Sunteți pe pagina 1din 9

Engineering Failure Analysis 18 (2011) 2179–2187

Contents lists available at ScienceDirect

Engineering Failure Analysis


journal homepage: www.elsevier.com/locate/engfailanal

Analysis on failure mechanisms of an interlock woven fabric


under ballistic impact
C. Ha-Minh a,b, F. Boussu a,b, T. Kanit a, D. Crépin a,b, A. Imad a,⇑
a
Univ. Lille Nord de France, Cité Scientifique, 59655 Villeneuve d’Ascq Cedex, France
b
ENSAIT–GEMTEX, 2 allée Louise et Victor Champier, BP 30329, 59056 Roubaix Cedex, France

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, damage mechanisms of a 3D interlock woven fabric subjected to ballistic
Received 6 June 2011 impact were analyzed using a numerical model. Two impact configurations were carried
Received in revised form 9 July 2011 out in order to validate the numerical model with experimental observations: perforation
Accepted 13 July 2011
(900 m/s) and no-perforation (90 m/s). Global deformation of the fabric during impact is
Available online 23 July 2011
determined continuously to detail fabric impact behavior. Also, in this study, the effects
of boundary conditions on failure mechanisms have been investigated. Boundary condi-
Keywords:
tions are divided into two cases: (1) only warp yarns fixed and (2) only weft yarns fixed.
Failure mechanisms
3D interlock fabric
Basing on continuous evolutions of global deformation, projectile velocity, different ener-
Ballistic impact gies and reaction force onto projectile, the influence of both these fixation conditions is
Boundary conditions investigated.
Finite element analysis Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Textile unidirectional laminates and 2D woven fabrics have been widely used in ballistic protection. Nowadays, 3D inter-
lock woven fabrics can be assumed as new challenging material in this application field. Compared to 2D woven fabrics, 3D
interlock woven fabrics have several advantages such as [1]:

– 3D woven fabric exhibit higher through the thickness and interlaminar properties because of their integrated structure in
the presence of orthogonal and/or angle interlock construction.
– 3D multilayer woven fabrics demonstrate high ballistic impact damage resistance and low velocity impact damages
which have been a major problem with 2D laminate in military aircraft structures. The through the thickness yarn
improves impact damages tolerance by suppressing delaminating.
– High performance fibers such as carbon, ceramic, aramid and polyethylene can be woven into multilayer fabrics.
– Multilayer woven fabrics can be given additional strength by insertion in each layer of stuffing yarns, which remains
straight and contribute to their full strength in that direction. Yarns that interlace between layers as binding yarns con-
tribute partially to the strength of their direction in orthogonal woven fabrics and contribute to the strength in the thick-
ness direction.
– 3D woven fabrics can reduce molding time and facilitate migration to vacuum infusion. Moreover, the z-yarn acts as cap-
illary channels to transfer resin inside the preform from the outer surface.

⇑ Corresponding author. +33 3 28 76 73 93; fax: +33 3 28 76 73 61.


E-mail address: abdellatif.imad@polytech-lille.fr (A. Imad).

1350-6307/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2011.07.011
2180 C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187

– Although these structures are typically more expansive than 2D fabrics and mats, they reduce labor, have higher perfor-
mances and improve process efficiency result in overall cost saving in a variety of applications. When we compare the
cost per square meter of finished composite structure, 3D woven fabric reinforcements consistently outperform tradi-
tional 2D materials.

Therefore, 3D interlock woven fabrics are very potential materials not only for military applications, but also marine,
automotive, aerospace and transportation industries. It is a new material; therefore, studies on failure mechanisms of these
fabrics under different loads have not been numerous. Chen and Hodgkinson [2] present a detailed investigation of the im-
pact damages of different fiber reinforced composites. A testing program of low and high-velocity impact was carried out to
evaluate the damage resistance and tolerance of laminates, non-crimp fabric reinforced and three-dimensional (3D) woven
fabric reinforced composites. Provost et al. [3] have inserted iron yarns on the front and back surfaces of a warp interlock
fabric with the objective to increase the stiffness of this type of fabric. These fabrics were shot by a 6 mm diameter steel ball
at the speed range of 600–700 m/s. This work indicated that the deformation in pyramid form on the back of the target, in
the case of the projectile is stopped, is more important than in the case of the target entirely perforated. Their results also
showed that these fabrics with iron yarns have a smaller deformation on the back face regarding to the ones only made of
textile yarns. Li et al. [4] have performed a unit-cell approach to construct finite element model for a woven fabric composite
subjected to ballistic impact. Other authors [5–10] carried out experimental tests to investigate mechanical properties of 3D
composites. Sánchez Gálvez and Sánchez Paradela [11] present a summary of analytical and numerical models for ceramic/
metal and ceramic/composite add-on armour used in ballistic protection. Enfedaque et al. [12] used a specific experimental
setup named as ‘‘one-stage light gas gun’’ to investigate ballistic impact on a composite type made up of a cement mortar
matrix and chopped glass fibers. In their tests, projectile residual velocity can be determined both using a high speed camera
with multiframe exposure and measuring the projectile’s penetration depth in molding clay blocks.
In this paper, damage mechanisms of 3D interlock woven fabric subjected to ballistic impact are analyzed. In order to
analyze the perforation and no-perforation configurations, two impact velocity values have been carried out (90 m/s and
900 m/s). The effects of boundary conditions are also investigated at the impact velocity of 200 m/s using two cases: (1) warp
yarns fixed and (2) weft yarns fixed. Indeed, the continuous evolutions of global deformation, projectile velocity, contact
force and different energies have been studied. A comparison between numerical results and experimental data is performed
to validate this numerical simulation.

2. Calculation conditions

Fig. 1 presents the initial configuration on the impact system studied using shell elements [13–16]. The model is calcu-
lated using explicit schema with finite element code Radioss. The material used is a Kevlar KM2Ò 3D warp interlock woven
fabric of 52.8  52.8 mm (Fig. 1a and b). The weave diagram of this fabric is a 3-layer-through-the-thickness-warp-interlock
structure [13,16]. Eighty-four thousand eight hundred and fifty three shell elements are used to describe the structure of this
fabric. In the fabric plane, the distance between bundles in warp direction is 1.1 mm and 0.9 mm in weft direction (Fig. 1b).
All weft yarns are assumed straight, whereas, a half of warp yarns has a strong crimp to consolidate the structure of this
fabric. This undulation of warp yarns also helps to ameliorate mechanical properties in thickness of the 3D fabric regarding

Fig. 1. (a) The initial configuration of the ballistic impact system simulated in the present study; (b) detailed illustration of Kevlar KM2Ò 3-layer-through-
the-thickness-interlock woven fabric.
C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187 2181

to the 2D ones. Difference on undulation level between weft and warp yarns conducts to specific impact behavior of this 3D
fabric that is mentioned in next sections.
It is assumed that the contact point between the fabric and the projectile is the crossover point at the fabric center. The
projectile has a spherical form with a diameter of 5.43 mm and a mass of 6.31  10 4 kg. In the case of a dry thin fabric stud-
ied, projectile deformation after impact is negligible, thus, this spherical projectile is assumed infinitely rigid. Static and dy-
namic tension tests show that Kevlar yarns are elastic until failure [14–16]. Due to constancy of mechanical properties along
yarn, a yarn can be assumed orthotropic. The different Kevlar KM2Ò yarn orthotropic material values are: longitudinal
Young’s modulus = 84.6 GPa, transverse modulus = 1.34 GPa, shear modulus = 24.4 GPa, Poisson’s coefficient = 0.6 [14–16].
These yarns have a dynamic failure strain = 4.58%. To simplify, two frictions: (1) between individual bundles; (2) between
fabric and the projectile can be taken into account by unique Coulomb’s coefficients; l = 0.23 and l = 0.20 respectively.

3. Results and discussions

3.1. Fabric damage mechanisms during impact

Two impact velocities 90 m/s and 900 m/s have been carried out to study the damage mechanisms of the 3D fabric in two
corresponding impact cases: no-perforation and perforation. In general, the fabric has a large global deformation in pyramid
form. Indeed, when the projectile penetrates through the target, strain waves propagate far as fabric edges. Moreover, if
yarns are not failed, the propagation and reflections of strain waves exist during the whole impact event. It is noted that
the pyramid deformation of fabric is not equivalent for warp and weft directions. The dimension of this pyramid in the weft
direction is greater than in the warp direction. This can be explained by the difference on yarn crimping between both direc-
tions. The yarn undulation in the weft direction is lower than in the warp direction (Fig. 1b). Thus, the path for the propa-
gation of strain waves on the weft direction is shorter than on the warp direction. This helps strain waves to go following the
weft direction more rapidly than the warp direction.
Fig. 2a and b shows a good agreement between the numerical model and the experimental result in the case of no-per-
foration. Fig. 2c details the fabric damage at the impact point. With 90 m/s, a slow impact velocity, bundles rupture does not
appear at this zone. Fabric damages is characterized by the sideway/lateral movement of individual yarns along the border of

Global deformation in pyramid


W
ef
td
ire
ct
io
n

on
(a) directi (b)
Warp

(c)
Weft direction

Warp direction

Fig. 2. (a) Configuration of the four-fixed-edge 3D fabric at 58.5 ls at 90 m/s impact; (b) experimental result of an identical 3D fabric [3]; (c) configuration
at the impact point in (a).
2182 C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187

the projectile head (Fig. 2c). We can see that the fabric has a large global deformation in the case of no-perforation (Fig. 2a
and b), but damages are apparent only at the contact zone between the fabric and projectile. Fig. 3 shows the evolution of
pyramid dimensions in weft and warp directions for the impact case of 90 m/s. With the low impact velocity, the pyramid
only appears after 10 ls. Just after appearance, the pyramid extends very strongly and then keeps a high expansion velocity
up to the edge (at 24 mm) in weft direction. However, in warp direction, the propagation velocity is absolutely slower and
it cannot reach the edge. Therefore, this figure confirms again higher propagation velocity of transverse strain wave in the
weft direction than the warp one.
Fig. 4a and b compares the configuration of the fabric in the perforation case between experimental observation and
numerical calculation. Both experimental and computational results state a very small global deformation in pyramid form
in this case. Fig. 5 shows that the pyramid develops very early at 1 ls during a short time 35 ls for the impact velocity of
900 m/s. Yarn failure causes strange vertical intervals on the curve ‘‘weft direction’’ in this figure because pyramid dimen-
sions are measured at nodes of weft yarns. Similar to the no perforation case, expansion velocity of the pyramid in the weft
direction is more rapid than the warp one. Thus, the pyramid always has a greater dimension in the weft direction regarding
to the warp direction. In fact, with a high impact velocity of 900 m/s, yarns are failed very early and just after 4 ls, no yarn
loads the projectile any more. Thus, transverse strain wave only has a very little time to propagate. Moreover, the impact
event is so fast; hence, yarns do not also have a lot time for sideway/lateral movement of on the surface of the projectile
head. Therefore, their successive fracture takes place rapidly and it is localized only at the contact zone between fabric
and projectile (Fig. 4). The pyramid is also enlarged with a very high velocity when it appears. However, both its dimensions
in weft and warp directions are not able to reach over 13 mm regarding to 24 mm (position of the fabric edge) (Fig. 5).
Delaminating is an impact phenomenon often observed for 2D plain-woven fabrics. However, in this study, this phenom-
enon does not exist because of the reinforcement of warp yarns in thickness direction (Figs. 2 and 4). This result can confirm
again the research works in the literature [1].
Fig. 6 shows the calculation times of both no-perforation and perforation cases. In the perforation case, the 900 m/s im-
pact event takes place very rapidly only within 35 ls and the associated CPU time is 1.05 h. However, with the same explicit
schema, the no-perforation case, CPU time is very large, 8.43 h. The reason is that the projectile penetration with initial
velocity of 90 m/s is slow (87 ls), while time step in explicit schema is always low to ensure the calculation precision.

3.2. Effect of boundary conditions

Figs. 7 and 8 show configurations of 3D fabric at 49.8 ls with an impact of 200 m/s in two cases: (1) warp bundles fixed at
two edges and (2) weft bundles fixed at two edges. Yarns do not fail when they are free at the two ends. These yarns slip on
other yarns and they are pulled out of the fabric by the projectile. This effect caused damage zones at two corresponding free
edges (Figs. 7b and 8b). However, yarns fixed at the two ends are ruptured due to tension and friction with the projectile. The
failure of these yarns takes place only at the impact point. The deformation pyramid in these cases seems to be developed
only in the direction of fixed yarns.
Fig. 9 shows the evolution of pyramid size in the warp direction in two cases: warp yarns fixed and weft yarns fixed. From
5 to 12 ls, both curves are the same, but, then, fabric transverse deformation is more strongly developed when warp yarns
are fixed. In fact, when warp yarns are free, they can be easily pulled out by the projectile. This mechanism restrains signif-
icantly development of the pyramid deformation in warp direction.
We can observe this phenomenon for the pyramid development in weft direction (Fig. 10). When weft yarns are fixed, the
propagation velocity of the pyramid is more rapid in this direction. Comparing Figs. 9 and 10 on curves of the same line type,

Fig. 3. Evolution of the global deformation of the fabric in the impact case of 90 m/s.
C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187 2183

Fig. 4. (a) Configuration of the four-fixed-edges 3D fabric after perforation at 900 m/s impact; (b) experimental result of an identical 3D fabric in the case of
perforation [3]; (c) configuration at the impact point in (a).

Fig. 5. Evolution of the global deformation of the fabric in the impact case of 900 m/s.
2184 C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187

Fig. 6. CPU time of two cases: perforation and no perforation.

(a) (b)
Zoom

Fixed
edge

edge
Fixed

Pyramid

Fig. 7. Configurations of 3D fabric at 49.8 ls of 200 m/s impact in the case of only warp yarns fixed: (a) global view; (b) damage at free edges.

Fixed edge

Zoom
Pyramid

Fixed edge
(a) (b)
Fig. 8. Configurations of 3D fabric at 49.8 ls of 200 m/s impact in the case of only weft yarns fixed: (a) global view; (b) damage at free edges.

we can note that the pyramid has always a larger dimension in the where yarns are fixed. This is also visible in Figs. 7 and 8.
Therefore, it can be though that the effect of boundary conditions is more important than yarn undulation on the global
behavior of this 3D fabric.
Fig. 11 illustrates the evolution of projectile velocity versus time in both fixation cases. The difference between two on
global deformation mechanism due to impact between two cases of boundary conditions is reflected obviously in this figure.
The projectile velocity decreases more strongly with fixed weft yarns than fixed warp yarns. Indeed, the density of weft yarns
is more compact regarding to warp yarns. When weft yarns are fixed, the pyramid deformation extends essentially on these
yarns, not warp ones. Therefore, the quantity of deformed yarns material for absorbing impact energy is more important
than the case of fixed warp yarns. Figs. 12 and 13 also show that contact and strain energies of yarns in 3D fabric are higher
when weft yarns are fixed. This confirms again that the development of pyramid in weft direction increases more yarns con-
tributing to stop projectile in comparison with the one in warp direction.
C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187 2185

Fig. 9. Evolution of the pyramid dimension in warp direction in the case of 200 m/s impact with two different boundary conditions.

Fig. 10. Evolution of the pyramid dimension in weft direction in the case of 200 m/s impact with two different boundary conditions.

Fig. 11. Evolution of the projectile velocity in the case of 200 m/s impact with two different boundary conditions.

Fig. 14 shows the evolution of reaction force on projectile versus time in both cases of boundary conditions. From the
impact moment (1 ls) to 9 ls, both curves are the same. At this period, yarns have not yet been tense, only a certain mass
of the fabric at the impact zone loads projectile. Therefore, difference on reaction force between two fixation cases does not
exist. However, from 10 ls to 20 ls, this force increases greatly for the case of fixed weft yarns, whereas it varies slightly
2186 C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187

Fig. 12. Evolution of the contact energy in the case of 200 m/s impact with two different boundary conditions.

Fig. 13. Evolution of the strain energy on the projectile in the case of 200 m/s impact with two different boundary conditions.

Fig. 14. Evolution of the reaction force on the projectile in the case of 200 m/s impact with two different boundary conditions.

for the case of fixed warp yarns. In fact, at this moment, weft yarns fixed suffers a certain tension that creates a strong reac-
tion force on the projectile. This tension is getting higher when the projectile penetrates through the fabric. It is noted that
unlike weft yarns (no crimp or no undulation), the undulation of warp yarns is significant. Therefore, these bundles cannot be
C. Ha-Minh et al. / Engineering Failure Analysis 18 (2011) 2179–2187 2187

Fig. 15. CPU time in the case of 200 m/s impact with two different boundary conditions.

tense immediately after impact of projectile although they are fixed at two edges. It means that warp yarns always have a de-
crimping process before tension period. Moreover, as mentioned in paragraphs above, weft yarns density is more numerous
than warp yarns. It is another important reason to explain why the reaction force in the case of weft yarns fixed is always so
higher than in the case of warp yarns fixed.
Fig. 15 illustrates the computation times in the case of 200 m/s impact with two different boundary conditions. In this
case, impact time, 60 ls is enough to compare on all aspects of impact phenomena between two these different boundary
conditions. With the same calculation scheme and impact velocity, CPU time in the case of fixed weft yarns is greater than
when warp yarns are fixed. The reason is that in the case of fixed weft yarns, projectile velocity is decreased more strongly.
Therefore, impact event in this case is slower in comparison with the case of fixed warp yarns.

4. Conclusion

A numerical model was performed to describe ballistic impact phenomena of a warp interlock fabric. The damage mech-
anisms of this fabric are analyzed and compared with experimental observations. A good agreement with experimental re-
sult on main damage phases of this fabric showed the validation of the numerical model. The impact behavior of this fabric is
characterized by pyramid formation with the apex being impact point. Dimensions of this pyramid depend on yarn crimping
in warp and weft directions. The fabric damage is dominated by the sideway/lateral movement of individual yarns along the
border of the projectile head and by their successive fracture. Delaminating does not exist due to the reinforcement of warp
yarns in thickness direction. This paper also indicates the effect of boundary conditions on fabric damage zones. Indeed,
yarns are failed only at the impact point if their two ends are fixed. The free edges of the fabric are damaged by the pull-
out phenomenon of yarns. Results also indicate that boundary conditions influence significantly damage mechanisms and
ballistic performance of the 3D fabric. CPU times of all calculations are presented. With the same explicit schema, these cases
where impact event goes on slowly, CPU time is great.

References

[1] Hu J. 3-D fibrous assemblies: Properties, applications and modeling of three-dimensional textile structures. Abington Hall, Abington, Cambridge, CB21
6AH, UK: Woodhead Publishing Limited; 2008.
[2] Chen F, Hodgkinson JM. Impact behaviour of composites with different fibre architecture. Proc Inst Mech Eng Part G: J Aerospace Eng
2009;223:1009–17.
[3] Provost B, Boussu F, Nussbaum J, Lefebvre M. Use of new warp interlock structures against high velocity impact. In: personal armour systems
symposium, Canada, September, 2010.
[4] Li Z, Sun B, Gu B. FEM simulation of 3D angle-interlock woven composite under ballistic impact from unit cell approach. Comput Mater Sci
2010;49:171–83.
[5] Gu H, Zhili Z. Tensile behaviour of 3D woven composites by using different fabric structures. Mater Des 2002;23:671–4.
[6] Wang Y, Zhao D. Effect of fabric structures on the mechanical properties of 3-D textile composites. J Ind Text 2006;35:239–56.
[7] Hosur MV, Adya M, Vaidya UK, Mayer A, Jeelani S. Effect of stitching and weave architecture on the high strain rate compression response of affordable
woven carbon/epoxy composites. Compos Struct 2003;59:507–23.
[8] Sun BZ, Gu BH, Ding X. Compressive behavior of 3D angle-interlock woven fabric composites at various strain rates. Polym Test 2005;24:447–54.
[9] Sun BZ, Yang L, Gu BH. Strain rate effect on four-step three dimensional braided composite compressive behaviour. AIAA J 2005;43:994–9.
[10] Sun BZ, Liu F, Gu BZ. Influence of strain rate on the uniaxial tensile behavior of 4-step 3D braided composites. Composites, Part A 2005;36:1477–85.
[11] Sánchez Gálvez V, Sánchez Paradela L. Analysis of failure of add-on armour for vehicle protection against ballistic impact. Eng Fail Anal
2009;16(6):1837–45.
[12] Enfedaque A, Cendón D, Gálvez F, Sánchez-Gálvez V. Failure and impact behavior of facade panels made of glass fiber reinforced cement (GRC).
Engineering Failure Analysis, in press. doi: 10.1016j.engfailanal.2011.01.004.
[13] Ha-Minh C, Boussu F, Thoral-Pierre K. In: Ktex_pattern: numerical tool for textile fabrics subjected to ballistic impact, Texcomp10, Lille, October, 2010.
[14] Ha-Minh C, Boussu F, Kanit T, Crépin D, Imad A. Effects of the transverse mechanical properties of bundles on the ballistic impact onto textile fabric:
numerical modeling. In: 19th DYMAT technical meeting, Strasbourg, France, December, 2010.
[15] Ha-Minh C, Kanit T, Boussu F, Imad A. Numerical multi-scale modeling for textile woven fabric against ballistic impact. Comput Mater Sci
2011;50(7):2172–84.
[16] Ha-Minh C, Kanit T, Boussu F, Crépin D, Imad A. Effect of yarns friction on the ballistic performance of a 3D warp interlock fabric: numerical analysis.
Appl Compos Mater 2011. doi:10.1007/s10443-011-9202-2.

S-ar putea să vă placă și