: Assessment of the small (a) and the big (b) artificial damages in AD-laminate using
phased array system - Images not to scale.
Figure 3: Assessment of the small (a) and the big (b) impact damages in ID-laminate using phased
array system - Images not to scale.
2.3 Experimental Procedure
2.3.1 NEWS Experiments
The experimental setup used in the NEWS experiments is shown in Figure 4. An
arbitrary waveform generator (TTi TGA12104) was used to send a continuous periodic
signal to the transmitter PZT, through a voltage amplifier (Falco Systems WMA-300)
with a 50x amplification factor. The receiver PZT was connected to an oscilloscope
(PicoScope 4424) which enabled monitoring of the time domain and the frequency
domain of the received signal, at a sampling frequency of 2 MHz with an acquisition
Page 6
6
time of 50 ms. Propagation of ultrasonic elastic waves through the AD- and the UD-
laminate was performed simultaneously and the two FFT spectrums were directly
compared (Figure 5). This allowed to distinguish the A2 harmonics generated due to
damage excitation from those generated due to noise (e.g. instrumentation noise). In
fact, the A2 harmonics detected in both FFT spectrums were considered as noise
whereas those presented only in the FFT spectrum of the AD-laminate were related to
the nonlinear response of the material due to excitation of the debonded layers.
Figure 4: Illustration of the set-up used in the NEWS experiments.
(a)
(b)
Figure 5: Frequency spectrum of the received signal in the UD-laminate (a) and in the AD-laminate
(b) - Input signal of 60 V at 184.7 kHz.
Page 7
7
The transmitted signal was swept from 20 kHz to 500 kHz in steps of 30-40 kHz,
and higher-harmonic generation due to damage excitation was detected at two input
signal frequency ranges; 102-106 kHz and 180-185 kHz. The 104.5 kHz and 184.7 kHz
frequencies corresponded to the A2 harmonics with the highest amplitude, and thus they
were chosen as the excitation frequencies of the two FEP patches. However, these two
frequencies did not necessarily correspond to the fundamental harmonics (A1) with the
highest amplitude (transmitter PZT/material excitation). The same experimental
procedure was repeated for the detection of the two impact damages in the ID-laminate.
The input signal frequency that caused excitation to the small damage was found to be
310 kHz, and for the big damage 128 kHz. In both the AD- and the ID-plates the
amplitude of the received signal was measured at the A1 and A2 frequencies for five
different input signal voltages (60 V, 70 V, 80 V, 90 V and 100 V). As it was expected,
the results obtained from the AD-plate (Figure 6) showed that for both input signal
frequencies (104.5 kHz and 184.7 kHz) the A1 and A2 amplitudes were rising with
increasing input signal voltage. The A2 amplitude was around two orders of magnitude
smaller than the A1 amplitude. These observations were also valid for the results
acquired from the ID-plate (Figure 7) at the driving frequencies of 128 kHz and 310
kHz.
(a)
(b)
Figure 6: Amplitude of the received signal in AD-laminate at the fundamental (a) and second (b)
harmonic frequencies - Input signals of 60, 70, 80, 90, and 100 V at 104.5 and 184.7 kHz
(a)
(b)
Figure 7: Amplitude of the received signal in ID-laminate at the fundamental (a) and second (b)
harmonic frequencies - Input signals of 60, 70, 80, 90, and 100 V at 128 and 310 kHz
Page 8
8
2.3.2 LV Experiments
The LV experiments were performed to verify that the chosen input signal
frequencies were indeed associated with the excitation of the damages. The
experimental setup of the LV is shown in Figure 8. The transmitter PZT was used for
the propagation of continuous periodic signals of 300 V at 104.5 kHz and 184.7 kHz in
the AD-laminate and at 128 kHz and 310 kHz in the ID-laminate. In each case, the out-
of-plane vibrational velocity of the plate surface at the A1 and A2 harmonic frequencies
was measured around the location of the damage using the LV scanning head (Polytec
PSV-400). Three-dimensional plots of the results (Figure 9) proved that in all cases the
vibrational velocity was higher at the damage position and A2 amplitude was
approximately an order of magnitude smaller relative to the A1 amplitude. The results
also revealed that the in the AD-laminate, the input signal at 104.5 kHz caused
excitation only to the big FEP patch whereas at 184.7 kHz only to the small FEP patch.
Similarly, the small and big impact damages in ID-laminate were only excited at the
128 kHz and 310 kHz respectively. This confirmed that the input signal frequencies
were chosen correctly.
Figure 8: Illustration of the set-up used in the LV experiments.
Page 9
9
(a) SmalldamageinAD-Laminate
A1freq.=184.7kHz
A1amp.=305μm/s
A2freq.=369.4kHz
A2amp.=43μm/s
(b) BigdamageinAD-Laminate
A1freq.=104.5kHz
A1amp.=257μm/s
A2freq.=209kHz
A2amp.=32μm/s
(c) SmalldamageinID-Laminate
A1freq.=310kHz
A1amp.=7038μm/s
A2freq.=620kHz
A2amp.=284μm/s
(d) BigdamageinID-Laminate
A1freq.=128kHz
A1amp.=2999μm/s
A2freq.=256kHz
A2amp.=108μm/s
Figure 9: 3D representation of the out-of-plane vibrational velocity at the location of the small (a)
and big (b) artificial damages, and the location of the small (c) and big (d) impact damages at the
fundamental and second harmonic frequencies.
3 Conclusions
This study demonstrated the capability of a novel configuration of embedded PZTs
in CFRP composites to detect material damage. This embedding technique involves
direct insertion of the PZTs between CFRP plies with the conductive surface of the
PZTs being covered by a single layer of woven E-glass fibre fabric for electrical
insulation. Pairs of embedded PZTs were used to perform NEWS experiments on two
CFRP plates of the same dimensions and lay-up. One plate included two artificial
damages and the other plate two impact damages. By propagating continuous periodic
ultrasonic waves through the laminates, excitation of each damage was achieved only at
a single input signal frequency. Damage excitation was detected based on the A2
harmonic generation in the frequency spectrum of the received signal, and A2 the
amplitude of these harmonics was found to increase with increasing input signal
voltage. In addition to the NEWS experiments, the material response at the chosen input
signal frequencies was scanned with an LV. In fact, the out-of-plane vibrational velocity
of the plate surface was measured at the A1 and A2 harmonic frequencies. The results
Page 10
10
verified that in all cases the area at which the vibrational velocity was higher matched
the position of the associated damage. The results of this study showed that the
proposed type of embedded PZTs can be used to detect multiple damages of different
size in composite plates which are excited at different frequencies. The experimental
results also proved the ability of these internal PZTs to detect in-plane delaminations
which are often caused due to manufacturing errors, as well as through-thickness
damages (fibre breakage and matrix cracking at multiple layers) that usually occur due
to impacts. Based on the above, this novel sensor embedding technique can be utilised
to provide nonlinear ultrasonic monitoring of spacecraft composite components, without
the risk of exposing the PZTs directly to the harsh outer space.
References
1. A Francesconi, C Giacomuzzo, S Kibe, Y Nagao and M Higashide, “Effects of
high-speed impacts on CFRP plates for space applications”, Advances in Space
Research 50(5), pp 539-548, 2012
2. WP Schonberg and EJ Walker, “Hypervelocity impact of dual-wall space structures
with graphite/epoxy inner wall”, Composites Engineering 4(10), pp 1045-1054,
1994.
3. WP Schonberg, “Studies of hypervelocity impact phenomena as applied to the
protection of spacecraft operating in the MMOD environment”, Procedia
Engineering 204(1) pp 4-42, 2017.
4. WP Schonberg, “Protecting Earth-orbiting spacecraft against micro-
meteoroid/orbital debris impact damage using composite structural systems and
materials: An overview”, Advances in Space Research 45(6), pp 709-720, 2010.
5. S Yashiro, K Ogi, T Nakamura and A Yoshimura, “Characterization of high-
velocity impact damage in CFRP laminates: Part I–Experiment”. Composites Part
A: Applied Science and Manufacturing 48(1), pp 93-100, 2013
6. EL Christiansen, “Meteoroid/debris shielding”, NASA Johnson Space Center, pp
75-76, 2003.
7. AS Birks, “Nondestructive Testing Handbook 7: Ultrasonic Testing”, ASNT
Handbook, 1991.
8. F Amerini and M Meo, “Structural health monitoring of bolted joints using linear
and nonlinear acoustic/ultrasound methods”, Structural Health Monitoring 10(6),
pp 659-672, 2011.
9. F Cesari, V Dal Re, G Minak and A Zucchelli, “Damage and residual strength of
laminated carbon–epoxy composite circular plates loaded at the centre”,
Composites Part A: Applied Science and Manufacturing 38(4), pp 1163-1173,
2007.
10. M R’Mili, M Moevus and N Godin, “Statistical fracture of E-glass fibres using a
bundle tensile test and acoustic emission monitoring”, Composites Science and
Technology 68(7), pp 1800-1808, 2008.
11. AR Oskouei and M Ahmadi, “Acoustic emission characteristics of mode I
delamination in glass/polyester composites”, Journal of composite materials 44(7),
pp 793-807, 2010.
12. MF Moura and AT Marques, “Prediction of low velocity impact damage in carbon-
epoxy laminates”, Composites Part A: Applied Science and Manufacturing 33(3),
pp 361-368, 2002.
Page 11
11
13. P Wright, X Fu and I Sinclair, “Spearing SM. Ultra high resolution computed
tomography of damage in notched carbon fiber-epoxy composites”, Journal of
composite materials 42(19) pp 1993-2002, 2008.
14. F Aymerich and WJ Staszewski, “Impact damage detection in composite laminates
using nonlinear acoustics”, Composites Part A: Applied Science and Manufacturing
41(9), pp 1084-1092, 2010.
15. O Buck, WL Morris and JM Richardson, “Acoustic harmonic generation at
unbonded interfaces and fatigue cracks”, Applied Physics Letters 33(5), pp 371-
372, 1978.
16. U Polimeno, M Meo, DP Almond and SL Angioni, “Detecting low velocity impact
damage in composite plate using nonlinear acoustic/ultrasound methods”, Applied
Composite Materials 17(5) pp 481-488, 2010.
17. I Solodov, “Resonant acoustic nonlinearity of defects for highly-efficient nonlinear
NDE”, Journal of Nondestructive Evaluation 33(2), pp 252-262, 2014.
18. F Ciampa, SG Pickering, G Scarselli and M Meo, “Nonlinear imaging of damage in
composite structures using sparse ultrasonic sensor arrays”, Structural Control and
Health Monitoring 24(5), 2017.
19. H Sohn, HW Park, KH Law and CR Farrar, “Damage detection in composite plates
by using an enhanced time reversal method”, Journal of Aerospace Engineering
20(3), pp141-151, 2007.
20. F Ciampa and M Meo, “Nonlinear elastic imaging using reciprocal time reversal
and third order symmetry analysis”, The Journal of the Acoustical Society of
America, 131(6) pp 4316-4323, 2012
21. Z Liu, H Yu, J Fan, Y Hu, C He and B Wu, “Baseline-free delamination inspection
in composite plates by synthesizing non-contact air-coupled Lamb wave scan
method and virtual time reversal algorithm”, Smart Materials and Structures 24(4),
045014, 2015.
22. KA Van Den Abeele, PA Johnson and A Sutin, “Nonlinear elastic wave
spectroscopy (NEWS) techniques to discern material damage, part I: nonlinear
wave modulation spectroscopy (NWMS)”, Journal of Research in Nondestructive
Evaluation 12(1), pp 17-30, 2000.
23. A Klepka, L Pieczonka, WJ Staszewski and F Aymerich, “Impact damage detection
in laminated composites by non-linear vibro-acoustic wave modulations”,
Composites Part B: Engineering 65(1), pp 99-108, 2014
24. D Dionysopoulos, GP Fierro, M Meo and Ciampa F, “Imaging of barely visible
impact damage on a composite panel using nonlinear wave modulation
thermography”, NDT & E International 95(1), pp 9-16, 2018.
25. Z Su, L Ye and Y Lu, “Guided Lamb waves for identification of damage in
composite structures: A review”, Journal of sound and vibration 295(3-5), pp 753-
780, 2006.
26. HP Konka, “Characterization of composite piezoelectric materials for smart joint
applications”, Master’s Thesis, Louisiana State University, Louisiana, USA, 2010.
27. S Masmoudi, A El Mahi, S Turki, and R El Guerjouma, “Mechanical behavior and
health monitoring by acoustic emission of unidirectional and cross-ply laminates
integrated by piezoelectric implant”, Applied Acoustics 31(86), pp 118-125, 2014.
28. A Klepka, M Staczkiewicz, L Pieczonka, WJ Staszewski, L Gelman, F Aymerich
and T Uhl, “Triple correlation for detection of damage-related nonlinearities in
composite structures”, Nonlinear dynamics 81(1-2), pp 453-468, 2015.
Page 12
12
29. X Liu, L Bo, K Yang, Y Liu, Y Zhao, J Zhang, N Hu and M Deng, “Locating and
imaging contact delamination based on chaotic detection of nonlinear Lamb
waves”, Mechanical Systems and Signal Processing, 109(1), pp 58-73, 2018.
30. EL Christiansen, J Arnold and B Corsaro B, “Handbook for designing MMOD
protection”, NASA Johnson Space Center, NASA/TM-2009-214785, 2009.
31. HY Tang, C Winkelmann, W Lestari and V La Saponara, “Composite structural
health monitoring through use of embedded PZT sensors”, Journal of Intelligent
Material Systems and Structures 22(8), pp 739-755, 2011.
32. S Masmoudi, A El Mahi and S Turki, “Use of piezoelectric as acoustic emission
sensor for in situ monitoring of composite structures”, Composites Part B:
Engineering 80(1), pp 307-320, 2015.
33. S Mall, “Integrity of graphite/epoxy laminate embedded with piezoelectric
sensor/actuator under monotonic and fatigue loads”, Smart Materials and Structures
11(4), pp 527-533, 2002.
34. CA Paget, K Levin and C Delebarre, “Actuation performance of embedded
piezoceramic transducer in mechanically loaded composites” Smart Materials and
Structures 11(6), pp 886-891, 2002.
35. Z Su, X Wang, Z Chen, L Ye and D Wang, “A built-in active sensor network for
health monitoring of composite structures”, Smart Materials and Structures 15(6),
pp 1939-1949, 2006.
36. SC Wooh and C Wei, “A high-fidelity ultrasonic pulse-echo scheme for detecting
delaminations in composite laminates”, Composites Part B: Engineering 30(5), pp
433-44, 1999.
37. C Andreades and F Ciampa, “Embedded Piezoelectric Transducers in Carbon Fibre
Composites for Nonlinear Ultrasonic Applications”, Structural Health Monitoring,
13930, 2017.
38. F Ciampa, G Scarselli and M Meo, “On the generation of nonlinear damage
resonance intermodulation for elastic wave spectroscopy”, The Journal of the
Acoustical Society of America, 141(4), pp 2364-2374, 2017.