You are on page 1of 25

ARTICLE IN PRESS

International Journal of Mechanical Sciences 47 (2005) 13011325


www.elsevier.com/locate/ijmecsci

The response of honeycomb sandwich panels under


low-velocity impact loading
M. Meoa,, R. Vignjevica, G. Marengob
a
Structures and Materials Group, College of Aeronautics, Cranfield University, Bedford, Mk 43 0AL, UK
b
GKN Aerospace Services, Cowes, Isle of Wight, PO31 6RH, UK
Received 30 January 2004; received in revised form 27 April 2005; accepted 12 May 2005
Available online 6 July 2005

Abstract

This paper describes the results of an experimental investigation and a numerical simulation on the
impact damage on a range of sandwich panels. The test panels are representative of the composite sandwich
structure of the engine nacelle Fan Cowl Doors of a large commercial aircraft. The low-velocity impact
response of the composites sandwich panels is studied at ve energy levels, ranging from 5 to 20 J, with the
intention of investigating damage initiation, damage propagation, and failure mechanisms. These impact
energy levels are typically causing barely visible impact damage (BVID) in the impacted composite
facesheet.
A numerical simulation was performed using LS-DYNA3D transient dynamic nite element analysis
code for calculating contact forces during impact along with a failure analysis for predicting the threshold
of impact damage and initiation of delaminations. Good agreement was obtained between numerical and
experimental results. In particular, the numerical simulation was able to predict the extent of impact
damage and impact energy absorbed by the structure. The results of this study is proving that a correct
numerical model can yield signicant information for the designer to understand the mechanism involved in
the low-velocity impact event, prior to conducting tests, and therefore to design a more efcient impact-
resistant aircraft structure.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Sandwich panel; Barely visible impact damage; Low-velocity impact

Corresponding author. Tel.: +44 1234750111x5220; fax: +44 1234752149.


E-mail address: m.meo@craneld.ac.uk (M. Meo).

0020-7403/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijmecsci.2005.05.006
ARTICLE IN PRESS

1302 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

1. Introduction

Sandwich constructions are nding increasing use in the aerospace, automotive and marine
industries because of their relative advantages over other structural materials in terms of
improved stability, high stiffness to weight and strength to weight ratios. Sandwich constructions
provide an efcient solution to increase bending stiffness without signicant increase in structural
weight. The sandwich are multi-layered material made by bonding stiff, high-strength, bre
reinforced laminated facings to low density core material. These structures can carry both in-plane
and out-of-plane loads and remain stable under compression without a signicant weight penalty.
While the preliminary design of aircraft sandwich structures have been investigated extensively
in the past years [118], there is a lack of understanding of low-speed impact-damage mechanisms,
and the effect of such damage, including penetration damage, on structural performance. The
presence of highly complex and transient dynamic failure modes in such materials and the
inaccessibility of internal damage to real-time monitoring have resulted in experimental studies
limited to only the nal impact damage characteristics of failure and to measurement of after
impact residual strengths.
Sandwich constructions are being considered for application to aircraft primary structures,
where durability and damage tolerance is a primary consideration, therefore, understanding the
adverse effect of in-service impact events (impact damage and penetration damage) has become
vital for applications where structural durability and damage tolerance are primary considera-
tions.
Certication authorities require that exposed aircraft components must be tested to prove their
capability to withstand low-velocity impact without suffering critical damage.
Low-velocity impact can be caused by a number of fairly common Discrete Sources of Damage
(DSD) such as bird-strike, hail stone, tool drops, debris thrown up from runways, etc. These
events may induce overall and localized damage in sandwich structures. The failure characteristics
of sandwich structures are signicantly different from conventional laminated structures and are
strongly dependent on the core and skin materials and their thickness. The core material is much
softer than the skin and will experience a much larger deformation. The localized damage is
usually conned to the top facing, the core-top facing interface, and the core material. The bottom
skin generally is left undamaged. In particular, permanent indentation in the impacted facesheet
accompanied with localized core crushing beneath and around the impact site is produced by a
901, low-velocity impact with a spherical impactor. The facing skin will typically rebound to some
degree after the impact event, therefore, the prole of the residual facesheet indentation does not
necessarily correspond to that of the underlying crushed core.
Depending on the size and mass of the impactor and impact energy, the various damage modes
that may develop failure are: (a) delamination in the impacted facesheet (b) skin-core debonding;
(c) core crushing and shear; (d) matrix cracking, and (e) bre breakage in the facings and (f) core
buckling [19]. The rst major type of failure is also called inter-layer failure. Generally, two
distinct inter-layer failure modes are observed. The rst one called delamination buckling
[2022] involves opening-dominated inter-layer cracking. The second one involves shear-
dominated inter-layer cracks or shear delaminations, and is typical of layered materials
subjected to out-of-plane impact [2325]. The second failure category is often referred to as
interfacial debonding and it is characterized by decohesion at skincore interface. The core
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1303

crushing and shear, matrix cracking and bre breakage are also referred to as intra-layer failure.
There are three possible intra-layer failure modes depending on the material constitution. The rst
one is called intra-layer cracking or matrix cracking. This type of cracking often occurs inside the
matrix of bre-reinforced composites or within the soft core of sandwich structures [26,27].
Another possible intra-layer failure is the failure of reinforcements such as bre breakage and
bre kinking within a layer [28,29].The fth possible intra-layer failure mode is interfacial
debonding between the matrix material and the reinforcement [30] such as debonding between
particle/bre and matrix occurring within a constituent layer.
Many researchers have investigated the impact response and damage characteristics of
sandwich structures [3134,3652], showing that localized impact loading on a sandwich structure
can result in the generation of localized damage which can lead to signicant reductions in its
load-bearing properties. A thorough and detailed review of the impact response of sandwich
structures is given by Abrate in Ref. [19].
Many authors [3840] have shown that both the skin conguration and the core density control
their impact behaviour. When tough skins are used, the behaviour of the core is less important
[39]; however, in most cases in order to obtain a sandwich structure with better crash
performances the energy absorbing behaviour of the core must be improved. Robinson studied
the relation between peak impact load and the area of impact damage [53]. Goldsmith and
Sackmann [54] have analysed panels with honeycomb cores with metal or polymer composite
skins and they have identied the energy absorbing modes as core crushing, bending and
stretching of the top face, wrinkling and punch through of the upper facing and global bending of
the structure. Rosenfeld and Gause conducted drop-weight impact tests on honeycomb sandwich
structures based on carbon reinforced/epoxy skins [55]. They showed that initial damage took the
form of localized buckling of the honeycomb core directly under the point of impact. Higher
energies were found to result in the generation of delamination damage in the composite face
sheets. They also showed that the support offered by the core material to the skins limits the
spread of damage to twice the diameter of the impactor even at the perforation threshold.
Horrigan et al. [56] conducted impact tests on a Nomex honeycomb sandwich structure with glass
bre reinforced epoxy facings. Their results indicate that a soft projectile generates shallow
crushing in the core whereas rigid impactors generate deeper damage that conforms to the shape
of the projectile. Rhodes [57] conducted impact tests on a number of sandwich structures and
showed that increasing the crush strength of the core material increases the impact resistance of
the sandwich structure.
According to Tomblin et al. [1] low-velocity blunt object impacts may produce fairly
widespread Nomex core crushing with small residual facing skin indentation and/or visibly
detectable facesheet damage and the backside facing skin is generally unaffected by the impact
event. Except for a small dent or blister in the surface, alongside a delamination within the
laminate, the impacted zone is often barely visible, but damage may grow under fatigue loading to
a critical size and putting in dangers the component [25]. Nettles and Hodge observed that for
honeycomb cores at very low energy levels, core buckling takes place under the impact point [58].
For higher-energy impacts, core cracking was observed when the glass bre reinforcement was
broken. Mines et al. studied the static and impact behavior of sandwich beams and identied four
mechanisms [59]. Mode I started with a top skin compressive failure, followed by either a stable
crushing of the core (mode Ia) or a core shear failure (mode Ib); mode II was where the upper
ARTICLE IN PRESS

1304 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

woven aramid skin failed in compression but remained intact, no sudden load drop was observed,
and nal failure resulted from tensile failure of the lower woven aramid skin; mode III was a core
shear failure; and mode IV corresponded to a skin failure that was loaded in tension. Cantwell et
al. showed that perforation occurred by shearing the skins, and that this process absorbed most of
the energy [60].
Mine et al. studied the modes of failure of various sandwich beams, and concluded that the best
mode of failure for energy absorption is upper skin failure and subsequent stable core crushing
[61]. The impact failure mode including upper skin compression failure, core crushing, lower skin
tensile failure is mainly applicable to high-density cores with a density greater than 200 kg/m3. For
low-density cores, localized core crushing at the point of application of the load occurs. However,
one of the primary disadvantages of sandwich structures is the dissimilarity of skin and core
materials. The skincore interface is a potential weakness for delamination under localized impact
loading, and causes relatively poor impact resistance [62]. Vayda et al. reported that the rst load
drop damage of the glass pile core material was associated with core failure [63]. Under low-
velocity impact, the threshold energy to cause core failure was found to be about 515 J for the
6 mm core thickness. An increase in core thickness resulted in lower damage thresholds under low-
velocity impact. This phenomenon is different from monolithic laminate, where the rst load drop
of the monolithic laminate is referred to as a Hertzian failure, which is probably due to the
initialization of delamination failure [64].
A limited number of studies have focused on modelling the dynamic response of sandwich
structures. A simple sprig-mass model was used by Nettles and Lance [65] to model the low-
velocity impact response of a honeycomb sandwich structure with carbon bre/epoxy facings.
Very good agreement was achieved between predictions and experimental data at energy levels
below that required to initiate signicant damage within the structure. Cantwell et al. [66] applied
an energy-balance model to predict the low-velocity impact response of a wide range of foam-
based sandwich beams with an accurate prediction of the maximum impact force at low impact
energies. They also showed that their approach can be used to investigate the effect of varying key
material properties on the impact response of a sandwich structure. Ambur and Cruz [67] studied
the contact mechanics between a sandwich pate and a spherical indentor to model the low-velocity
impact response of composite sandwich panels. Davies et al. [68] used nite element analysis
(FEA) to simulate the low-velocity impact response of a PVC foam sandwich structure with glass
bre/polyester skins. The maximum impact force was overestimated but the loss of contact during
the impact event was observed. Nemes and Simmonds also used a FEA to model the low-velocity
impact response of a foam core sandwich structure [69]. The FE analysis resulted very expensive
in terms of computing time, and the target displacements was overestimated by up to 30%.
The current approach to study impact design problem is either to conduct an experimental
campaign to study impact behaviour of the studied material under different loading conditions
and sample geometry or to perform simulation of the impact phenomena using nite element
methods (FEM) and requires very powerful hardware and software resources. This work tries to
integrate these two approaches, presenting data about the experimental campaign coupled to the
computer simulation of a low-velocity impact phenomena.
The present paper presents the results of an experimental and analytical study into low-velocity
impact on sandwich engine nacelle panels. The test panels are representative of the structure that
will be used in a Fan Cowl Door of an underwing engine nacelle of a large commercial aircraft.
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1305

Table 1
List of test panels

Design Panel Type Size (mm) Description

Existing Fan Cowl Door design 1 Sandwich 381  305 Basic door sandwich
2 Sandwich 381  305 Basic pandown sandwich
3 Sandwich 381  305 Basic door reinforced sandwich
Modied Fan Cowl Door design 5 Sandwich 381  305 Basic door construction
6 Sandwich 381  305 Reinforced door construction
7 Sandwich 381  305 Basic door pandown

The objectives of the study was to understand the impact damage mechanism, the load
distribution in the composite face sheets of the sandwich panel, and to determine panel
failure mechanisms. The impact-damage condition simulated in the experimental study
corresponds to BVID. Experimental data was generated using a steel ball indentor to
create the impact damage according to the test conditions specied by the certication authorities
(Table 1).
A nite element model was developed to evaluate the damage on a panel due to low-velocity
impact in order to provide signicant information, prior to conducting tests, to design an impact-
resistant aircraft structure and to show that numerical simulation tools can be used successfully to
reduce the amount of testing to be conducted.

2. Experimental details

2.1. Panel manufacturing

A total of three panels was prepared to compare the impact resistance of the Trent Modied
Fan Cowl Door with the existing design (Fig. 1).
A more complete description of the panels is presented in Tables 2 and 3. The resin content of
the panels was 3773%. In Fig. 2, the location on the Fan Cowl Doors of the structure
congurations considered by this test programme are presented.

2.2. Impact testing apparatus

The testing was intended for development purposes and for investigating the effects of impact
testing on the various door designs. The panels were subjected to a range of impact energies
enabling a direct comparison of the resulting damage. The test specimens used in this
experimental study were fabricated by GKN Aerospace Services. A dropped-weight impact test
machine described in Fig. 3 was used for impacting each test panel at 5 energy levels ranging from
5 to 20 J. These impact energy level were chosen in order to inict damage in the composite test
panel facesheet corresponding to a barely visible impact damage (BVID). The instrumented
dropped-weight impactor with a hemispherical tip was dropped from different heights to generate
ARTICLE IN PRESS

1306 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

Fig. 1. (a) Existing design panels 1 and 3, (b) existing design panel 2, (c) modied panels 4 and 5 and (d) modied
design panel 6.

Table 2
Trent Existing Fan Cowl Door test panel description

Panel 1 Trent Existing Fan Cowl Skin2 plies of carbon epoxy Cycom 69/GFE 3105 H prepreg
Door basic sandwich Core Nomex 0.600 thick HRH10-1/8-1.8
Skin2 plies of carbon epoxy Cycom 69/GFE 3105 H prepreg
Panel 2 Trent Existing Fan Cowl Skin4 plies of carbon epoxy Cycom 69/GFE 3105 H prepreg
Door pandown sandwich Core Nomex 0.2500 thick HRH-10-1/8-6.0
Skin4 plies of carbon epoxy Cycom 69/GFE 3105 H prepreg
Panel 3 Trent Existing Fan Cowl Skin2 plies of carbon epoxy Cycom 69/GFE 3105 H prepreg
Door reinforced sandwich Core Nomex 0.600 thick HRH-10-1/8-6.0
Skin2 plies of carbon epoxy Cycom 69/GFE 3105 H prepreg

different impact-energy levels. Each individual honeycomb panel was subjected to impacts at the
locations shown in Fig. 4. The impact energy to be applied to each test panel is summarized in
Table 4.
Two different types of test were conducted. The rst batch of testing (Results AJ inclusive)
used impact energies levels with a 16 mm diameter tip while the second batch of tests used the
same prescribe energy levels with a 50 mm diameter tip. These tests were performed to ensure that
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1307

Table 3
Trent Modied Fan Cowl Door test panel description

Panel 4 Trent Modied Fan Cowl Skin3 plies of carbon epoxy Hexcel AS4-8552
Door basic sandwich Core Nomex 1.000 thick HRH-10-1/8-4
Skin2 plies of carbon epoxy Hexcel AS4-8552
Panel 5 Trent Modied Fan Cowl Skin4 plies of carbon epoxy Hexcel AS4-8552
Door pandown sandwich Core Nomex 0.2500 thick HRH-10-1/8-6
Skin3 plies of carbon epoxy Hexcel AS4-8552
Panel 6 Trent Modied Fan Cowl Skin4 plies of carbon epoxy Hexcel AS4-8552
Door reinforced sandwich Core Nomex 1.000 thick HRH-10-1/8-4
Skin3 plies of carbon epoxy Hexcel AS4-8552

Fig. 2. Engine modied left and right hand Fan Cowl Door schematic view.

the damage levels specied certication requirements were achieved (20 mm diameter hole or
30 mm diameter delamination).
During the tests, each panel was placed on a cylindrical ring with an outer radius of 0.1016 m
centered at the impact locations.

2.3. Test results

This series of test activities provided the results to carry out a direct comparison between the
materials and the sandwich congurations used in two different designs of the Fan Cowl Door,
with regards to impact damage. The results provided by this series of tests were:

 Impact dent depth


 C-scan image of the damage
 Area affected by the impact damage
ARTICLE IN PRESS

1308 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

Fig. 3. Schematic of the impact test apparatus.

G H J M

.305m
0.05m D E F L

A B C K
0.05m

0.08m

0.0762 m 0.0508 m 0.0508 m 0.0508 m

0.381 m

Fig. 4. Details of impact locations.

The dent depth was measured by inserting a needle in the impact crater and by measuring the
insertion length. The needle has been inserted consistently in the deepest point of the impact crater
taking care of not going through the damaged skin.
For the delamination pattern visualization a C-scan NDI ultrasonic technology with a
pulse-echo and an immersion scanning method, Physical Acoustics, was used, with the following
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1309

C-scan data:
 Scan frequency: 5 MHz;
 Scan speed (X-axis): 30 mm/s;
 Index axis: Y-axis;
 Scan interval index: 0.05 mm.

There are two kinds of data processing in pulse-echo C-scanning: echo level view amplitude and
time-of-ight view. The latter data were converted into a delamination depth view through the
transverse sound velocity in CFRP laminates. Therefore the identication of delamination
interfaces was easily identied and the measurement of the respective area was carried out with an
image processing equipment.
Prior to conducting the impact testing, each panel was subjected to a C-scan examination in
order to clear the presence of manufacturing damages. Each test panel was tested at room
temperature ambient conditions using the impact test rig shown in Fig. 3. On completion of the
impact testing, each panel was subjected to a C-scan examination. The delamination area was
evaluated by considering them always circular and having a diameter equal to the maximum size

Table 4
Impact test requirements for honeycomb panels

Test panel no. Impact energy (J) Location of impact

1, 2, 3, 4, 5 & 6 5 A, B & C
1, 2, 3, 4, 5 & 6 8 D, E & F
1, 2, 3, 4, 5 & 6 10 G, H & J
1, 2, 3, 4, 5 & 6 5 K, L & M
1, 2, 3, 4, 5 & 6 10 N, O & P
1, 2, 3, 4, 5 & 6 20 Q, R & S

Fig. 5. Damage area evaluation.


ARTICLE IN PRESS

1310 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

of the delamination coming from the C-scan images. This approach is considered to be
conservative and it is illustrated in Fig. 5.
The diagrams of Figs. 6 and 7 show the impact damages recorded for panels 1 and 4,
representing the basic sandwich conguration used on the existing and modied design of the Fan
Cowl Door. All the data sets exhibit a linear trend between the impact energy levels and the
damage areas, expect for panel 1 with a 50 mm tup. The position of the curves connecting impact
test data recorded for the Modied Fan Cowl Door (Panel 4) shows considerable less damage at
the same energy level if compared with the Existing Fan Cowl Door (Panel 1). In particular the
curves for the 16 mm impacts show less penetration in the case of modied design, while the curves
for the 50 mm tup impacts show a smaller delaminated area for the modied design. The improved
performance is due to the higher strength material used in the skin, the extra layer of carbon-
epoxy prepreg, and to the higher density Nomex core used in this sandwich skin design (Table 5).

6000.0
Delamination Area, mm2

5000.0 Panel 1- Tup 16 mm

4000.0 Panel 4 - Tup 16 mm


Panel 1 - Tup 50 mm
3000.0
Panel 4 - Tup 50 mm
2000.0

1000.0

0.0
4 9 14 19 24
Impact Energy, J

Fig. 6. Experimental delamination area for panels 1 and 4 for impact with increasing energy levels and with different
impactor weights (basic sandwich structure).

7.0

6.0 Panel 1 - Tup 16 mm


Panel 4 - Tup 16 mm
Dent Depth , mm

5.0

4.0 Panel 1 - Tup 50 mm


Panel 4 - Tup 50 mm
3.0

2.0

1.0

0.0
4 9 14 19 24
Impact Energy, J

Fig. 7. Experimental dent depth for panels 1 and 4 for impact with increasing energy levels and with different impactor
weights (basic sandwich structure).
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1311

Table 5
Dent depth and delamination area comparison for panels 1 and 4

Impact Required Tup size Dent depth (mm) Difference Damage area (mm2) Difference
location drop energy (mm) (%) (%)
(J)
Panel 1 Panel 4 Panel 1 Panel 4

A 5 16 3.7 0.96 74.05 1017.9 400.3 60.67


B 5 16 3.84 0.94 75.52 1128.2 455.4 59.63
C 5 16 3.96 0.36 90.91 984.2 480.1 51.22
D 8 16 6.06 3 50.50 1425.3 613.6 56.95
E 8 16 5.2 2.16 58.46 1244.1 661.7 46.81
F 8 16 5.31 2.16 59.32 1365.7 711.6 47.89
G 10 16 6.8 3.5 48.53 1493 763.3 48.87
H 10 16 6.78 3.38 50.15 1626 681.4 58.09
J 10 16 6.26 2.66 57.51 1244.1 838.8 32.58
K 5 50 0.15 0.1 33.33 2214.5 1111.8 49.79
L 5 50 0.09 0.08 11.11 2706.2 1049.2 61.23
M 5 50 0.11 0.12 9.09 3068 1049.2 65.80
W+N 20 50 1.05 0.52 50.48 5714.6 1310.6 77.07
O 20 50 3.7 0.41 88.92 5714.6 1283.2 77.55
P 20 50 1.4 0.68 51.43 4976.4 1310.6 73.66
Q 10 50 0.97 1.4 44.33 4512.6 1920.5 57.44
R 10 50 1.07 1.82 70.09 4548.4 2004.9 55.92
S 10 50 1.04 1.67 60.58 4488.8 1757.2 60.85

The diagrams of Figs. 8 and 9 show the direct comparison between the impact damages
recorded on panel representing the reinforced sandwich structure of the Existing (Panel 2) and
Modied (Panel 5) Fan Cowl Door. When impacted with a 16 mm tup the panel representing the
Modied Fan Cowl Door experienced a damage, which was not as deep as that seen on the
existing design, but had a comparable size of the delaminated area. The cause of this behaviour is
due the material with higher strength used on the skin of the modied door panel and to the
different thickness of the sandwich skin (two extra layers ), which in the case of the modied
design, absorbs the impact energy by global deformation of the plate without deep penetration
(Table 6).
This is conrmed by the fact that when impacting the same panel with a much larger tup
(50 mm), the damage are very similar both in depth and size of the delaminated area. In this
particular case, the lower density of the Nomex core used in the modied design plays a role in
reducing the difference between the impact resistance performances of the two sandwich designs.
The diagrams of Figs. 10 and 11 present the direct comparison between the impact damages
recorded on panel representing the pandown sandwich of the Existing (Panel 3) and Modied
(Panel 6) Fan Cowl Door. In this case the two panels have both the skin made of four plies of
carbon epoxy and the are built using the same Nomex core type. The results shows very little
difference in the delamination area between the two panel congurations when impacted with a
16 mm tup. However, signicant difference in the delamination area was recorded between panel 3
(569.4 mm) and 6 (737.7 mm) when impacted with a 50 mm tup for an impact energy of 10 J.
ARTICLE IN PRESS

1312 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

Panel 2- Tup 16 mm
Panel 5 - Tup 16 mm
Panel 2 - Tup 50 mm

1400.0 Panel 5 - Tup 50 mm

1200.0
Delamination Area, mm2

1000.0

800.0

600.0

400.0

200.0

0.0
4 9 14 19 24
Impact Energy, J

Fig. 8. Experimental delamination area for panels 2 and 5 for impact with increasing energy levels and with different
impactor weights (reinforced sandwich structure).

Panel 2 - Tup 16 mm
Panel 5 - Tup 16 mm
Panel 2 - Tup 50 mm

2.5 Panel 5 - Tup 50 mm

2.0
Dent Depth , mm

1.5

1.0

0.5

0.0
4 9 14 19 24
Impact Energy, J

Fig. 9. Experimental dent depth for panels 2 and 5 for impact with increasing energy levels and with different impactor
weights (reinforced sandwich structure).

Fig. 11 shows that signicant dent depth was experienced by panel 3 when impacted with a 16 mm
and 50 mm tup compared with panel 6. These results showed that the modied panel possessed
better impact damage resistance of panel 3 (Table 7).
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1313

Table 6
Dent depth and delamination area comparison for panels 2 and 5

Impact Required Tup Size Dent depth (mm) Difference Damage area (mm2) Difference
location drop energy (mm) (%) (%)
(J)
Panel 2 Panel 5 Panel 2 Panel 5

A 5 16 0.2 0.32 60.00 147.7 259.9 75.96


B 5 16 0.1 0.4 300.00 125.9 291.3 131.37
C 5 16 0.02 0.8 3900.00 117.6 291.3 147.70
D 8 16 0.72 1.32 83.33 349.7 324.6 7.18
E 8 16 0.86 1.12 30.23 283.2 435.2 53.67
F 8 16 0.88 1.48 68.18 349.7 396.5 13.38
G 10 16 1.3 1.8 38.46 462.4 475.7 2.88
H 10 16 1.18 2.1 77.97 400.3 517.9 29.38
J 10 16 1.38 2.14 55.07 503.5 475.7 5.52
K 5 50 0.39 0.01 97.44 546.4 517.9 5.22
L 5 50 0.32 0.04 87.50 462.4 562 21.54
M 5 50 0.41 0.02 95.12 590.9 562 4.89
W+N 20 50 0.77 0.12 84.42 952 809.3 14.99
O 20 50 0.74 0.08 89.19 1010.5 864.1 14.49
P 20 50 0.73 0.09 87.67 1034.5 920.8 10.99
Q 10 50 2.12 0.98 53.77 1262.3 1039.5 17.65
R 10 50 2.2 0.96 56.36 1329.6 1165.4 12.35
S 10 50 2.17 0.97 55.30 1398.7 1101.5 21.25

Panel 3- Tup 16 mm
Panel 6 - Tup 16 mm
Panel 3 - Tup 50 mm

2000.0 Panel 6 - Tup 50 mm


1800.0
1600.0
Delamination Area, mm2

1400.0
1200.0
1000.0
800.0
600.0
400.0
200.0
0.0
4 9 14 19 24
Impact Energy, J

Fig. 10. Experimental delamination area for panels 3 and 6 for impact with increasing energy levels and with different
impactor weights (pandown sandwich structure).
ARTICLE IN PRESS

1314 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

Panel 3 - Tup 16 mm
Panel 6 - Tup 16 mm
Panel 3 - Tup 50 mm

8.0 Panel 6 - Tup 50 mm

7.0
6.0
Dent Depth , mm

5.0
4.0
3.0
2.0
1.0
0.0
4 9 14 19 24
Impact Energy, J

Fig. 11. Experimental dent depth for panels 3 and 6 for impact with increasing energy levels and with different
impactor weights (pandown sandwich structure).

Table 7
Dent depth and delamination area comparison for panels 3 and 6

Impact Required Tup size Dent depth (mm) Difference Damage area (mm2) Difference
location drop energy (mm) (%) (%)
(J)
Panel 3 Panel 6 Panel 3 Panel 6

A 5 16 1.96 0.32 83.67 329.3 259.9 21.08


B 5 16 2.36 0.4 83.05 293.7 291.3 0.82
C 5 16 1.92 0.8 58.33 260.2 291.3 11.95
D 8 16 4.08 1.32 67.65 406.5 324.6 20.15
E 8 16 4.02 1.12 72.14 448.2 435.2 2.90
F 8 16 4.02 1.48 63.18 406.5 396.5 2.46
G 10 16 8 1.8 77.50 585.3 475.7 18.73
H 10 16 6.8 2.1 69.12 537.6 517.9 3.66
J 10 16 6.42 2.14 66.67 585.3 475.7 18.73
K 5 50 0.07 0.01 85.71 687 517.9 24.61
L 5 50 0.45 0.04 91.11 796.7 562 29.46
M 5 50 0.11 0.02 81.82 687 562 18.20
W+N 20 50 0.42 0.12 71.43 1545.7 809.3 47.64
O 20 50 0.37 0.08 78.38 1467.4 864.1 41.11
P 20 50 0.35 0.09 74.29 1317 920.8 30.08
Q 10 50 1.32 0.98 25.76 1792.6 1039.5 42.01
R 10 50 1.43 0.96 32.87 1967.4 1165.4 40.76
S 10 50 1.37 0.97 29.20 1708.3 1101.5 35.52
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1315

The damage area from the C-scan and dent depth results are presented in Figs. 1217 for
increasing energy levels and impactor masses. Damage areas are qualitative indications of
damage, showing in-plane extension of delaminations, matrix cracks and bre fracture as a
whitened region. In these tests, the damage in specimen was almost entirely interlaminar
delamination as evidenced by the small permanent indentation and minimal back face damage

Fig. 12. C-scan image of damages on panel 5.

Fig. 13. Exploded view of FE Model.


ARTICLE IN PRESS

1316 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

D
Stress 

pl

Strain 

Fig. 14. Schematic stressstrain curve for honeycomb.

Fig. 15. Strain energy honeycomb and composite skin.

Fig. 16. (a) Final overall deformation and (b) localized deformation.
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1317

Fig. 17. (a) Transverse displacement (m) at impact location versus time and (b) through-the-thickness displacement (m)
at impact location.

Fig. 18. Damage area: (a) matrix cracking mode(0 failed1 elastic) and (b) experimental versus FE results.

(not shown). It is worth pointing out that each of the failure mechanisms absorb different
amounts of energy and therefore the damage area alone does not accurately characterize the
damage level or the amount of energy absorbed in the fracture process. However, the presence of
a signicant damage, detectable by C-scanning, regardless of the failure mechanism, is sufcient
to give good indication of delamination threshold levels. However, these results cannot be used to
interpret directly the compressive residual strength of the plate (Fig. 18).

3. Finite element analysis

Considering the high cost and the long time delay to obtain test results, the computer
simulation (virtual design) can be an effective aid to design aircraft sandwich structures. FEM
ARTICLE IN PRESS

1318 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

offers the possibility to evaluate and to optimize aircraft structures in the preliminary design stage
and to determine the inuence of each design parameter on the structural behavior under certain
conditions. FEA can be used to compare a proven construction with an alternate construction, to
determine where the differences occur and to predict distribution of internal stress and strains that
are difcult to be measured experimentally. In this work, FEA was employed to understand how
sandwich structures fail and to identify critical parameters. This offered the opportunity to
evaluate the capability of numerical simulation tools to reduce the cost of manufacturing and
design of future sandwich structures.
During impact of a projectile on a sandwich panel, the target experiences localized deformation
at impact locations as well as overall structural deformation. Classical beam, plate, and shell
theories generally do not account for the local deformation near the impact site. The localized
deformations in the studied sandwich panel were estimated from a large-scale 3D numerical
analysis of the dynamic contact problem and by introducing contact law that accounts for the
effect of inelastic deformation and damage induced in the impact zone.
Transient dynamic FEA was performed using the LS-DYNA3D nite element code to analyse
the impact behaviour of the sandwich panels. Results relative to panel 5 are presented. The nite
element mesh used for the calculations is presented in Fig. 13. A ner mesh was employed at the
impact location in order to better capture the local damaged state of the sandwich panel. The
mesh was made progressively coarser towards the boundary. The panel was modelled as a
combination of solid and shell elements. A four-node Belytschko-Tsai shell element was used for
modelling the composite facing skin while eight node solid element were employed to model
honeycomb core. A total of 55,605 elements were used. Loads were applied to the model by
specifying an initial velocity of the impactor. The impacting speed assumed was 30.5 m/s (Table 4,
Energy 10 J, impact location J). Lay-up sequence, ply material and sandwich properties are shown
in Tables 810. Simple supported boundary conditions were specied at the bottom of the support
ring. An elasto-plastic with isotropic hardening constitutive material model was assumed for the

Table 8
Ply layup and geometrical dimensions of the plate

Lay-up sequence No. of plies Thickness (mm)

[90, 45, 45, 90] 4 0.3 mm

Table 9
Mechanical properties of laminate

Moduli parameter Strength parameters

Longitudianal modulus, E x (GPa) 47 Long. tension, X t (MPa) 247


Transverse modulus, E y (GPa) 47 Long. compression, X c (MPa) 151
Shear modulus, G xy (GPa) 14 Trans. tension, Y t (MPa) 247
Poissons ratio, nxy 0.054 Trans. compression, Y c (MPa) 151
In-plane shear, S c (MPa) 99
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1319

Table 10
Mechanical properties of honeycomb core

Moduli parameter Strength parameters

Young modulus compacted material, E z (GPa) 1 Stabilised compression, Z c (Plateau stress) (MPa) 7.76
Poissons ratio, compacted material, nxy 0.25
Transverse modulus, E z (MPa) 414 In-plane shear, Sl (expansion direction) (MPa) 2.66
Shear modulus direction of expansion, G l (MPa) 89 In-plane shear, Sr (ribbon direction) (MPa) 1.38
Shear modulus ribbon direction, G r (MPa) 44
Densication strain 0.8 Yield stress fully compacted material (MPa) 3.19

impactor and support ring. Surface contact algorithms were introduced into the model to simulate
contact between the impacting body and the top facing composite skin and bottom facing
composite skin sandwich and support ring.

4. Failure analysis

The damage of composite laminates due to transverse impact loading generally consists of
several modes such as matrix tensile cracking, matrix compressive/shear failure bre matrix
debonding, bre brakeage (tensile or compressive), delamination, etc. Among these damage
modes, delamination is the most important mode. However, visible delamination was not
observed in the preliminary tests. Therefore, the Carbon Epoxy AS4/8552 composite laminate was
modelled as an orthotropic material with damage occurring by matrix cracking, compressive
failure and bre breakage. Moreover, transverse impact rst cause matrix crack in laminate layer
with brittle matrix. The Chang and Chang failure criterion [2423] was utilised to predict damage
in the composite faceskin.

4.1. Matrix cracking criterion

Two different failure criteria were used depending on whether transverse normal stress, s22, is in
tension or in compression. The failure index, em is dened in terms of ply transverse tensile
strength, Yt, ply transverse compressive strength Yc, and in-plane shear strength S , and is
expressed as
e2m n s22 =n Y c n Y c =2n S2  1 n s22 =n S2 n t12 =n S2 for s22 o0 (1)
and
e2m n s22 =n Y t 2 n t12 =n S2 for s22 40 , (2)
where the subscripts of 1 and 2 are the local coordinates of the nth layer normal to the bre
direction and the out-of plane direction, respectively, and t12 and s22 are the averaged in-plane
shear and transverse stresses, respectively, within the nth ply. Matrix failure was assumed to occur
when the failure index em exceeds unity.
ARTICLE IN PRESS

1320 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

4.2. Fibrematrix shearing failure criterion

Fibrematrix shearing failure is assumed to be dependent on a combination of axial stress, s11


and shear stress t12 and is expressed as follows:
e2s n s11 =n X t 2 n t12 =n S2 for s11 40 and n s11 =n X t on t12 =n S2 (3)

e2s n s11 =n X c 2 n t12 =n S2 for s11 o0 and n s11 =n X c on t12 =n S2 , (4)


where es is the failure index for brematrix shearing, X t the tensile strength along the bre
direction and X c the compressive strength along the bre direction. Fibrematrix shearing
dominated failure occurs when the failure index es exceeds unity.

4.3. Fibre breakage failure criterion

Fibre breakage failure occurs in tension due to the combination of axial stress and shear stress
while the failure in compression is governed by buckling and expressed in terms of only axial
stress. The criterion for bre breakage failure is expressed as follows:
e2f n s11 =n X t 2 n t12 =n S2 for s11 40 , (5)

e2f n s11 =n X c 2 n t12 =n S2 for s11 o0 . (6)

5. Honeycomb core material properties

The stressstrain curve of the Nomex core material is shown schematically in Fig. 14. Material
properties are summarized in Table 10. Three different regimes can be observed on this curve, and
each regime is associated with a mechanism of deformation: (1) At low strains (0.55%) a linearly
elastic region, (2) a region corresponding to progressive crushing at nearly constant stress level
(the plateau stress spl ), and (3) a region of rapidly increasing stresses with further deformation
due to the fact that the cell walls are forced into contact with each other. For high density
honeycombs, the linearly elastic region ends when the honeycomb cell walls start fracturing. The
initiation of elastic buckling does not cause a complete loss of stiffness since after the initial
fracture of the cell walls is observed, a sudden drop from a collapse stress to a steady crush
strength occurs as shown in Fig. 4. Small oscillations may occur about the mean stress value due
to consecutive local buckling. The densication strain D typically ranges from 50% to 80%. The
area under the stressstrain curve up to densication is the useful energy that can be absorbed per
unit volume.

6. Results and discussion

The structural response of sandwich panel to foreign object damage is distinct from that of
laminated composites. Impact response and energy absorption of sandwich composites is largely
dominated by crushing of the sandwich core [24]. This is conrmed by Fig. 15, where the strain
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1321

Table 11
Comparison of experimental and numerical results

Impact tests Finite element method Error (%)

Dent depth (mm) 1.48 1.52 2.6


Area of delamination (mm2) 396.5 354.8 10.5

Fig. 19. Impact force versus indentation displacement.

energy of the honeycomb is eight times higher than the impact energy absorbed by the composite
facing skin. The target experienced localized deformation in the vicinity of the point of impact as
well as overall structural deformation as shown in Fig. 16.
Fig. 17 show through the thickness displacement and transverse displacement (indentation) at
impact location dened as the relative displacement between the indentor and sandwich panel.
The duration of the impact is 0.0004 s since at this time the impactor started bouncing back. Once
the impactor was bounced back, no permanent deformation can occur therefore the results will
just show the elastic motion of the structure. In order to reach a state of static equilibrium to nd
the nal dent depth, dynamic relaxation was used [24]. Articial mechanism were introduced to
damp out the vibration present inside the material. Damping [70] was applied to all the nodes so
as to achieve zero velocity for all them.The nal dent depth was 1.52 mm.
The nal damaged state is shown in Fig. 18. In particular, failure indicator of tensile matrix
mode (cracking) is shown. A value of one indicates no damage within the composite plies, while a
value less than unity show damage within the composite plies. Since delamination is always
associated with initial matrix cracking, this failure criteria indicator was compared with
experimental C-Scan damage state. The calculated damaged area was 354.8 mm2. In Table 11
numerical simulation and experimental results are compared. Good agreement was obtained. The
nal dent depth was obtained with an error of 2.6%. The delamination area was estimated with an
error of 10.5%.
ARTICLE IN PRESS

1322 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

These results showed that the LS-Dyna3d code simulation agreed fairly well with experimental
tests at low-velocity impact. Therefore, this allowed for calculation of the contact force versus
indentation relationship during impact (Fig. 19). In particular, it shows that sandwich structures
response is signicantly different from laminated composites [9,10], showing a nonlinear
behaviour with different loading and unloading characteristics.

7. Conclusion

The present paper describes the results from an experimental and numerical study of low-
velocity impact responses of composite sandwich panels representing the Trent Existing Fan Cowl
Door structure and the Modied Fan Cowl Door structure. The main objective was to understand
the load distribution in damaged sandwich structure and to study the failure mechanisms of such
a structure in the presence of impact damage.
The panels are representative of a Fan Cowl Door of a large transport aircraft underwing
engine nacelle. Several impact energy level were investigated in order to inict damage conditions
in the test panels corresponding to a BVID in the impacted composite facesheet.
Results from the impact-damage C-scan test indicate that signicant internal damage occurs at
relatively low impact-energy levels, which can signicantly reduce the residual strength of the
panel. Dent depth and delamination area within the plies were measured experimentally. Good
agreement was obtained by comparing numerical and experimental results, in particular, the
numerical simulation was able to predict dent depth with an error of 2.6% and delamination area
with an error of 10.5%. This proved that a correct mathematical model can yield signicant
information for the designer to understand the mechanism involved in the low-velocity impact
event, prior to conducting tests, and therefore to design an impact-resistant aircraft structure.
The present work will be benecial to our further research about the damage extension
sandwich structures with different skin lay-ups under low-velocity impact and compression after
impact behaviour.

Acknowledgements

We would like to thank Westland Aerospace for supporting this work, and for conducting the
tests at Cowes, Isle of Wight, UK

References

[1] Tomblin J, Lacy T, Smith B, Hoopes S, Vizzini A, Lee S. Review of Damage Tolerance for Composite Sandwich
Airframe Structures. Final Report, DOT/FAA/AR-99/49, August 1999.
[2] Krajcinovic D. Damage Mechanics. Amsterdam: North Holland/Elsevier; 1996.
[3] McGowan DM, Ambur DR. Compression Response of a Sandwich Fuselage Keel Panel With and Without
Damage. NASA Technical Memorandum 110302.
[4] McGowan DM, Ambur DR. Damage-tolerance characteristics of composite fuselage sandwich structures with
thick facesheets. NASA Technical Memorandum 110303.
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1323

[5] Rouse M, Ambur DR, Bodine J, Dopker B. Evaluation of a composite sandwich fuselage side panel with damage
and subjected to internal pressure. NASA Technical Memorandum 110309.
[6] Abrate S. Localized impact on sandwich structures with laminated facings. Applied Mechanics Review
1997;50(2):6981.
[7] Chang FK, Chang KY. A progressive damage model for laminated composite containing stress concentrations.
Journal of Composite Materials 1987;21:83455.
[8] Choi HT, Chang FK. A model for predicting damage in graphite/epoxy laminated composites resulting from low
DYNA3D subjected to low velocity point impact. Journal of Composite Materials 1992;26:213469.
[9] Abrate S. Impact on laminated composite materials. Applied Mechanics Review 1991;44(4):15590.
[10] Abrate S. Impact on laminated composites: recent advances. Applied Mechanics Review 1994;47(11):51744.
[11] LS-DYNA3D.Users Manual. Ver. 950 Livermore Software Technology Corporation, Livermore, California;
1999.
[12] Allen HG. Analysis and design of structured sandwich panels. Oxford: Pergamon Press; 1969.
[13] Plantema FJ. Sandwich construction. Chichester: Wiley; 1966.
[14] Vinson JR. The behavior of sandwich structures of isotropic and composite materials. Lancaster: Technomic;
1999.
[15] Dorey G. Impact damage in compositesdevelopment, consequences, and prevention. In: Proc 6th Int Conf on
Composite Materials and 2nd European Conf on Composite Materials, vol. 3. Imperial College, London, 1988.
p. 3.13.26.
[16] Abrate S. Impact on laminated composites: recent advances. Applied Mechanics Review 1994;47:51744.
[17] Li X, Carlsson LA. The tilted sandwich debond (TSD) specimen for face/core interface fracture characterization.
Journal of Sandwich Structural Materials 1999;1(1):6575.
[18] Singh RP, Shukla A. Subsonic and intersonic crack growth along a bimaterial surface. Journal of Applied
Mechanics 1996;63:91924.
[19] Abrate S. Impact on composite structures. Cambridge: Cambridge University Press; 1998. p. 24257.
[20] Vizzini AJ, Lagace PA. The buckling of a delaminated sublaminate on an elastic foundation. Journal of Composite
Material 1987;21:110617.
[21] Gioia G, Ortiz M. Delamination of compressed thin lms. Advanced Applied Mechanics 1997;33:11992.
[22] Kadomateas GA. Post-buckling and growth behavior of face-sheet delaminations in sandwich composites. In:
Rajapakse YDS, Kadomateas GA, editors. Thick composites for Load Bearing Structures, AMD-vol. 235, 1999.
p. 5160.
[23] Wu HT, Springer GS. Measurements of matrix cracking and delamination caused by impacted on composite
plates. Journal of Composite Materials 1988;22:51832.
[24] Choi HY, Wu HT, Chang F-K. A new approach toward understanding damage mechanisms and mechanics of
laminated composites due to low-velocity impact: part IIanalysis. Journal of Composite Materials
1991;25:101238.
[25] Lambros J, Rosakis AJ. An experimental study of the dynamic delamination of thick ber reinforced polymeric
matrix composite laminates. Experimantal Mechanics 1997;37:3606.
[26] Lee J-W, Daniel IM. Progressive transverse cracking of crossply composite laminates. Journal of Composite
Materials 1990;24:122543.
[27] Karandikar P, Chou T-W. Characterization and modeling of microcracking and elastic moduli changes in nicalon-
CAS composites. Composite Science Technology 1992;46:2539.
[28] Hutchinson JW, Suo Z. Mixed mode cracking in layered materials. Advanced Applied Mechanics 1992;29:63191.
[29] Shenoy VB, Schwartzman AF, Freund LB. Crack patterns in brittle thin lms. International Journal of Fracture
2000;103:17.
[30] Ju JW. A micromechanical damage model for uniaxially reinforced composites weakened by interfacial arc
microcracks. Journal of Applied Mechanics 1991;58:92330.
[31] Geubelle PH, Baylor JS. Impact-induced delamination of composites: a 2D simulation. Composites Part B
1998;29B:589602.
[32] Vinson JR, Sierakowski RL. The behavior of structures composed of composite materials. Dordrecht: Martinus,
Nijhoff Publishers; 1986.
ARTICLE IN PRESS

1324 M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325

[33] Williams ML. Stress singularities resulting from various boundary conditions in angular corners in extension.
Journal of Applied Mechanics 1952;19:5268.
[34] Cantwell WJ, Morton J. The impact resistance of composite materialsa review. Composites 1991;22:
34762.
[36] Mines RAW, Roach AM, Jones N. High velocity perforation behaviour of polymer composite laminates.
International Journal of Impact Engineering 1999;22:56188.
[37] Needleman A, Rosakis AJ. The effect of bond strength and loading rate on the conditions governing the
attainment of intersonic crack growth along interfaces. Journal of the Mechanics and Physics of Solids
1999;47:241149.
[38] Shih WK, Jang BZ. Instrumented impact testing of composite sandwich panels. Journal of Reinforced Plastics
Composite 1989;8:27198.
[39] Kim CG, Jun EJ. Journal of Composite Materials 1992;26(15):2247.
[40] Chun LU, Lam KY. Behavior of uniform anisotropic beams of rectangular section under transverse impact of a
mass. Shock and Vibration 1997;4(2):12541.
[41] Ravi-Chandar K, Knauss WG. An experimental investigation into dynamic fracture: III. On steady-state crack
propagation and crack branching. International Journal of Fracture 1984;26:14154.
[42] Riley WF, Dally JW. A photoelastic analysis of stress wave propagation in a layered model. Geophysics
1966;31:88199.
[43] Rosakis AJ, Samudrala O, Singh RP, Shukla A. Intersonic crack propagation in bimaterial systems. Journal of the
Mechanics and Physics of Solids 1998;46:1789813.
[44] Semenski D, Rosakis AJ. Dynamic crack initiation and growth in light-core sandwich composite materials. In:
Proceedings of the 17th Danubia-Adria Symposium on Experimental Mechanics in Solid Mechanics, Prague; 1999.
p. 297300.
[45] Sun CT, Rechak S. Effect of adhesive layers on impact damage in composite laminates. In: Whitcomb JD, editor.
Composite materials: testing and design (eighth conference). Philadelphia: ASTM STP 972, American Society for
Testing and Materials; 1988. p. 97123.
[46] Takeda N, Sierakowski RL, Ross CA, Malvern LE. Delamination-crack propagation in ballistically impacted
glass/epoxy composite laminates. Experimental Mechanics 1982;22:1965.
[47] Valenti M. Stealth on the water. ASME Mechanical Engineering 2001;123:5661.
[48] Walter ME, Ravichandran G. Experimental simulation of matrix cracking and debonding in a model brittle matrix
composite. Experimental Mechanics 1997;37:1305.
[49] Wen HM, Reddy TY, Reid SR, Soden PD. Indentation, penetration and perforation of composite laminates
and sandwich panels under quasi-static and projectile loading. Key Engineering Materials 1998;141143:50152.
[50] Xu LR, Rosakis AJ. Impact failure characteristics of sandwich structures; part II: effects of impact speeds and
interfacial bonding strengths. International Journal of Solids Structures 2002;39(16):423748.
[51] Xu LR, Rosakis AJ. An experimental study on impact-induced failure events in homogeneous layered materials
using dynamic photoelasticity and high-speed photography. Optics and Lasers Engineering 2003;40(4):26388.
[52] Xu LR, Rosakis AJ. International Journal of Solids and Structures 2002;39:421535.
[53] Robinson P, Davies GAO. Impactor mass and specimen geometry effects in low velocity impact of laminated
composites. International Journal of Impact Engineering 1992;12(2):189207.
[54] Goldsmith W, Sackmann JL. An experimental study of energy absorption in impact of sandwich plates.
International Journal of Impact Engineering 1992;12(2):24162.
[55] Rosenfeld MS, Gause LW. Compression fatigue behavior of graphite/epoxy in the presence of stress raisers. In:
Fatigue of brous composite materials, ASTM STP723; 1981. p. 17496.
[56] Horrigan DPW, Aitken RR, Moltschaniwskyj G. Modelling of crushing due to impact in honeycomb sandwiches.
Journal of Sandwich Structural Materials 2000;2:13151.
[57] Rhodes MD. Impact fracture of composite sandwich structures. Proceedings of ASME/AIAA/SAE, 16th
Structure, and Structural Dynamic Material Conference 1975:3116.
[58] Nettles AT, Hodge AJ. Impact testing of glass/phenolic honeycomb panels with graphite/epoxy facesheets.
In: Proceedings of the 35th International SAMPE Symposium and Exhibition, Anaheim, CA, April 25, 1990.
p. 143040.
ARTICLE IN PRESS

M. Meo et al. / International Journal of Mechanical Sciences 47 (2005) 13011325 1325

[59] Mines RAW, Worrall CM, Gibson AG. The static and impact behaviour of polymer composite sandwich beams.
Composites 1994;25(2):95110.
[60] Akil Hazizan MD, Cantwell WJ. The low velocity impact response of foam-based sandwich structures.
Composites, Part B 2002;33:193204.
[61] Mines RAW. Impact energy absorption of polymer composite sandwich beams. Key Engineering Materials
1998;141143:55372.
[62] Herup EJ, Palazotto AN. Low-velocity impact damage initiation in graphite/epoxy/nomex honeycomb-sandwich
plates. Composite Science Technology 1997;57:158198.
[63] Vaidya UK, Hosur MV, Earl D, Jeelani S. Impact response of integrated hollow core sandwich composite panels.
Composites: Part A 2000;31:76172.
[64] Davies GAO, Zhang X. Impact damage prediction in carbon composite structure. International Journal of Impact
Engineering 1994;16(1):14970.
[65] Nettles AT, Lance DG. Enhancement of impact damage tolerance of composite laminates. Composite Engineering
1993;3:38394.
[66] Hazizan MA, Cantwell WJ. The low velocity impact response of foam-based sandwich structures. Composites:
Part B 2002;33:193204.
[67] Ambur DR, Cruz J. Low speed impact response characteristics of composite sandwich panels. Proceedings of the
36th AIAA/ASME/ASCE/AHS/ASC Structure, Structural Dynamics Material Conference 1995:26819.
[68] Davies P, Choqueuse D, Verniolle P, Prevosto M, Genin D, Hamelin P. Impact behaviour of composite sandwich
panels. In: Proceedings of the European Symposium on Dynamic Fracture Polymer Composite, Sardinia 1993;
paper D3.
[69] Nemes JA, Simmonds KE. Low velocity impact response of foamcore sandwich composites. Journal of Composite
Materials 1992;26:50019.
[70] Rao SS. Mechanical vibrations. New York: Pearson Prentice-Hall; 2004.

You might also like