Available online at www.sciencedirect.

com

Composite Structures 83 (2008) 335–340

www.elsevier.com/locate/compstruct

A new hybrid concept for sandwich structures

a,*
A.G. Mamalis , K.N. Spentzas b, N.G. Pantelelis b, D.E. Manolakos a, M.B. Ioannidis a

a
Laboratory of Manufacturing Technology, National Technical University of Athens, 9, Iroon Polytechniou Avenue, 15780 Athens, Greece
b
Vehicles’ Laboratory, National Technical University of Athens, 9, Iroon Polytechniou Avenue, 15780 Athens, Greece

Available online 17 May 2007

Abstract

Sandwich structures are considered as optimal designs for carrying bending loads and can be either metal (aluminium faces and hon-
eycomb or metal foam cores) or polymer structures (composite faces with polymer foam cores). In this paper, a new hybrid sandwich
structure has been developed by combining most of the advantages of metallic and polymeric materials while avoiding some of their main
disadvantages. For this new concept metal sheets are used at the outer surfaces to maximize rigidity while introducing in between light-
weight cores adhesively bonded to keep the whole structure together. Furthermore, composite or wood layers may be used as interme-
diate layers to improve impact resistance. Potential methods for the manufacturing of this new structure are based on compression under
vacuum. The results include the study of several panel configurations theoretically based on Finite element analysis and on the modified
simplified equations and experimental results in the most representative cases of the study.
 2007 Elsevier Ltd. All rights reserved.

Keywords: Sandwich structure; Finite element analysis; Three-point bending; Metal skins; Foam core

1. Introduction erably the structural performance of either of their materials

Although sandwich structures are considered as optimal
set-ups for carrying bending loads, their structural design
and manufacturing entail knowledge and experience [1].
Many researchers have studied the variety of the failure
mechanisms of the sandwich structures either under static
loads, e.g. [1,3,6,7], or under dynamic loads, e.g. impact
[7,8], where the importance of the materials’ choice is
prevailing.
In order to tackle some of the weaknesses of the existing
materials for sandwich structures some researchers have stud-
ied successfully certain material combinations either in the
core [8–10] or introducing internal layers [11] or even by mix-
ing of completely different materials such as metals and poly-
mers with applications that may be found in the building
sector [12] (wall and roof panels) and in the automotive sector
(thin steel or aluminium face sheets and polymer cores) [13].
These hybrid sandwich structures although improving consid-
can not be used as primary structures.
In this paper, a new hybrid sandwich structure has been
developed by combining most of the advantages of metallic
and polymeric materials while avoiding some of their major
disadvantages. The design concept is to use metals at the
face sheets in order to maximize rigidity and extremely
lightweight cores while introducing an intermediate layer
from composite materials or wood between the face sheets
and the core appropriately bonded to keep the whole struc-
ture together [14,15].
The rest of the paper is organised as follows: in Section 2
the classic sandwich theory is analysed and in Section 3 the
new hybrid sandwich concept is analysed while in Section 4
other structural issues such as impact behaviour are
addressed. In Section 5 theoretical and experimental results
are presented for the proof of concept and in Section 6 con-
clusions are drawn.
2. Classic sandwich theory

*
Corresponding author. Tel.: +30 210 7723688. In general, to study and optimise a structure its basic
E-mail address: mamalis@central.ntua.gr (A.G. Mamalis). features need to be modelled and analysed. For this reason,

0263-8223/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compstruct.2007.05.002
336 A.G. Mamalis et al. / Composite Structures 83 (2008) 335–340
a narrower version of a panel (Fig. 1a) with equivalent
loads and conditions will be studied extensively in theoret-
ical simulations and experimental tests.
Considering the basic beam theory in a typical three-
point bending situation the mid-point deflection is given as
PL3 PL
d¼ þ ð1Þ
48ðEIÞeq 4ðAGÞeq
where P is the load, L and b are the beam’s span and width,
E and G are the materials’ Young’s and shear modulus. In
the conventional sandwich structure there are two materi-
als composing the structure: the faces and the core so the
contribution of each material can be analysed straightfor-
ward. As in the present case we will study sandwich struc-
tures with
(a) face thickness considerably smaller than the core
thickness (tf  tc)
(b) core stiffness significantly smaller than the faces stiff-
ness (Ef  Ec)
(c) length considerably larger than thickness (L  tc)
the mid-point deflection for long beams in three-point
bending (Eq. (1)) becomes:
PL3
d¼ ð2Þ
24Ef tf t2c b
The failure of a sandwich structure is a very complicated
phenomenon and may be due to various failure mecha-
nisms in one of the materials that composes the structure.
These mechanisms have been analysed and tested by vari-
ous researchers [1,2,12] and the allowable mid-point load-
ings in a three-point test with respect to each failure
mode can be summarised as [2]:
Fig. 1. Sandwich structure section showing the thin outer faces (top and
bottom) and the inner thick core. (a) Classic concept. (b) Hybrid concept
showing also the intermediate layers.
P 4tf tc rf
Face microbuckling ðFMBÞ : ¼ ð3Þ
b L
P 2tf tc p
3
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Face wrinkling ðFWÞ : ¼ E f E c Gc ð4Þ
b L
P
Core shear ðCSÞ : ¼ 2tc sc ð5Þ
b s ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P 2 2 3
3 p Ef r t tc P 16E E t 3
t
c f f c f c
IndentationðIndÞ : ¼ or ¼
b L b 3L2
ð6Þ
where b; L; tf ; tc ; P ; Ef ; Ec ; Gc ; rf ; rc ; sc are, respectively, the
width and length of the beam, the face sheets’ and core’s
thicknesses, the maximum point load the Young’s modulus
for the skins and the core, the core shear modulus, the max-
imum stress for the face sheets and the maximum compres-
sion and shear stresses for the core.
Eqs. (3)–(6) give a rough idea of how each material con-
tributes to the failure of the whole structure. For example if
a low cost foam core is used it is very likely that the beams
will fail either due to core shear or due to indentation espe-
cially in three-point loading.
However, this theoretical analysis does not consider
three basic issues: the interaction between failure mecha-
nisms, the role of the adhesion between the materials and
the contribution of a third material, between the face sheet
and the core. The first two issues are very complicated and
need separate studies but at the present paper the role of
the third intermediate material will be addressed.
3. The hybrid sandwich concept
The classic sandwich structure consists of two thin and
stiff face sheets and a thick, lightweight and stiff enough
material in between. Although this structure seems optimal
in practice the large differences between the structural prop-
erties of the materials in contact create many problems.
Moreover the cost of the existing core materials is very high
pushing towards the selection of the lowest possible perfor-
mance cores. Of course the final selection of the core mate-
rial is a compromise between these limits, the cost, the
weigh but also other characteristics of the core such as
the impact resistance, fatigue, moisture, etc. Because the
homogeneity of the core materials makes this compromise
very tough, some attempts to develop either graded cores
or two stacked foam cores [4,9–11] have been presented.
A compromise between the homogeneous and graded
cores is the introduction of an intermediate layer as can
be seen in Fig. 1b. This layer of thickness ti should be much
stiffer than the core material, lightweight enough and pref-
erably much thicker than the face sheet. This intermediate
layer will allow the use of very thin face sheets, e.g. metals,
and very cheap cores, e.g. XPS or PUR, at the expense of a
slightly higher weight. Furthermore, if a common material
is chosen, e.g. wood, this intermediate layer will also
decrease the cost considerably. However, a very good adhe-
sion between all these three materials has to be ensured.
 A.G. Mamalis et al. / Composite Structures 83 (2008) 335–340 337

The advantage of this intermediate layer may not be
seen in face microbuckling or core shear which are practi-
cally properties that do not include interaction with other
materials but it will improve face wrinkling and indenta-
tion which are of equal importance especially in the case
of cheap foam cores. In order to quantify this improvement
Eqs. (4) and (6) can still hold if the mechanical properties
of the foam are substituted by the corresponding properties
of the intermediate layer so Eqs. (4) and (6) are trans-
formed to
P 2tf tc p ffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3
Face wrinkling : ¼ Ef Ei Gi ð7Þ
brffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L
2 3
P 2
3 p Ef r t tc
Indentation : ¼ i f
ð8Þ
b L
4. Other design issues
Besides the rigidity of the panel other issues such as the
impact and fatigue resistance should also be taken into
consideration. The sandwich structure, as such, is vulnera-
ble to impact, so special consideration for these loads is
required. The impact vulnerability of the sandwich panel
is due mainly to the debonding between the skins and the
core. This debonding is due to the large deformations
between the flexible foam core and the skin leading at least
to the destruction of the adhesive layer and most of the
times the internal failure of the outer surface of the core.
In order to moderate this difference at the present design
an intermediate layer has been introduced. Initially, a thin
glassfibre/epoxy layer was used, but after several impact
tests, a thicker but lighter plywood layer was chosen.
Fatigue is also an important issue when designing for
the transportation sector. The most vulnerable material
with the most critical role is the adhesive which for the
present study was an ordinary epoxy resin and not a special
adhesive. However, the foam core is also very vulnerable to
fatigue and its characteristics may fall by 50% after hun-
dreds of thousand cycles. So it is very important to
decrease as much as possible the loads of the foam core
in order to extend its life expectancy. Towards the improve-
ment of the fatigue behaviour of the core contributes the
addition of the plywood layer.
5. Results
5.1. Simulations
In order to explore the potential of the new sandwich
concept, a typical panel, 1 m long and 0.5 m wide, has been
analysed using finite elements under linear central bending
load of 6000 N in total in a typical three-point bending
case. The sandwich composite structure has been modelled
using a single surface with classical shell elements with a
series of material layers including the foam core. The accu-
racy of the simulations would be increased only marginally
by modelling the foam core with solid elements in contrast
to the large solution overhead. In Table 1 all the studied
configurations has been analysed together with their struc-
tural performance characteristics and limit loads while in
Table 2 their corresponding mass, indicative cost are also
provided.
Looking at the single material configurations (cases 1
and 2) we can observe that deflections and stresses are very
high meaning that these panels cannot withstand the
applied bending load. Moving to sandwich structures, the
first case is to study the glass epoxy composite skins with
PVC core of 25 mm thick (case 3) where deflection has
dropped but not to acceptable level while stresses at the
core material are still high. Furthermore, the limit loads
for the sandwich failure modes are low not allowing any
safety margins. If we triple the thickness of the skin and
increase by 50% the foam core thickness (case 4), we end
up with significantly improved structural performance,
i.e. acceptable deflection at 11.5 mm and significantly lower
stresses and higher limit loads with respect to the sandwich
failure modes but with two disadvantages: the first is the
much higher cost and the 0/90 biaxial glass fabric that

Table 1
Performance comparison among various 1 · 0.5 sandwich panel configurations for a mid span linear load of 6000 N (St: stainless steel, Al: standard
aluminium, GE: glass (0/90) epoxy composite and W: plywood)
Case Layers + thickness Structural performance
FEA simulations Failure modes
Face Intermediate Core Deflection (mm) Von Mises stresses in MPa Maximum mid-point loads (kN) per failure mode
(mm) (top/bottom faces) Face Intermediate Core FMB FW CS Inde
1 St: 6 67.8 503 – – – –
2 Al: 8 80.1 283.3 – – – –
3 GE:1.2 25 57.0 97.2 1.22 18 6 18 5
4 GE:3.6 38 11.5 34 0.26 82 30 27 35
5 St: 0.5 GE:1.2 25 12.9 183.6 18.6 0.52 8 178 18 78
6 Al: 2/1 W:4 25 7.4 44.1 0.56 0.35 23 244 18 125
7 St: 0.5 GE:1.2 38 7.8 126 12.9 0.27 11 270 27 96
8 St: 0.8 GE:1.2 38 6.8 82 8.38 0.26 18 432 27 195
9 St: 0.5 GE:1.2 50 4.3 61.2 6.2 0.14 15 356 35 111
10 Al: 2/1 W:4 60 3.3 29.5 1.47 0.19 58 624 45 201
338 A.G. Mamalis et al. / Composite Structures 83 (2008) 335–340

Table 2 the thickness of the core as can be seen in case 9 where

Mass, indicative material cost and maximum deflections at experimental its thickness was increased to 50 mm. In this case the cor-
three-point bending tests for the 10 panel configurations
responding deflection was dropped to 4.3 mm but, because
Case Mass (kg) Cost (€) Yield (mm @ N) of the PVC core, this configuration is still expensive. In the
1 23.4 82 last case (case 10) two critical design parameters have been
2 11.2 45 modified: the first is the further thickening of the core
3 2.7 27 23 @ 2900
4 7.3 61
allowing for the use of cheaper cores such as the extruded
5 6.7 42 11 @ 5650 polystyrene foam core as the maximum core stress has
6 6.4 27 9 @ 4800 dropped below 0.2 MPa and the second is the replacement
7 7.0 52 of the glass fibre epoxy layer with a thicker plywood sheet.
8 9.3 61 The existence of the composite layer under the metal sheet
9 7.2 62
10 7.8 32 15 @ 22,750
had a foam core protection role from impact loads provid-
ing an intermediate structural zone between the very thin
and stiff metal sheet and the very flexible foam core. The
has been used. As the loads in a panel are not oriented at replacement of the composite layer with plywood contrib-
all, the use of oriented fabrics are not recommended mean- utes towards the improvement and the simplification of
ing that we should use a near-isotropic lay-up in order to the production process maintaining the structural role of
cope with all the possible load cases. the composite (Fig. 2).
In order to optimise the structural performance of the
panel, metal skins were introduced. By bonding properly 5.2. Experiments
a 0.5 mm steel sheet on the outer surfaces of the case 3
panel the deflection drops more than four times and the In order to validate the theoretical simulations some of
stresses at the glass layer more than five times where the the manufactured panel were tested in actual size (1.0 m by
core stresses decreased to less than half resulting in an 0.5 m) three-point bending test. The most representative
impressive improvement as can be seen in Table 1. The configurations, cases 3, 5, 6 and 10, were tested and their
drawback is that the weight and the cost (Table 2) bending, impact and fatigue behaviour can be seen in
increased significantly, but both kept lower than those of Fig. 4. In the case 3 panel test (Fig. 3), the panel yielded
case 4 of a similar performance non metallic sandwich. very suddenly at 2900 N. When two 0.5 mm thick steel
When the core thickness is also increased (case 7), the sheets added to the skins (case 5), its rigidity increased
structural performance improved further at the expense more than four times while the failure limit load increased
of a higher cost thicker foam core. As expected, a key fac- almost 100% and the failure came from the shear failure of
tor for the improvement of the structural performance is the foam core. Although the use of aluminium instead of

Fig. 2. Three-point bending simulation of case 8: Von Mises stresses at top face.
 A.G. Mamalis et al. / Composite Structures 83 (2008) 335–340 339

Fig. 3. Experimental three-point bending of case 3.

steel and, especially, of the use of plywood (4 mm) instead

of epoxy/glass (case 6) present no significant advantage
over the more conventional case 5, as can be seen at the
overall diagram of the tested panels (Fig. 4) and do present
a large difference between the simulated and the real cases
(simulation: deflection 7.4 mm @ 6000 N, test: 9 mm @
4800 N). This difference should be due to the nonlinear
behaviour of the foam core but also that this panel had
been tested for impact resistance before doing the bending
test. Furthermore, the bending test was not concluded, so
the indicated maximum load is not a failure load. Finally,
a thicker panel was tested (case 10) with impressive results.
Even if this panel had been tested for fatigue (in a small
area of the panel, though) the rigidity of the panel was
enormous (Fig. 5) and the failure load reach almost
23,000 N at comparatively small deflection (15 mm). How-
ever, the potential of the panel was not fully revealed as the
foam core was the failed material, as can be seen in Fig. 6.
Fig. 5. Case 10: three-point bending test.
If a more advanced core was used, an even better structural
performance would be achieved.
It is very important to highlight that all four tests were behaviour relieves the core so much higher performance
performed with the same foam core material (crosslinked can be achieved.
PVC) so it is evident that improving the skins’ structural For the impact testing, the weights of the impactors
were from 0.5 to 1.0 kg and were let from 3.6 m high

Fig. 4. Load/deflection diagram of cases 3, 5, 6 and 10. Case 6 had been

tested in impact with drop weights and was stopped before yielding. Case
10 had been tested before in point fatigue at the area of failure. Fig. 6. The failure of case 10 panel was mainly due to indentation.
340 A.G. Mamalis et al. / Composite Structures 83 (2008) 335–340
resulting in impact energy from 18 J to 36 J. These tests
indicated that cases 3 (only glassfibre/epoxy skin) and 6
(aluminium and plywood skins) presented acceptable
impact resistance whereas case 5 (steel on top with glassfi-
bre/epoxy intermediate layer) presented large debonding
areas. This behaviour can be easily explained as the ply-
wood layer is much thicker than the composite layer and
because of its lower modulus of elasticity provides an
improved behaviour to low energy impact loads.
The fatigue resistance was studied again in actual size
panels using a cam mechanism for a large number of cycles
(up to 2.5 million cycles). Although this measurement facil-
ity was not standardized by any means, it did reveal the
basic phenomena of the fatigue of the panels. As expected
the more vulnerable material in the present design was the
PVC core and especially the thinner cores. This is not a sur-
prise as based on the simulation results presented in Table
1 the core stresses are much closer to the yield stress of the
PVC foam core which is 0.6 MPa resulting in core failure in
cases 3 and 5. However, the reinforcement of the core with
the plywood reduces the maximum stress as can be seen in
case 6 making the panel much more resistant to cyclic fati-
gue. The fatigue resistance is further improved with the
thickening of the foam core as the maximum load is line-
arly proportional to this thickness as can be seen from
Eq. (5). Thus if the thickness of the core is further
increased, the maximum Von Mises stresses drop below
0.2 MPa (case 10) improving considerably the fatigue
behaviour of the panel even if a cheaper/lower performance
foam core is used.
With respect to the weight and cost of the panels,
some very interesting observations can be extracted from
Table 2. For the fair comparison of the polymer compos-
ite sandwich with the other panels the case 4 panel is
used as the case 3 panel is very flexible and of low per-
formance. So comparing the case 4 panel to the case 5
panel we can observe that the former is heavier and
much more expensive for comparable performance. If
we substitute the glass/epoxy layer with plywood (case
6) the panel is even lighter and the cost drops dramati-
cally to more than half of the polymer composite panel
(case 4). For case 10 which is the best panel configura-
tion achieved, the cost and weight increased marginally
as for foam core in real applications common extruded
EPS can be also used instead of the much more expen-
sive xPVC.
6. Conclusions
A new concept of a lightweight sandwich structure for
high performance and reasonable cost has been presented.
The new concept employs thin metallic skins, low density
foams and intermediate layers to improve impact and
bonding behaviour. Finite element simulation and other
structural analysis tools were used for the optimal sand-
wich design and for the optimization of the materials’
topology of the hybrid structure. The new concept was also
proved in laboratory scale testing and further tests in fati-
gue and impact resistance are inline.
References
[1] Zenkert D. An introduction to sandwich construction. London,
UK: Chameleon Press Ltd.; 1995.
[2] Steeves CA, Fleck NA. Collapse mechanisms of sandwich beams with
composite faces and a foam core, loaded in three-point bending. Part
I: analytical models and minimum weight design. Int J Mech Sci
2004;46:561–83.
[3] Rizov V, Shipsha A, Zenkert D. Indentation study of foam core
sandwich composite panels. Compos Struct 2005;69:95–102.
[4] Borsellino C, Calabrese L, Valenza A. Experimental and numerical
evaluation of sandwich composite structures. Compos Sci Technol
2004;64(10–11):1709–15.
[6] Abrate S. Impact on composite structures. Cambridge, UK: Cam-
bridge University Press; 1998.
[7] Mamalis AG, Manolakos DE, Ioannidis MB, Papapostolou DP. On
the crushing response of composite sandwich panels subjected to
edgewise compression: experimental. Compos Struct 2005;71:246–57.
[8] Pitarresi G, Carruthers JJ, Robinson AM, Torre G, Kenny JM,
Ingleton S, et al. A comparative evaluation of crashworthy compos-
ite sandwich structures. Compos Struct 2007;78:34–44.
[9] Zhu H, Sankar BV. Analysis of sandwich TPS panel with functionally
graded foam core by Galerkin method. Compos Struct 2007;77:280–7.
[10] Hosur MV, Abdullah M, Jeelani S. Manufacturing and low-velocity
impact characterization of foam filled 3-D integrated core sandwich
composites with hybrid face sheets. Compos Struct 2005;69:167–81.
[11] Jiang D, Shu D. Local displacement of core in two-layer sandwich
composite structures subjected to low velocity impact. Compos Struct
2005;71:53–60.
[12] Davies JM. Lightweight sandwich construction. Oxford, UK: Black-
well Science, Ltd.; 2001.
[13] Gresham J, Cantwell W, Cardew-Hall MJ, Compston P, Kalyanas-
undaram S. Drawing behaviour of metal–composite sandwich struc-
tures. Compos Struct 2006;75:305–12.
[14] Pantelelis NG, European Patent No EP1770227 (4/2007).
[15] Pantelelis NG, Vrouvakis T, Spentzas K. A new concept for low cost
composite trailer. In: Proceedings of the 12th international conference
in composites engineering, Tenerife, August, 2005.