Finite Element Simulation and Experimental Study On Mechanical Behavior of 3D Woven Glass Ber Composite Sandwich Panels
Finite Element Simulation and Experimental Study On Mechanical Behavior of 3D Woven Glass Ber Composite Sandwich Panels
Finite Element Simulation and Experimental Study On Mechanical Behavior of 3D Woven Glass Ber Composite Sandwich Panels
Composites: Part B
journal homepage: www.elsevier.com/locate/compositesb
a r t i c l e
i n f o
Article history:
Received 26 September 2012
Received in revised form 18 April 2013
Accepted 12 June 2013
Available online 25 June 2013
Keywords:
A. Polymermatrix composites (PMCs)
A. Fabrics/textiles
C. Finite element analysis (FEA)
D. Mechanical testing
Composite sandwich
a b s t r a c t
The results of nite element simulation followed by an experimental study are presented in order to
investigate the mechanical behavior of three-dimensional woven glass-ber sandwich composites using
FE method. Experimental loaddisplacement curves were obtained for atwise compressive, edgewise
compressive, shear, three-point bending and four-point bending loads on the specimens with three different core thicknesses in two principal directions of the sandwich panels, called warp and weft. A 3D
nite element model is employed consisting of glass fabric and surrounding epoxy resin matrix in order
to predict the mechanical behavior of such complex structures. Comparison between the nite element
predictions and experimental data showed good agreement which implies that the FE simulation can be
used instead of time-consuming experimental procedures to study the effect of different parameters on
mechanical properties of the 3D woven sandwich composites.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
Sandwich structures are increasingly used in structural applications where materials with high mechanical strength and stiffness
at low specic weight are required. Thermal and acoustic insulation, high energy absorption and low cost are some other benets
of sandwich materials which make them more suitable than isotropic materials for many advanced constructions in marine, aerospace, automotive and building industries. Despite all mentioned
benets, the connection between the skin and core is one of the
major weak spots in traditional sandwich materials which could
cause damage initiation during impact, shear and bending loads
[1,2]. Although some techniques have been used to improve the
skincore debonding problem [3,4], each method is likely to enforce some restrictions and extra cost during the production process. High skincore debonding resistance of 3D woven sandwich
composites established a new era in the eld of sandwich structures which is very benecial in increasing their lifetime and damage tolerance. This kind of sandwich composite is a fabric woven
out of a glass yarn and consists of two parallel deck-layers bonded
together by vertical threads, called piles. These piles are woven
into deck-layers thus forming an integral sandwich structure.
According to the core geometry, two principal material directions
could be considered; warp and weft which both are shown in
Fig. 1. Velvet weaving technique is the method which is used to
produce these fabrics. The idea of using velvet weaving techniques
Corresponding author. Tel.: +98 21 66405844; fax: +98 21 66419736.
E-mail address: mojtaba@aut.ac.ir (M. Sadighi).
1359-8368/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compositesb.2013.06.030
159
Fig. 1. Cross section of 3D woven sandwich composites: (a) warp view; and (b) weft view.
2. Mechanical tests
All of the mechanical tests in the current study were performed
by a Zwick testing machine with axial actuators. The jack has a static capacity of 250 kN, with a maximum stroke of 300 mm and a
linear variable differential transformer (LVDT) mounted on it. A
constant movement rate of 0.5 mm/min of the movable head of
the testing machine was applied through all types of mechanical
tests.
The specimens made of a Parabeam 3D woven E-glass fabric
and impregnated with DERAKANE 411-45 epoxy vinyl ester resin
using hand lay-up process. The resin content of the nal panels
was 55% in weight. Specimens with thicknesses of 8 mm, 10 mm
and 12 mm were used for each type of tests. Area density of the
panels were 1775 g/m2, 2583 g/m2 and 2866 g/m2 for 8 mm,
10 mm and 12 mm panels, respectively. The thickness of panels
facesheets were about 0.60 mm and average distance between
neighboring piles were 6.02 mm in warp and 3.73 mm in weft
directions. The piles had an average diameter of 0.70 mm and
made an average angle of 75 to the facesheets and had an average
of 86% degrees of stretching. Each of the presented data is obtained
by at least ve specimens in order to get maximum accuracy and to
avoid probable errors.
160
Fig. 6. Experimental stressstrain curves for panels with three different thicknesses
under atwise compression.
Fig. 5. Specimen under bending loading: (a) three-point bending; and (b) four-point bending.
161
Fig. 7. Experimental shear stressstrain curves for the 3D-fabric sandwich composites of three different thicknesses: (a) warp direction; and (b) weft direction.
8.7 mm thickness and 2046 g/m2 area density, with a distance between piles of 4.2 mm in warp direction. Li et al. [8] presented a
atwise compressive strength of 2 MPa, 2.8 MPa and 3.4 MPa for
panels of 10 mm thickness with three different pile angles, in
which the distance between piles were 4.8 mm in warp direction
and 4.1 mm in weft direction. Fan et al. [10] reported an average
atwise compressive strength of 4.4 MPa for panel with 10.5 mm
thickness and 3000 g/m2 area density, in which distance between
piles were 6.7 mm in warp direction and piles had a thickness of
3 mm. Different yield strength of panels with similar thickness reveals high effect of structure parameters such as piles tilt angle and
stretching ratio, distance between neighboring piles, and resin
ratio.
2.2.2. Shear test results
The experimental stressstrain curves for the 3D-fabric sandwich composites of three different thicknesses, under shear loads
are depicted in Fig. 7. Overlay, the load increases linearly until
the resin yield starts at the intersections of the piles with the facesheets and the piles begin to incline toward the direction of shear
and the load reaches a plateau due to resin piles plastic deformation. This process continues until the piles become straightened
and glass strings inside the piles start to resist against tensile load,
causing the load to increase with a sharp rate.
It can be deduced from the experiments that the panels with
lower thicknesses have higher yield stress and shear modulus. In
addition, the warp direction shows higher yield stress and shear
modulus respect to the weft direction. Li et al. [8] reported higher
yield stress and shear modulus in weft direction compared to warp
direction. van Vuure et al. [7] implemented an investigation on the
effect of piles stretching degree on the shear behavior of 3D woven
composite structures. Their experiments revealed that warp direction shows higher shear performance than the weft direction in
Fig. 8. Loaddeection diagrams based on three-point bending test: (a) warp direction; and (b) weft direction.
162
Fig. 9. Loaddeection diagrams based on four-point bending test: (a) warp direction; and (b) weft direction.
Fig. 10. Bending stiffness of 3D woven composite panels with different thicknesses,
based on experimental results.
Fig. 11. Experimental loaddisplacement curves for edgewise compression of the 3D-fabric sandwich composites of three different thicknesses: (a) warp direction; and (b)
weft direction.
163
details, including the pile shape and the distance between the piles
in the core, the piles diameter, the distance between the bers inside the facesheets and the thickness of the facesheets are considered as were described in Section 2. The modeled panels were
used as an input to the nite element package, ABAQUS.
Fig. 12 depicts the fabric embedded in resin, modeled for atwise compression FE simulation. The embedded region technique
makes it possible to embed glass fabric within resin matrix, since
there is no distinct boundary between the regions. Dimensions of
the modeled specimen for each test simulation were equal to the
same test specimen.
3.1.1. Glass fabric
A circular cross-section of 0.47 mm diameter with Youngs
modulus of 72 GPa and Poisson ratio of 0.27 was used for glass
strings. Considering the fact that glass strings do not bear bending
moments, as was discussed before, T3D3 element type which is a
3-node 3D truss with three translational degrees of freedom per
node was chosen for woven glass fabric which is appropriate for
the curved geometrical shape of the bers inside the core. van
Vuure et al. [19] used beam elements for the piles and implemented linear elastic analysis to predict the core shear modulus
of 10 mm panels. Their simulation showed high errors in the weft
direction. The presented data are before the fracture of the glass bers because as was explained, the panels failures start due to the
fracture of the resin.
3.1.2. Resin
A C3D10 element type, which is a 10-node quadratic tetrahedron with three translational degrees of freedom per node, was
used for resin. This element is suitable for the complex shape of
the piles inside the core, because it has a high order shape function
due to usage of intermediate nodes on each edge of the element.
Fig. 13. Deformed shape of 12 mm panels under atwise compressive loading: (a) nite element; and (b) experiment.
164
Fig. 15. Predicted and measured atwise compressive properties of panels with different thicknesses: (a) compressive strength; and (b) compressive modulus.
Fig. 16. Predicted and real deformed shape of 12 mm panel under shear loading: (a) warp; and (b) weft.
the displacement was applied through the reference node of the rigid plate. Fig. 13 depicts the deformed shape of the specimens with
12 mm thickness under atwise compressive loading, predicted by
nite element simulation and observed through mechanical tests.
3.2.2. Shear
In order to get the maximum resemblance to the test conditions, two rigid surfaces were modeled and their nodes were tied
to the surface of the upper and lower facesheets. The upper skin
displacement was applied through the reference node of the upper
rigid surface, while the lower rigid surface was only free to rotate
along its edge. The predicted and real deformed shape of 12 mm
Fig. 18. Predicted and experimental shear properties for panels with three different thicknesses: (a) shear strength; and (b) shear modulus.
165
Fig. 19. Predicted and experimental deformed shape of 12 mm panels: (a) three point bending; and (b) four point bending.
Fig. 20. Predicted and experimental bending loaddeection curves of 12 mm panels in warp and weft directions: (a) three point bending; and (b) four point bending.
Fig. 21. Predicted and measured bending stiffness of panels with three different
thicknesses in warp and weft directions.
panel under shear loading are shown in Fig. 16. The matrix failure
starts at intersections of the piles with the facesheets. Fig. 17 depicts the experimental and nite element shear stressstrain
curves for 12 mm panel in warp and weft directions. The nite element predictions are in good agreement with the experimental
data as is compared in Fig. 18 and it is possible to predict the accurate values of shear properties for panels with more than 8 mm
thickness.
3.2.3. Bending
The load in the nite element simulations was applied through
rigid cylinders with the same diameters as the test apparatus. The
specimens were free to skid on lower cylinders. The predicted and
experimental deformed shape of 12 mm panels under three point
bending and four point bending load are shown in Fig. 19. The failure usually starts due to shear failure of the panels, which starts at
the intersection of the piles with the facesheets.
The predicted and experimental loaddeection curves of
12 mm panels in warp and weft directions are depicted in
Fig. 20a and b for three point bending and four point bending,
respectively. Based on the experimental and FE results for panels
with 8 mm, 10 mm and 12 mm thicknesses, the predicted and
measured bending stiffness of panels in warp and weft direction
are shown in Fig. 21. There is a satisfactory correspondence between the experimental and FE results, especially for the panels
with higher thickness which is due to the reasons mentioned
before.
3.2.4. Edgewise compression
The translational degrees of freedom of rigid plate adjacent to
one of model ends was set to zero, while the rigid plate through
which the load was applied, was free to translate and rotate. The
simulated and experimental fracture mode and stressstrain
Fig. 22. Simulated and experimental fracture mode of 12 mm panel under edgewise compression: (a) warp; and (b) weft.
166
accomplished according to related ASTM test standards. Finite element results are in good agreement with the experimental data,
especially in the laminates with higher thicknesses in which the
modeled resin distribution are more similar to the real resin distribution shape.
References
Fig. 23. Experimental and nite element stressstrain curves for 12 mm panel
under edgewise compression.
Fig. 24. Predicted and measured facesheets maximum stress under edgewise load
for three different panels in warp and weft directions.
curves of panels with 12 mm thickness under edgewise compressive load are depicted in Figs. 22 and 23, respectively. Fig. 24 compares the facesheets stress at buckling load. The experimental
maximum load is lower than the predicted critical load for the panel with 8 mm thickness in both warp and weft directions. Little
eccentric loads, which are unavoidable, may cause the panels to
buckle under predicted loads. Another reason is that the specimens
may not be perfectly straight because they are made through a
hand layup process, so buckling under loads lower than the theoretical failure load is probable.
4. Conclusion
The present paper concerns the prediction of the mechanical
behavior of 3D woven glass ber sandwich composites under different mechanical loads using nite element simulations. For this
purpose, a 3D nite element model is constructed where the glass
fabric is considered as a material which is not able to carry bending
moments, while the surrounding resin behaves as a homogenous
solid. In order to evaluate the nite element predictions,
mechanical tests including atwise compression, shear, three point
bending, four point bending and edgewise compression were
[1] Vaidya UK, Nelson S, Sinn B, Mathew B. Processing and high strain rate impact
response of multi-functional sandwich composites. Compos Struct 2001;52(3
4):42940.
[2] Zenkert D, Falk L. Interface debondings in foam core sandwich beams and
panels. In: Proceedings of ICCM-8 Conference Honolulu; July 1991. p. 3H1
3H10.
[3] Potluri P, Kusak E, Reddy TY. Novel stitch-bonded sandwich composite
structures. Compos Struct 2003;59(2):2519.
[4] Cartie DD, Fleck NA. The effect of pin reinforcement upon the throughthickness compressive strength of foam-cored sandwich panels. Compos Sci
Technol 2003;63(16):24019.
[5] Verpoest I, Wevers M. 3D-fabrics for compression and impact resistant
composite sandwich structure. In: Proceedings of 35th international SAMPE
symposium. Anaheim, May; 1990. p. 296307.
[6] van Vuure AW, Ivens JA, Verpoest I. Sandwich-fabric panels. In: Proceedings of
40th international SAMPE symposium. Anaheim; May 1995. p. 128192.
[7] van Vuure AW, Ivens JA, Verpoest I. Mechanical properties of composite panels
based on woven sandwich-fabric preforms. Compos A 2000;31:67180.
[8] Li M, Wang S, Zhang Z, Wu B. Effect of structure on the mechanical behaviors of
three-dimensional spacer fabric composites. Appl Compos Mater
2009;16:114.
[9] Fan H, Zhou Q, Yang W, Jingjing Z. An experiment study on the failure
mechanisms of woven textile sandwich panels under quasi-static loading.
Compos B 2010;41:68692.
[10] Fan H, Zhao L, Chen H, Kuang N, Yang Ch, Huang Sh, et al. Ductile deformation
mechanisms and designing instruction for integrated woven textile sandwich
composites. Compos Sci Technol 2012;72:133843.
[11] Zheng J, Zhao L, Fan H. Energy absorption mechanisms of hierarchical woven
lattice composites. Compos B 2012;43:151622.
[12] Fan H, Yang W, Zhou Q. Experimental research of compressive response of
multi-layered woven textile sandwich panels under quasi-static loading.
Compos B 2011;42:11516.
[13] Bannister MK, Braemar R, Crothers PJ. The mechanical performance of 3D
woven sandwich composites. Compos Struct 1999;47:68790.
[14] Van Vuure AW, Verpoest I, Ko FK. Sandwich-fabric panels as spacers in a
constrained layer structural damping application. Compos B 2001;32:119.
[15] Judawisastra H, Ivens J, Verpoest I. Fatigue behavior and damage development
of 3d woven sandwich composites. Compos Struct 1998;43(1):3545.
[16] Vaidya AS, Vaidya UK, Uddin N. Impact response of three-dimensional
multifunctional sandwich composite. Mater Sci Eng A 2008;472:528.
[17] Hazizan AM, Cantwell WJ. The low velocity impact response of foam based
sandwich structures. Compos B 2002;33:193204.
[18] Corigliano A, Rizzi E, Papa E. Experimental characterization and numerical
simulations of a syntactic-foam/glass-ber composite sandwich. Compos Sci
Technol 2000;60:216980.
[19] Van Vuure AW, Pug J, Ivens JA, Verpoest I. Modeling the core properties of
composite panels based on woven sandwich-fabric preforms. Compos Sci
Technol 2000;60:126376.
[20] Anon. Standard test method for atwise compressive properties of sandwich
cores. Annual book of ASTM Vol. 15.03, space simulation; aerospace and
aircraft; high modulus bers and composites, (ASTM C 365-00). New
York: ASTM; 2000.
[21] Anon. Standard test method for edgewise compressive strength of sandwich
constructions. Annual book of ASTM Vol. 15.03, space simulation; aerospace
and aircraft; high modulus bers and composites, (ASTM C 364-99). New
York: ASTM; 2000.
[22] Anon. Standard test method for shear properties of sandwich core materials.
Annual book of ASTM Vol. 15.03, space simulation; aerospace and aircraft;
high modulus bers and composites, (ASTM C 273-00). New York: ASTM;
2000.
[23] Anon. Standard test method for exural properties of sandwich constructions.
Annual book of ASTM Vol. 15.03, space simulation; aerospace and aircraft;
high modulus bers and composites, (ASTM C 393-00). New York: ASTM;
2000.