Composite Structures 184 (2018) 352–361

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Analysis of the impact and compression after impact behavior of tufted MARK
laminated composites
⁎
A.T. Martins , Z. Aboura, W. Harizi, A. Laksimi, K. Khellil
Sorbonne Universités, Université de Technologie de Compiègne, CNRS, Laboratoire de Mécanique Roberval UMR 7337, Dept GM, CS 60319, 60203 Compiègne Cedex,
France

A R T I C L E I N F O A B S T R A C T

Keywords: The effect of trough-the-thickness reinforcement by tufting on impact and after impact behavior of woven carbon
Tufting fiber composites is investigated. The density and angle of the carbon fiber threads during tufting process were
Trough-the-thickness reinforcement varied to achieve better results on damage tolerance. The samples were submitted to out-of-plane impact with
Impact two different energies to compare their response under loading. Then, they were analyzed by ultrasonic C-Scan
Compression after impact
technique in order to evaluate the damaged zone. Further, compression after impact (CAI) assisted by acoustic
Multi-instrumented testing
emission and Digital Image Correlation compared the compressive strength and materials response according to
their tufting parameters. Also, compression before impact (CBI) were performed to evaluate the in-plane
properties of the tufted composites. The transversal tufting reinforcement achieved the best results when
compared to angular and non-tufted laminate composites. Moreover, when increasing the tufting density, the
damaged area is decreased and the ultimate compressive strength improved. Tufting reinforcements were found
to decrease the damage area up to 4 times (transversal tufting) when compared to non-tufted laminates.
Additionally, the residual ratio (CAI/CBI) increased up to 30% and 28% for the transversal and angular re-
inforcements respectively in comparison to the reference.

1. Introduction loads, owing to local instability [6–8].

Carbon fiber reinforced composites has been increasingly applied to
aircraft structures. The well known higher strength-to-weight as com-
pared to metallic alloys helps the aircraft to reduce weight, conse-
quently improving fuel efficiency. They also can offer the advantage to
reduce parts counts and their longer life cycle can decrease the main-
tenance frequency [1]. However, despite great in-plane properties as
strength and stiffness of the laminate composite materials, they can
show weakness, particularly in the transversal direction when sub-
mitted to impact loading.
Laminated composite materials are liable to fatal damage under
impact on service conditions, maintenance operations or even in part
manufacturing [2–4].Therefore, when a foreign object impacts a lami-
nate, several damage modes as delaminations, fiber breakage, and
matrix cracks occur in the composite structure [5]. Particularly, dela-
mination is one of the principal damage mechanisms on impact, espe-
cially in low-velocity under transversal loading. This failure mode re-
duces the impact strength due to the low interlaminar strength of the
composites materials. In addition, this damage can seriously reduce the
load bearing capacity of the laminate, especially under compressive
Hence, considerable improvements in damage tolerance and dela-
mination growth resistance can be achieved using both the constituents
of the composite as tougher resin or reinforcement in z-direction [9,10].
For instance, the composites can be toughened by incorporation of
micro-phase dispersed rubbery or thermoplastic polymers, rigid parti-
cles or ‘hybrid-toughened’ epoxy polymers by combining both rubber
toughening and silica nanoparticles. However, these methods have
difficulties in the proper particles distribution (rubber or thermoplastic
polymers) in the matrix. Moreover, to achieve significant toughness an
important amount of particle is necessary which typically increases the
resin viscosity, usually unacceptable for resin-infusion process [11].
Furthermore, the tougher resins provide only moderate improvements
on impact damage resistance and the usage in large practical composite
structures is still being studied [9,12].
The interleaving approach has also been used to improve the pe-
netration resistance and damage tolerance of carbon/epoxy composites
by increasing their fracture toughness [13–18]. This method involves
the incorporation of thin layers of a high shear strength resin along the
interface of the laminas. However, it is not practical to add an adhesive
layer along every interface of the laminate. This may cause the laminate

⁎
Corresponding author.
E-mail address: alan-tulio.martins@utc.fr (A.T. Martins).

http://dx.doi.org/10.1016/j.compstruct.2017.09.096
Received 16 May 2017; Received in revised form 28 August 2017; Accepted 28 September 2017
Available online 05 October 2017
0263-8223/ © 2017 Elsevier Ltd. All rights reserved.
A.T. Martins et al.
a great weight penalty and potentially decrease of glass transition (Tg)
[14,16]. Moreover, because the toughened resin layers have relatively
lower stiffness and strength, their application has to be limited in order
not to alter the overall composite performance [19,20].
Interesting enhancements in the transversal mechanical properties
of the laminate composites can be achieved using three-dimensional
(3D) reinforcements. Three-dimensional textile composites contain a
3D network of planar and through-the-thickness reinforcement (TTR)
fibers. There are several types of 3D textile composites, and they are
classified according to the method by which the 3D fibers reinforcement
are formed within the material [21]. The main methods are 3D
weaving, stitching, tufting, z-anchoring and z-pinning which are well
dependent on the applied material (ex. dry perform or pre-im-
pregnated) and end application.
In the present work, the tufting process has been used. It is a one-
sided stitching process in which the needle pushes high-strength yarn
into the fabric and tufts are held in place when the needles are with-
drawn by friction resistance imposed through the fabric [21]. The ad-
vantage of tufting is the low tension under which the thread is inserted.
This results in a reduction of the stitching effect on the in-plane prop-
erties of polymer matrix composites [9]. TTR in the form of stitching
has been shown to increase the fracture toughness under both mode I
and II loading [22]. In general, for mode I, the interlaminar loading
increases the delamination resistance by reducing the crack opening
displacement, while in mode II, it increases the delamination resistance
by resisting crack sliding displacement [23].
Karuppannan et al. [9] compared unidirectional and quasi-isotropic
tufted carbon fabric composites and showed that in-plane properties of
quasi-isotropic and notched composites are not well affected with
transversal tufting reinforcement. Also, the mode I fracture toughness
obtained values more than 16 times greater compared with non-tufted
composites. Deconinck et al. [24] studied the high-velocity impact-in-
duced behavior on tufted carbon fiber composites laminates. They
found that delamination area was decreased by 24% when the tufting
density was increased compared with specimens without tufting. Colin
de Verdiere et al. [25] studied the tufting effect on mode II fracture
toughness of the carbon non crimp fabric composites. It showed about 2
times higher than the non-tufted samples. Dell'Anno et al. [26] in-
vestigated the Compression After Impact (CAI) strength on carbon fiber
composites laminates reinforced by tufting with carbon and glass
threads. The authors reported an increase of CAI strength by 25% and
27% for carbon and glass threads respectively. They also reported 10%
of the decrease in static tensile modulus and strength for the glass tufted
laminate. The improvements on CAI strength were also reported by
Scarponi et al. [27], when compared to non-tufted composites, em-
ploying tufted aramid fibers reinforcements. They studied the techni-
ques of low and high tensioned lock stitch, tufting and z-pinning to
reinforce trough-the-thickness carbon fiber preforms. The tufted lami-
nates showed CAI strength superior to the others techniques and
especially 16% higher than the non-tufted laminates.
Also, the improvements on out-of-plane properties were reported for
sandwich structures reinforced by stitching. Lascoup et al. [28] ob-
tained improvements under 4-point bending tests in the bending
module (278%) and maximum stress (9 times greater) when compared
with the non-tufted specimens. Moreover, great values on transversal
flatwise compression tests were achieved with intrinsic modulus
growing by a coefficient 14 and the ultimate stress by a factor 8. The
impact resistance was also enhanced and reported by different authors
[29–32].In general, the stitched sandwich composites were capable of
bearing greater the impact load, absorb more the impact energy, reduce
damaged area and penetration depth.
This introduction reviews, in a non-exhaustive manner, various
works about the through-the-thickness reinforcements. In general, these
reinforcements provide out-of-plane improvements and in-plane re-
duction of the mechanical properties. Consequently, it is important to
well understand this subject in order to better control the tufting
353
Composite Structures 184 (2018) 352–361
technique which contributes along with other techniques [9–20] to
improve the delamination behavior of composite structures.
The research works based on tufted composites have not been ex-
pressively presented in the literature, especially concerning to increase
the mode II fracture toughness. This case is very important to the out-
of-plane impact loading because the delamination cracks resulting to
the impact are driven especially by mode II shear [33]. Therefore, this
research investigates the behavior of inclined tufting reinforcements in
order to increase the damage tolerance of the carbon fiber laminate
composites under impact and evaluating their CAI strength.
Moreover, in the majority of works presented on impact resistance
and compression after impact, there is a lack of studies which deal with
the detection of the first major damage. Indeed, it is this load level that
must be taken into account in a design project, while the maximum
stress reached, although interesting to evaluate, is not important in a
project. Therefore, it has seemed interesting to focus on the detection of
this first level of damage. Nevertheless, it is quite difficult to visualize
this event using the stress curve alone. In this study, two techniques
were used to accurately detect the major damage initiation point: the
acoustic emission (AE) technique by means of clustering analysis as-
sociated with digital image correlation (DIC). It contributes to esti-
mating the effect of tufting parameters (density and angle) in the in-
itiation of the first major damage on compression after impact.
Moreover, it is possible to identify a class of acoustic signals related to
this point which will cause the failure on CAI.
2. Materials and methods
2.1. Materials
Woven carbon fabric/epoxy composites were manufactured using a
5HS woven fabric from 6K with an areal density of 364 g/m2. For the
tufting process, 2K Tenax-J HTA 40 carbon thread wrapped by two PBO
yarns was employed to reinforce the carbon fabric preform. Two la-
minates with the layup [0]12 were processed according to the angle of
the reinforcement insertion. The transversal and angular tufting threads
were inserted parallel and at ± 30° to the normal plane (Fig. 1) of the
preform respectively. The insertion at ± 30° was chosen as the max-
imum angle possible due to the machine limits. Also, the preforms had
different tufting densities divided by zones for the reference (non-
tufted), 5 × 5 and 10 × 10 mm tufted square patterns.
The tufting reinforcements were performed by KUKA 6-axis robot
arm (KR 100-2 HA 2000). Pressure foot to compact the dry preforms
during the process was develop and manufacture by a 3D printer. It was
made with a flat surface in order to avoid misalignments on fabric tows,
homogenizing the force applied. The process used polyurethane foam as
well as a nylon film placed under dry preform to hold the tuft loops.
The composite plates were molded by VARTM with EPOLAM 5015
epoxy resin system. During infusion process, the vacuum pressure was
about −1 bar at room temperature. The cure cycle was performed at
room temperature for 24 h and post cured at 80 °C for 16 h. The final
thickness was about 5 mm. Then, samples with 100 × 150 mm di-
mensions were prepared for the impact and CAI tests. The specimens
were designated according to the tufting density and angle as REF for
non-tufted, T10 and T5 for 10 × 10 and 5 × 5 mm transversal square
patterns respectively. A10 and A5 were designated to angular tufting
( ± 30°) with 10 × 10 and 5 × 5 mm square patterns respectively.
2.2. Compression before impact
Compression Before Impact (CBI) was performed to three samples of
the REF, A5 and T5 configuration. They were tested according to the
standard test method ASTM D6641/D6641M-01 [34].
A.T. Martins et al.
Square paƩern (x=y)
x
TuŌing path
TuŌing direcƟon
2.3. Drop-weight impact tests
Three specimens of each configuration were tested in the Instron
Dynatup 9250HV drop-weight impact machine under 25 and 60J im-
pact energies. The low-velocity impact at 25J was evaluated before to
be chosen in order to create barely visible impact damage (BVID) in the
composites. Differently, medium velocity impact at 60J was submitted
to create visible damages in the specimens. The main goal was to
evaluate the tufted composites behavior in the two classes of impact
velocity (low and medium) which are intended to create different da-
mages levels, under impact and compression after impact. The impact
energy was controlled by varying the velocity while keeping the same
mass to the tests. The specimens were 100 × 150 mm according ASTM
D7136-15 [35]. Further, a laser sensor was used on the bottom side of
the samples, placed in the middle of the laminate to evaluate the out-of-
plane displacement.
The damaged area from impact tests was observed on ultrasound C-
Scan using a pulse-echo immersion mode with a 10 MHz focal trans-
ducer. Moreover, the characterization of the damaged area was made
from the images built with eco rebound from the back surface. It was
performed in an 80 × 80 mm area from the center of the sample sur-
face. The C-Scan images were analyzed on ImageJ software to calculate
the damaged area after conversion on binary images.
2.4. Compression after impact
Compression After Impact (CAI) tests were performed with a sup-
port according to the specified in ASTM D7137-12 [36] to avoid
buckling. These tests were monitored on Digital Image Correlation
(DIC) to evaluate the out-of-plane displacements (w) of a single side
surface (impacted surface) taking series of digital photographs from a
stereo-system with two stereo cameras (camera CC-0003 by ALLIED
Vision Technologies) throughout the loading history and correlating
them with the software VIC 3D®. The parameters of subset and step size
used to correlate were 29 and 7 respectively. The area chosen to ana-
lyze the mean out-of-plane displacement (w) was 30 × 30 mm cen-
tralized in the impacted zone.
Also, it was compared the dent depth among the sample config-
urations from the DIC analysis before CAI test initiation (Fig. 2). The
dent depth measured by DIC has been used by different authors
354
Composite Structures 184 (2018) 352–361
Fig. 1. Angular tufting manufacturing and the tufted pre-
Preform
form schema.
60°
[37–39]. This method allows measuring with precision, especially
smaller dent depth frequently produced by low-velocity impacts.
Additionally, two wideband sensors (Micro80 - 200–900 kHz) for
Acoustic Emission (AE) activity were longitudinally placed 45 mm from
the center of the sample, one in each side of the transversal axis. Fig. 3
exhibits the multi-instrumented CAI test with cameras and AE sensors
after sample failure.
3. Results and discussions
3.1. Drop-weight impact
The curves load vs. deflection were plotted in Figs. 4 and 5 to the 25
and 60J impact energies respectively. Initially, they show firstly a linear
part followed by a first load drop. Then, the second part is characterized
by a stiffness loss before the maximum load. The general behavior
confirmed the enhancement of composite toughness by the tufted
composites during tests, showing a maximum deflection to the non-
tufted (REF) and lower to the higher angular density (A5).
It is important to note that in both energies the delamination onset
force (Ponset) was higher in the angular tufting samples than in the
transversal, especially to the densest samples as summarized in Table 1.
This onset point means the first significant damages which results in the
stiffness loss. Then, in the case of out-of-plane impact, the angled
tufting increases the delamination resistance by resisting crack sliding
displacement.
Moreover, during the impact tests, the deflection was measured on
the bottom side of the samples aided by a laser system. The results of
maximum deflection are compared for both energies in Table 1.
Transversal tufting presented a little improvement in comparison to
angular reinforcement to resist the transversal deflection in the both
impacted energies. The results can be attributed to the carbon tufting
threads aligned transversally which enhance the trough-the-thickness
stiffness. Also, as expected, the maximum deflection was inversely
proportional to the tufting density as a result of the enhancement in the
transversal stiffness.
Further, Fig. 6 shows the typical damaged area of the different
sample configurations to the two impact energies studied. The images
obtained by C-Scan characterization were prepared and analyzed on
software ImageJ. It is seen that the tufting reinforcements change the
A.T. Martins et al. Composite Structures 184 (2018) 352–361

Fig. 2. The contour plot and the equivalent 3D image of the

z,w impacted specimen on Z field.

y
Dent depth

Fig. 5. Typical load-deflection curves during impact at 60J.

Table 1
Ponset values and maximum deflection at both impacted energies.
Fig. 3. Multi-instrumented CAI test with AE sensors and DIC cameras.
Sample Ponset (N ± SD) Max. deflection (mm ± SD)

25J 60J 25J 60J

REF 6.31 ± 0.08 6.52 ± 0.08 4.78 ± 0.24 7.03 ± 0.15

T10 7.00 ± 0.07 7.15 ± 0.01 3.66 ± 0.06 6.39 ± 0.22
T5 7.98 ± 0.09 8.99 ± 0.01 3.47 ± 0.16 5.59 ± 0.11
A10 6.85 ± 0.62 6.71 ± 0.22 3.84 ± 0.13 6.61 ± 0.14
A5 8.18 ± 0.74 11.30 ± 0.01 3.33 ± 0.13 6.02 ± 0.07

Fig. 4. Typical load-deflection curves during impact at 25J.
damaged area as well as their parameters (angle and density).
The damaged area values according to the tufting parameters are
shown to the both impact energies in Fig. 7. The results showed that the
damaged area is inversely proportional to the tufting density. Ad-
ditionally, the transversal tufted samples impacted at 25J were the most
efficient to restrain the delamination progress. However, in the samples
impacted at 60J, the T10 configuration had larger damaged area than
A10, which did not follow the behavior as in the lower energy. Also, in
this impact energy, the angular tufting becomes more efficient to re-
strain fissures development, especially delamination, which is shown
from the lower increase in the damaged area from 25 to 60J when
compared with transversal tufting. Moreover, when comparing the
maximum deflection at the bottom side of impact (Table 1) with the
damaged area (Fig. 7), it is seen that the damaged area is directly
proportional to the deflection. For example, the reference deflects the
most among the configurations under impact because their lower out-
of-plane toughness which creates and develops more the fissures,
especially as delamination when compared with the other samples.
Further, the T10 and A10 configurations decrease the deflection under
impact when compared with the reference, and it is the cause of the
reduction in the damaged area. Finally, the optimal results are achieved
for the densest tufted specimens.
Therefore, it is possible to conclude that tufting reinforcement re-
duces the crack development by concentrating the damage in a small
area. In relation to the angular parameter, among the specimens espe-
cially at 60J, there are no expressively differences between angular and
transversal tufting.

355
A.T. Martins et al. Composite Structures 184 (2018) 352–361

Fig. 6. C-Scan images from the impacted samples config-

REF T10 T5 A10 A5 ured and analyzed on ImageJ.

25J

60J

fabric interface and then continue propagating as delamination. The

Fig. 7. Average of damaged area for: REF; T10; T5; A10 and A5 samples at 25J and 60 J.
Fig. 8a–c show optical micrographs from REF, T5 and A5 specimens
impacted at 25J, respectively. The principal cracks are highlighted in
the figures. Fig. 8a presents the trough-the-thickness crack evolution for
the non-tufted sample. It is possible to note that matrix cracks at ap-
proximately 45° due to transverse shear stress develop until reach the
damages on the bottom layer are due to high tensile stress induced by
the bending on impact loading. The sample behaves as reported in the
literature to impacted fabric composites. Fig. 8b shows the crack de-
velopment to the laminate reinforced transversally, T5. Differently, to
the non-tufted specimen, delamination is no longer evident and the
cracks are in the majority from the matrix at 45°. The tufting threads
inhibit the crack development by displacing the crack tips to around
them, decreasing the damage energy. Also, improving the through-the-
thickness stiffness means lower tensile stress response due to bending
by the impact. This behavior can be verified by the decrease of damage
in the bottom layer when compared with the reference. From the Fig. 8c
it is seen that delamination is more important than in the transversal
samples but less than in the reference. The tuft threads had the role to
deviates the cracks and especially stop the delamination progress.
Moreover, significant cracks appear from the surface to inside of the
composite around the threads. Besides, differently, that was planned
the threads did not cross near the middle plane of the thickness. The
first reason was the applied angle of ± 30° to the normal of the surface.
Also, it was seen porosity concentration near to the threads which is the
counter effect of the angular tufting process.
Furthermore, to the both tufting configurations, the threads were
not straight as expected. It is caused by the preform compaction during
VARTM process. This fact must be taken into account especially in fu-
ture finite element analysis. In addition, a resin layer caused by the
tufting threads loops is created on the surface of the laminate which

Fig. 8. Optical micrographs for: a) reference; b) T5 and c)

A5 impacted at 25J.

356
A.T. Martins et al.
Fig. 9. Schematization of the impact damages for the sample configurations: a) A5; b)
A10; c) T5; d) T10 and e) reference.
helps crack initiation and development as one can see in Fig. 8b and c.
From the microscopy observations, a typical damage scenario is
schematized for each sample and shown in Fig. 9. In the non-tufted
sample (Fig. 9e), the damage develops from the impacted point to the
distal face in a conical shape as well described in the literature. In the
case of the tufted composite laminates it has been remarked that:
• By increasing the tufting density the dent depth after impact is also
increased, regardless of the tufting angle. The dent depth measure
was performed by DIC and shown in Figs. 10 and 11.
• The damages are mainly developed at 45° (matrix cracks) to the
transversal tufted laminates (T5 and T10 in Fig. 9c and d respec-
tively) while they are in the majority at 0° (delamination) to the
laminates with inclined tufts (A5 and A10 in Fig. 9a and b respec-
tively).
• The damaged area diminishes with the increase in the tufting den-
sity.
• The angular tufted laminates show impact damages larger than the
composites reinforced with traversal tufts. These damages on an-
gular tufted samples are mainly for delamination, which should
Fig. 10. Dent depth from impacted surface and its comparison to the damaged area at
25J.
357
Composite Structures 184 (2018) 352–361
Fig. 11. Dent depth from impacted surface and its comparison to the damaged area at
60J.
weaken the structure on compression unless that the tufting density
helps to reduce this phenomenon.
3.2. Compression after impact
The images took before CAI tests were analyzed on VIC 3D for the Z
displacement field. Figs. 10 and 11 exhibit the dent depth obtained
from the impacted surfaces of the different configurations impacted at
25 and 60J respectively. Moreover, the damaged area images obtained
by C-Scan were compared with these results. The increase of the impact
energy also increases the dent depth. In general, the tufting reinforce-
ment increases the dent depth by improving the trough-the-thickness
stiffness. It is also related to the maximum deflection under impact
which was already discussed. In comparison to the damaged area, the
dent depth shows inversely proportional, especially to the 60J results. It
can be explained by the tufting role in the composites.
The tufting threads prevent the failure development in the lami-
nates, especially for delamination, when impacted. The delamination is
enclosed in a smaller zone when compared to the reference due to the
energy concentration by the threads, which creates a larger dent depth
in the tufted samples. Otherwise, reference samples dissipate the impact
energy through the failure mechanisms, as delamination.
The typical behavior of the samples on CAI tests are shown in the
Figs. 12 and 13 to the specimens submitted at 25 and 60J impact en-
ergies respectively. In both energies, the reference sample curves dis-
tinguish from the others with a visible drop load near to the ultimate
Fig. 12. Typical behavior on CAI tests to the samples impacted at 25J.
A.T. Martins et al. Composite Structures 184 (2018) 352–361

Table 2
Ultimate compressive strength on CBI and CAI tests, and the residual ratio (CAI/CBI).

Ultimate strength (MPa ± SD) Residual Residual

ratio (%) ratio (%)
CBI CAI 25J CAI 60J 25J 60J

REF 453.1 ± 40.9 176.1 ± 1.3 137.4 ± 1.1 38.9 30.3

T5 323.9 ± 28.9 223.5 ± 2.5 171.5 ± 2.9 69.0 52.9
A5 271.9 ± 19.7 178.9 ± 4.9 157.8 ± 4.9 65.8 58.0

35000
Class 0
Class 1
30000 Class 2
Class 3
25000

20000

Counts
Fig. 13. Typical behavior on CAI tests to the samples impacted at 60J.
15000

10000

5000

0
0 20000 40000 60000
Energy (J)
Fig. 16. Clustered signals for the sample configuration A5 impacted at 25J under CAI.

tufting density. Additionally, the transversal tufted laminates were

Fig. 14. Ultimate compressive strength comparison for the different sample configura-
tions.
strength. The delaminations generated during impact weak the samples
on CAI. They cause buckling deflection reverse on the impact side and
reduces the load carrying capacity of the delaminated plates [39]. This
failure is reduced for the tufted specimens, especially increasing the
more capable to contain the buckling in comparison to the angular
tufted composites. Furthermore, the difference between the reference
and the tufted composites are better seen for those impacted at 25J.
This response is related to the lower amount of damaged tuft threads in
this energy which makes them more able to restrain the opening me-
chanism and unstable compression than the samples impacted at 60J.
Fig. 14 summarizes the mean ultimate compressive strength
achieved in the samples under CAI tests. It is possible to note that the
only configuration which showed lower CAI ultimate strength than
reference was A10. In general, the tufting reinforcement plays a role to
increase the transversal strength of the samples. The tufted samples
showed a better behavior to retain the failure, especially avoiding de-
lamination progress by means of the opening mode (mode I) under CAI,
which was considered the principal factor to buckling the samples and

Fig. 15. Schematization of the damage on CAI of the com-

posite laminates: a) A10 and b) A5.

358
A.T. Martins et al. Composite Structures 184 (2018) 352–361

Fig. 17. Clustered signals, compressive stress and w as time

function of the sample configuration A5 on CAI impacted at
25J.

Fig. 18. Comparison of Class 1 from clustered signals of the

different sample configurations on CAI impacted at 25J.

Table 3 that significant damage created on impact tests is concentrated around

Normalized stress obtained from the cumulative energy curves (Class 1) at the P0 point. the tufting threads as exemplified in the Fig. 9a and b for A5 and A10
configuration respectively. The cracks in this region are initiated on
Configuration Stress at P0
impact and advance on CAI tests as schematized in Fig. 15. It can be due
25J 60J to the maximum shear stress that is produced on CAI tests at 45° which
is closer to the tufting insertion angle. This becomes very critical for the
REF 1000 1000 A10 configuration (Fig. 15a) where it will generate an early failure
T10 1197 1130
T5 1791 1379
when compare with T10. Additionally, the results of the damaged area
A10 0722 1037 by C-Scan image analysis as a planar view of the through-the-thickness
A5 1272 1109 damages are not capable of distinguish the cracks around the threads,
small in a plan visualization, which can decrease significantly the
compressive strength on CAI. It will be necessary for the future to
earlier failure. The CAI strength achieved improvements of 6% (T10), analyze this kind of damage by microtomography. This is evident
27% (T5) and 2% (A5) when impacted at 25J in comparison with the especially to the A10 samples impacted at 60J which had the damaged
reference. To the samples impacted at 60J, the improvements were 19% area lower than the T10 configuration. These cracks also appear for the
(T10), 25% (T5) and 15% (A5). Unfortunately, the A10 configuration A5 configuration, but due to the crossing of the threads near to the
decreased the CAI strength about 21% and 10% (25 and 60J respec- surface the cracks are not propagate in this region. Also, the fissures
tively). which propagate from the bottom side of the samples (opposite to the
It is remarkable that for the angular reinforcement configurations impacted surface) around the tufted threads, are deflected and bridged

359
A.T. Martins et al.
by the threads which avoid the loss of resistance when compared with
the A10 configuration, as shown in Fig. 15b.
The specimens which achieved the best results on CAI were com-
pared with those on CBI as manner to evaluate in-plane properties of
the tufted composites (Table 2). The loss of in-plane ultimate com-
pressive strength arrived until 29% to the T5 configuration and 40% to
A5. This reduction has been reported to many authors which research
trough-the-thickness reinforcements [9,18,40,41]. The development of
the numerical methods for modeling mechanical properties of the
transversal reinforcements in the composite should diminish their im-
pact on the in-plane properties. However, the tufted specimens showed
greater residual ratio (CAI/CBI) in comparison with the non-tufted
specimens (approximately 30% superior to T5 at 25J and 28% to A5 at
60J).
The out-of-plane displacement characterized by DIC on CAI tests
evidenced the sudden displacement which is due to the local buckling.
This behavior was also seen on acoustic emission due to an important
increase in the cumulative energy. Then, non-supervised clustering
from acoustic emission signals was performed as a means of distinguish
a class of acoustic signals which initiates with the major damages and
results in the sudden failure of the laminates. The classification used the
parameters of signal amplitude, counts to peak, counts and energy, as
input in the k-means method for classify the signals into 4 classes. This
method applied the parameters of Euclidian distance and random initial
partitioning. Then, it was realized that a specific class of signal appears
at the moment which the out-of-plane displacement is higher and
abrupt failure happens. This class is shown in Fig. 16, named Class 1. It
is possible to note that this class is composed of signals from lower to
medium energy as well their counts.
Fig. 17 exhibits the out-of-plane displacement related to the com-
pressive stress and the four classes obtained. The Class 1 appears near to
the crossing of the tangential curves in the out-of-plane displacement
(w), named P0. This point is described from the beginning of the non-
linearity in the time-stress graph. However, it is difficult to define from
the curve, which can cause imprecise results. The AE method helps to
well localize the P0, where the cumulative acoustic energy increases
grace to the important damages caused by buckling on CAI.
Further, from the Class 1 it was obtained for each sample config-
uration. The Fig. 18 shows a typical behavior of this class on CAI to the
samples impacted at 25J under compressive load. The main goal is to
compare the initial point (P0) between them. The Table 3 summarizes
the normalized stress obtained at P0 for the different samples. In general
the tufting changes the P0, it means increasing the stress which the
considerable damages start occurring. Also, the density plays an im-
portant function increasing this critical stress. It is possible to note that
transversal tufting is more efficient than angular. The T5 sample con-
figuration increased P0 79% and 38% to 25 and 60J respectively in
comparison to the reference. The A5 samples enhanced 27% and 11% to
25 and 60J respectively and can be compared to T10 configuration
impacted at 60J (increasing 13% in the P0). Moreover, it is significant
the loss of mechanical property of the A10 configuration (−28% at
25J), which was already described in others tests results. Therefore,
these results validate the efficiency of tufting on CAI, especially to
transversal tufting. In addition, they are in accordance with the results
of ultimate strength on CAI tests, which by increasing the critical stress
(P0) leads the material to a later failure with a higher compressive
stress.
4. Conclusions
The tufting composites achieved their role to increase the damage
resistance under impact tests which restrained the crack development
in a smaller area when compared with the reference. This response was
better seen on 60J impact when transversal tufting (T5) decreased
about 4 times the damaged area. Additionally, the dent depth results
confirmed the tufting function to restrain the damage development. The
360
Composite Structures 184 (2018) 352–361
relation between damage area and dent depth is inversely proportional.
The results on CAI tests showed superiority to in the majority of the
tufting reinforcements. Unfortunately, angular tufting A10 samples
showed CAI strength lower than the reference which becomes necessary
a higher tufting density as A5 to improve this property. This sample
configuration (A10) should be investigated in the future to better un-
derstanding. Also, despite the greater results on delamination onset in
the impact tests as well impact deflection and damaged areas similar to
the transversal tufting, especially compared to 5 × 5 mm squared
patterns, angular sample configurations showed the lower increase. The
CAI strength was 27% and 2% at 25J while at 60J they were 25% and
14% to T5 and A5 respectively.
The data acquired from the multi-instrumented characterization
during CAI tests showed that the tufting density and angle change the
cumulative energy of acoustic emission signals. Then, with non-su-
pervised signal classification on AE, it was possible to predict the first
significant damages from a class of AE and correlate them with the
beginning of the suddenly out-of-plane displacement measured by DIC,
especially focused in the impacted zone. It is possible to note that
tufting reinforcement avoids plies opening as result of delamination
process under compression loads and the transversal tufting was more
efficient to increase this fracture toughness.
Acknowledgement
The author A. T. Martins would like to acknowledge the PhD
scholarship awarded by Conselho Nacional de Desenvolvimento
Científico e Tecnológico (CNPq) – Brazil.
References
[1] Jin BC, Li X, Mier R, Pun A, Joshi S, Nutt S. Parametric modeling, higher order FEA
and experimental investigation of hat-stiffened composite panels. Compos Struct
2015;128:207–20.
[2] Byun JH, Song SW, Lee CH, Um MK, Hwang BS. Impact properties of laminated
composites with stitching fibers. Compos Struct 2006 Oct 31;76(1):21–7.
[3] Hosur MV, Karim MR, Jeelani S. Experimental investigations on the response of
stitched/unstitched woven S2-glass/SC15 epoxy composites under single and re-
peated low velocity impact loading. Compos Struct 2003;61(1):89–102.
[4] Petit S, Bouvet C, Bergerot A, Barrau JJ. Impact and compression after impact ex-
perimental study of a composite laminate with a cork thermal shield. Compos Sci
Technol 2007;67(15):3286–99.
[5] Aktaş M, Atas C, İçten BM, Karakuzu R. An experimental investigation of the impact
response of composite laminates. Compos Struct 2009;87(4):307–13.
[6] Larsson F. Damage tolerance of a stitched carbon/epoxy laminate. Compos A Appl
Sci Manuf 1997;28(11):923–34.
[7] Dorey, G., 1986. Impact damage tolerance in advanced composite materials. In
Seminar on Advanced Composites, Cranfield Institute of Technology, Cranfield, UK.
[8] Wu HYT, Springer GS. Impact induced stresses, strains, and delaminations in
composite plates. J Compos Mater 1988;22(6):533–60.
[9] Karuppannan, D., Sivaraman, V., Gaddikeri, K.M., Sundaram, R. and Ramesh, A.,
Effect of tufting on mechanical properties of laminated composites. In ISAMPE
National Conference on composite Materials, pp. 2–3.
[10] Masters JE. Improved impact and delamination resistance through interleafing. Key
Engineering Materials. Trans Tech Publications; 1989. pp. 317.
[11] Kinloch AJ, Masania K, Taylor AC, Sprenger S, Egan D. The fracture of glass-fibre-
reinforced epoxy composites using nanoparticle-modified matrices. J Mater Sci
2008;43(3):1151–4.
[12] Sela N, Ishai O. Interlaminar fracture toughness and toughening of laminated
composite materials: a review. Composites 1989;20(5):423–35.
[13] Peijs AAJM, Venderbosch RW, Lemstra PJ. Hybrid composites based on poly-
ethylene and carbon fibres Part 3: impact resistant structural composites through
damage management. Composites 1990;21(6):522–30.
[14] Rechak S, Sun CT. Optimal use of adhesive layers in reducing impact damage in
composite laminates. J Reinf Plast Compos 1990;9(6):569–82.
[15] Hojo M, Matsuda S, Tanaka M, Ochiai S, Murakami A. Mode I delamination fatigue
properties of interlayer-toughened CF/epoxy laminates. Compos Sci Technol
2006;66(5):665–75.
[16] Shivakumar K, Lingaiah S, Chen H, Akangah P, Swaminathan G, Russell Jr L.
Polymer nanofabric interleaved composite laminates. AIAA J 2009;47(7):1723.
[17] Wong DW, Zhang H, Bilotti E, Peijs T. Interlaminar toughening of woven fabric
carbon/epoxy composite laminates using hybrid aramid/phenoxy interleaves.
Compos A Appl Sci Manufact 2017.
[18] Ning H, Li Y, Li J, Hu N, Liu Y, Wu L, et al. Toughening effect of CB-epoxy interleaf
on the interlaminar mechanical properties of CFRP laminates. Compos A Appl Sci
Manuf 2015;68:226–34.
A.T. Martins et al.
[19] Sohn MS, Hu XZ. Mode II delamination toughness of carbon-fibre/epoxy composites
with chopped Kevlar fibre reinforcement. Compos Sci Technol 1994;52(3):439–48.
[20] Maples HA, Smith O, Burgstaller C, Robinson P, Bismarck A. Improving the ply/
interleaf interface in carbon fibre reinforced composites with variable stiffness.
Composites Science and Technology 2016. pp. 128, 185–192.
[21] Mouritz AP. Fatigue of 3D textile-reinforced composites. Fatigue Textile Compos
2015:255.
[22] Plain KP, Tong L. The effect of stitch incline angle on mode I fracture
toughness–Experimental and modelling. Compos Struct 2010;92(7):1620–30.
[23] Tan KT, Watanabe N, Iwahori Y, Ishikawa T. Effect of stitch density and stitch
thread thickness on compression after impact strength and response of stitched
composites. Compos Sci Technol 2012;72(5):587–98.
[24] Deconinck P, Capelle J, Bouchart V, Chevrier P, Ravailler F. Delamination propa-
gation analysis in tufted carbon fibre-reinforced plastic composites subjected to
high-velocity impact. J Reinf Plast Compos 2014;33(14):1353–63.
[25] Colin de Verdiere M, Skordos AA, Walton AC, May M. Influence of loading rate on
the delamination response of untufted and tufted carbon epoxy non-crimp fabric
composites/Mode II. Eng Fract Mech 2012;96:1–10.
[26] Dell’Anno G, Cartié DD, Partridge IK, Rezai A. Exploring mechanical property
balance in tufted carbon fabric/epoxy composites. Compos A Appl Sci Manuf
2007;38(11):2366–73.
[27] Scarponi C, Perillo AM, Cutillo L, Foglio C. Advanced TTT composite materials for
aeronautical purposes: Compression after impact (CAI) behaviour. Compos B Eng
2007;38(2):258–64.
[28] Lascoup B, Aboura Z, Khellil K, Benzeggagh M. On the mechanical effect of stitch
addition in sandwich panel. Compos Sci Technol 2006;66(10):1385–98.
[29] Lascoup B, Aboura Z, Khellil K, Benzeggagh M. Impact response of three-dimen-
sional stitched sandwich composite. Compos Struct 2010;92(2):347–53.
[30] Xia F, Wu XQ. Study on impact properties of through-thickness stitched foam
sandwich composites. Compos Struct 2010;92(2):412–21.
361
Composite Structures 184 (2018) 352–361
[31] Samlal S, Paulson V, Santhanakrishnan R. Effect of stitching angle on impact
characteristics of sandwich panels. Int J Innov Res Sci Eng Technol
2015;4(12):30–4.
[32] Potluri P, Kusak E, Reddy TY. Novel stitch-bonded sandwich composite structures.
Compos Struct 2003;59(2):251–9.
[33] M'membe B, Gannon S, Yasaee M, Hallett SR, Partridge IK. Mode II delamination
resistance of composites reinforced with inclined Z-pins. Mater Des
2016;94:565–72.
[34] Standard ASTM. D6641/ D6641M: Standard Test Method for Determining the
Compressive Properties of Polymer Matrix Composite Laminates Using a Combined
Loading Compression (CLC) Test Fixture. West Conshohocken: ASTM International;
2001.
[35] Standard ASTM. D7136: Standard test method for measuring the damage resistance
of a fiber-reinforced polymer matrix composite to a drop-weight impact event.
ASTM International: West Conshohocken; 2015.
[36] Standard, A.S.T.M., 2012. D7137, Standard Test Method for Compressive Residual
Strength Properties of Damaged Polymer Matrix Composite Plates. ASTM Int., West
Conshohocken, PA.
[37] Mcquigg, T. D, 2011. Compression after impact experiments and analysis on hon-
eycomb core sandwich panels with thin facesheets.
[38] Vieille B, Casado VM, Bouvet C. Influence of matrix toughness and ductility on the
compression-after-impact behavior of woven-ply thermoplastic-and thermosetting-
composites: a comparative study. Compos Struct 2014;110:207–18.
[39] Rhead AT, Hua S, Butler R. Damage resistance and damage tolerance of hybrid
carbon-glass laminates. Compos A Appl Sci Manuf 2015;76:224–32.
[40] Colin de Verdiere M, Pickett AK, Skordos AA., Witzel V. Effect of tufting on the
response of non crimp fabric composites. Porto Portugal 2007;12–14.
[41] Colin de Verdiere M, Pickett AK, Skordos AA, Witzel V. Evaluation of the me-
chanical and damage behaviour of tufted non crimped fabric composites using full
field measurements. Compos Sci Technol 2009;69(2):131–8.