Mechanics of Advanced Materials and Structures

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/umcm20

Numerical simulation on mechanical properties

and damage behavior of CFRP with self-healing
microvascular channels

Yang Zhao, Zhongqi Wang, Peng Jiang, Anyang Wang, Zhengping Chang &
Yonggang Kang

To cite this article: Yang Zhao, Zhongqi Wang, Peng Jiang, Anyang Wang, Zhengping Chang &
Yonggang Kang (2024) Numerical simulation on mechanical properties and damage behavior
of CFRP with self-healing microvascular channels, Mechanics of Advanced Materials and
Structures, 31:2, 297-313, DOI: 10.1080/15376494.2022.2114040

To link to this article: https://doi.org/10.1080/15376494.2022.2114040

Published online: 03 Sep 2022.

Submit your article to this journal

Article views: 324

View related articles

View Crossmark data

Citing articles: 11 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=umcm20
MECHANICS OF ADVANCED MATERIALS AND STRUCTURES
2024, VOL. 31, NO. 2, 297–313
https://doi.org/10.1080/15376494.2022.2114040

ORIGINAL ARTICLE

Numerical simulation on mechanical properties and damage behavior of CFRP

with self-healing microvascular channels
Yang Zhaoa, Zhongqi Wanga, Peng Jiangb, Anyang Wanga, Zhengping Changa, and Yonggang Kanga
a
School of Mechanical Engineering, Northwestern Polytechnical University, Xi’an, China; bAVIC Chengdu Aircraft Industrial (Group) Co., Ltd,
Chengdu, China

ABSTRACT ARTICLE HISTORY

In this study, 3D representative volume element models with microchannels were established Received 29 March 2022
based on measured data. And the periodic boundary conditions were moderately modified to Accepted 12 August 2022
adapt to the actual constraints of the studied object. The effects of different microvascular parameters
KEYWORDS
on the tensile and compressive properties of composites were studied. And the damage pattern of
CFRP; microvascular
carbon fiber as the main load-bearing phase under the force was analyzed. The models were verified channel; simulation;
by tensile testing and micromorphology experiments. The research results are of application to the mechanical properties;
design and analysis of CFRP laminates containing similar microchannel structures. damage behavior

1. Introduction an external self-healing system, the healing process is realized

by placing microchannels or microcapsules containing healing
Owing to their unique and superior properties, including a
agents in the matrix [13]. The healing process can be con-
high specific strength and modulus, corrosion resistance,
ducted under different conditions by selecting different healing
fatigue resistance, good damping/vibration reduction, and
agents. Multiple effective healing processes can be achieved
designability, interest in carbon fiber-reinforced polymer
through supplementation with these external healing agents.
(CFRP) has grown very rapidly [1–3]. However, micro dam- However, the addition of an external self-healing structure
ages to such composites inevitably occur during the process affects the mechanical properties. Therefore, it is of great sig-
of preparation and service [4]. In particular, traditional com- nificance to engineering applications to study the effects of
posite structures can easily undergo delamination, micro- self-healing systems on the mechanical properties and failure
cracks, fiber-matrix debonding, and other damage when behaviors of composites, so as to provide better mechanical
subjected to even a low-energy impact [5]. It is difficult to properties simultaneously with the healing performance.
detect and repair this type of microscale damage from the After Toohey et al. proposed self-healing systems with
outside; hence, it is called “barely visible impact damage,” single microchannels [14] and double microchannels [15] in
which may lead to serious degradations of material mechan- composites, research on microchannels entered a rapid
ical properties [6]. Aiming to solve this problem, the U.S. development stage. Nowadays, microvascular approaches
military initially proposed the concept of self-healing com- include hollow glass fibers [6, 14–20], fugitive ink dissol-
posites [7]. The main idea of self-healing composites is to ution [21], nylon fiber removal [22], solder evacuation and
imitate the self-healing process of an organism to heal polytetrafluoroethylene-coated steel removal [23–25], and
internal damages to a material, avoid further damage to the vaporization of sacrificial components (VaSC) [26–28].
material, and prolong the service life of the structure [8]. However, compared with the experimental research of com-
Self-healing methods can be divided into intrinsic self- posites with microchannels, the research on simulation
healing [9,10] and external self-healing [11,12] approaches. methods is comparatively limited due to the limitations of
An intrinsic self-healing system does not require an external corresponding technologies. Based on the reference of repre-
healing agent. It can become self-healing based on a revers- sentative volume element (RVE), researchers have carried
ible chemical reaction when cracks appear in the matrix. out some work on the mechanical properties of composites
There is no need to consider the compatibility of the addi- with microchannels combined with the finite element (FE)
tive with the matrix in an intrinsic self-healing system, as it method, which is proven to be a powerful technique [29].
does not degrade the performance of the original material. The RVE should be sufficiently small relative to the whole
However, the healing efficiency decreases as the number of structure to ensure that it suffers a uniform load, while the
healings increases due to the deactivation of the catalyst. For RVE should be sufficiently large relative to the

CONTACT Zhengping Chang [email protected] School of Mechanical Engineering, Northwestern Polytechnical University, Xi’an 710072, China;
Yonggang Kang [email protected] School of Mechanical Engineering, Northwestern Polytechnical University, Xi’an 710072, China.
Supplemental materials are available for this article at https://doi.org/10.1080/15376494.2022.2114040.
ß 2022 Taylor & Francis Group, LLC
298 Y. ZHAO ET AL.

Figure 1. Cross-sectional morphologies of CFRP with 600 lm diameter microchannel of two arrangement patterns.

microstructure and contain as much micro-information as In summary, the current research on the mechanical
possible [30]. The component phases of composite have a properties of CFRP with microchannels is mainly based on
statistical periodicity. When the periodically arranged RVEs experiments, and there are still some imperfections in the
in the composite structure suffer a uniform remote load, existing simulation studies. The current researches mainly
theoretically all the RVEs are subjected to the same stress- focus on the local stress of the resin-rich zone and strength
strain field. The stress-strain field on a random RVE can be prediction of the composite structure, and the effects of the
used to reflect the meso stress-strain field of the composite. important microvascular parameters on the mechanical
The equivalent elastic modulus and mechanical properties of properties have not been discussed systematically and
the whole composites can be obtained by calculating the summarized regularly. And the existing researches on the
mechanical response of a random RVE [31,32]. The arrange- mechanical properties and the damage in weak areas of
ment of microchannels in composites can be considered composites with microchannels are insufficient. In this
periodic, and the mechanical properties and damage analysis study, RVE models with microchannels are established based
based on RVE models have been carried out. on measured data. The commonly periodic boundary condi-
By combining experimental data and progressive damage tions are improved by freeing the upper and lower planes to
methods, Nguyen et al. [5] established RVE models consid- make them more consistent with the actual boundary con-
straints of the studied object. The effects of microvascular
ering in-plane and interlayer damages, aiming to investigate
channels with different parameters (diameter, spacing, and
the effects of the microvascular diameter and shape on the
arrangement) on the tensile and compressive properties of
strength and modulus of CFRP. The results showed that the
the CFRP are analyzed. The effects of microvascular parame-
models predicted with good accuracy (often within 5%).
ters on mechanical properties and damage behavior of compo-
Huang et al. [33] observed the effects of microvascular
sites under different loading conditions are explored. The
diameters on the morphologies of CFRP laminates by
study aims at recommending proper microvascular parameters
microscopy, and established plane strain FE models for ana- for further research on self-healing performance and ensuring
lyzing the mechanical responses and failure modes under that the mechanical properties will not be greatly degraded.
compression loads. The results showed that the error in the
simulation results was approximately 10%, and that the deg-
radation of the compressive strength of the composite with 2. Experimental procedure
the microchannel varied from 13% to 70%. Hartl et al. [34]
The 5015 resin and 5015 hardener were provided from Sino
established RVE models to study the effects of microchan- Composites Beijing Co., Ltd. The T700 carbon fabrics were
nels on the elastic response, damage initiation, and failure provided by Zhongfu Shenying Carbon Fibre Co., Ltd.
mode of a laminate under complex loading. The results Biodegradable polylactic acid (PLA) toughened using the nano-
showed that the strength and stiffness decreased significantly droplets method [36] in the laboratory was used to generate
when the loading direction and fibers near the channel were the microchannels in the composites. The PLA pellets were
transverse to the microchannel. A larger spacing delayed the provided by Suzhou Jiangcangfa Plasticising Co., Ltd.
occurrence of damage, yet accelerated the ultimate failure. Composite laminates with the stacking sequence ([90/0]6s,
Al-Shawk et al. [35] investigated the effects of the micro- [0/90]6s) were fabricated by vacuum assisted resin infusion
channel on the stress concentrations and failure modes of (VARI) processing, as shown in Figure S1 (supplementary
laminates with different stacking sequences. The results material). The purpose of the operation was to observe the
revealed that the lowest normalized stresses were observed effects of the microvascular arrangement, diameter (400 lm,
with a 0
stacking configuration, whereas the normalized 600 lm, 800 lm, and 1000 lm), and spacing (2 mm, 4 mm,
stresses were the highest at 90
. The stacking direction of 6 mm, and 8 mm) on the number of influenced layers. The
the fabric primarily affected the failures in resin-rich zones. PLA fibers were placed in the middle of the preform. The
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 299

Table 1. Cross-sectional parameters of RVE with a transverse microchannel.

Microvascular Resin-rich zone Fiber fabric waviness Assumed transition Stacking
diameter /d (lm) length /LRPa(mm) height /hd (mm) thickness /t (mm) sequence
400 1.50 1.06 0.013 [90/0]6s
600 2.04 1.20 0.020
800 2.45 1.46 0.027
1000 2.84 1.58 0.033
a
Here, LRP represents the resin-rich zone length, which is equal to half of the total resin-rich zone length.

Table 2. Cross-sectional parameters of RVE with a longitudinal microchannel.

Microvascular Fiber disturbance Fiber disturbance Assumed transition Stacking
diameter/d (lm) zone length /LRPa (mm) height /hd (mm) thickness /t (mm) sequence
400 0.51 0.28 0.013 [0/90]6s
600 0.87 0.46 0.020
800 1.27 0.68 0.027
1000 1.80 0.95 0.033
a
Here, LRP represents the fiber disturbance zone length, which is equal to half of the total fiber disturbance zone length.

arrangement patterns for the microchannels were realized by the microchannels are oriented parallel to the adjacent 90

changing the stacking sequences of the fabric layers. fiber layers, that is, the longitudinal arrangement of the micro-
Metallographic samples of cured CFRP with microchannels of channel or ’’longitudinal microchannel’’. Here, the fibers are
various diameters were observed to obtain the cross-sectional squeezed by the microchannel, but remain straight. The add-
parameters of the samples. Four kinds of laminates, the two ition of the microchannel causes fiber disturbances, but with-
with stacking sequence [0/90]3S and the two with stacking out producing a resin-rich zone in cross-section of CFRP. The
sequences [90/0]3S, considering a 4 mm spacing at a diameter longitudinal microchannel has less of an effect on the fiber
of 600 lm and a blank group with stacking sequence [0/90]3S structure at the same diameter as the transverse microchannel,
were prepared by the VARI process. To verify the accuracy of consistent with the observations in [37]. The cross-sectional
RVE models, transverse tensile tests of the blank group (Group morphology changed after the PLA fiber degraded.
A), transverse tensile tests of the laminates with transverse Considering that the microvascular shape has little effect on
microchannels (Group B), longitudinal tensile tests of the lami- the mechanical properties of CFRP [5], the microchannel is
nates with transverse microchannels (Group C), transverse ten- assumed to be circular to simplify the RVE models.
sile tests of the laminates with longitudinal microchannels Tables 1 and 2 list the geometric parameters of a cross-sec-
(Group D), and longitudinal tensile tests of the laminates with tion of CFRP with various diameters of microchannels under
longitudinal microchannels (Group E) were carried out. The two arrangement patterns. According to Al-Shawk et al. [35],
laminates were cut into specimens with dimensions of the cross-sectional geometric parameters of the microchannel
250  25 mm2 according to the ASTM D3039. The tension in CFRP are defined as shown in Figure 2. The differences
tests were conducted on an electronic tensile machine between the two cross-sectional theoretical models are whether
(ETM104C) at a speed of 1 mm/min. The micromorphology of a resin-rich zone appears, and the number of layers being
the tensile fracture surrounding the microchannel was observed influenced. Figure 1a shows that the actual thickness of the
using a scanning electron microscope (SEM, TESCAN VEGA transitional zone directly above the microchannel may be t ¼
3 LMU). More detailed experimental information is described 0, indicating a sharp point. It is disadvantageous to perform
in the supplementary material. mesh generation and simulation analysis in the RVE. Based on
Ref. [34], it is assumed that t ¼ d=30, d represents the micro-
vascular diameter. The disturbed zone decays smoothly to a
3. Geometrical parameters and material properties small thickness toward both ends.
3.1. Dimension parameters of the model
3.2. Properties of component materials
The cross-sectional morphologies of CFRP with self-healing
microchannels were observed by optical microscopy to In this study, composites with fiber volume fraction of
obtain the parameters for establishing the RVE models. The approximately 55% were fabricated. The properties of com-
microchannel left after PLA degradation served as the loca- ponent materials are listed in Table 3. The data come from
tion for filling the healing agent, therefore was considered as technical documents provided by the merchant or laboratory
a cavity. Figure 1 shows the cross-sectional morphologies of experiments. The properties of the cohesive elements are
CFRP with a 600 lm diameter microchannel under two presented in Table 4.
arrangement patterns. As shown in Figure 1a, the micro-
channel is oriented perpendicular to the adjacent 0
fiber 4. Finite element models development
layers, that is, the transverse arrangement of the microchan-
4.1. Generation of RVE
nel or “transverse microchannel.” The addition of the micro-
channel causes the formation of a resin-rich zone, resulting At present, the RVE approach is commonly used for theor-
in the waviness of the 0
fiber layers. As shown in Figure 1b, etical and numerical analyses of inhomogeneous materials
300 Y. ZHAO ET AL.

Figure 2. Cross-sectional parameters of RVE with a microchannel.

Table 3. Properties for component materials.

Component materials Property Value
T700/5015 Young’s modulus E11 ¼ 1:39  105 MPa, E22 ¼ E33 ¼ 8:15  103 MPa
Shear modulus G12 ¼ G13 ¼ 4:43  103 MPa, G23 ¼ 3:55  103 MPa
Poisson’s ratio l12 ¼ 0:321
Tensile strength XT ¼ 1935 MPa, YT ¼ 31 MPa
Compressive strength XC ¼ 1272 MPa, YC ¼ 125 MPa
Shear strength S12 ¼ S23 ¼ 111 MPa
Density q ¼ 1:59  109 T=mm3
In-plane fracture toughness Gft ¼ 91 mJ=mm2 , Gfc ¼ 79 mJ=mm2
Gmt ¼ 0:15 mJ=mm2 , Gmc ¼ 0:45 mJ=mm2
5015 resin Modulus E ¼ 3:03  103 MPa
Poisson’s ratio l ¼ 0:35
Density q ¼ 1:1  109 T=mm3

Table 4. Properties for cohesive element. 2ðy0  y1 Þ 3ðy1  y0 Þ

a¼ , b¼ , c ¼ 0, d ¼ y0 : (2)
Property Value x13 x12
Interface stiffness KI ¼ 106 N=mm KII ¼ 106 N=mm KIII ¼ 106 N=mm
Interface peak strength Nmax ¼ 61 MPa Smax ¼ Tmax ¼ 68 MPa For the transverse microchannel, the coordinates of the
Interlaminar fracture GI ¼ 0:445 mJ=mm2 GII ¼ GIII ¼ 1:856 mJ=mm2 two points ð0, r þ tÞ and ðLRP , tÞ are substituted into Eq. (2)
toughness
Density q ¼ 1:4  109 T=mm3 Then, the function f ðxÞ of the fiber fabric waviness curve is
expressed as follows:
8
< 2r x3  3r x2 þ r þ t, x < L ,
RP
[37]. Here, to realize the parametric modeling of RVE, the f ðxÞ ¼ L3RP L2RP (3)
:
fiber disturbance or the fiber fabric waviness caused by the t , x  LRP :
microchannel are described mathematically, according to For the longitudinal microchannel, the coordinates of
[34]. In the coordinate system shown in Figure 3, several two points ð0, r þ tÞ and ðLRP , tlam Þ are substituted into Eq.
assumptions are introduced to obtain the mathematical
(2). Accordingly, the function f ðxÞ of the fiber disturbance
curve: (1) f ðxÞ  0, 8x 2 ½0, L; (2) f ðxÞ ¼ f ðxÞ, 8x 2
curve is expressed as follows:
½0, L; (3) f 0 ð0Þ ¼ f 0 ðLRP Þ ¼ 0; (4) when x 2 ½0, LRP , the (
2ðrþttlam Þ 3 lam Þ 2
curve of fiber fabric waviness or fiber disturbance is x  3ðrþtt x þ r þ t, x < LRP ,
defined as a cubic function, whereas when x 2 ½LRP , L, f ðxÞ ¼ L3RP L2RP
tlam , x  LRP ,
the curve of the unaffected fiber fabric is defined as a
line; and (5) the whole curve can be defined as a periodic (4)
function, and the minimum positive period is the micro- where tlam is the monolayer thickness in laminates.
vascular spacing 2 L. Based on the above assumptions, the After the function of the curves was defined, the parame-
curve is expressed as follows: terized models were implemented as a Python script for exe-
(
ax3 þ bx2 þ cx þ d, x < LRP, cution by Abaqus. Then, RVE models with different
f ðxÞ ¼ (1) parameters were established by changing the program. The
C , x  LRP:
effects of parameters such as the arrangement (transverse
The cubic function passes through two fixed points (set and longitudinal), spacing (2 mm, 4 mm, 6 mm, and 8 mm),
to ð0, y0 Þ and ðx1 , y1 Þ), and the values of the first derivatives diameter (400 lm, 600 lm, 800 lm, and 1000 lm), and load
at the two points are 0, meeting assumption (3). The four direction (transverse tensile, transverse compressive, longitu-
coefficients are determined as follows: dinal tensile, and longitudinal compressive) on the
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 301

Figure 3. Curves of fiber disturbance or fiber fabric waviness caused by microchannel.

Figure 4. 3 D RVE model with a microchannel.

one RVE with a microchannel in the thickness direction of

CFRP, whereas the structure was periodic in the horizontal
X–Z plane. According to the selected method for the RVE
model, periodic boundary conditions were applied in the
four surrounding planes of the RVE, and no constraint was
applied on the top and bottom planes, as shown in Figure 5.
The periodic boundary conditions were subject to two con-
straints: (1) coordinated deformation, that is, there was no
clearance or interference on the boundary of adjacent ele-
ments during deformation; and (2) continuous stress, that
is, the stress on the associated boundary of the RVE
remained the same during deformation [38, 40]. The same
Figure 5. Boundary conditions of RVE. deformation pattern was ensured for every RVE by applying
periodic boundary conditions to obtain accurate mechanical
mechanical properties of CFRP with microchannels were responses of the composites [41]. As stated by Suquet [42],
analyzed. A schematic diagram of the generated 3D RVE the expression for the periodic displacement field is given as
model containing a microchannel is shown in Figure 4. In follows:
the coordinate system, the load parallel to the X direction is
ui ¼ e ik xk þ ui , (5)
called the transverse load perpendicular to the microchannel,
and the load parallel to the Z direction is called the longitu- where e ik represents a globally average strain, xk represents
dinal load parallel to the microchannel. the coordinates of any node, and ui represents the displace-
ment correction, which is related to the applied global load.
The values of ui are the same for associated surfaces, edges
4.2. Boundary conditions and loading conditions
or corner. ui represents the displacement of any node. Xia
For accurate results, FE models must apply reasonable et al. [43] derived unified periodic boundary conditions for
boundary conditions [38,39]. In this study, there was only RVE with parallel boundary surfaces, as follows:
302 Y. ZHAO ET AL.

 
jþ j jþ j j
ui  ui ¼ e ik xk  xk ¼ e ik Dxk , (6) cohesive elements, which were meshed with a 3 D eight-
node solid element (COH3D8) between the layers of CFRP.
jþ j The bilinear constitutive model was adopted for the cohesive
where ui and ui are the displacements of the correspond-
jþ elements, and the secondary stress criterion was adopted for
ing nodes on two associated surfaces, respectively. xk and
j the damage initiation criterion. The thickness of the cohe-
xk are the coordinates of the corresponding nodes on the sive layers was assumed as 0.01 mm. In general, proper
j
associated surfaces, respectively. Dxk is the coordinate differ- meshing is of great importance to avoid distortion of the
ence between the corresponding nodes. elements. Here, continuum shell elements (SC8R) were
The boundary condition settings for RVE are shown in chosen to model the composite laminates. The resin-rich
Figure 5, where a local coordinate system X  Y  Z is zone was meshed with wedge-shaped solid elements (C3D6).
defined, with point A as the origin of the coordinate system. The RVE model was loaded with different directions of tension
The length, width, and height of RVE model are Dx, Dz, or compression displacement through the reference point with
and Dy, respectively. Assuming that constant average in- a coupling plane. Nguyen et al. [5] found that high mesh dens-
 
plane strains e0xx , e0yy , e0zz , c0xy , c0xz , c0yz are imposed on the ity can accurately capture the realistic response post-ultimate
load of CFRP, but has little effect on the stress and modulus of
RVE, the periodic boundary conditions defined in Eq. (6) CFRP. The strength and modulus of CFRP and the damage
can be transformed into constraint equations, as follows. trends of the main load-bearing phase fibers at the ultimate
The displacement boundary conditions for the corre-
load were the focus of this study, so a low mesh density was
sponding nodes on the associated surfaces x ¼ 0 and x ¼ Dx
used to shorten the calculation time.
or z ¼ 0 and z ¼ Dz of RVE are given as follows:
P  P0 : ui ðDx, y, zÞ
5. Results and discussion
0 y Dy,
 ui ð0, y, zÞ ¼ Dxe0i1 , i ¼ 13 5.1. Comparison of the simulation result with the
0 z Dz,
experimental result
M  M0 : ui ðx, y, DzÞ
0 x Dx, Figure 6 depicts five groups of simulation (FE) and their veri-
 ui ðx, y, 0Þ ¼ Dze0i3 , i ¼ 13 fied experimental (EXP) stress-strain curves. As shown in
0 y Dy:
Figure 6a, the load of the composite decreases sharply during
(7) the experiment. However, the load first drops gently and then
The displacement boundary conditions for corresponding drops suddenly, delaying the fracture process of CFRP in the
nodes on the associated edges of RVE (taking the four edges simulation result; this similar to the result shown in a previous
AB, EF, HG, and DC parallel to the X-axis as examples) are study [5]. Table 5 presents more details experimental and
denoted as follows: simulation results of mechanical properties of CFRP with
microchannels. It can be seen from Table 5 that the simula-
EF  AB : ui ðx, Dy, 0Þ  ui ðx, 0, 0Þ
tion results are often slightly larger than the experimental
¼ Dye0i2 , results, that is, because original defects such as pores and
HG  AB : ui ðx, Dy, DzÞ  ui ðx, 0, 0Þ cracks are inevitably introduced into the CFRP specimens in
, i ¼ 1, 3, 0 x Dx,
¼ Dye0i2 þ Dze0i3 the experiment, whereas the microdefects are not considered
DC  AB : ui ðx, 0, DzÞ  ui ðx, 0, 0Þ in the models; thus, the simulation values are greater than the
experimental values. The errors between the experiment and
¼ Dze0i3 :
simulation results are less than 5%, proving that the estab-
(8) lished RVE models can effectively predict the tensile and com-
The displacement boundary conditions for the associated pressive properties of CFRP. Figure 6b shows the SEM images
corner nodes of RVE (taking the four nodes A, D, H, and G of the transverse tensile fracture morphology of CFRP with
as examples) are expressed as follows: longitudinal microchannels considering a 4 mm spacing at a
DA: ui ð0, 0, DzÞ  ui ð0, 0, 0Þ ¼ Dze0i3 ,
diameter of 600 lm. The fiber fracture initiates at the side of
HA: ui ð0, Dy, DzÞ  ui ð0, 0, 0Þ ¼ Dye0i2 þ Dze0i3 , i ¼ 1, 3,
the hole and progresses to the surface of the laminate in a dir-
ection approximately 45
off the vertical direction; this is simi-
GA: ui ðDx, Dy, DzÞ  ui ð0, 0, 0Þ ¼ Dxe0i1 þ Dye0i2 þ Dze0i3 :
lar to the simulation result shown in Figure 18b of Section
(9) 5.4, where the fiber damage extends at the sides of the hole by
For other edges and corner nodes, the constraint equa- deflecting at an angle to the surface of the laminate.
tions can be derived by referring to Eq. (7)–(9).
5.2. Effect of transverse microchannels on tensile and
4.3. Other settings of RVEs compressive properties of CFRP
The in-plane damage was considered based on the Hashin 5.2.1. Effect of microvascular diameter
progressive damage criterion in the RVE with a microchan- The effects of the microvascular diameters (400 lm, 600 lm,
nel. The interlaminar damage was considered by introducing 800 lm, and 1000 lm) on the CFRP with the transverse
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 303

Figure 6. (a) Stress-strain curves of simulation and experimental results for CFRP, (b) SEM images of the damage pattern for CFRP with longitudinal microchannels
with 4 mm spacing at diameter of 600 lm.

Table 5. Experimental and simulation results of mechanical properties of CFRP with microchannels.
EXP (S.D.) FE Error (%)
Group Strength (MPa) Modulus (GPa) Strength (MPa) Modulus (GPa) Strength Modulus
A 908.6 (18.6) 65.3 (0.2) 935.4 67.8 2.95 3.83
B 720.6 (36.7) 54.8 (0.3) 745.6 55.9 3.47 2.01
C 794.8 (33.2) 55.6 (0.2) 808.1 58.0 1.67 4.32
D 808.6 (29.3) 60.5 (0.5) 843.0 60.6 4.25 0.17
E 882.5 (16.0) 63.1 (0.6) 911.3 65.7 3.26 4.12
Note: S.D. indicates the standard deviation.

microchannel were studied while considering only a 4 mm increases, causing a decrease in the strength and modulus of
spacing under transverse or longitudinal loads. According to CFRP. In particular, the resin-rich zones begin to overlap
the data in Table 1, there are different parameters for the when the diameter is larger than 600 lm. For the stacking
resin-rich zones formed by the transverse microchannels sequence of [90/0]3S, the main load-bearing zone of CFRP
with the various diameters. There is no overlapping zone under a longitudinal load comprises the 90
layers oriented
when the total length of the resin-rich zone is less than the parallel to the microchannel. The fibers in the 90
layers
microvascular spacing, as shown in Figure 2Sa. When the remain straight in the direction of the load; thus, their
total length of a resin-rich zone is greater than the micro- mechanical properties are not affected. While the main load-
vascular spacing, there is an all fiber fabric waviness zone in bearing zone of CFRP under a transverse load comprises the
the cross-section, as shown in Figure 2Sb. There are overlap- 0
layers oriented perpendicular to the microchannel, there
ping resin-rich zones in cross-section of CFRP with trans- is fiber fabric waviness zone. The mechanical properties are
verse microchannels at diameters of 600 lm, 800 lm, and affected because the fibers are bent in the direction of the
1000 lm. The mathematical expression of the waviness curve load. Therefore, the degradations in the strength and modu-
(Eq. (1)) satisfies the cubic function, and there is no linear lus of CFRP under longitudinal loads are less than those
part. As mentioned above, RVE models were established to under transverse loads.
study the mechanical responses of CFRP with transverse
microchannels under different loads, according to the effects
mentioned above from the microvascular diameters on the 5.2.2. Effect of microvascular spacing
geometry of the cross-section of CFRP. The results are pre- The effects of microvascular spacings (2 mm, 4 mm, 6 mm,
sented in Figure 7. and 8 mm) on CFRP with a transverse microchannel are
The stress-strain curves of CFRP without a microchannel described here while considering only a 600 lm diameter
under tensile or compressive loads are basically the same under transverse or longitudinal loads. According to the
regardless of the direction of the loads, as shown in Figure data in Table 1, the length of the resin-rich zone caused by
7a and b or Figure 7c and d. Based on the simulation results the transverse microchannel with a diameter of 600 lm is
in Figure 7, the effects of the diameter on the tensile and 2.04 mm, that is, the total length of the resin-rich zone
compressive properties of CFRP with transverse microchan- is 4.08 mm. The resin-rich zones overlap when the spacing
nel can be calculated, as shown in Figure 8. In general, the is less than the total length of the resin-rich zone, and the
mechanical properties decrease as the diameter increases. A waviness of the fiber fabric occurs in the entire cross-section
microchannel with a higher diameter forms a larger resin- of CFRP, as shown in Figure 3S(a). There is an overlapping
rich zone and the number of fiber fabric waviness layers resin-rich zone when the spacing is 2 or 4 mm. The curve of
304 Y. ZHAO ET AL.

Figure 7. Stress-strain curves of CFRP with transverse microchannels of various diameters under different loads.

Figure 8. Effect of diameters on tensile and compressive properties of CFRP with transverse microchannels.

the fiber fabric waviness can be expressed mathematically as According to the effects mentioned above of microvascular
Eq. (1) satisfies the cubic function change, and there is no spacings on the geometry of cross-section in CFRP, the
linear part. There is no overlapping resin-rich zone when RVEs are established to study the mechanical responses of
the microvascular spacing is 6 mm or 8 mm, as shown in CFRP with transverse microchannels under different loads.
Figure 3S(b), and the curve can be expressed using Eq. (3). The results are given in Figure 9.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 305

Figure 9. Stress-strain curves of CFRP with transverse microchannels of various spacings under different loads.

Based on the simulation results in Figure 9, the effects of the main load-bearing zones, whereas no waviness is
the microvascular spacing on the tensile and compressive observed in the fiber fabric; thus, the strength and modulus
properties of CFRP with transverse microchannels can be of CFRP remain the same.
calculated, as shown in Figure 10. The strength and modulus
of CFRP show different regularities with spacing under
5.3. Effect of longitudinal microchannels on tensile and
transverse and longitudinal loads, but show the same regu-
compressive properties of CFRP
larities with spacing under tensile and compressive loads.
Under the transverse loads, the strength and modulus of 5.3.1. Effect of microvascular diameter
CFRP show decreasing trends with a reducing spacing. The diameter of the microchannel was respectively set to
Because the overlapping resin-rich zone near the microchan- 400 lm, 600 lm, 800 lm, and 1000 lm at a spacing of 4 mm.
nels increases as the spacing reduces, the main load-bearing According to the data in Table 2, the total length of the
zone of the 0
layers is more significantly influenced by the fiber disturbance formed by the longitudinal microchannel
fiber fabric waviness. This is the main reason for the with a diameter of 1000 lm is only 3.6 mm, which is smaller
decrease in the strength and modulus of CFRP with trans- than the considered spacing of 4 mm. Therefore, there is no
verse microchannels. Under the longitudinal loads, there is overlapping fiber disturbance zone in the cross-section of
also no overlap between the resin-rich zones when the spac- CFRP within the range of microvascular diameters studied.
ing is reduced from 8 mm to 6 mm. The fiber fabric wavi- The curve of the fiber disturbance was established according
ness zone is not affected by the reduced spacing; thus, the to Eq. (4). Then, RVE models were established to study the
strength and modulus of CFRP decrease slightly. The resin- mechanical responses of CFRP with longitudinal microchan-
rich zones begin to overlap when the spacing decreases from nels under different load modes. The results are presented
6 mm to 4 mm, causing a significant decrease in the strength in Figure 11.
and modulus of CFRP. The resin-rich zones overlap further Based on the simulation results in Figure 11, the effects
when the spacing is reduced from 4 mm to 2 mm. Fiber dis- of the diameter on the tensile and compressive properties of
turbances are caused by microchannels in the 90
layers of CFRP with longitudinal microchannels can be calculated, as
306 Y. ZHAO ET AL.

Figure 10. Effect of spacings on tensile and compressive properties of CFRP with transverse microchannels.

shown in Figure 12. Under longitudinal loads, the values of Eq. (4), and RVE models were established to study the mechan-
strength and modulus fluctuate within 4.4% compared with ical responses of CFRP with longitudinal microchannels under
the control group, which can be considered as ’’no change’’ different loads. The results are presented in Figure 13.
in the mechanical properties. The 90
layers oriented parallel Based on the simulation results in Figure 13, the effects of
to the microchannel play the main role of load bearing, and the spacing on the tensile and compressive properties of CFRP
the loaded fibers remain straight in the direction of the with longitudinal microchannels were calculated, as shown in
load. Therefore, the effect of the diameter on the mechanical Figure 14. The strength and modulus of CFRP increase slowly
properties of CFRP with longitudinal microchannels is small. with increasing spacing considering the four loads studied. In
Under transverse loads, the strength and modulus of CFRP particular, the values of the tensile strength and modulus fluc-
hardly decrease relative to the control group when the diam- tuate within 5% compared with the control group under longi-
eter is 400 lm. However, the mechanical properties of CFRP tudinal loads; this can also be seen as ’’no change’’ in the
decrease significantly as the diameter increases. According mechanical properties. This is because the effect of the longitu-
to the data in Table 2, the height of the fiber disturbance dinal microchannel at a diameter of 600 lm on the fiber dis-
caused by a longitudinal microchannel with a diameter of turbance of CFRP is small, as also confirmed by the data in
400 lm is only 0.28 mm, which is slightly thicker than that Table 2. An overlap of the fiber disturbance zone is not
of a single layer of CFRP. The upper and lower 90
layers formed by the longitudinal microchannel at a diameter of
near the microchannel are affected. The basically unaffected 600 lm in the studied spacing range.
0
layers oriented perpendicular to the microchannel play With the results discussing above, the degradation rates
the main role of load bearing. Therefore, the tensile and for the tensile and compressive properties of CFRP with
compressive properties of CFRP with microchannels at a various microchannels were calculated, as shown in Table 6.
diameter of 400 lm are almost the same as that without The results reveal that the effect of the microvascular diam-
microchannels under transverse loads. The fiber fabric wavi- eter on the cross-sectional morphology of CFRP is greater
ness occurs in the main load-bearing zone of the 0
layers than that of spacing, and further reduces the mechanical
when the diameter is greater than 400 lm. The number of properties. Regarding the microvascular arrangement pat-
influenced layers increases with the gradually increasing tern, the mechanical properties of CFRP with longitudinal
diameter. In particular, the bending angle of the fiber in the microchannels deteriorate less. This is because the longitu-
main load-bearing zone is larger, leading to a significant dinal microchannel (oriented parallel to the adjacent fiber
decrease in the mechanical properties of CFRP. layers) can be more compatible with the laminate, resulting
in no resin-rich zone between the interlayers. The effect on
5.3.2. Effect of microvascular spacing the fiber structure is small, so the mechanical properties of
The microvascular spacings were respectively set to 2 mm, CFRP are significantly improved. With respect to the load
4 mm, 6 mm, and 8 mm at a diameter of 600 lm. According to direction, the CFRP has a better load-bearing capacity under
the data in Table 2, the total length of the fiber disturbance longitudinal loads than under transverse loads. The 90
fiber
formed by a longitudinal microchannel at a diameter of 600 lm layers oriented parallel to the microchannel play the main
is 1.74 mm, which is smaller than the minimum spacing of role of load bearing under the longitudinal loads. There is
2 mm. Therefore, there is no overlapping fiber disturbance zone no fiber fabric waviness for the 90
layers whose fibers
in CFRP in the range of the microvascular diameters studied. remain straight in the load direction, which can effectively
The curve of the fiber disturbance was established according to inhibit the degradation of mechanical properties of CFRP.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 307

Figure 11. Stress-strain curves of CFRP with longitudinal microchannels of various diameters under different loads.

Figure 12. Effect of diameters on tensile and compressive properties of CFRP with longitudinal microchannels.

5.4. Progressive damage analysis

analysis focused on the fiber fracture damage index of RVE
The laminate stacking sequences [90/0]3s and [0/90]3s were models with microchannels of various diameters at a spacing
adopted, and the main load-bearing components were the of 4 mm, including the fiber tensile damage and fiber com-
carbon fibers under the four loads studied. The damage pression damage.
308 Y. ZHAO ET AL.

Figure 13. Stress-strain curves of CFRP with longitudinal microchannels of various spacings under different loads.

Figure 14. Effect of spacing on the tensile and compressive properties of CFRP with longitudinal microchannels.

Figure 15 shows the fiber fracture damage index of RVE edges under a transverse tensile load, as shown in Figure
with a transverse microchannel at a diameter of 600 lm 15a. The fiber damage path propagates from the top of the
under ultimate loads. They show characteristics of the pro- microchannel along the direction of approximately 45
off
gressive damage under transverse loads. The fiber damage the vertical direction to the laminate edges, followed by
path propagates mainly from the upper and lower sides of propagating to both ends in every fiber layer of the RVE
the microchannel along the vertical direction to the laminate model under the transverse compression load, as shown in
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 309

Table 6. Effect of microvascular parameters on tensile and compressive properties of CFRP.

Microvascular diameter/lma Microvascular spacing/mmb
Degradation rate% Property 400 600 800 1000 2 4 6 8
Transverse Strength 11.9 20.0 27.9 35.8 25.6 20.0 13.6 9.3
tensile Modulus 9.2 17.1 26.9 34.4 26.5 17.0 13.7 11.6
Transverse Strength 10.5 17.3 22.3 28.6 26.0 17.3 11.2 6.1
compression Modulus 10.1 16.4 26.7 33.6 29.3 16.3 14.2 11.7
Transverse arrangement Longitudinal Strength 5.0 12.8 14.9 20.8 12.8 12.8 5.5 4.3
Horizontal arrangement tensile Modulus 5.0 12.9 15.0 21.1 13.0 12.9 5.6 4.4
Longitudinal Strength 4.9 12.9 14.6 20.4 12.9 12.8 5.5 4.2
compression Modulus 5.0 13.1 15.1 21.0 13.4 13.1 5.9 4.5
Transverse Strength 0.0 9.3 13.0 22.5 10.4 9.1 5.3 2.0
tensile Modulus 0.0 3.5 7.7 17.1 4.8 3.0 2.6 2.6
Longitudinal arrangement Transverse Strength 0.7 8.9 13.4 22.1 11.3 9.1 4.6 0.6
compression Modulus 0.3 3.8 9.0 18.2 6.9 4.1 3.2 2.9
Longitudinal Strength 1.9 2.4 4.3 1.4 4.7 2.4 1.5 1.2
tensile Modulus 2.0 2.4 4.4 1.5 5.0 2.4 1.7 1.2
Longitudinal Strength 1.8 2.4 4.3 2.2 4.7 2.4 1.5 1.0
compression Modulus 1.8 2.4 4.3 1.9 5.0 2.5 1.7 1.3
a
4 mm of the microvascular spacing.
b
600 lm of the microvascular diameter.

Figure 15. Fiber fracture damage index of RVE with a transverse microchannel at diameter of 600 lm under ultimate loads.

Figure 15b. The fiber damage to the main load-bearing zone the RVE, as shown in Figure 16a. When the diameter is
of the 0
layers occurs simultaneously as the loads approach larger than 400 lm, the fiber damage mainly propagates
the ultimate longitudinal loads, as shown in Figure 15c and d. from the upper and lower sides of the microchannel along
Figure 16 shows the fiber fracture damage index of RVE the vertical direction throughout the whole thickness. The
with a transverse microchannel at various diameters under a larger diameter often causes the more concentrated fiber
transverse tensile ultimate load. The fiber fabric waviness is damage, as shown in Figure 16b–d. In particular, Figure 16c
more significantly influenced by the diameter with the inclu- and d shows that delamination damage occurs between the
sion of a transverse microchannel than with a longitudinal two outermost layers in the upper and lower sides of CFRP.
one. Under the transverse tensile load, the influence of fiber Moreover, the debonding damage between the resin-rich
fabric waviness is small at a diameter of 400 lm, and the zone and fiber layer is more evident with the increase of the
damage propagates to both ends in every 90
fiber layer of microvascular diameter, as shown in Figure 16.
310 Y. ZHAO ET AL.

Figure 16. Fiber fracture damage index of RVE with a transverse microchannel with various diameters under a transverse tensile ultimate load.

Figure 17. Fiber fracture damage index of RVE with a longitudinal microchannel at a diameter of 600 lm under ultimate loads.

Figure 17 shows the fiber fracture damage index of RVE the cross-section is the weakest zone of the composites
with a longitudinal microchannel at a diameter of 600 lm owing to the existence of the microchannel, as shown in
under ultimate loads. Under transverse loads, the fiber dam- Figure 17a and b. The fiber damage has already initiated
age first initiates on the both sides of the microchannel, as before reaching the ultimate loads. Moreover, the fiber
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 311

Figure 18. Fiber fracture damage index of RVE with a longitudinal microchannel at various diameters under a transverse tensile ultimate load.

damage path progresses to the laminate surfaces in a direc- the vertical direction is the main damage pattern, as shown
tion approximately 30
off the vertical direction on both in Figure 18c. When the diameter increases to 1000 lm, the
sides of the microchannel, indicating the progressive pattern number of fiber layers affected increases, and the fiber dam-
of the fiber damage. Under the longitudinal loads, the load age paths progress along the upper and lower sides of the
path is parallel to the direction of the unaffected 90
fiber microchannel, as shown in Figure 18d. The fiber damage is
layers. The fiber damage in the main load-bearing zone also more concentrated on both sides of the microchannel
occurs almost simultaneously. The process of the fiber dam- as the diameter increases.
age is almost the same for CFRP with longitudinal micro-
channels as that with transverse one (Figure 15(c) and (d)),
6. Conclusion
as shown in Figure 17c and d. The uniformity of the fiber
damage significantly improves the load-bearing capacity of In this study, the 3D RVE model with a microchannel con-
CFRP under longitudinal loads. sidering in-plane and interlaminar damages are established
Figure 18 shows the fiber fracture damage indexes of based on measured parameters. The effects of different
RVE with a longitudinal microchannel at various diameters parameters on the tensile and compressive properties of the
under a transverse tensile ultimate load. As shown in Figure CFRP with microchannels are studied. The results from the
18a, the fiber damage is more uniform and the load-bearing numerical approach are close to the experimental results,
capacity of CFRP is correspondingly better than under a confirming the analyses regarding the strength, modulus,
larger diameter. This is attributed to the fact that the load- and damage behavior.
bearing fiber layers are not affected when the diameter of
the microchannel is 400 lm. When the microvascular diam- For the transverse microchannel, the mechanical proper-
eter increases to 600 lm, the fiber disturbance of the first ties of CFRP decrease as the diameter increases. The deg-
layer closest to the microchannel begins to affect the second radations in the strength and modulus of CFRP under
fiber fabric layer. The fiber damage paths progress to the longitudinal loads are less than those under transverse
surface of the laminate in a direction approximately 30
off loads. A decrease of strength and modulus is predicted
the vertical direction, as shown in Figure 18b. When the when the microvascular spacing is reduced. However,
diameter increases to 800 lm, the fiber damage first appears reducing the spacing from 4 mm to 2 mm has a negligible
at the upper and lower sides of the microchannel, which effect on the strength and modulus of CFRP under trans-
represent the sites of the maximum stress concentrations verse loads.
[35, 44]. The fiber damage paths progress throughout the For the longitudinal microchannel, the effects of the
thickness along two directions, i.e. the vertical direction and diameter on the mechanical properties of CFRP under
approximately 45
off the vertical direction. The path along longitudinal loads are small. The properties of CFRP
312 Y. ZHAO ET AL.

with microchannels at a diameter of 400 lm are the same [7] R.E. Newnham and G.R. Ruschau, Electromechanical properties
as that without microchannels under transverse loads. of smart materials, J Intel. Mat. Syst. Struct., vol. 4, no. 3, pp.
289–294, 1993. DOI: 10.1177/1045389X9300400301.
The strength and modulus of CFRP increase slowly with [8] Y. Shields, N.D. Belie, A. Jefferson, and K.V. Tittelboom, A
increasing spacing. The spacing has little effect on the review of vascular networks for self-healing applications, Smart
mechanical properties of CFRP with longitudinal micro- Mater. Struct., vol. 30, no. 6, pp. 63001–63022, 2021. DOI: 10.
channels at a diameter of 600 lm in the range of spac- 1088/1361-665X/abf41d.
[9] S.A. Mcdonald, S.B. Coban, N.R. Sottos, and P.J. Withers,
ings studied.
Tracking capsule activation and crack healing in a microcap-
The fiber damage shows characteristics of progressive sule-based self-healing polymer, Sci. Rep., vol. 9, no. 1, pp.
damage under transverse loads, whereas damage occurs 17773, 2019. DOI: 10.1038/s41598-019-54242-7.
simultaneously when the ultimate longitudinal load is [10] J.H. Liu, H.Z. Song, Z.H. Wang, J.X. Zhang, J. Zhang, and
almost reached. The debonding and delamination dam- X.W. Ba, Stretchable, self-healable, and reprocessable chemical
cross-linked ionogels electrolytes based on gelatin for flexible
ages become more evident with an increasing diameter supercapacitors, J. Mater. Sci., vol. 55, no. 11, pp. 1–14, 2020.
of the transverse microchannel. The fiber damage is DOI: 10.1007/s10853-019-04271-4.
more uniform when CFRP is subjected to longitudinal [11] W.R.K. Weihermann, M.M. Meier, and S.H. Pezzin,
loads than transverse loads. The more uniform damage Microencapsulated amino-functional polydimethylsiloxane as
can significantly improve the load-carrying capacity autonomous external self-healing agent for epoxy systems, J
Appl Polym Sci., vol. 136, no. 23, pp. 47627, 2019. DOI: 10.
of CFRP. 1002/app.47627.
[12] R. Luterbacher, T.S. Coope, R.S. Trask, and I.P. Bond, Vascular
The results are of application to the design and analysis self-healing within carbon fibre reinforced polymer stringer
of CFRP structures containing microchannels, recommend- run-out configurations, Compos. Sci. Technol., vol. 136, pp.
67–75, 2016. DOI: 10.1016/j.compscitech.2016.10.007.
ing proper microvascular parameters for further research on [13] S.K. Ghosh, Self-healing materials: fundamentals. Design
self-healing performance premised on ensuring the mechan- Strategies, and Applications, Weinheim, Wiley, 2009.
ical properties of the laminates. [14] K.S. Toohey, N.R. Sottos, J.A. Lewis, J.S. Moore, and S.R. White,
Self-healing materials with microvascular networks, Nat Mater.,
vol. 6, no. 8, pp. 581–585, 2007. DOI: 10.1038/nmat1934.
Funding [15] K.S. Toohey, C.J. Hansen, J.A. Lewis, S.R. White, and N.R.
Sottos, Delivery of two-part self-healing chemistry via micro-
This research is supported by the National Natural Science Foundation vascular networks, Adv. Funct. Mater., vol. 19, no. 9, pp.
of China (grant nos. 51905443 and 51875476), Fundamental Research 1399–1405, 2009. DOI: 10.1002/adfm.200801824.
Funds for the Central Universities (grant no. 31020200503003), [16] S.M. Bleay, C.B. Loader, V.J. Hawyes, L. Humberstone, and
Aviation Foundation (grant no. ASFC-201916053001). P.T. Curtis, A smart repair system for polymer matrix compo-
sites, Compos. A: Appl. Sci. Manuf., vol. 32, no. 12, pp.
1767–1776, 2001. DOI: 10.1016/S1359-835X(01)00020-3.
[17] J.W.C. Pang and I.P. Bond, Bleeding composites’—damage
References detection and self-repair using a biomimetic approach,
Compos. A: Appl. Sci. Manuf., vol. 36, no. 2, pp. 183–188,
[1] G. Yamamoto, K. Koizumi, T. Nakamura, N. Hirano, and T. 2005. DOI: 10.1016/j.compositesa.2004.06.016.
Okabe, Tensile-strength-controlling factors in unidirectional [18] J.W.C. Pang and I.P. Bond, A hollow fibre reinforced polymer
carbon fiber reinforced plastic composites, Compos. A: Appl. composite encompassing self-healing and enhanced damage
Sci. Manuf., vol. 140, pp. 106140, 2021. DOI: 10.1016/j.compo- visibility, Compos. Sci. Technol., vol. 65, no. 11-12, pp.
sitesa.2020.106140. 1791–1799, 2005. DOI: 10.1016/j.compscitech.2005.03.008.
[2] J. Hohe, A. Neubrand, S. Fliegener, C. Beckmann, M. Schober, K. [19] G. Williams, R. Trask, and I. Bond, A self-healing carbon fibre
P. Weiss and S. Appel, Performance of fiber reinforced materials reinforced polymer for aerospace applications, Compos. A:
under cryogenic conditions–a review, Compos. A: Appl. Sci. Appl. Sci. Manuf., vol. 38, no. 6, pp. 1525–1532, 2007. DOI: 10.
Manuf., vol. 141, pp. 106226, 2021. DOI: 10.1016/j.compositesa. 1016/j.compositesa.2007.01.013.
2020.106226. [20] S. Zainuddin, T. Arefin, A. Fahim, M.V. Hosur, J.D. Tyson, A.
[3] A. Ramji, Y.G. Xu, M. Grasso, M. Yasaee, and P. Webb, Effect Kumar, J. Trovillion and S. Jeelani, Recovery and improvement
of interfacial fibre orientation and PPS veil density on delamin- in low-velocity impact properties of e-glass/epoxy composites
ation resistance of 5HS woven CFRP laminates under mode II through novel self-healing technique, Compos. Struct., vol. 108,
no. 1, pp. 277–286, 2014. DOI: 10.1016/j.compstruct.2013.09.
loading, Compos. Sci. Technol., vol. 207, pp. 108735, 2021.
023.
DOI: 10.1016/j.compscitech.2021.108735.
[21] A.R. Hamilton, N.R. Sottos, and S.R. White, Self-healing of internal
[4] D.J. Bull, S.M. Spearing, and I. Sinclair, Image-enhanced model-
damage in synthetic vascular materials, Adv Mater., vol. 22, no. 45,
ling of residual compressive after impact strength in laminated
pp. 5159–5163, 2010. DOI: 10.1002/adma.201002561.
composites, Compos. Struct., vol. 192, pp. 20–27, 2018. DOI: [22] A.R. Hamilton, N.R. Sottos, and S.R. White, Pressurized vascu-
10.1016/j.compstruct.2018.02.047. lar systems for self-healing materials, J R Soc Interface., vol. 9,
[5] A.T.T. Nguyen and A.C. Orifici, Structural assessment of no. 70, pp. 1020–1028, 2012. DOI: 10.1098/rsif.2011.0508.
microvascular self-healing laminates using progressive damage [23] C.J. Norris, G.J. Meadway, M.J. O’Sullivan, I.P. Bond, and R.S.
finite element analysis, Compos. Part A: Appl. Sci. Manuf., vol. Trask, Self-healing fibre reinforced composites via a bioinspired
43, no. 11, pp. 1886–1894, 2012. DOI: 10.1016/j.compositesa. vasculature, Adv. Funct. Mater., vol. 21, no. 19, pp. 3624–3633,
2012.06.005. 2011. DOI: 10.1002/adfm.201101100.
[6] G.J. Williams, I.P. Bond, and R.S. Trask, Compression after [24] C.J. Norris, J.A.P. White, G. Mccombe, P. Chatterjee, I.P. Bond,
impact assessment of self-healing CFRP, Compos. A: Appl. Sci. and R.S. Trask, Autonomous stimulus triggered self-healing in
Manuf., vol. 40, no. 9, pp. 1399–1406, 2009. DOI: 10.1016/j. smart structural composites, Smart Mater. Struct., vol. 21, no. 9,
compositesa.2008.05.021. pp. 094027–094312, 2012. DOI: 10.1088/0964-1726/21/9/094027.
 MECHANICS OF ADVANCED MATERIALS AND STRUCTURES 313

[25] C.J. Norris, I.P. Bond, and R.S. Trask, Healing of low-velocity impact in glass fiber composites with microvascular channels under
damage in vascularised composites, Compos. Part A: Appl. Sci. tensile loading, Compos. Struct., vol. 192, pp. 101–114, 2018.
Manuf., vol. 44, pp. 78–85, 2013. DOI: 10.1016/j.compositesa.2012. DOI: 10.1016/j.compstruct.2018.02.061.
08.022. [36] Y. Wang, K. Li, X. Zhao, H. Tekinalp, T. Li, and S. Ozcan,
[26] Y.J. Wang, D.T. Pham, Z.C. Zhang, J.J Li, C.Q. Ji, Y.J. Liu and Toughening by nanodroplets: Polymer-Droplet biocomposite
J.S. Leng, Sustainable self-healing at ultra-low temperatures in with anomalous toughness, Macromolecules., vol. 53, no. 11,
structural composites incorporating hollow vessels and heating pp. 4568–4576, 2020. DOI: 10.1021/acs.macromol.9b02677.
elements, R Soc Open Sci., vol. 3, no. 9, pp. 160488, 2016. DOI: [37] C.J. Norris, I.P. Bond, and R.S. Trask, The role of embedded
10.1098/rsos.160488. bioinspired vasculature on damage formation in self-healing
[27] K.R. Hart, S.M. Lankford, I.A. Freund, J.F. Patrick, B.P. Krull, carbon fibre reinforced composites, Compos. Part A: Appl. Sci.
E.D. Wetzel, N.R. Sottos and S.R. White, Repeated healing of Manuf., vol. 42, no. 6, pp. 639–648, 2011. DOI: 10.1016/j.com-
delamination damage in vascular composites by pressurized positesa.2011.02.003.
delivery of reactive agents, Compos. Sci. Technol., vol. 151, no. [38] L.H. Qi, W.L. Tian, and J.M. Zhou, Numerical evaluation of
20, pp. 1–9, 2017. DOI: 10.1016/j.compscitech.2017.07.027. effective elastic properties of composites reinforced by spatially
[28] K.R. Hart, E.D. Wetzel, N.R. Sottos, and S.R. White, Self-heal- randomly distributed short fibers with certain aspect ratio,
ing of impact damage in fiber-reinforced composites, Compos. Compos. Struct., vol. 131, pp. 843–851, 2015. DOI: 10.1016/j.
B: Eng., vol. 173, pp. 106808, 2019. DOI: 10.1016/j.compositesb. compstruct.2015.06.045.
2019.05.019. [39] Z.Q. Zhao, H.Y. Dang, C. Zhang, G.J. Yun, and Y.L. Li, A
[29] M. W€ urkner, H. Berger, and U. Gabbert, Numerical investiga- multi-scale modeling framework for impact damage simulation
tions of effective properties of fiber reinforced composites with
of triaxially braided composites, Compos. A: Appl. Sci. Manuf.,
parallelogram arrangements and imperfect interface, Compos.
vol. 110, pp. 113–125, 2018. DOI: 10.1016/j.compositesa.2018.
Struct., vol. 116, no. 1, pp. 388–394, 2014. DOI: 10.1016/j.
04.020.
compstruct.2014.05.012.
[40] M. Hori and S. Nemat-Nasser, On two micromechanics theories
[30] S. Bargmann, B. Klusemann, J. Markmann, J.E. Schnabel, K.
for determining micro-macro relations in heterogeneous solids,
Schneider, C. Soyarslan and J. Wilmers, Generation of 3D repre-
Mech. Mater., vol. 31, no. 10, pp. 667–682, 1999. DOI: 10.1016/
sentative volume elements for heterogeneous materials: A review,
Prog. Mater. Sci., vol. 96, pp. 322–384, 2018. DOI: 10.1016/j. S0167-6636(99)00020-4.
pmatsci.2018.02.003. [41] Z.H. Xia, C.W. Zhou, Q.L. Yong, and X.W. Wang, On selection
[31] L. Wang, J.Y. Wu, C.Y. Chen, C.X. Zheng, B. Li, S.C. Joshi and of repeated unit cell model and application of unified periodic
K. Zhou Progressive failure analysis of 2D woven composites at boundary conditions in micromechanical analysis of compo-
the meso-micro scale. Compos. Struct., vol. 178, pp. 395–405, sites, Int. J. Solids. Struct., vol. 43, no. 2, pp. 266–278, 2006.
2017. DOI: 10.1016/j.compstruct.2017.07.023. DOI: 10.1016/j.ijsolstr.2005.03.055.
[32] P. Sharma and S.S. Mulay, Damage modeling of unidirectional [42] P. Suquet, Elements of homogenization for inelastic solid
laminated composites, Mech. Adv. Mater. Struc., vol. 26, no. 1, mechanics. Homogenization Techniques for Composite Media,
pp. 93–99, 2019. DOI: 10.1080/15376494.2018.1534173. Springer, Berlin, pp. 193–278, 1987.
[33] C.Y. Huang, R.S. Trask, and I.P. Bond, Characterization and [43] Z.H. Xia, Y.F. Zhang, and F. Ellyin, A unified periodical bound-
analysis of carbon fibre-reinforced polymer composite laminates ary conditions for representative volume elements of compo-
with embedded circular vasculature, J R Soc Interface., vol. 7, sites and applications, Int. J. Solids. Struct., vol. 40, no. 8, pp.
no. 49, pp. 1229–1241, 2010. DOI: 10.1098/rsif.2009.0534. 1907–1921, 2003. DOI: 10.1016/S0020-7683(03)00024-6.
[34] D.J. Hartl, G.J. Frank, and J.W. Baur, Effects of microchannels [44] H. Tanabi, A. Al-Shawk, and B. Sabuncuoglu, Stress concentrations
on the mechanical performance of multifunctional composite in composites with microvascular channels, XXVII International
laminates with unidirectional laminae, Compos. Struct., vol. Conference Mathematical and Computer Simulation in Mechanics
143, pp. 242–254, 2016. DOI: 10.1016/j.compstruct.2016.01.106. of Solids and Structures, Fundamentals of Static and Dynamic
[35] A. Al-Shawk, H. Tanabi, and B. Sabuncuoglu, Investigation of Fracture (MCM), vol. 6, pp. 56–63, Saint-Petersburg, 2017. DOI:
stress distributions in the resin rich region and failure behavior 10.1016/j.prostr.2017.11.009.