You are on page 1of 12

Mechanical Characterization and Finite Element Modeling of

Composites Sandwich Structures

M. Aiman Iqbal, Zaffar M Khan, M. Bilal Khan


School of Chemical and Materials Engineering
National University of Sciences and Technology, Pakistan
zaffark@yahoo.com*

Abstract

The need for light weight materials with high flexural stiffness for use in aircraft and
aerospace vehicles has led to the development of sandwich structures. Sandwich structures
are preferred over conventional materials in various aerospace applications because of its
high flexural resistance, stiffness and cost effectiveness. In sandwich structure, the face sheets
carry the in-plane and bending loads whereas the core is responsible for transmitting shear
forces between the face sheets.

The aim of the present work is to extend the knowledge of mechanical properties both on
laminates and sandwich structures, focusing on the effects induced by different kind of skin
arrangements (carbon and glass fibers). Flexural test was performed for a complete static
mechanical characterization of the sandwich structure according to ASTM standards. The
mechanical results were used as input data in order to implement the FEM analysis by the
commercial ANSYS code. A simplified model is proposed to simulate the flexural test of
laminates and sandwich structures. The experimental and computational results are correlated
for verification of the model.

1. Introduction

The quest for reduced weight and superior mechanical performances, in the last few years,
has led to progressive development of sandwich structures [1]. Sandwich construction is one
of the composite forms to have achieved wide acceptance in aeronautical, automobile, infra-
structural and other engineering industries. Extensive usage of sandwich structure is
employed by the state of art automobile, space and marine vehicles. Sandwich composites are
used not only in the field of aeronautical constructions, but also in the fields of land
transports and marine constructions. Sandwich structures typically consist of two thin face
sheets or skins and a light weight thicker core. Commonly used materials for face sheets are
composite laminates and metals, while cores are made of metallic and non-metallic
honeycombs, cellular foams, balsa wood or trusses. The face sheets are typically bonded to
the core with an adhesive, and carry most of the bending and in-plane loads. The core
provides the flexural stiffness and out-of-plane shear and compressive strength.
The structural performance of sandwich panels depends not only on the properties of the
skins, but also on those of the core, the adhesive bonding the core to the skins, and the
geometrical dimensions of the components. Because of their main features, such as the high
flexural resistance and stiffness [2], the high impact strength [3,4], the high corrosion
resistance [5] and the low thermal and acoustics conductivity [6-8], sandwich structures are
preferred over conventional materials in various industrial applications. Although large
number of research projects have been performed by various authors, the design of structural
elements made from sandwich composites is often a difficult task.This is mainly because a
reliable strength prediction models, preliminary knowledge of the mechanical behaviour of
skins and core, as well as of the peculiar damage mechanisms [9-11] and failure criteria of
sandwich structure do not exist.

The aim of the present work is to investigate the mechanical properties of sandwich
structures, focusing on the effects of carbon and glass fibers skins. Therefore two different
types of sandwich structures were fabricated using carbon face sheets with nomax
honeycomb as core material. Flexural tests were performed for static mechanical
characterization of the virgin laminates and sandwich structures. The mechanical results of
laminates and sandwich structures were used as input data in order to implement a simplified
model, developed by a commercial finite element analysis code.

The samples have been verified by simulating flexural tests of laminates and sandwich
structures. In addition their mechanical behaviour has been compared with experimental data
by the aforesaid static tests of complex sandwich structure.

2. Experimental Procedure

2.1 Design Considerations

The design of sandwich panel is greatly dependent on manufacturing options, material


configurations and laboratory/shop environments. The face sheets are manufactured by
impregnating carbon and glass 00/900 with in 50:50 ratio by weight and cured at 25 0C for 8
hours by vacuum assisted process as shown in Figure 1. The honeycomb is placed on the
lower face sheet and upper face sheet is placed over honeycomb. The entire assembly is again
vacuum bagged and co cured.

Figure 1: Vacuum assisted process

2.2 Mechanical Testing

Static mechanical tests were preliminary conducted to acquire mechanical characterization of


the sandwich structure and to implement the numerical simulation. The three point bending
test, according to ASTM D-790 for laminates and ASTM C-393, was used to determine the
behaviour of laminates and sandwich beams subjected to flat wise flexure. Samples for three
point flexural testing were 80 mm long with span length of 50 mm. The three point bending
tests were carried out until core failure occurred.
2.3 Finite Element Analysis

Numerical simulations were conducted using the ANSYS 9.0 finite element software. An 8
node, 3-D shell element was used. Load (N) is used as an input parameter to ANSYS model.
Composite modeling which is based on the knowledge of the materials data obtained via
experimental test. The post elastic behavior is intentionally neglected due to non-linear
behavior of material properties.The numerical analysis is carried out to compare the
theoretical model with the experimental results.

3. Results and Discussion

3.1Mechanical results

3.1.1 Carbon Laminates Samples

Flexural test were performed in three point bending under displacement control, in which
laminates were loaded until failure, at a constant rate of 0.02 mm/sec. It is observed that the
fracture in carbon laminates starts at loads much smaller than the composite failure loads. i.e.
the fracture start taking place ply by ply. Typical damage seen in a sample of a carbon skin as
shown in Figure 2. Thus the fiber fractures as the load increases. This results in weakening of
the cross section of the composite and the structure fails completely.

Figure 2: Typical deformation behaviour of a Carbon Laminate Sample

Micrograph of upper surface of specimen, which under went compressive loads is shown in
Figure 3. Fiber kinks were evident which appeared as a result of fiber micro buckling, kink
band formation and fiber. In both Figures 3 and 4, it can be observed that the majority of
fibers have fractures while very few are intact. This indicates that in case of three point bend
test, failure is originating in the fibers instead of matrix.

Kink Band
Figure 3: 00 ply at the top surface 1000X

The out brush of 00 fiber at the bottom surface indicate tensile fracture mode in Figure 4. The
flexural bending imposed tensile force at the bottom surface. During flexural test of cross-ply
laminates, the failure is generally associated with tensile failure of the outer 0 0 ply. Due to
these observations, it is revealed that failure in case of flexural testing was fiber dominated.

Figure 4: 00 ply at the bottom surface 200X

3.1.2 Glass Laminates Samples


Typical damage seen in a sample of glass skin is shown in Figure 5. The glass sample
appeared to fail primarily with skin wrinkling and fiber fracture. Compression/tension failure
mode is the dominated mode in these samples. i.e. the upper layer was under compression
force and the lower layer was under tensile loading. When the load reached to its maximum
level the lower layer fractured that was followed upward and ultimately the sample failed.

Figure 5: Typical deformation behaviour of a Glass Laminate Sample


In the case of the glass laminate samples, post-failure examination of the fracture surfaces
revealed that the failure was always primarily due to delamination followed by micro
buckling of 00 plies. It can be inferred from micrograph that buckling results due to micro-
cracks of the matrix. Microbuckling is a failure mode in which the composite suffers
localized collapse within a narrow zone, often called the Kink Band (KB). This kink band is
illustrated in Figure 6. It is known that the compressive side of CP laminates is controlled by
fiber micro buckling which leads to Kink Band formation. Due to the progression of micro-
buckling, the fibers that are adjacent to the buckled fibers fail due to shear deformation of the
recently micro buckled fibers.
Figure 6: 00 ply (Laminate Upper ply)

In this particular case, the initiation of Kink Band was caused by fiber bending and matrix
deformation. Matrix cracking in the 90 plies and fracture in some of the adjacent 0 fibers is
observed which ultimately developed into an out-of-plane (through the thickness of the
specimen) kink-band.

Figure 7: Glass Laminate Lower ply

The fracture initiated in the lower ply which followed upward without any shear among the
layers is shown in figure 7. Interfacial delamination may be caused due to initial flaws within
the specimens as a result of weak bonding of lamina, voids or defects developed during the
manufacturing process.

3.1.3 Carbon Sandwich Sample

Typical damage seen in a carbon sample with a honeycomb core is shown in Figure 8. As the
load is applied on the carbon sandwich samples, initial load is carried by the upper skin.
Further loading lead to a top skin failure by bending under the load application point and this
phenomenon is followed by a large decrease in load.
Figure 8: Typical deformation behaviour of a Carbon Sandwich Sample

The function of the core is to support the thin skins so that they don't buckle and stay fixed
relative to each other. The core experiences mostly shear stresses (sliding) as well as some
degree of vertical tension and compression. Figure 8 shows the core shear behaviour in
carbon sandwich sample which can also seen in Figure 9.

Core Shear

Figure 9: Core shear in Carbon Sandwich Sample

3.1.4 Glass Sandwich Sample

Typical damage seen in a glass sample with a honeycomb core is shown in Figure 10. This
structure appeared to fail primarily in the skins through skin wrinkling and fiber fracture.
Figure 10: Typical deformation behaviour of a Glass Honeycomb Sample

There were some minor cracking in the core along with crushing within the core structure due
to shear. This agrees with the shape of the load displacement curve with the drop in load
related to skin fracture. No delamination was observed between the honeycomb and glass
skin; as resin flows in the honeycomb structures cavities and thus making a strong inter-phase
and making it difficult to leave the skin without rupturing itself. The core shear behaviour in a
glass laminate sample is shown in figure 10. This behavior can also be seen in Figure 11.

Figure 11: Core shear in Glass Sandwich Sample

3.2 Experimental and Finite element analysis results

It has been observed that ANSYS software can be effectively used as useful finite element
analysis tool for structural examination of composite and sandwich specimens. The percent
difference between experimentally observed displacement and that predicted by the finite
element model for the laminate structure samples was 2.8 % and that for the sandwich
structure was 4.8 %.

3.2.1 Carbon Laminate Sample

Carbon laminate samples can bear 2000 N load with the experimental displacement of 1.756
mm and finite element displacement of 1.798 mm as recorded in Table 1. Figure 12 shows the
graphical comparison of experimental and finite element displacements of carbon laminate
samples. Figure 13 shows the computer based simulation of carbon laminate sample in three
point bent test at 2000 N force. The central area of the sample (blue band) shows the
displacement of 1.708 mm which is recorded in Table 1.

Table 1: Experimental and FEM results of carbon laminate sample


Figure 12: Graph showing comparison of experimental and FEM results of carbon laminate
sample

Figure 13: Computer based Simulation of Carbon Laminate Sample in 3PBT at 2000 N force

3.2.2 Glass Laminate Sample

Glass laminate samples can only bear 1400 N load before fracturing with the experimental
displacement of 2.421 mm and finite element displacement of 2.352 mm as recorded in Table
2. Figure 14 shows the graphical comparison of experimental and finite element
displacements of glass laminate samples. Figure 15 shows the computer based simulation of
glass laminate sample in three point bent test at 1400 N force. The central area of the sample
(blue band) shows the displacement of 2.421 mm which is recorded in Table 2.
Table 2: Experimental and FEM results of glass laminate sample

Figure 14: Graph showing comparison of experimental and FEM results of glass laminate
sample

Figure 15: Computer based Simulation of Glass Laminate Sample in 3PBT at 1400 N force
3.2.3 Carbon Sandwich Sample

Carbon sandwich samples can bear load of 1200 N load with the experimental displacement
of 1.231 mm and finite element displacement of 1.171 mm as recorded in Table 3. Figure 16
shows the graphical comparison of experimental and finite element displacements of carbon
sandwich samples which shows the deviation of 4.8 %. Figure 17 shows the computer based
simulation of carbon sandwich sample in three point bent test at 1200 N force. The central
area of the sample (blue band) shows the displacement of 1.171 mm which is recorded in
Table 3.
Table 3: Experimental and FEM results of carbon sandwich sample

Figure 16: Graph showing comparison of experimental and FEM results of carbon sandwich
sample

Figure 17: Computer based Simulation of Carbon-Nomex honeycomb Sandwich Sample in


3PBT at 1200 N force

3.2.4 Glass Sandwich Sample

Glass sandwich samples can bear load of 800 N load with the experimental displacement of
1.158 mm and finite element displacement of 1.102 mm as recorded in Table 4. Figure 18
shows the graphical comparison of experimental and finite element displacements of glass
sandwich samples which shows the deviation of 4.8 %.. Figure 19 shows the computer based
simulation of carbon sandwich sample in three point bent test at 800 N force. The central area
of the sample (blue band) shows the displacement of 1.158 mm which is recorded in Table 4.
Table 4: Experimental and FEM results of glass sandwich sample

Figure 18: Graph showing comparison of experimental and FEM results of glass sandwich
sample

Figure 19: Computer based Simulation of Glass-Nomex honeycomb Sandwich Sample in


3PBT at 800 N force

Comparison of area under the load displacement curve is a measure of toughness. The area
under the load displacement curve of carbon sandwich sample is 1440 N-mm and for glass
sandwich sample is 960 N-mm. Results of carbon sandwich samples and glass sandwich
samples clearly indicates the superiority of carbon over glass fabric in the elastic region refers
to the fact that it can handle higher loads before fracturing.

4. Conclusions
In the present work the actual mechanical behaviour of two types of sandwich structures
widely have been studied by means of experimental investigations and successive numerical
analyses under different loading conditions. On the basis of the experimental flexural tests
performed the mechanical characteristics of the laminates and sandwich structures are
obtained in bending, results show the superiority of the carbon over glass fabric. The bending
behaviour of the sandwich beams is governed by the face sheets. The presence of different
fiber skin induces relevant effects on the strength of the sandwich structure. The finite
element analysis, performed by employing the commercial ANSYS code, was carried out to
simulate the static mechanical tests. The static mechanical behaviour of the composite
structure is well approximated by numerical simulations in elastic zone. The laminate and
sandwich model is suitable to perform a proper engineering design of sandwich composite
structures. This is helpful to choose the correct geometry as well as the correct layer
arrangement at varying loading conditions that are not often experimentally available.

References

[1] Atckinson R. Innovative uses for sandwich constructions. Reinf Plast 1997; 41(2):30-3
[2] Vinson J R .The behaviour of sandwich structures of isotropic and composite materials.Westport: Technomic; 1999
[3] Mines RAW, Worrall CM,Gibson AG.Low velocity perforation behaviour of polymer composite sandwich panels.Int J
Impact Engng 1998;21(10):885-79
[4] Torre L, Kenny JM. Impact testing and simulation of composite sandwich structures for civil transportation.Compos
Struct 2000; 50:257-67
[5] Kootsookos A, Burchill PJ. The effect of the degree of cure on the corrosion resistance of vinyl ester/glass fiber
composites. Composites A 2004; 35:501-8
[6] Allard JF. Propogation of sound in porous media: modeling sound absorbing materials. Elsevier Applied Science; 1993
[7] Sahraoui S, Mariez E,Etchessahar M. Mechanical testing of polymeric foams at low frequency. Polym Test 2001;20:93-6
[8] Kalaprasad G,Pradeep P,George Mathew,Pavithran C,Sabu Thomas. Thermal conductivity and thermal diffusivity
analyses of low density polyethylene composites reinforced with sisal,glass and initimately mixed sisal/glass fibers. Compos
Sci Technol 2000;60:2967-77
[9] Anderson,Melvin S.Optimum proportions of truss core and web core sandwich plates loaded in compression.NASA TN
D-98;1959
[10] Steeves CA,Fleck NA,Collapse mechanisms of sandwich beams with composite faces and a foam core,loadied in three
point bending.Part I: analytical models and minimum weight design. Mech Sci 2003; 46:561-83
[11] Steeves CA,Fleck NA. Material selction in sandwich beam construction.Scripta Mater 2004; 50:1335-9

You might also like