ScienceDirect

Available online at www.sciencedirect.com

StructuralScienceDirect
Integrity Procedia 00 (2019) 000–000
www.elsevier.com/locate/procedia
Available online at www.sciencedirect.com
Structural Integrity Procedia 00 (2019) 000–000
www.elsevier.com/locate/procedia
ScienceDirect
Procedia Structural Integrity 21 (2019) 206–214

1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials

1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials
Simulation of Drop-Weight Impact Test on Composite Laminates
using Finite
Simulation of Drop-Weight Element
Impact TestMethod
on Composite Laminates
using Finite Element Method
Mirac Onur Bozkurta,b, Levend Parnasc, Demirkan Cokera,b,*
Mirac Onur Bozkurt
Department
a a,b
, Levend
of Aerospace Parnas
Engineering, , Demirkan
METU, cAnkara Cokera,b,*
06800, Turkey
b
METU Center for Wind Energy, METU, Ankara 06800, Turkey
c
Department
a of Mechanical
Department Engineering,
of Aerospace TEDMETU,
Engineering, University,
AnkaraAnkara 06800,
06800, Turkey
Turkey
b
METU Center for Wind Energy, METU, Ankara 06800, Turkey
c
Department of Mechanical Engineering, TED University, Ankara 06800, Turkey

Abstract

Abstract
This study presents the simulation of standard drop-weight impact test on a [0/90/0]s composite laminate. For this purpose, a three-
dimensional virtual test setup is developed in ABAQUS/Explicit finite element tool. Hemispherical impactor and specimen fixture
This study presents
are modeled thebodies.
as rigid simulation of standard
Composite platedrop-weight
is modeledimpact
as a 3-Dtest deformable
on a [0/90/0]solid
s composite laminate. using
and discretized For thisa purpose, a three-
biased mesh for
dimensional virtual
computational test setup
efficiency. For is developedofinply
simulation ABAQUS/Explicit finite element
damage in the composite tool.aHemispherical
laminate, continuum damageimpactor and specimen
mechanics fixture
based damage
are modeled
model as rigidand
is developed bodies. Composite
implemented into plate is modeled
the analysis via a as a 3-D deformable
user-written subroutinesolid and discretized
VUMAT. usingdamage
Delamination a biased mesh for
is simulated
computational
by efficiency.
inserting cohesive For simulation
elements of plyofdamage
at the interfaces in thedifferent
plies having composite laminate, aResults
orientations. continuum
showdamage
that themechanics based
initial failure damage
mechanism
model
in is developed
the 3-D low-velocityand impact
implemented
event isinto
thethe analysis
matrix via a in
cracking user-written
the lowermostsubroutine VUMAT. Delamination
plies independent damage
from the stacking is simulated
sequence of the
by inserting
laminate. cohesive elements
Furthermore, at the interfaces
the simulation accuratelyof plies having
predict that different orientations.
delaminated Results mainly
regions expand show that
in the initial failure mechanism
same direction as of the
in the of
fibers 3-Dthelow-velocity impact
lower adjacent layerevent is the matrix
in accordance cracking
with in thestiffness
the bending lowermost plies independent
mismatching concept.from the stacking sequence of the
laminate. Furthermore, the simulation accurately predict that delaminated regions expand mainly in the same direction as of the
fibers
© 2019ofThe
the Authors.
lower adjacent layerby
Published inElsevier
accordanceB.V.with the bending stiffness mismatching concept.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/
©
© 2019
2019TheTheAuthors.
Authors.Published by Elsevier
Published by B.V. B.V.
Peer-review under responsibility of Elsevier
the 1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/
organizers
Peer-review under responsibility of the 1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials organizers
Peer-review under responsibility of the 1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials
organizers
Keywords: Drop-weight impact; finite element analysis; composite; matrix crack; delamination

Keywords: Drop-weight impact; finite element analysis; composite; matrix crack; delamination

* Corresponding author. Tel.: +90-312-210-4257; fax: +90-312-210-4250.

E-mail address: [email protected]
* Corresponding author. Tel.: +90-312-210-4257; fax: +90-312-210-4250.
E-mail address:
2452-3216 [email protected]
© 2019 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/
Peer-review underThe
2452-3216 © 2019 responsibility of theby
Authors. Published 1stElsevier
International
B.V. Workshop on Plasticity, Damage and Fracture of Engineering Materials
organizers
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/
Peer-review under responsibility of the 1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials
organizers

2452-3216 © 2019 The Authors. Published by Elsevier B.V.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Peer-review under responsibility of the 1st International Workshop on Plasticity, Damage and Fracture of Engineering Materials organizers
10.1016/j.prostr.2019.12.103
 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214 207
2 Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000

1. Introduction

Use of fiber-reinforced plastics (FRP) in engineering applications is favorable because they offer numerous
advantages such as light weight, high stiffness and strength. However, laminated FRP composites are susceptible to
delamination damage which leads to considerable losses in the residual strength. In the design phase, damage tolerance
analysis consisting of sequential transverse impact and compression after impact (CAI) analyses are conducted with
certified methods. Since testing alone is highly expensive due to the manufacturing and testing of large number of
coupons required to verify every geometry, loading, environment and failure mode, development of virtual test setups
which accurately predicts impact damage and CAI strength is of great interest.
In the literature, several studies are conducted on developing virtual impact test setups. Lopes et al. (2009)
investigated the effect of dispersed stacking sequences on impact response of fiber-reinforced polymers by use of
constitutive models which take into account the physical progressive failure behaviour of fibres, matrix, and interfaces
between plies. Later, Gonzalez et al. (2012) conducted simulations of sequential impact and compression-after-impact
tests using finite element method. Topac et al. (2017) simulated their line impact experiments on [0/90]s beams by
implementing the constitutive material models into the finite element model and predict damage formation process in
agreement with their experiments. More recently, Soto et al. (2018) simulated low velocity impact and compression
after impact in large composite stiffened panels using continuum damage mechanics based material models.
In this study, a virtual drop-weight impact test setup is developed to simulate standard large mass - low velocity
impact tests (ASTM, 2012). 3D finite element model is generated in ABAQUS/Explicit. Hemispherical impactor and
specimen fixture are modeled as rigid bodies. 3D solid elements are used in modeling of the cross-ply composite
laminate. Constitutive material model accounts for both interlaminar and intralaminar failure modes and it is
implemented via a user-written VUMAT subroutine. Damage formation process is investigated in detail. In order to
assess the accuracy of the simulations, comparisons with experimental results are performed.

2. Modelling of low-velocity impact damage

Two distinct material model is used for simulation of low-velocity impact damage in composite laminates: (i) a
continuum damage mechanics based material model accounting for ply level damage modes including fiber and matrix
failure, (ii) a cohesive damage model for simulation of delamination.

2.1. Intralaminar damage model

A continuum damage mechanics based intralaminar damage model predicting initiation and evolution of composite
damage is developed with 3-D stress formulation. The model accounts for fiber and matrix damages in tension and
compression modes.
Damage initiation in tensile fiber (FT), compressive fiber (FC), tensile matrix (MT) and compressive matrix (MC)
modes are controlled by failure indexes FIN. When the FIN reaches unity, damage in corresponding mode initiates.
Maximum stress and Hashin Failure Criteria (Hashin, 1980) are used for longitudinal and transverse failure indexes,
respectively, which are given for each damage mode as

 11
FI FT  (1)
XT

 11
FI FC   (2)
XC
208 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214
Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000 3

2 2 2
  22   33   232   22 33  12   13 
FI
MT        (3)
 YT  S232  S12   S13 

 Y  2     2
  22   33   23
2 2
  22 33  12   13 
2

FI MC    1      
C 22 33
2  (4)
 2 S 23   YC  2 S 23  S 23  S12   S13 

In equations (1) – (4), XT and XC are the longitudinal tensile and compressive, YT and YC are the transverse tensile
and compressive strengths, S12, S13 and S23 are the in-plane longitudinal, out-of-plane longitudinal and transverse shear
strengths of a unidirectional ply.
The evolution of the damage is modelled by a linear softening response with equivalent stress-strain approach
shown in Fig. 1a. The area under the curve corresponds to the energy dissipated per unit volume and is defined as

g N  GcN / L* (5)

where L* is the characteristic length of finite element and GCN is the fracture toughness for the damage mode N.
The damage variable dN shows a non-linear saturation type behavior, as expressed in Eq. (6), to provide the linear
softening response of damaged material.

 eqf   eqmax   eq0 

dN  (6)
 eqmax   eqf   eq0 

where εeq0 and εeqf are the equivalent strains at the initiation of damage and complete failure, respectively. The
maximum value of equivalent strain, in time history, εeqmax, is used to satisfy the irreversibility of the existent damage.

2.2. Interlaminar Damage Model

The interlaminar damage model used in the analysis to simulate delamination damage is the one offered by
ABAQUS finite element package (Simulia, 2012). Fracture mechanics based bilinear traction-separation response,
seen in Fig. 1b, is assigned to each cohesive element. Since delamination is usually caused by multi-axial stress state
subjecting to the interface, mode-mixity is taken account when modeling the cohesive material response.
The initial response of the cohesive element is assumed to be linear until a damage initiation where the slope of the
line is called penalty stiffness, Ki (i = I,II,III). The value of the penalty stiffness must be high enough to prevent
interpenetration of the crack faces and to prevent artificial compliance from being introduced into the model by the
cohesive elements. Initiation of interlaminar damage is controlled by quadratic nominal stress criterion which is given
as

2 2 2
 TI   TII   TIII 
       1
  (7)
 To , I   To , II   To , III 

In this equation, Ti and To,i (i=I,II,II) are the tractions applying to the surface and interlaminar strengths for
corresponding fracture modes, respectively. Once interlaminar damage initiates, Benzeggagh-Kenane criterion, given
in equation (8), is used for modeling mixed-mode propagation of damage (Benzeggagh and Kenane, 1996).
 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214 209
4 Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000


 GII  GIII 
Gc GIc  (GIIc  GIIIc  GIc )   (8)
 GI  GII  GIII 

where GI and GIc are the work done by tractions and the components of the fracture toughness in mode i fracture,
respectively. η is the mode interaction parameter. Linear softening response is defined to the cohesive elements in the
plane where damage initiation occurs. The cohesive damage variable d shows a non-linear saturation type behavior,
as expressed in equation (9), to provide the linear softening response of damaged cohesive elements.

Fig. 1. (a) Linear softening response of ply material with equivalent stress-strain approach, (b) Mixed-mode bilinear traction-separation of cohesive
material.

 eqf  eqmax   eq0 

d (9)
 eqmax  eqf   eq0 

where δeq0, δeqf are the equivalent displacements at the initiation of damage and complete failure, respectively, while
δeq is the maximum value of equivalent displacement in time history.
max

3. Virtual test setup

To simulate low-velocity impact experiments on composite plates, three-dimensional finite element model is
generated in ABAQUS/Explicit. The model consists of a 3-D deformable composite plate of 150×100 mm2, a rigid
fixture base having a 125×75 mm2 window, four rigid cylindrical clamps with 10 mm diameter and a rigid semi-
spherical impactor with 16 mm diameter. The assembly of the virtual test setup consisting of these parts is shown in
Fig. 2.
[04/904/02]s composite plate of which the geometry is shown in Fig. 2 is modeled as a three dimensional deformable
solid body. The plate consists of 20 unidirectional composite layers having equal thicknesses of 0.125 mm. In the
model, clustered plies are considered as a single homogenized thick ply. Plate geometry is meshed with 8-noded linear
brick elements with reduced integration and hourglass control (C3D8R in ABAQUS library). Each two ply of
composite beam is modeled with one element in through-the-thickness direction. In-plane mesh size is set variable to
reduce the computational cost. A central region of 50×50 mm2 including the impact zone is meshed uniformly with
elements of 0.25×0.25×0.25 mm3. A biased mesh is used outside of this region such that each element has a 1.1 times
greater element size than the previous one while going towards plate boundaries. Material model accounting initiation
and propagation of composite ply damage is defined to the finite elements in the central zone while only elastic
210 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214
Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000 5

behavior is found adequate for the remaining elements. In order to prevent unrealistic element deformations due to
numerical issues, enhanced hourglass control is introduced to the finite elements of the plate. Additionally, distortion
of the elements is limited to the 10% of the original size for preventing excessive distortion due to the issues like
negative element volume or material degradation. Cohesive regions are modeled at interfaces of plies with different
orientations inside the central region and are composed of 0.25×0.25 mm 2 cohesive elements with zero thickness.
Carbon/epoxy material of which the mechanical and interface properties are given in Table 1 and Table 2 is used
as the ply material of composite beam in the analysis. Most of the material properties are provided by the material
supplier and measured using standard test methods. Only G1+, G1-, and η values are taken from Lopes et al. (2009)
since the elastic and strength properties of the materials are similar.

Table 1. Mechanical properties of carbon/epoxy material.

Density 1520 kg/m3

Elastic E1 = 140 GPa; E2 = 9 GPa; E3 = 9 GPa
ν12 = 0.35; ν13 = 0.35; ν23 = 0.48
G12 = 5 GPa; G13 = 5 GPa; G23 = 4 GPa
Strength XT = 2000 MPa; XC = 1500 MPa
YT = 65 MPa; YC = 220 MPa
S12 = 110 MPa; S13 = 110 MPa; S23 = 83 MPa;
Toughness G1+ = 81500 N/m ; G1- = 106300 N/m
G2+ = 270 N/m; G6 = 570 N/m

Table 2. Interface properties of carbon/epoxy material.

Interface strength To,I = 65 MPa; To,II = To,III = 110 MPa

Fracture toughness GI,c = 270 N/m; GII,c = 570 N/m
B-K criterion constant η = 1.45
Penalty stiffness KI = 3.6 x 1014 N/m3; KII = KIII = 4.9 x 1014 N/m3

Fixture base, which is a rectangular steel part having a rectangular window of 125×75 mm2, is modeled as a discrete
rigid body using with outer dimensions seen in Fig. 2. A total of 1220 Quadratic rigid elements (R3D4 in ABAQUS
library) are used for discretization of the fixture base. All degrees of freedoms of the reference point associated with
the fixture base are constrained in accordance with the standard experiment. In the assembly, the composite plate is
located on the fixture base. Four discrete rigid clamps are positioned on the top surface of the plate at the start of the
simulation, as seen in Fig. 2. Rigid clamps are also discretized using quadratic rigid elements (R3D4).
Hemi-spherical steel impactor is modeled as a discrete rigid body with 16 mm diameter and 3.387 kg mass. The
discretization of the hemi-spherical impactor is made by spherified cube method and quadratic rigid elements (R3D4)
are used. An initial velocity of 2.976 m/s corresponding to a 15 J impact is given to the impactor and it is located
above the center of the composite plate. All degree of freedoms of the impactor except translation in vertical direction
are restricted.
In the model, different contact interactions are defined between mating parts using general contact algorithm of the
ABAQUS/Explicit: (i) between the impactor and top surface of the plate, (ii) between the bottom surface of the plate
and top surface of the fixture base, (iii) between the clamp tips and the top surface of the plate, and (iv) inside the
beam. The reason why a contact is defined inside the beam is opposing free surfaces form at the interfaces following
a delamination damage. In all cases, hard contact with separation allowance is defined for interactions in normal
direction. For tangential motion, contact is defined with Coulomb friction model with friction coefficients 0.3 and 0.5
for metal-to-composite and composite-to-composite contacts, respectively. Rough contact is defined between the
rubber clamp tip and the composite plate meaning that no relative tangential motion occurs between these parts.
 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214 211
6 Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000

The analysis consists of two consequent solution steps. In the first step, with the aim of squeezing the plate between
the fixture base and the clamps, rigid clamps are moved 0.005 mm downward with a smooth step during 0.0001 s.
The resultant reaction force in each clamp is measured about 2700 N which is more than twice of the minimum
clamping force specified in ASTM D7136 standard (2012). During the second solution step of the analysis, clamps
are held stationary and the impactor moves downward with constant velocity until the initial contact between the
impactor and the plate occurs. Analysis of the impact event is performed following the initial contact.

Fig. 2. Geometry and boundary conditions of the virtual impact test setup.

4. Results

Finite element analysis of the 15 J - impact experiment on the [04/904/02]s CFRP plate specimen was performed in
a high performance workstation using 30 central processing unit. Variable mass scaling technique is used to prevent
decrease of the stable time increment under a predefined value. The analysis ran for 3 days until it was terminated 5
milliseconds after the initial contact due to the rebound of the impactor. In this section, results obtained from the finite
element analysis are presented.
Fig. 3 shows the matrix damage distribution under the impact zone at different contact times, t. One quarter of the
plate is removed for clear visualization of the damage states in both y-z and x-z planes under the impactor. The
sequence of the damage process can be summarized as follows,
 The first frame is taken at ti = 0 at which the initial contact between the impactor and the beam happens.
 At ti = 0.06 ms, initial matrix crack forms in the bottom layers where the maximum elongation due to bending
occurs. This form of the failure is similar to the one observed in the line impact experiments of [90/0] s beams.
Although the stacking sequence of the plate is [0 4/904/02]s in the plate coordinate system, matrix cracking of the
bottom plies is expected initial failure mode independent from the stacking sequence due to the spherical shape of
the impactor, the plate geometry and the boundary conditions.
 At ti = 0.13 ms, the initial matrix crack propagates in the bottom group of 0° plies. A damaged region starts to form
in the lowermost ply of the laminate. Some shear and bending cracks and delaminations are also observed through
the thickness of the plate.
 At ti = 0.30 ms, the initial matrix crack propagates in the bottom ply along the principal axis of the plate.
Delamination initiation is completed at all 0/90 interface at this instant. Additionally, matrix damage occurs in the
top 0° plies around the boundary of the contact between the impactor and the plate.
 At ti = 0.40 ms, further matrix crack propagation is observed in the bottom ply along the principal axis of the plate.
212 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214
Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000 7

 At ti = 1.00 ms, matrix damage in the bottom ply expands out from the impact zone. A significant increase in sizes
of the existent delaminations in two interfaces of the beam are observed in this frame.
 At ti = 2.00 ms, the existent matrix damage and delaminations grows outwards from the center of the plate. Multiple
vertical matrix cracks occur in the top 0° plies. It can be observed that delamination growth occurs mainly in the
orientation of the bottom layer.
 The last frame shows the distribution of the complicated matrix damage state inside the laminate when the impactor
reaches the maximum displacement at ti = 4.10 ms.

Fig. 3. Through-the-thickness views matrix damage distribution under the impact zone at different contact times (Images are taken with multiple
cut planes: x-z and y-z).

Fig. 4a shows the delamination damage at the end of the analysis (ti = 5.00 ms) at each four 0/90 interfaces starting
from the uppermost. The pictures encircle the 100 × 100 mm2 central region of the plate. This result show that
delaminations propagate forming a peanut shape with a major axis in the same orientation as of the lower adjacent ply
in accordance with the experiments presented in the literature (Abrate, 1991). Fig. 4b compares the projected
delaminations obtained in the experiment and the simulation. It should be noted that the experiment is conducted with
the identical conditions for comparison purposes. Although overall projected delaminated area is larger in case of
simulations, a good agreement was obtained in terms of the extent of the delamination in the direction of plate length.
8 Mirac Onur
Bozkurt et al. /Bozkurt et al.
Structural / Procedia
Integrity Structural
Procedia Integrity
00 (2019) 21 (2019) 206–214
000–000 213

Fig. 4. (a) Delamination damage at 0/90 interfaces of the laminate at the end of the simulation and (b) comparison of the projected delaminations
obtained in the simulation and the experiment.

5. Conclusions

In this study, a virtual impact test setup was modeled and finite element analysis of the 15 J – impact event on the
[04/904/04]s CFRP plate specimen was conducted in ABAQUS/Explicit. The material model accounting matrix and
fiber failure modes of the fiber reinforced composites was developed and implemented into the model via a user-
written subroutine VUMAT. Delamination damage in the plate was simulated by inserting cohesive elements at the
interfaces of plies with different orientations.
The results of the present analysis show that the initial failure mechanism in the 3-D low-velocity impact event is
the matrix cracking in the lowermost plies independent from the stacking sequence of the laminate. Although matrix
cracking does not lead to a considerable drop in the impact load, it should be taken into account in the analyses since
it promotes formation of delamination which is one of the most energy dissipative failure modes of composites. It is
also observed that delaminated regions expand in the same direction as of the fibers of the lower adjacent layer in
accordance with the bending stiffness mismatching concept.

Acknowledgements

This work is supported by Turkish Aerospace – Rotary Wing Technology Center under contract DKTM/2015/05.
The authors of the paper would like to thank METU Center for Wind Energy for the use of their facilities.

References

Abrate, S., 1991. Impact on laminated composite materials. Appl.Mech. Rev., 44(4), 155-190.
ASTM International, 2012. Standard D7136 - standard test method for measuring the damage resistance of a fiber reinforced polymer matrix
composite to a drop weight impact event. http://dx.doi.org/10.1520/D7136.
Benzeggagh M.L., Kenane M., 1996. Measurement of mixed-mode delamination fracture toughness of unidirectional glass/epoxy composites with
mixed-mode bending apparatus. Composite Science and Technology; 56:439–49.
González, E. V., Maimí, P., Camanho, P. P., Turon, A., and Mayugo, J. A., 2012. Simulation of drop-weight impact and compression after impact
tests on composite laminates. Composite Structures, Vol. 94, (11), pp. 3364-3378.
Hashin, Z., 1980. Failure criteria for unidirectional fiber composites. Journal of applied mechanics, Vol. 47, (2), pp. 329-334.
Joshi, S.P. and Sun, C.T., 1985. Impact induced fracture in a laminated composite. J. Compos. Muter., 19, 51-66.
Lopes, C. S., Camanho, P. P., Gürdal, Z., Maimí, P., and González, E. V., 2009. Low-velocity impact damage on dispersed stacking sequence
laminates. Part II: Numerical simulations. Composites Science and Technology, Vol. 69, (7-8), pp. 937-947.
214 Mirac Onur Bozkurt et al. / Procedia Structural Integrity 21 (2019) 206–214
Bozkurt et al. / Structural Integrity Procedia 00 (2019) 000–000 9

Lopes, C. S., Sádaba, S., González, C., Llorca, J., and Camanho, P. P., 2016. Physically-sound simulation of low-velocity impact on fiber-reinforced
laminates. International Journal of Impact Engineering, Vol. 92, pp. 3-17.
Simulia, 2012. ABAQUS 6.12-2 - analysis user manual. Providence (RI, USA).
Soto, A., González, E. V., Maimí, P., de la Escalera, F. M., de Aja, J. S., and Alvarez, E., 2018. Low velocity impact and compression after impact
simulation of thin ply laminates. Composites Part A: Applied Science and Manufacturing, Vol. 109, pp. 413-427.
Topac, O. T., Gozluklu, B., Gurses, E., and Coker, D., 2017. Experimental and computational study of the damage process in CFRP composite
beams under low-velocity impact. Composites Part A: Applied Science and Manufacturing, Vol. 92, pp. 167-182.