Composites Part B 163 (2019) 642–654

Contents lists available at ScienceDirect

Composites Part B
journal homepage: www.elsevier.com/locate/compositesb

Damage and failure mechanism of thin composite laminates under T

low-velocity impact and compression-after-impact loading conditions
Hongliang Tuoa,∗, Zhixian Lua, Xiaoping Maa,b, Jun Xinga,c, Chao Zhanga,d,∗∗
a
School of Aeronautics, Northwestern Polytechnical University, Xi'an, 710072, China
b
Institute of Engineering Thermophysics, Chinese Academy of Sciences, Beijing, 100190, China
c
Airworthiness Certification Center, Civil Aviation Administration of China, Beijing, 100102, China
d
Shaanxi Key Laboratory of Impact Dynamics and its Engineering Applications, Xi'an, 710072, China

ARTICLE INFO ABSTRACT

Keywords: Impact resistance and damage tolerance are of great significance in the design of composite structures. This
Polymer-matrix composites (PMCs) study investigated the damage and failure mechanism of thin composite laminates under low-velocity impact
Impact behavior and compression-after-impact (CAI) loading conditions. Four levels of impact energy were included in the test
Damage mechanics matrix. Delamination induced by low-velocity impact was captured using ultrasonic C-scan, and a three-di-
Finite element analysis (FEA)
mensional (3D) digital image correlation (DIC) system was employed to measure full-field displacement during
Thin composite laminate
the CAI tests. Infrared thermography was also used to online monitor the thermal field variation of the test
specimen during the impact and CAI process. The cross sections of typical tested specimens were inspected using
an optical microscope and a scanning electron microscope (SEM). A 3D damage model that considers both
interlaminar and intralaminar damage was proposed to study the complex damage and failure mechanism.
Excellent correlation was obtained between the experimental results and the numerical results. The experimental
results obtained from various tests and the results from the numerical simulation were combined to provide a
new and deep insight of damage evolution and failure mechanisms under low-velocity impact and CAI loading
conditions.

1. Introduction investigate impact resistance and damage tolerance of composite

structures and it is widely accepted that low-velocity impact events lead
Advanced composite laminates such as carbon fiber reinforced to delamination, matrix damage and fiber damage. Delamination, ma-
plastics (CFRP) have been widely used in aircraft structures due to their trix damage and fiber failure are common damage modes under low-
high specific strength, high specific stiffness and good fatigue and velocity impact events depending on the levels of impact energy and
corrosion resistance [1–4]. Its applications have ranged from secondary may greatly reduce the residual strength of composite structures.
load-bearing structures to primary load-bearing structures and the use Compared with the residual tensile performance, compression strength
of composite materials in aircraft structures has greatly increased. and stiffness would be severely weakened due to impact damage. As a
Composite aircraft structures are likely to be exposed to impact consequence, low-velocity impact is a critical load condition for com-
loads (e.g., bird strike, hail and tool-drop) during their service life. posite structures in aircraft. Because low-velocity impact and com-
Impact events can be distinguished as low-velocity impact, high velo- pression-after-impact (CAI) behaviors of laminated structures are very
city impact and ballistic impact [5], of which low-velocity impact for important in the design of composite structures, the damage modes and
composite structure is defined as “a falling weight with a mass of 10–20 failure mechanisms during impact and CAI loadings should be thor-
pounds dropped from a few feet” according to the Composite Materials oughly investigated and understood [8,9].
Handbook (CMH-17) [6]. Under a low-velocity impact, Barely Visible Studying damage in composite laminates can be challenging, due to
Impact Damage (BVID) will be induced, which will reduce significantly the difficulty in characterizing the internal damage behavior.
the damage tolerance and integrity of composite structures [7]. Nondestructive techniques such as ultrasonic C-scan, X-radiography
In recent decades, considerable efforts have been made to and infrared thermography must be used to determine the type of

∗
Corresponding author.
∗∗
Corresponding author. School of Aeronautics, Northwestern Polytechnical University, Xi'an, 710072, China.
E-mail addresses: [email protected] (H. Tuo), [email protected] (C. Zhang).

https://doi.org/10.1016/j.compositesb.2019.01.006
Received 20 August 2018; Received in revised form 14 December 2018; Accepted 1 January 2019
Available online 02 January 2019
1359-8368/ © 2019 Elsevier Ltd. All rights reserved.
H. Tuo et al. Composites Part B 163 (2019) 642–654

damage and the related distribution laws to apply [3,10–13]. The ultimate compressive strength as the tufting density increased. Lisle
duration of a low-velocity impact event is instantaneous (usually be- et al. [48] employed an infrared camera to monitor the thermal re-
tween 1-10 ms), which makes it difficult to observe the damage in- sponse during impact tests and CAI tests in order to investigate the
itiation and evolution process. Ultrasonic C-scan has been proved to be impact damage and compressive failure mechanism as well as to eval-
an efficient tool for detecting and quantifying delamination damage. uate the fracture toughness of compressive fiber failure. Liu et al. [49]
However, there remains a lack of efficient tools for online monitoring of performed CAI tests on two different composite laminate lay-ups to
the fiber and matrix damage. Acoustic emission techniques and infrared investigate the failure mechanisms of hybrid UD/woven carbon-fiber/
thermography are two promising nondestructive online inspection epoxy composite laminates under compression load. The results showed
methods for low-velocity impact events [14–16]. that the use of woven plies on the surfaces of a laminate had a small,
In the early decades, Choi's research [17–19] reported that low- but measurable, positive influence on the residual strength of composite
velocity impact events generated matrix cracks, which lead to delami- laminates. García-Rodríguez et al. [50] interleaved non-woven veils
nation damage. Further studies by Moura et al. [20–23] determined into the interlaminar regions to improve the damage resistance and
that delamination and matrix cracks were the main damage modes and tolerance of thin laminates manufactured with thin-ply non-crimp-
that delamination damage with different shapes and sizes occurred fabrics. They used X-ray tomography and C-scan to investigate impact
during the interlamination between different oriented layers. Wang and CAI behavior of thin-ply non-crimp fabrics. The results showed that
et al. [24] conducted low-velocity impact tests on thick composite la- delamination and matrix damage causes significant stiffness reduction,
minates with three different stacking sequences at three energy levels which led to premature local buckling and a reduction of compressive
and investigated the distribution of the resulting delamination along strength. Caminero et al. [51] studied the influence of thickness and
the thickness direction. Several low-velocity impact tests on glass/ ply-stacking sequence on the CAI strength of carbon fiber-reinforced
epoxy composite laminates conducted by Aktaş et al. [25] showed that epoxy laminates and showed that the lay-up configuration had an effect
delamination and matrix cracking appeared following impact events on the CAI strength, with cross-ply laminates exhibiting lower CAI
with lower impact energy, while fiber damage resulted from higher strength than angle-ply laminates.
impact energy. Artero-Guerrero et al. [26] studied the mass effects on Despite the extensive research conducted in low-velocity impact and
impact response of carbon/epoxy woven laminates under low-velocity CAI behavior of composite laminates, a considerable amount of re-
impact, employing a C-scan and a phased array ultrasonic system to search is still needed to reveal the complex damage and failure char-
detect the impact damage. The authors concluded that no influence of acteristics of thin composite laminates under impact and CAI loading
the impactor mass could be observed until a relatively high impact conditions. In this paper, a series of low-velocity impact tests and CAI
energy level was attained. In addition, it was reported that peak contact tests on thin composite laminates were conducted. Four levels of impact
force, coefficient of restitution, maximum displacement and residual energy were included in the test matrix. The delamination damage was
stiffness plots were functions of impact energy alone and not velocity determined by ultrasonic C-scan, the full-field displacement and stain
impact or impactor mass separately. Comparative experiments con- distribution were measured using a three-dimensional (3D) DIC system
ducted by Evci and Gülgeç [27] reported that woven composite lami- during CAI loading, and an infrared camera was employed for online
nates had superior damage resistance than unidirectional laminates. A monitoring of the thermal field on the surface of the specimens. A 3D
similar comparative study on the low-velocity impact behavior of damage model that considers interlaminar and intralaminar behavior
carbon woven-ply reinforced polymer laminates was carried out by was developed to reveal the complex damage and failure mechanism in
Vieille et al. [28,29], where damage mechanics were investigated by composite laminates. An intralaminar damage model containing a
ultrasonic C-scan inspections and microscopic observations. Single im- maximum strain failure criterion and an improved Puck criterion was
pact and repeated impact tests under various energy levels were con- built, in which the interlaminar delamination damage was simulated by
ducted by Arikan and Sayman [30] to study impact behavior of glass interfacial cohesive elements. Excellent correlation was obtained be-
reinforced and epoxy composite laminates. Nash et al. [31] demon- tween experimental results and numerical results. The results from the
strated that inclusion of a thermoplastic phase can improve the impact experiments and the numerical simulation were used to analyze the
damage resistance and residual strength performance when subjected to damage evolution and failure mechanism of thin composite laminates
low-velocity impact loadings. Liu et al. [32] conducted experiments to under low-velocity impact and CAI loading conditions.
compare the behavior of hybrid UD/woven carbon-fiber reinforced
composite laminates and pure UD carbon-fiber reinforced composite 2. Experimental setup
laminates, and it was reported that the use of woven plies on the top
and bottom layers can reduce the extent of damage that occurs during 2.1. Materials and specimens
an impact event. Panettieri et al. [33] presented a benchmark study on
quasi-isotropic carbon/epoxy laminates under low-velocity impact with Rectangular specimens (150 mm × 100 mm) having a nominal
different impact energy levels, focusing on delamination threshold thickness of 2 mm were prepared using a carbon/epoxy prepreg with a
loads as well as the shape and propagation of the delamination damage. curing temperature of 120 °C. The material system was T700/3234 with
Other studies of low-velocity impact behavior of sandwich composite a stacking sequence of (45/0/-45/90/45/-45/90/0)s.
laminates have also been reported [34–38].
A number of researchers have used experimental methods to study 2.2. Instrumentation
the behavior of composite laminates under compression-after-impact
(CAI) loading conditions [39–45], where failure was reported to occur The following equipment was used during the tests:
instantaneously due to local instability and uniaxial compression. Im-
pact and CAI tests on thin laminates with rectangular and circular • Two V711 high-speed cameras with an ARAMIS 3D digital image
geometries performed by Ghelli and Minak [46] proved that impact correlation non-contact system (GOM Company, Germany) were
damage and buckling mode had a noticeable effect on compressive used to measure displacement and surface strain fields on the spe-
behavior and strength. Martins et al. [47] investigated the effects of cimens during axial compressive loading.
through-the-thickness reinforcement by tufting on impact and com- • The thermal field on the surface of specimens during testing (for
pression-after-impact behavior of woven carbon fiber composites. The both impact tests and CAI tests) was captured using an Infra Tec VH-
researchers reported that transversal tufting reinforcement achieved the 680HS infrared camera, which obtains images at a resolution of
best results when compared to angular and non-tufted laminate com- 640 × 480 pixels. The temperature resolution of this camera was
posites, with a decrease in the area of damage and an increase in 0.03 K with temperature measurement range from −40 K to 1200 K.

643
H. Tuo et al. Composites Part B 163 (2019) 642–654

The CAI strength was calculated according to Eq. (1).

F CAI = Pmax / A (1)

where Pmax is the maximum force prior to failure and A is the cross-
sectional area of the specimen.

4. Numerical Simulation

The damage behavior of composite laminates under low-velocity

impact and CAI loading conditions can mainly be classified into two
categories: intralaminar damage (which are fiber damage and matrix
cracking) and interlaminar damage (which is delamination).

Fig. 1. Impact setup for low-velocity impact tests. 4.1. Intralaminar damage model

• A POCKET-UT ultrasonic C-scanner (Physical Acoustics Corporation, An energy-based continuum damage mechanics model was im-
plemented in the finite element model to predict intralaminar damage.
USA) was employed to characterize the delamination damage. The
frequency of the ultrasonic transducer was 5 MHz, and a XY double- The following subsections describe how the model characterizes fiber
shaft automatic scanning device was used to realize a scan step damage and matrix damage.
length of 0.3 mm.
• A Sony SMZ645 optical microscope was used to obtain images of the 4.1.1. Fiber damage
damage in the specimens. To observe damage clearly, a Zeiss SUPRA Maximum strain criterion was selected in order to identify the in-
55 field emission scanning electron microscope (SEM) was used to itiation of tensile and compressive fiber damage in the composite la-
detect fiber and matrix damage; prior to the scanning, the specimens minates:
were processed by metal spraying to enhance their electrical con- Fiber tension mode( 11 > 0):
ductivity.
2
t 11
f11 ( 11) = 1 0
3. Experimental equipment and procedures
ot
11 (2)

3.1. Low-velocity impact tests Fiber compression mode( 11 < 0):

2
11
Low-velocity impact tests on composite laminates were conducted f11c ( 11) = oc
1 0
according to ASTM D7136 [52] using an Instron Dynatup 9250HV 11 (3)
Impact Testing Machine (Fig. 1) equipped with a 5.607 kg steel im- where and
ot
are failure initiation strain for fiber tension and
oc
11 11
pactor. The specimens were placed into a picture frame fixture having a compression, respectively, and are calculated according to 11 ot
= R t / E11 ,
125 mm × 75 mm open window in the center; four columnar rubber oc
= R c
/ E11 . R t
and R c
denote the longitudinal tensile and compressive
11
heads were used to clamp the specimen to prevent out-of-plane motion. strengths of the unidirectional laminates.
The tests were conducted by releasing the impactor from a specific After fiber damage initiation, the tensile and compressive fiber da-
height, allowing it to drop and impact the specimen. Four levels of mage factors were calculated according to the bilinear constitutive re-
impact energy (4 J, 6 J, 10 J and 15 J) were included in the testing lation in Eq. (4):
program. The impact energy was adjusted by changing the height of the
impactor. Each test group contained at minimum of three specimens to
ft (c ) ot (c )
11 ( 11 11 )
dtf ( 11) =
reduce data dispersion and improve data reliability, as shown in the test 11 (
ft (c )
11
ot (c )
11 ) (4)
matrix presented in Table 1.
where 11ft (c ) is the failure strain of fiber damage, expressed in terms of
fracture toughness Gft (c ) , associated with longitudinal strength and
3.2. Compression-after-impact tests
characteristic length lfib .
CAI tests were performed according to ASTM D 7137 [53] using an ft (c ) 2Gft (c )
=
Instron 8803 testing machine with displacement control of 1 mm/min. 11
R t (c ) lfib (5)
DIC was employed to monitor the displacement of the specimen during
compressive loading, and an infrared camera was used to monitor the
thermal field.
4.1.2. Matrix damage
The criterion proposed by Puck and Schurmann [54] was used for
Table 1 matrix damage initiation and evolution. This physical-based failure
Test matrix for impact and CAI tests. criterion is proposed as follows:
Matrix tension mode:
Case ID Impactor shape Impactor Impact No. of specimens
diameter energy 2
1 pt nt ( )
2
nl ( )
2

H16-4 Hemispherical 16 mm 4J 5 fE ( ) = n( ) + +
Rt R R R
H16-6 6J 4
H16-10 10 J 3
pt
H16-15 15 J 3 + n( ) for n( ) 0
R (6)
Note: H denoted hemispherical, the number 16 after ‘H’ denoted the diameter of
the hemispherical impactor, the last number after ‘-’ denoted impact energy. Matrix compression mode:

644
H. Tuo et al. Composites Part B 163 (2019) 642–654

pc
2 2 2 with n = n
, nt = nt
, nl = nl
.
nt ( ) nl ( ) eq eq eq
fE ( ) = n( ) + +
R R R Combine Eq. (11) and Eq. (12), the equivalent failure strain f
eq can
be expressed as:
pc
+ n( ) for n( ) <0 1/
R (7) f 2
o
n n
o
nl nl
o
nt nt
eq = + +
and lmat Gmt (c ) G12c G23c
(13)
t ,c t,c
p p t,c p R c
Next, a matrix damage factor is defined according to the bilinear
= cos2 + sin2 R =
R R R 2(1 + pc ) relation defined in Eq. (14).

2 2 f o
nt nl eq ( eq eq)
cos2 = 2 2
sin2 = 2 2 dmat =
nt + nl nt + nl
f
eq ( eq
o
eq) (14)
where n ( ) , nt ( ) and nl ( ) are stress components on the potential
Fiber damage and matrix damage irreversibility were taken into
fracture plane. The stress components lnt = [T ( )] 123 [T ( )]T can be
account by using the “max” function.
calculated using standard transformation matrix T ( ) from the material
coordinated system to the specified fracture plane. Rt is the transverse
tensile strength, Rc is the transverse compressive strength, and R is the 4.2. Interlaminar damage model
shear strength of composite laminates. The variable is a nonlinear
shear parameter that is equal to 1.58 × 10−8 and is determined by To effectively predict interlaminar delamination damage, a cohesive
experiments. The inclination parameters ( pt , pc , pt and pc ) element having a traction–separation constitutive relation was in-
characterize the influence of normal stress at the fracture surface on the corporated into the numerical model. A quadratic stress criterion was
matrix damage. The recommended values for these parameters are employed to determine the initiation of delamination damage under
listed in Table 2. In this model, the extended golden section search mixed mode loadings.
(EGSS) [55,56] algorithm was used to find the maximum value of the
stress exposure factor. When the global maximum value of fE reaches or 3
2
1
2
2
2

exceeds 1, the potential fracture plane is actual fracture plane and it + + =1

N S T (15)
indicates the initiation of matrix damage.
Once matrix damage initiates, the stress and strain components on where i (i = 1,2,3) are the in-plane shear stresses and normal stress
the fracture plane are recorded. The equivalent stress and equivalent respectively. N, S and T denote the interface tensile and shear strengths.
strain acting on fracture plane are defined as: A mixed-mode relationship proposed by Benzeggagh and Kenane (B–K
2 2 propagation criterion) [57] was applied to determine delamination
(9)
= 2 + +
eq n nl nt
propagation.
2 2
(10)
= 2 + +
eq n nl nt Gshear
GC = GIC + (GIIC GIC )
where < • > is McCauley operator and where n , nt and nl are strain GT (16)
components acting on fracture plane. The parameter eqo
is defined as the
where Gshear = GII + GIII and GT = GI + Gshear (in which GIC , GIIC and
equivalent stain of damage initiation on the fracture plane; eqf is the
GIIIC are the fracture toughness of modes I, II and III, respectively); GC is
equivalent strain of complete matrix failure, which is determined by
the mixed-mode fracture toughness; and is the mixed-mode para-
mixed-mode critical strain energy release in the element characteristic
meter, which is determined from experiments.
length.
The energy-based damage law under mixed-mode is built as:
4.3. Finite element model
gn gnl gnt
+ + =1
Gmt (c )/ lmat G12c / lmat G23c / lmat (11) The low-velocity impact model and the CAI finite element model
where gn , gnl and gnt are strain energy density of corresponding stress were implemented in ABAQUS/Explicit software. The damage con-
components; Gmt (c ) is transverse critical energy release rate; G12c and stitutive models were implemented in VUMAT subroutines. The overall
G23c are the in-plane and out-of-plane critical transverse energy release FEM model includes impactor, supporting plate and specimen. Both the
rates, respectively; lmat is the characteristic length; and is a material impactor and supporting plate were simulated using R3D4 rigid ele-
parameter having a value of 2 for this model. ments. A mass of 5.607 kg was assigned to the impactor through point
The strain energy density associated with each stress component mass definition. The general contact algorithm was used to simulate
due to complete matrix damage is given by contact behavior, and the friction coefficient was set as 0.25. In the FEM
model, the geometry of the specimens was rectangular
f
j 1 0 f 1 (150 mm × 100 mm) with a total thickness of 2 mm. Each ply has a
gjf = jd j = 0 f
j = n , nl, nt
0 2
j j
2
j j eq
(12) thickness of 0.125 mm and was discretized as linear brick elements
where is a stress component on the fracture plane at the initiation of
o (C3D8R), and a material orientation is defined for each ply to represent
j
matrix damage. j denotes the mix-mode ratio and can be expressed the laminate layup. Zero-thickness 8-node 3D cohesive elements
(COH3D8) were inserted between layers to model the interface. The
Table 2 cohesive elements and the plies were connected by shared nodes
Recommended inclination parameters. methods.
The boundary conditions of the low-velocity impact model and the
Material pt pc pt pc
CAI model are shown in Fig. 2. Tables 3–5 present the properties of the
unidirectional laminates: material properties are listed in Table 3,
GFRP 0.30 0.25 0.20–0.25 0.20–0.25
CFRP 0.35 0.30 0.25–0.30 0.25–0.30 fracture toughness properties are listed in Table 4, and interface
properties are included in Table 5.

645
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 2. Boundary conditions for (a) low-velocity impact model; (b) CAI model.

Table 3
Material properties of unidirectional laminate.
Elastic modulus/GPa Poisson's ratio Strength/MPa

E1 E2 = E3 G12 = G13 G23 12 13 23 Rt Rc Rt Rc R

140 8.8 4.6 3.3 0.27 0.27 0.36 2186 1240 87.1 229 165

Table 4
Fracture toughness properties of the unidirectional laminate.
Parameter Gft /(kJ/m2) Gfc /(kJ/m2) Gmt /(kJ/m2) Gmc /(kJ/m2) G12c /(kJ/m2) G23c /(kJ/m2)

Value 90 82 0.52 1.61 0.92 0.92

Table 5
Material properties of interface cohesive elements.
Parameter G IC /(J/m2) G IIC (J/m2) G IIIC (J/m2) N/MPa S/MPa T/MPa

Value 560 920 920 1.46 30 80 80

5. Results and discussion using an optical microscope and SEM are presented in Fig. 4. As shown
in Fig. 4(a) and (b), delamination and matrix damage in case H16-10
5.1. Impact behavior occurred in the lower section near the impact region; the numerical
results also predict that a small area with matrix damage would be
For impact damage in composite laminates, delamination damage is observed (Fig. 5(a)). For case H16-15, as shown in Fig. 4(c), severe
one of the main damage modes. Fig. 3 shows the delamination shapes interlaminar delamination damage occurred in the upper and lower
and area obtained by experimental testing and numerical simulation section near the impact region, while less severe damage occurred in
under four levels of impact energy, and the results show good corre- the middle section. The phenomenon can be explained that under low-
lation. The dimensions of captured square regions are velocity impact loadings, the local contact led to visible bending de-
90 mm × 60 mm. The delamination damage was observed at locations formation, and the upper and lower sections had larger bending de-
around the impact indentations in the form of an irregular circle under formation and stresses, which may induce delamination and matrix
low impact energy. As the impact energy increased, the delamination damage. From Fig. 4(d), it can be observed that a small number of fibers
area increased and the shapes gradually changed from an irregular ruptured. In addition, fiber-matrix debonding and matrix cracks were
circle to an irregular ellipse. No matrix damage was observed via op- located in close proximity, which will have a significant influence on
tical microscope or SEM for samples in impact cases H16-4 (impact the residual compressive strength of the composite. Severe matrix da-
energy of 4 J) and H16-6 (impact energy of 6 J). The numerical results mage was also predicted in the numerical simulation (Fig. 5(b)); the
also predict that only delamination damage will occur in these two simulation also predicted a small area of fiber damage located in the
cases. impact region (Fig. 5(c)).
In order to obtain more detailed damage information for samples in The experimental results and numerical results showed that the
cases with moderate and high impact energy, the specimens in cases damage states can be divided into three stages. In low levels of impact
H16–10 and H16-15 were cut using abrasive water jet machining energy (4 J and 6 J), only delamination damage occurred. Under
[58,59]. The damage details from the central cut section obtained by moderate impact energy (10 J), delamination damage and matrix

646
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 3. Comparison between delamination damage obtained by C-scan and FEM: (a) delamination shapes; (b) delamination area.

damage were observed. For high impact energy (15 J), severe delami- Fig. 6(b) presents the impact force–displacement curves under four
nation, matrix damage and fiber damage occurred. levels of impact energy. The curves of cases with impact energy of 4 J
The experimental impact force-time curves of H16 impact cases and 6 J were smooth, while curves for cases with impact energy of 10 J
under four levels of impact energy are shown in Fig. 6(a). In cases and 15 J showed oscillation. The maximum and final displacement of
where the impact energy was low (4 J and 6 J), the curves were sym- the impactor increased with the increase of the impact energy. The area
metric and exhibited sinusoidal shapes with smooth loading and un- enclosed by the curves indicated the energy absorbed by the laminate,
loading portion. For moderate impact energy (10 J), delamination da- which increased with the increase of impact energy.
mage and matrix damage were observed. The force–time curve for these The absorbed energy–time histories for the four levels of impact
cases revealed strong oscillations when the contact force approached energy are presented in Fig. 6(c). In the impact tests, the kinetic energy
the maximum value, which indicated the initiation of matrix damage. of the impactor began to transfer to the composite laminates as soon as
For cases where the impact energy was high (15 J), violent oscillations contact was made. The energy began to be absorbed through the pro-
were observed near peak force, and a sudden force drop occurs with the cesses of elastic deformation, damage dissipation and other vibrations
appearance of fiber damage. For cases with 4 J, 6 J and 10 J impact, the of the laminates. When the velocity of impactor turned to zero and the
impact durations were almost same (8.7 ms), but a longer duration displacement reached the maximum, the elastic energy in the laminates
(9.6 ms) was observed when the impact energy was 15 J. It can be ex- drove the impactor to rebound. Because the energy absorbed by vi-
plained that the serious matrix damage and fiber damage in impact case bration of the laminates was small enough to be considered as negli-
will lead the significant stiffness degradation in the contact area. By gible, the final absorbed energy can be mainly attributed to the damage
comparing the curves of four impact energy levels, we can conclude mechanism. At the lowest level of impact energy (4 J), the ratio of
that the different damage states show different impact force-time curve absorbed energy was 11.94%; this ratio increased significantly to
characteristics. 67.08% at the highest impact energy (15 J). We can conclude that with

647
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 4. Images of impacted specimens: (a) optical microscope image of impact case H16-10; (b) SEM image of impact case H16-10; (c) optical microscope image of
impact case H16-15; (d) SEM image of impact case H16-15.

Fig. 5. The damage obtained by numerical simulation: (a) matrix damage in impact case H16-10; (b) matrix damage in impact case H16-15; (c) fiber damage in
impact case H16-15.

Fig. 6. Impact response of specimens in H16 impact cases: (a) impact force versus time; (b) impact force versus impactor displacement; (c) absorbed energy versus
time.

648
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 7. Comparison between numerical simulations and experimental results for (a) maximum impact force; (b) maximum displacement; (c) absorbed energy.

Fig. 8. Infrared thermography images from a time sequence taken in impact case H16-15: (a) t = 0.02 s; (b) t = 0.04 s; (c) t = 0.22 s; (d) t = 2.2 s.

Fig. 9. Infrared thermography images with the maximum temperature in impact cases: (a) H16-4; (b) H16-6; (c) H16-10; and (d) H16-15.

the increase of impact energy, the laminates absorbed more energy, Fig. 10, which shows that the time to reach the maximum temperature
which also meant that the damage state became more severe. decreased with an increase in impact energy; the slope of the ΔTmax-t
Fig. 7 shows the comparison of maximum impact force, maximum curves also increased with the increase impact energy. From the shape
displacement and absorbed energy between the numerical results and of the curve, it can be concluded that the curve gradually changed from
experimental results at the four impact energy levels. As can be noticed smooth to steep as the impact energy increased. In impact case H16-4,
in these figures, the maximum impact force, maximum displacement of the warm area was very small and uniform with a maximum ΔT of only
impactor and absorbed energy delivered by the numerical model cor-
related well with the experimental results.
Infrared thermography was employed to study the thermal field
variation during the impact process. Some ΔT thermal images taken
during impact process in the H16-15 impact cases are shown in Fig. 8. It
can be seen that hot spots or lines appeared immediately during the
impact process, which indicated the initiation of damage. The warm
area rapidly enlarged, and ΔT also reached the maximum value quickly.
After that, the surface temperature began to degrade slowly.
Fig. 9 presents thermal images obtained at the maximum tempera-
ture under four levels of impact energy. It can be observed that the area
and maximum ΔT increased as the impact energy increased from 4 J to
15 J; this result indicates that the area of damage and the severity of
damage under low-velocity impact loading increased with an increase
in impact energy.
To perform a quantitative study, a circular section with a diameter
of 20 mm was selected in the impact area (Fig. 9). In this section, the
point marked as P1 (the point of maximum ΔT) can be automatically
located using commercial software IRBIS3. Next, the ΔT of P1 was ex-
tracted for four levels of impact energy and was plotted against time in
Fig. 10. ΔTmax distribution versus time in H16 impact cases.

649
H. Tuo et al. Composites Part B 163 (2019) 642–654

largest reduction of CAI strength was observed for impact case H16-15.
This demonstrates that the fiber damage induced by relatively high
impact energy will significantly affect the CAI strength.
A typical compressive load–displacement curve in impact case H16-
15 is presented in Fig. 12 as an example. As shown in Fig. 12(a), the
global compressive load–displacement curves can be divided into three
loading stages. At the beginning stage (Stage I), which includes the
segment between the origin (O) to Point A, showed a pure linear elastic
relation. When the compressive load reached Point A, the initial stiff-
ness KI degraded to a lower stiffness KII, indicating the onset of local
buckling. With the evolution of local buckling in Stage II (the segment
between Point A and Point D), delamination damage and matrix da-
mage would be further induced and would eventually lead to fiber
damage in the central area of the impact region. In Stage III (Point D to
Point E), the specimens failed completely with a sudden drop in load.
As shown in Fig. 12(b), the strain-compressive load curves of im-
pacted laminates (where F denotes the impact side of the test specimen,
B represents the back side of the specimen, mean denotes the average
strain, and bending represents the bending strain calculated by
Fig. 11. Comparison between numerical and experimental results for CAI tests. ( B1 + B 2 F1 F 2 )/4 )) remained linear during Stage I. The strain
values of four gauges in the initial loading stage remained nearly same
1.49 K. The slope of the curve remained small, and after it reached the until 25 kN, which indicates that local buckling occurred. The initial
maximum the temperature remained stable for a relatively long period stiffness KI had degraded to KII for local bucking; local bucking devel-
of time before degrading slowly. Considering the results obtained from oped gradually before the composite laminate finally collapsed. The
the experiments, the simulation and the damage analysis, we can know mean values of the four gauges remained linear; the bending strain
that only delamination damage occurred in the impact case and the started at 25 kN and developed gradually, which showed that the
delamination damage released very little energy [15]. For H16-10 case, bending deformation was initiated in Stage II.
both matrix and delamination damage were present. A larger warm To investigate the out-of-plane displacement and strain field dis-
area, a larger slope of temperature increase in the plots and a more tributions during compressive loading and to further determine the
rapid temperature degradation speed were observed. We also can compressive failure mechanisms, a 3D DIC system was employed to
conclude that for matrix damage, the rate of damage evolution was observe the impacted surface during the compressive tests. Each sub-
higher, and more energy was released. For impact case H16-15, the figure in Fig. 13 contains five subplots representing Stages A, B, C, D
warm area reached the largest size and located along the direction of and E in the compressive load–displacement curve. In these figures, a
−45°; in addition, the maximum temperature and curve slope were the positive value (marked in red) indicates that a discrete point on the
highest of all the test cases, due to the appearance of fiber damage. For laminate moves from the original position to one that is closer to the
fiber damage, a large quantity of energy was released instantaneously DIC cameras, while a negative value (marked in blue) indicates
during brittle rupture [60,61]. Thus, the curve shows three character- movement to a position farther from the DIC system.
istics: a more rapid increase in temperature, a steeper slope for the Fig. 13(a) shows the DIC results for out-of-plane displacement of the
curve, and more rapid degradation. impacted specimens under compressive loading. The local out-of-plane
displacement originated in impacted regions and became obvious at
5.2. CAI strength and damage mechanism Point A. The phenomenon may result from local nonlinearity and local
buckling in the impacted region. The deformation area then increased,
Fig. 11 presents a comparison of the numerical predictions and and its amplitude became relatively stable in Stage II of the loading.
experimental CAI values for four impact energy levels. As can be no- Finally, sudden and complete failure occurred along the transverse
ticed from this figure, the experimental results and predicted values cross section of the impact area due to global buckling.
show good correlation. It can also be noticed that for all impact cases, Fig. 13(b) shows the in-plane longitudinal compressive strain y on
the CAI strength decreased with the increase in impact energy, and the the impact side under different stages of loading. An interesting

Fig. 12. CAI behavior in impact case H16-15: (a) load versus displacement curves; (b) strain versus load curves.

650
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 13. DIC results at different stages of compressive loading: (a) out-of-plane displacement in impacted side; (b) y in impacted side; (c) x in impacted side; and (d)
xy in impacted side.

observation was made regarding the distribution feature: when the load Fig. 13(d). From the figure, it can be concluded that the larger shear
reached point A, the impact region began to bear a greater amount of strain first appeared in the 45° and −45° directions of the impact re-
longitudinal compressive load. The strain component y became larger gion. As the damage accumulated, the values became larger and ex-
(in terms of absolute value), and it gradually extended to the two outer tended in the direction perpendicular to the loading as the compressive
free edges along the transverse direction. At the last loading stage, the load increased. The area in the last contour plot, which showed the final
area with maximum absolute values extended across the laminates, failure of the composite, was the same area that was covered by the
following a narrow and curved path. distributions of x and y .
Fig. 13(c) presents the distribution of the strain component x under Fig. 14 shows longitudinal compressive contours for impact case
a compressive loading condition. From this figure, it can be noticed that H16-15 at failure. It can be noticed from this figure that the long-
on the left and right sides of the impact region, the transverse strains itudinal compressive strain contours predicted by the numerical model
were high. Some strains were in the form of transverse tensile strain are in good agreement with the DIC results.
(positive values, which are marked in red) and some were in the form of To further investigate the damage and failure mechanism during
compressive strain (negative values, which are marked in blue), which compression-after-impact loadings, the CAI specimens in the H16-15
exhibited an irregular distribution. test group were cut into small pieces using a water-jet method [58,59]
The distribution of the shear strain component xy is shown in and were inspected by optical microscope and SEM. In the side view of

651
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 14. Comparison of longitudinal strain contours obtained from (a) DIC results; (b) simulation results.

Fig. 15. Images for samples in CAI tests obtained by (a) optical microscope; (b) SEM.

Fig. 16. Numerical simulation of CAI damage contours for impact case H16-15: (a) delamination damage; (b) matrix damage; (c) fiber damage.

the sample obtained from the optical microscope (Fig. 15(a)), it can be noted for the CAI behavior, and some matrix crush along the fracture
seen that the damage mechanisms were complex for various damage section was detected. The largest area of delamination damage was also
types and interactions. A larger amount of fiber rupture was found found in this fracture section. From the SEM images in Fig. 15(b), it can
along the thickness direction, which was the main damage pattern clearly be seen that severe fiber damage was found in the fracture

652
H. Tuo et al. Composites Part B 163 (2019) 642–654

Fig. 17. Images obtained under compressive load in impact case H16-15: ΔT images from a time sequence taken at (a) t = 0.02 s; (b) t = 0.08 s; (c) t = 0.12 s; (d)
t = 0.51 s; (e) macroscopic fracture.

the matrix damage contour obtained in the numerical simulation

(Fig. 16(b)). Thus, it can be concluded that Section 1 exhibited fiber
compressive failure and Section 2 exhibited the matrix failure.

6. Conclusions

This work presented an experimental and numerical investigation

on the impact and CAI failure behavior of thin composite laminates. A
detailed assessment of the damage and failure mechanism was carried
out by various testing methods, such as ultrasonic C-scan, SEM, DIC and
infrared thermography, together with numerical simulations. Based on
this systematic study, the following conclusions can be drawn:

(a) For the case with impact energy of 15J, due to the existence of fiber
damage, a sudden load drop was observed followed by a longer
impact duration, which resulted in the largest ratio of energy ab-
Fig. 18. Contours of different damage types in CAI tests for impact case H16- sorption.
15: (a) Section 1; (b) Section 2. (b) The CAI process of thin composite laminates can be divided into
three stages: an elastic phase (Stage I), initiation of local buckling
section and that matrix damage was found nearby. and propagation (Stage II) and sudden collapse (Stage III). It was
Fig. 16 presents the numerical results regarding the different types concluded that damage always started from the impact point and
of damage produced by CAI loading conditions. Fiber damage, matrix extended instantaneously across the laminate to the two outer
damage and delamination damage were predicted across the entire edges. The area of fiber damage was found to be narrow, and the
width of laminate and, in each case, were aligned with the central matrix-damage area was large enough to contain the section with
impact region. The predicted area with fiber damage was narrow, the fiber damage. The area of delamination damage was the largest and
area for matrix damage was somewhat larger, and the area for dela- exhibited an irregular shape.
mination was largest. This trend, which can be noticed in the images (c) A good correlation was found between the thermal field and the
obtained via optical microscopy, was also in accordance with the con- damage initiation and evolution of composite laminates during the
clusions made by Abir et al. [40]. impact and CAI tests. It can be a promising method to study the
Fig. 17 presents ΔT images from a time sequence in impact case damage and failure mechanism of composite materials.
H16-15 during compressive loading. The last frame before the damage (d) The finite element model based on continuum damage mechanics
initiation was defined as the starting point (and was set as t = 0 s). It can predict accurately the impact and CAI behavior of the thin
can be observed that a slender line with the highest temperature in- composite laminate.
crement extended across the cross section of the indentation with some
irregular short branches. From the experimental and numerical results,
we can conclude that the area with largest temperature increment is Acknowledgement
consistent with the macroscopic fracture types in the failed specimen.
We can also see that the area with the largest temperature change was This work was supported by the National Natural Science
related to the presence of fiber compressive damage. Foundation of China (Grant No. NSFC 11772267) and Fundamental
Additional quantitative analysis was conducted to determine the Research Funds for the Central Universities. The authors give sincere
damage and failure mechanism for the thin composite laminates under thanks to Dr. Wenzhi Wang, Xiaoyu Zhang, Jianwu Zhou and Huifang
low-velocity impact. Fiber damage was found to release a large quantity Liu in the School of Aeronautics at Northwestern Polytechnical
of energy, and the temperature increment was the largest. It can be University for their assistance in the experimental work.
observed from Fig. 18(a) that there is a temperature increase ranging
from 4 K to 9 K, which was consistent with the fiber damage contour
obtained in the numerical simulation (Fig. 16(c)). Matrix damage re- Appendix A. Supplementary data
sulted in a release of less energy over a smaller temperature range;
Section 2 (marked in yellow), which had a temperature increase ran- Supplementary data to this article can be found online at https://
ging from 0.5 K to 3.99 K (Fig. 18(b)), which was also consistent with doi.org/10.1016/j.compositesb.2019.01.006.

653
H. Tuo et al. Composites Part B 163 (2019) 642–654

References improve impact and post-impact performances of carbon fibre reinforced thermo-
setting composites — a review. Mater Des 2015;85:582–97.
[32] Liu H, Falzon BG, Tan W. Experimental and numerical studies on the impact re-
[1] Wang J, Wang H, Chen B, Huang H, Liu S. A failure mechanism based model for sponse of damage-tolerant hybrid unidirectional/woven carbon-fibre reinforced
numerical modeling the compression-after-impact of foam-core sandwich panels. composite laminates. Compos B Eng 2018;136:101–18.
Compos Sci Technol 2017;151:258–67. [33] Panettieri E, Fanteria D, Montemurro M, Froustey C. Low-velocity impact tests on
[2] Guinard S, Allix O, Guédra-Degeorges D, Vinet A. A 3D damage analysis of low- carbon/epoxy composite laminates: a benchmark study. Compos B Eng
velocity impacts on laminated composites. Compos Sci Technol 2002;62:585–9. 2016;107:9–21.
[3] Léonard F, Stein J, Soutis C, Withers PJ. The quantification of impact damage [34] Dogan A, Arikan V. Low-velocity impact response of E-glass reinforced thermoset
distribution in composite laminates by analysis of X-ray computed tomograms. and thermoplastic based sandwich composites. Compos B Eng 2017;127:63–9.
Compos Sci Technol 2017;152:139–48. [35] Kurşun A, Şenel M, Enginsoy HM, Bayraktar E. Effect of impactor shapes on the low
[4] Liu H, Nie H, Zhang C, Li Y. Loading rate dependency of Mode I interlaminar velocity impact damage of sandwich composite plate: experimental study and
fracture toughness for unidirectional composite laminates. Compos Sci Technol modelling. Compos B Eng 2016;86:143–51.
2018;167:215–23. [36] Chen Y, Hou S, Fu K, Han X, Ye L. Low-velocity impact response of composite
[5] Bogenfeld R, Kreikemeier J, Wille T. Review and benchmark study on the analysis of sandwich structures: modelling and experiment. Compos Struct 2017;168:322–34.
low-velocity impact on composite laminates. Eng Fail Anal 2018;86:72–99. [37] Arachchige B, Ghasemnejad H. Effect of variable core stiffness on the impact re-
[6] Polymer matrix composites: guidelines for characterization of structural materials. sponse of curved sandwich plates. Compos Struct 2018;200:565–78.
Washington, DC: Military Handbook MIL-HDBK-17-1F; June 2002. 16. [38] Kyner A, Dharmasena K, Williams K, Deshpande V, Wadley H. Response of square
[7] Zhang X, Hounslow L, Grassi M. Improvement of low-velocity impact and com- honeycomb core sandwich panels to granular matter impact. Int J Impact Eng
pression-after-impact performance by z-fibre pinning. Compos Sci Technol 2018;117:13–31.
2006;66:2785–94. [39] Sun XC, Hallett SR. Failure mechanisms and damage evolution of laminated com-
[8] Aymerich F, Priolo P. Characterization of fracture modes in stitched and unstitched posites under compression after impact (CAI): experimental and numerical study.
cross-ply laminates subjected to low-velocity impact and compression after impact Compos Part A Appl Sci Manuf 2018;104:41–59.
loading. Int J Impact Eng 2008;35:591–608. [40] Abir MR, Tay TE, Ridha M, Lee HP. Modelling damage growth in composites sub-
[9] Williams GJ, Bond IP, Trask RS. Compression after impact assessment of self-healing jected to impact and compression after impact. Compos Struct 2017;168:13–25.
CFRP. Compos Part A Appl Sci Manuf 2009;40:1399–406. [41] Li N, Chen PH. Experimental investigation on edge impact damage and
[10] Djabali A, Toubal L, Zitoune R, Rechak S. An experimental investigation of the Compression-After-Impact (CAI) behavior of stiffened composite panels. Compos
mechanical behavior and damage of thick laminated carbon/epoxy composite. Struct 2016;138:134–50.
Compos Struct 2018;184:178–90. [42] Tan W, Falzon BG, Chiu LNS, Price M. Predicting low velocity impact damage and
[11] Bull DJ, Spearing SM, Sinclair I. Observations of damage development from com- Compression-After-Impact (CAI) behaviour of composite laminates. Compos Part A
pression-after-impact experiments using ex situ micro-focus computed tomography. Appl Sci Manuf 2015;71:212–26.
Compos Sci Technol 2014;97:106–14. [43] Debski H, Rozylo P, Gliszczynski A. Effect of low-velocity impact damage location
[12] Li X, Ma D, Liu H, Tan W, Gong X, Zhang C, et al. Assessment of failure criteria and on the stability and post-critical state of composite columns under compression.
damage evolution methods for composite laminates under low-velocity impact. Compos Struct 2018;184:883–93.
Compos Struct 2019;207:727–39. [44] Sun W, Guan Z, Ouyang T, Tan R, Zhong X. Effect of stiffener damage caused by low
[13] Abisset E, Daghia F, Sun XC, Wisnom MR, Hallett SR. Interaction of inter- and in- velocity impact on compressive buckling and failure modes of T-stiffened composite
tralaminar damage in scaled quasi-static indentation tests: Part 1 – Experiments. panels. Compos Struct 2018;184:198–210.
Compos Struct 2016;136:712–26. [45] Aktaş M, Karakuzu R, Arman Y. Compression-after impact behavior of laminated
[14] Hongkarnjanakul N, Bouvet C, Rivallant S. Validation of low velocity impact composite plates subjected to low velocity impact in high temperatures. Compos
modelling on different stacking sequences of CFRP laminates and influence of fibre Struct 2009;89(1):77–82.
failure. Compos Struct 2013;106:549–59. [46] Ghelli D, Minak G. Low velocity impact and compression after impact tests on thin
[15] Meola C, Boccardi S, Carlomagno GM, Boffa ND, Monaco E, Ricci F. Nondestructive carbon/epoxy laminates. Compos B Eng 2011;42:2067–79.
evaluation of carbon fibre reinforced composites with infrared thermography and [47] Martins AT, Aboura Z, Harizi W, Laksimi A, Khellil K. Analysis of the impact and
ultrasonics. Compos Struct 2015;134:845–53. compression after impact behavior of tufted laminated composites. Compos Struct
[16] Petit S, Bouvet C, Bergerot A, Barrau J. Impact and compression after impact ex- 2018;184:352–61.
perimental study of a composite laminate with a cork thermal shield. Compos Sci [48] Lisle T, Bouvet C, Hongkarnjanakul N, Pastor M, Rivallant S, Margueres P. Measure
Technol 2007;67:3286–99. of fracture toughness of compressive fiber failure in composite structures using
[17] Choi HY, Wang HS, Chang F. Damage of graphite/epoxy composite plates due to infrared thermography. Compos Sci Technol 2015;112:22–33.
line-loading impact. J Compos Mater 1992;26:804–27. [49] Liu H, Falzon BG, Tan W. Predicting the Compression-After-Impact (CAI) strength of
[18] Choi HY, Wang HS, Chang F. A new approach toward understanding damage me- damage-tolerant hybrid unidirectional/woven carbon-fibre reinforced composite
chanisms and mechanics of laminated composites due to low-velocity impact: Part laminates. Compos Part A Appl Sci Manuf 2018;105:189–202.
II-analysis. J Compos Mater 1991;25:1012–38. [50] García-Rodríguez SM, Costa J, Singery V, Boada I, Mayugo JA. The effect inter-
[19] Choi HY, Wang HS, Chang F. A new approach toward understanding damage me- leaving has on thin-ply non-crimp fabric laminate impact response: X-ray tomo-
chanisms and mechanics of laminated composites due to low-velocity impact:Part I- graphy investigation. Compos Part A Appl Sci Manuf 2018;107:409–20.
experiments. J Compos Mater 1991;25:992–1011. [51] Caminero MA, García-Moreno I, Rodríguez GP. Experimental study of the influence
[20] de Moura MFSF, Goncalves JPM, Marques AT, et al. Modelling compression failure of thickness and ply-stacking sequence on the compression after impact strength of
after low velocity impact on laminated composites using interface elements. J carbon fibre reinforced epoxy laminates. Polym Test 2018;66:360–70.
Compos Mater 1997;31:1462–79. [52] ASTM D7136/D7136M-15 standard test method for measuring the damage re-
[21] de Moura MFSF, Gonçalves JPM, Marques AT, de Castro PMST. Prediction of sistance of a fiber-reinforced polymer matrix composite to a drop-weight impact
compressive strength of carbon–epoxy laminates containing delamination by using event.
a mixed-mode damage model. Compos Struct 2000;50:151–7. [53] ASTM D7137/D7137M - 12 standard test method for compressive residual strength
[22] de Moura MFSF, Marques AT. Prediction of low velocity impact damage in carbo- properties of damaged polymer matrix composite.
n–epoxy laminates. Compos Part A Appl Sci Manuf 2002;33:361–8. [54] Puck A, Schürmann H. Failure analysis of FRP laminates by means of physically
[23] de Moura MFSF, Gonçalves JPM. Modelling the interaction between matrix based phenomenological models. Compos Sci Technol 2002;62:1633–62.
cracking and delamination in carbon–epoxy laminates under low velocity impact. [55] Wiegand J, Petrinic N, Elliott B. An algorithm for determination of the fracture
Compos Sci Technol 2004;64:1021–7. angle for the three-dimensional Puck matrix failure criterion for UD composites.
[24] Wang HR, Long SC, Zhang XQ, Yao XH. Study on the delamination behavior of thick Compos Sci Technol 2008;68:2511–7.
composite laminates under low-energy impact. Compos Struct 2018;184:461–73. [56] Schirmaier FJ, Weiland J, Kärger L, Henning F. A new efficient and reliable algo-
[25] Aktaş A, Aktaş M, Turan F. Impact and post impact (CAI) behavior of stitched rithm to determine the fracture angle for Puck's 3D matrix failure criterion for UD
woven–knit hybrid composites. Compos Struct 2014;116:243–53. composites. Compos Sci Technol 2014;100:19–25.
[26] Artero-Guerrero JA, Pernas-Sánchez J, López-Puente J, Varas D. Experimental study [57] Benzeggagh ML, Kenane M. Measurement of mixed-mode delamination fracture
of the impactor mass effect on the low velocity impact of carbon/epoxy woven toughness of unidirectional glass/epoxy composites with mixed mode bending ap-
laminates. Compos Struct 2015;133:774–81. paratus. Compos Sci Technol 1996;49:439–49.
[27] Evci C, Gülgeç M. An experimental investigation on the impact response of com- [58] El-Hofy M, Helmy MO, Escobar-Palafox G, Kerrigan K, Scaife R, El-Hofy H. Abrasive
posite materials. Int J Impact Eng 2012;43:40–51. water jet machining of multidirectional CFRP laminates. Procedia CIRP
[28] Vieille B, Casado VM, Bouvet C. Influence of matrix toughness and ductility on the 2018;68:535–40.
compression-after-impact behavior of woven-ply thermoplastic- and thermosetting- [59] Alberdi A, Suárez A, Artaza T, Escobar-Palafox GA, Ridgway K. Composite cutting
composites: a comparative study. Compos Struct 2014;110:207–18. with abrasive water jet. Procedia Engineering 2013;63:421–9.
[29] Vieille B, Casado VM, Bouvet C. About the impact behavior of woven-ply carbon [60] Meola C, Carlomagno GM. Infrared thermography to evaluate impact damage in
fiber-reinforced thermoplastic- and thermosetting-composites: a comparative study. glass/epoxy with manufacturing defects. Int J Impact Eng 2014;67:1–11.
Compos Struct 2013;101:9–21. [61] Meola C, Boccardi S, Carlomagno GM, Boffa ND, Ricci F, Simeoli G, et al. Impact
[30] Arikan V, Sayman O. Comparative study on repeated impact response of E-glass damaging of composites through online monitoring and non-destructive evaluation
fiber reinforced polypropylene & epoxy matrix composites. Compos B Eng with infrared thermography. NDT&E Int 2017;85:34–42.
2015;83:1–6.
[31] Nash NH, Young TM, McGrail PT, Stanley WF. Inclusion of a thermoplastic phase to

654