Mechanical properties and

design of sandwich materials

AF. JOHNSON and G.D. SIMS
(National Physical Laboratory, UK)
Sandwich materials consisting of a low density core with stiff skins
offer considerable potential for weight saving in panel applications,
where the main loads are flexural. Sandwich materials of interest for
car and van body panels, seat shells, etc, include steel/plastic
laminates, integral skinned plastic foams and glass fibre-reinforced
polyester skins with foamed plastic cores. In this paper, basic design
formulae for the flexural stiffness and strength of such sandwich
materials are reviewed and a method for designing optimum sandwich
structures for least weight or cost is given. Mechanical property data
are presented on a range of sandwich materials of potential interest for
vehicle panel applications. It is then shown how use of the least-
weight design method enables core and skin thicl~nesses to be
determined and gives a means of improving the flexural properties of
existing sandwich constructions.

Key words: sandwich materials; design calculations; flexural

stiffness; strength; weight saving; automotive components

Sandwich materials consising of a low density core example Reference 1. Such detailed analysis methods
with stiff skins offer considerable potential for weight are not appropriate for vehicle applications because of
saving in panel applications, where the main loads are their complexity and because design criteria are less
flexural. For aerospace applications, new materials critical. For these applications a simpler design
such as carbon fibre-reinforced epoxy skins with approach may be used in which the materials design is
aluminium honeycomb cores have been developed considered independently from the structural design.
along with relevant design methods. Because of their The materials design problem is concerned with the
high cost and difficulties in fabrication, such materials calculation and optimization of the sandwich stiffness
are not appropriate for use in panel applications in car and strength properties in terms of skin and core
and van bodywork or container vehicles. Sandwich properties. Details of the sandwich construction are
materials of interest for vehicle applications are then neglected in the structural design problem which
integral skinned plastic foams for interior components is concerned with calculating the effect of support
such as air filter covers, seat shells and floor panels, conditions and loads on a panel of known stiffness
and steel/plastic laminates or glass-fibre reinforced and strength. Thus the structural design problem can
plastics (GRP) with foamed plastic cores for low-weight follow traditional design methods for loaded panels.
body panels with good surface finish. Double skinned
In this paper attention is restricted to the materials
metal or thermoplastic panels with foamed plastic
design of sandwich panels under flexural loads. A brief
cores are also of interest for the energy absorbing or
discussion is given of the classification of sandwich
thermal insulation properties provided by the core
material. This paper is concerned with the materials for design purposes based on the definition
of a critical panel length. This is followed by basic
development of simplified design methods for such
design formulae for the flexural stiffness and strength
materials and with the provision of relevant materials
of sandwich materials and a method for designing
property data to enable optimum sandwich
optimum sandwich materials for least weight or cost.
constructions to be selected.
Mechanical property data are then presented on a
In aerospace sandwich structures, design analysis is range of sandwich materials of potential interest for
carried out using finite element and other computer- vehicle panel applications. These include a steel/plastic
aided design methods in which details of the stress laminate, thermoplastic structural foams and GRP/
distribution through the core and skins of the structure plastic foam sandwiches. Classification data are
are examined in terms of the applied loads, see for presented along with a comparison of measured and

0010-4361/86/040321-08 $03.00 © 1986 Butterworth ~ Co (Publishers) Ltd

COMPOSITES. VOLUME 17. NO 4. OCTOBER 1986 321
predicted stiffness and strength properties. Next, the and flexible in shear if:
design formulae for least-weight design of a sandwich
material are demonstrated and least-weight D/12Q > 0.01 (3)
constructions for several combinations of material
determined. These least-weight sandwich materials are It is apparent that the relative importance of shear
then compared with 1 mm thick steel sheet to show the depends on beam length I, and a critical length l* may
potential weight savings that are possible. be defined from Equations (2) and (3):

l* = IO(D/Q)1~ (4)
DESIGN METHODS FOR SANDWICH MATERIALS
Classification of sandwich materials such that when beam length l > l*, shear effects are
small.
An ideal structural sandwich material consists of a
thick sheet of lightweight material, with a thin sheet of A series of three-point bend tests on sandwich beams
much stiffer, stronger material attached to each face. of different lengths l enables both D and Q to be
The skin resists in-plane forces and bending moments, measured and hence the critical length l* to be
whilst the core resists transverse shear forces in the determined. This procedure has the advantage of
panel. The core must be stiff enough in shear to classifying sandwich materials without having to
prevent the skins sliding over each other, and in measure skin and core properties separately. If the
compression to keep the skins the correct distance measured centre point deflections w for several beams
apart. The design methods needed to analyse sandwich of length l and given load P are plotted in the form
panels under transverse loads depend on the relative w/Pl against l 2, Equation (1) shows that the results
importance of core shear deformations during bending. should be a straight line with gradient 1/48D and
In this paper, sandwich materials are classified into intercept 1/4Q. Alternatively, a plot of w/Pl3 against 1/l2
two types depending on the core flexibility. When the should again give a straight line with gradient 1/4Q
core is stiff in shear, shear deformations in bending are and intercept 1/48D. Thus D, Q and hence the critical
negligible and classical bending theory may be used length l* may be determined. In order to use classical
for sandwich panels, with an appropriate value for the beam and plate design formulae for sandwich panel
flexural rigidity. When the core isflexible in shear, shear analysis, typical panel length dimensions are then
deformations may be significant and contribute to required to be greater than the. critical length l*.
panel deflections during transverse loading. In this Classification data on a range of materials obtained by
case classical bending analysis for beams and panels these methods will be presented below.
requires modification to include thickness shear effects.
Calculation of stiffness and strength
In this paper attention is restricted to the stiff core
materials and data are presented to show that the Consider the symmetric three-ply sandwich beam
sandwich materials of interest for vehicle applications shown schematically in Fig. 1. The beam has width b,
thickness h and consists of a central core with
fall into this category. For these materials conventional
structural design methods based on conventional thickness he, and skins of thickness hs, so that h = hc
beam, plate or shell theory, which neglect through- + 2hs. The core and skin Young's moduli along the
thickness shear effects, are valid to a good beam axis are Ec and Es respectively and it is assumed
approximation. For a more rigorous analysis, modified that transverse strains arising from differences in
Poisson's ratio are negligible. The core and skin
design analysis methods which include core shear
flexibility are available, 2 and these methods are densities are denoted by Pc and Ps respectively.
necessary for certain aerospace sandwich materials If p is the beam density, consideration of the total mass
with very stiff skins. of the beam gives:
A convenient method of classifying materials is to bhlp = bl(hcP c + 2hsPs)
consider their behaviour in a three-point beam bend
test. Consider a beam of length l, with flexural rigidity whence
D and shear stiffness Q, simply supported at the ends
and subjected to a transverse central load P. When p = pc" +p,(1 - v ) (5)
deflections due to both bending and shear are taken
into account, the centre point deflection w is given by: 2

p13 [ 12/)] (1)

w=a- M
t3
where the second bracketed term gives the contribution
to the beam deflection from shear. It can be seen that
ifD/lZQ ~ 1, Equation (1) reduces to the classical beam
bending relation in which shear effects are neglected. h©
In sandwich materials these shear deformations arise
from the core. In order to classify materials, we define hi ~..] ,olP¢
the core to be stiff in shear if: b

D/12Q < 0.01 (2) Fig. 1 Schematic of sandwich beam

322 COMPOSITES . OCTOBER 1986

where the core/sandwich thickness ratio parameter v is For linearly elastic skins we may set eu = au/Es, where
introduced: O'u is the skin material failure stress, and using
Equation (10) obtain:
v = hc/h (6)
M u = bh 2 ouEf/6E s (14)
Using a similar 'rule-of-mixtures' argument to the
In Equation (14) cru is the lowest of the skin material
beam tensile stiffness, the tensile modulus of the
tensile or compressive strength and the skin wrinkling
sandwich material Et is given by the formula:
stress. The latter may be estimated from a buckling
analysis, and for thin skins is given approximately by: 2
Et = Ecv+Es(1-v) (7)
% = ½E~s13E2c13 (15)
which shows that the skins and core contribute to the
tensile stiffness in proportion to their thicknesses. Design formulae for other modes of failure such as
In order to calculate the flexural rigidity D, assume that core shear failure or skin/core debonding are also
classical bending theory may be applied to the discussed by Allen. 2
sandwich beam. Thus cross-sections which are plane
and perpendicular to the neutral axis of the unloaded Design o f s a n d w i c h materials
beam remain so when bending takes place. Also In the design analysis of sandwich panels, the basic
assume that the beam is narrow so that transverse formulae - - Equations (10), (11) and (14) - - are used
stresses parallel to the faces may be neglected. It can for calculating the flexural rigidity D and ultimate
then be shown that D is given by: bending m o m e n t Mu in terms of sandwich geometry.
The critical length l* is also estimated from Equation
D = Eclc + 2Esls (8) (4) or measured in a series of beam bending tests.
Provided that typical panel dimensions are greater
than l*, we may use standard design formulae (see, for
where Ic and Is are the moments of inertia of the core example, Roark and Young 3) or computer programs for
and skin cross-sections about the centroidal axis of the
plates and beams to determine quantities such as panel
beam. With the notation of Fig. 1, Ic = bh3c/12 and
deflections and bending moments. The m a x i m u m
Is=bh]/12 + (hc + hs)2bhs/4, which gives the result:
deflections may then be compared with imposed
limiting deflections and the m a x i m u m bending
D = bEs[h3s/6 +hs(he +hs)2/2] +bEch3c/12 (9) moments with the ultimate bending moment Mu for the
sandwich material.
Let Ef be the effective flexural modulus of the However, in order to make best use of sandwich
sandwich beam, then: materials, guidelines are needed for the materials
design problem in which core and skin thicknesses are
D = bhaEf/12 (10) specified. Note that the flexural strength and flexural
modulus Ef for a sandwich material are lower than
those for a homogeneous beam of the stiffer skin
After some algebraic manipulation, Equation (9) may
material. The main advantage of a sandwich
be expressed in the more simple and useful form as a
construction arises because the low density core
formula for El:
material enables the panel thickness h to be increased
for the same weight. It is then the increase in h 3 in
E f = Ec v3 + E s ( 1 - v a) (11) Equation (10) which offsets the lower flexural modulus
and results in a higher sandwich panel flexural rigidity
Several modes of failure may occur in sandwich beams D. Similarly, the h E term in Equation (14) puts up the
subjected to flexural loading, depending on sandwich limiting bending moment Mu. Increasing the core
geometry and core and skin properties. Two important thickness to enhance flexural properties also increases
failure modes are tensile (or compressive) failure of the panel weight and attention may need to be given to
skin material at the b e a m surface, where stresses in least-weight design. The choice of sandwich geometry
flexure are highest, and by skin wrinkling at the for a least-weight panel is considered by Johnson 4 and
compressive face. Assuming that the skin and core design charts are given for optimizing core thickness
materials are linearly elastic to failure, flexure beam ratio.
theory gives the strain e at a distance z from the For a sandwich material with specified skin and core
neutral axis of a sandwich beam subjected to a moduli, strengths and densities, the main design
bending m o m e n t M as: variables become the total thickness h and the core
thickness ratio, v = hc/h. In terms of h and v we have:
e = Mz/D (12)
he = hv, h s = h(1 - v ) / 2 (16)
where D is given by Equation (9) or (10). In the case of
a symmetric sandwich, in which the skin material fails
at the beam surface z = +-- h/2, the surface strain Johnson 4 considers the design of a least-weight
reaches the skin material failure strain eu at the sandwich material for a given flexural rigidity D or
ultimate bending moment Mu, given from Equation ultimate bending m o m e n t Mu. It is shown that in both
(12) by: the stiffness and strength limited cases there is an
optimum choice of the core thickness ratio v.
Mu = 2Deu/h (13) In the stiffness limited case it is found that the panel

COMPOSITES. OCTOBER 1986 323

weight/unit area, IV, given by: Pc/Ps
0.1 0.2 0.4 0.6 0.8 0.95
W = hp = h [pc v + Ps(l - v)]
1.0
from Equation (5), is minimized for a sandwich
material of given flexural rigidity D, when the thickness
ratio v is given by:
0.8
.--// //
[1-pc/p, (17) / /. / E
v = [ 1 - E cIE s
J 0.6 /// / i
To obtain the least-weight sandwich construction for a
given flexural rigidity D, v is first calculated from
: 1f" //
Equation (17) in terms of skin and core moduli and
densities. The sandwich flexural modulus Ef is then
0.4 ~~ ~//
given by Equation (11), and the required panel
thickness is determined from Equation (10) written in
the form: 0.2
Stiffness limited
h = (12D/bEf) '13 (18) - - - - - - - - - S t r e n g t h limited

in terms of the required flexural rigidity/unit width, i I I I I I I I i

0.2 0.4 0.6 0.8 1.0
D/b. Equation (14) may then be used to determine the Ec/E s
ultimate bending moment for the sandwich material.
Fig. 2 Design chart for optimum core thickness ratio hc/h (= v) for
In the strength limited case, it is found that the sandwich materials
optimum core thickness ratio v satisfies the cubic
equation:
modulus ratios Ec/Es. The continuous lines on Fig. 2
v a - 3v 2/(1 - Pc IPs) + 21(1 - EclEs) =0 (19) refer to the stiffness limited values from Equation (17)
and the broken lines are the corresponding strength
For given values of Pc/Ps and E J E s , the root v in the limited values derived from the appropriate root of the
range 0 < v < 1 may be determined graphically or by cubic Equation (19).
numerical computation. In the strength limited case, v
is obtained from Equation (19) and the sandwich If Pc and Ps are taken to be the cost/unit volume of the
thickness h is then given from Equation (14) in terms core and skin materials, then the design formulae given
of the required ultimate bending moment/unit width, here for the optimum core thickness ratio v may be
Mu/b: used to design a least-cost sandwich material for a
given flexural rigidity or flexural strength.
h = (6Mu Es/b ouEf) y2 (20)
SANDWICH MATERIAL PROPERTIES
[n order to facilitate the design method, Equations (17)
Skin a n d core p r o p e r t y data
and (19) for the least-weight core thickness ratios may
be put onto a design chart, see Fig. 2. Using the figure In order to use the sandwich materials design formulae
it is possible to read off the optimum core thickness presented above, materials property data on skin and
ratios v for given values of the density ratios Pc/Ps and core materials are needed. Table 1 lists such data for a

Table 1. Mechanical property data on skin and core materials

Material Density, Young's Shear Tensile

p (kg m -3) modulus, modulus, strength,
E (GPa) G (GPa) ~ (MPa)

Steel 7900 207 88 250*

GRP fabric 1900 18.8 -- 265
GR nylon 1390 7.2 1.3 150
PP 910 1.23 0.45 27*
PP (Skin 860 1.73 -- 36*
structural
foam t Core 580 0.73
PU foam [ 72 O.019 0.O05 0.52
155 0.056 0,016 2.2

*Yield stress

324 COMPOSITES. OCTOBER 1986

T a b l e 2. C l a s s i f i c a t i o n d a t a on s a n d w i c h m a t e r i a l s

Material Thickness, h Flexural rigidity Shear stiffness Critical I*/h

(mm) per mm width, per mm width, length, I*
D/mm (kN mm -1) Q/mm (kN mm -1) (mm)

PP structural foam 8 56 1.6 59 7.4

PC structural foam 8 86 2.8 55 6.9
PPO structural foam 8 80 3.2 50 6.3
RIM PU 3.7 2.7 -- < 35 <10
GRP/PU sandwich 10.7 820 0.53 390 37
Steel/PP sandwich 0.95 11.9 0.25 69 73
Steel 1 17.3 88 4.4 4.4

range of skin and core materials of potential interest in basic materials data of Table 1 (see also below).
vehicle applications. Values are given for density p, Assuming that all the shear deformations take place in
Young's modulus E, shear modulus G and tensile the plastics core, Q may be estimated from the relation:
strength tr; these were measured at the National
Physical Laboratory (NPL) or collated from materials Q =bhcGc (21)
suppliers' data sheets. The N P L data were obtained
under slow loading rates ('~ 20 mm rain -~) at 20°C and
50% relative humidity (RH). Note that the properties of where Gc is the core shear modulus. In a similar way it
the foam core materials are strongly dependent on the is possible to use Equations (10) and (21) to obtain the
degree of foaming, which may be characterized by the D and Q values given for the steel panel material.
density. The table shows typical values for these The main point to notice from the table is that all the
materials at a particular density. The materials listed sandwich materials have much greater critical lengths
are steel for the skin material in steel/plastic laminates; than the value 4.4 mm obtained for the steel. The
polypropylene (PP) as a typical core material for such structural foams and steel/PP laminate are seen to have
laminates; a glass fabric-reinforced polyester (GRP critical lengths in the range 50-70 mm, whilst the RIM
fabric) and short glass fibre-reinforced nylon (GR PU value, which could not be measured accurately, is
nylon) as typical reinforced plastics skin materials for below 35 mm. It follows that classical design methods
use with low density cores such as rigid polyurethane for calculating panel flexural response are valid, and
(PU) foam. The PP structural foam is an integral shear effects small, provided that panel dimensions are
skinned polypropylene foam, and is a typical greater than 70 mm. This will be the case for typical
thermoplastic structural foam fabricated in a single panel components in vehicles. However, it is apparent
shot injection moulding process. Note that the PP that care is needed with these materials at fixing points
structural foam contains a different grade to that used and under local loads, where the effective panel length
in the metal/plastic laminate. The GRP fabric and GR may be small. In these situations shear effects may
nylon have anisotropic stiffness and strength become significant in increasing panel deflections and
properties; the values quoted refer to a principal fibre thus reducing panel stiffness below the assumed
direction. flexural rigidity. The classification data on the GRP
fabric/PU foam core sandwich give a critical length of
Classification data 390 mm for this high performance, low-weight
sandwich material. Thus, although it has high flexural
Table 2 gives classification data on a range of rigidity, significant attention would need to be given to
sandwich materials in the form of flexural rigidity per shear effects in this type of material if used in small
mm width (D/mm), shear stiffness per mm width automotive structures.
(Q/mm), critical length (l*) and critical length/thickness
ratio (l*/h). Data are given on integral skinned Because of the wide range of sandwich thicknesses and
thermoplastic structural foams of PP, polycarbonate hence flexural rigidities, the materials in Table 2 are
(PC) and polyphenylene oxide (PPO); short glass fibre- not all equivalent in their flexural properties to a 1 mm
filled reaction injection moulded PU (RIM PU); GRP steel panel. To allow for this, it is informative to
fabric/Pu foam sandwich and a standard steel/PP express the critical length relative to the panel
sandwich laminate (Hercules, Liteplate S). Comparative thickness; ie, as l*/h. The integral skinned plastics
data on a 1 mm thick steel panel are also given. foams all have l*/h < 10 and are not much higher than
Variable length three-point beam bend tests, as the steel value. However, the steel/PP laminate value of
described above, were carried out at N P L to obtain the 73 and GRP/PU value of 39 indicate that sandwich
classification data on all the polymeric sandwich materials with very stiff skins and low modulus cores
materials. For the structural foams and RIM PU there may need special design methods for critical areas near
are no well-defined core and skins, hence this test load and fixing points.
method is useful in providing a rapid means of
obtaining sandwich stiffness properties without Measured stiffness and strength properties
separating the core and skins. For the steel/PP laminate
the core and skins are well defined and it is possible to In this section the measured stiffness and strength
determine l* by calculation. Thus flexural rigidity D properties of sandwich materials are compared with
was calculated using Equations (10) and (11) with the values predicted from the design formulae

COMPOSITES . OCTOBER 1986 325

Table 3. Comparison of measured and predicted sandwich properties

Thicknesses (ram) Moduli (GPa) Strength, Density,

Material Mu/b p
h hc hs Et Ef (N mmmm -1) (kgm -3)

/ Experiment 0.95 0.55 0.2 88 170 -- 3720

Steel/PP
/ Theory 86 167 30.3 3850
GRP/PU foam / Experiment 10.7 8.9 0.9 3.5 8.0 1200 440
/ Theory 3.3 8.1 1605 452
f Experiment 8 -- -- 0.96 1.25 416 650
PP structural foam
/ Theory 6.1 0.95 0.96 1.29 280

(Equations (7), (10), (11) and (14), thus establishing the In the integral skinned PP structural foam, it is difficult
validity of these simple formulae for sandwich to measure the skin thickness with any accuracy, since
materials design. The materials considered are a the skins and core merge together. However, core, skin
standard steel/PP laminate (Hercules, Liteplate S), a and sandwich densities are readily determined, and
GRP fabric/PU foam sandwich material and a PP these can be used to calculate an effective hc/h value
structural foam. Table 3 gives core and skin for use in the design formulae. Rearranging
thicknesses along with measured and predicted values Equation (5) gives:
of tensile modulus Et, flexural modulus Ef, density p
and ultimate bending moment/unit width Mu/b. v = hc/h = (p,- p ) / ( p , - Pc) (22)
The standard steel/PP laminate has well-defined core
and skins, whose thickness could be easily measured. which enables v to be determined in terms of the
Table 3 shows very good agreement between measured densities p, Pc and Ps. Using the density data on PP
and predicted tensile and flexural moduli. Note that structural foam from Tables 1 and 3, this gives a value
the flexural modulus is about 80% of the steel value, v -- 0.76. Hence from Equation (16), h¢ = 6.1 mm and
whilst the tensile modulus is much lower, being about hs = 0.95 mm are the effective core and skin
40% of the steel value. There is, however, considerable thicknesses for use in the design formulae. Table 3
weight saving, with the density being about 50% of shows very good agreement between measured and
that of steel. It was not possible to measure accurately predicted tensile and flexural moduli based on
the bending moment at failure of this material because Equations (7) and (11)with v ----0.76. Equation (14)
of the difficulty of detecting the onset of yield in the now predicts a low value for Mu compared with that
steel skins. measured in a three-point bend test. The calculated Mu
The GRP fabric/Pu foam sandwich also had clearly value was based on an assumed yield stress for tru, ie,
defined skins and again reasonable agreement is seen Mu is the bending moment at yield, whereas the
between measured and predicted modulus values. Note measured Mu is at ultimate failure, which explains the
that the PU foam core in this sandwich material has a higher measured value. Compared with the solid PP
density of 155 kg m -3 so that the higher density data skin material, the structural foam sandwich retains 72%
from Table 1 are used. Because of the importance of of its flexural modulus with a weight saving of 25%.
core shear effects in these materials, discussed above,
the flexural modulus was obtained from the flexural LEAST-WEIGHT SANDWICH MATERIALS
rigidity D measured in the classification tests and given
in Table 2. In calculating the ultimate bending moment Examples of least-weight design method
Mu, Equation (14) was used. From Equation (15) it was The design methods presented above are now
found that the skin wrinkling stress, au, was 195 MPa illustrated by calculating the stiffness limited least-
for this sandwich material. This is below the GRP weight sandwich construction for a range of materials.
fabric strength given in Table 1. Failure is therefore
The calculations are based on using Equation (17) or
expected to occur by buckling of the compressive skin,
the design chart (Fig. 2) to obtain the optimum core
and a failure stress of 195 MPa was used for tru in
thickness ratio v for given core and skin density and
Equation (14). A three-point beam bend test was
modulus ratios. It is of interest to compare a range of
carried out on this material to failure and a sandwich materials with an equivalent steel panel. A
compressive buckling failure was observed. However,
typical vehicle panel of 1 mm thick mild steel has a
the measured ultimate bending moment Mu was about
flexural rigidity per mm width (D/mm) of 17.3 kNmm
25% lower than the predicted value. This is probably as indicated in Table 2. The results of using the design
due to the difficulty in observing the onset of skin formulae to determine least-weight sandwich geometry
buckling and the fact that Equation (15) gives only an
for this value of flexural rigidity are given in Table 4.
approximate lower bound for the skin wrinkling stress.
Note that this GRP/PU sandwich material has very low For the steel/PP laminate, the data from Table 1 give
density compared with the steel/PP sandwich and, Pc/Ps = 0.115 and Ec/Es = 0.0059; hence, from
because of its thickness, a high ultimate bending Equation (17) or reading from Fig. 2, v = 0.94. The
moment.

326 COMPOSITES. OCTOBER 1986

Table 4. Least-weight sandwich constructions

Thicknesses (mm) Stiffnesses Strength, Panel Normalized

Material Mu/b weight, W weight
h hc hs D/b T/b (N mm mm -1) (kg m -2)
(kN mm) (kN mm -1)

Steel 1 -- -- 17.3 207 42 7.9 1

Steel/PP 1.8 1.7 0.05 17.3 24.4 23 2.4 0.30
PP structural 5.5 4.1 0.7 17.3 5.4 132 3.6 0.45
foam
GRP/PU foam 4.5 4.3 0.1 17.3 3.9 87 1.0 0.13
GR nylon/PU 5.7 5.4 0.15 17.3 2.6 117 1.2 0.16
foam

flexural modulus of this least-weight laminate is given optimum laminate requires 0.05 m m skins on a 1.7 m m
from Equation (11) as Ef = 36 GPa. The thickness h to core. Similarly, the reinforced plastics skins
give a flexural rigidity D / m m of 17.3 k N m m is obtained sandwiching the PU foam need only be 0.1 or 0.15 mm.
from Equation (18) as h = 1.8 mm. Hence However, the PP structural foam in Table 3 is close to
Equation (16) gives for the core and skin thicknesses the least-weight construction with a core thickness ratio
he = 1.7 m m and hs = 0.05 mm. In order to evaluate v = 0.76, compared with an optimum value v = 0.75.
other properties of the laminate, the tensile stiffness T,
For applications in vehicles it is probable that these
defined by:
least-weight constructions require skins that are too
thin and hence susceptible to damage or corrosion.
T = bh Et (23) However, the design procedure given here shows the
potential weight savings that are possible with
sandwich panels and in practice thicker skins than the
is also given in Table 4, along with the ultimate optimum may be required. The main reason for the
bending m o m e n t Mu calculated from Equation (14) very thin skins shown in Table 4 is that the sandwich
and the panel weight/unit area W given by the formula materials were designed to be equivalent to 1 m m steel.
presented above. In applications such as vehicle doors and bonnets, the
In a similar m a n n e r the least-weight sandwich construction is a double skin panel or a steel skin on a
constructions for PP structural foam, GRP fabric/PU subframe. A plastics sandwich construction to meet the
foam and GR nylon/PU foam, for the same flexural component design specifications would thus be much
rigidity, may be determined using the materials data thicker than the idealized panels in Table 4, with
from Table 1 and these results are also shown in correspondingly thicker skins. The thicker core
Table 4. required in, say, a door would also allow other benefits
to be designed into the component, such as impact
Discussion of weight saving energy absorption, thermal insulation or vibration
damping.
Table 4 shows that considerable weight savings are
possible with sandwich materials designed for an
equivalent flexural rigidity as 1 m m steel. The optimum CONCLUSIONS
steel/Pp laminate is only 30% and the PP structural The classification of sandwich materials for design
foam only 45% of the steel weight. The two reinforced purposes has been considered in this paper. If core
plastics panels with the very lightweight PU foam core shear thickness is too low then shear deflections during
have weights only about 15% of that of steel. Note also flexural loading are significant. It is shown that a
that most of the sandwich panels have ultimate critical panel length may be defined, above which
bending moments Mu comparable to or greater than shear effects are small and where design may be based
the value at yield in a steel panel. The major on conventional beam, plate or shell analysis. Data on
disadvantage of these constructions compared with sandwich materials of interest for automotive
steel is the significant reduction in tensile stiffness T, applications show that they are generally in this latter
with the lowest weight panels being less than 2% of the category.
steel tensile stiffness. This could be a significant
disadvantage for sandwich panels subjected to both in- Simple design formulae are given for calculating the
plane and bending loads, or under compressive flexural stiffness and strength of sandwich materials.
buckling loads. These are supported by test data on a range of different
material combinations. Design formulae are also given
It is interesting to compare the least-weight sandwiches for selecting skin and core thicknesses to give a
with the more typical constructions described in sandwich material of least weight or least cost for a
Table 3. The main difference is that the least-weight required flexural stiffness or strength. Some least-
materials have thicker cores and thinner skins. Thus weight sandwich configurations are given and the
the commercially available steel/PP laminate in Table 3 weight saving compared with an equivalent steel panel
has 0.2 m m skins on a 0.55 m m core, whilst the is calculated.

COMPOSITES. OCTOBER 1986 327

It is shown that sandwich materials have high potential 2 Allen, H.G. 'Analysis and Design of Structural Sandwich Panels'
for application in low weight automotive panels and (Pergamon Press, qxford, UK, 1969)
components. Possible problems in their use may be 3 Rnark, R.J. and Yeling, W.C. 'Formulasfor Stress and Strain"
(McGraw-Hill, New York, NY, USA, 1975)
associated with fixing and loading points, where shear 4 Johnson, A.F. 'Optimum design of laminated GRP materials'
effects may be significant and where the stiff skin in 'Advances in Composite Materials" edited by A.R. Bunsell
materials are susceptible to local damage. et al (Pergamon Press Oxford, UK, 1980) pp 683-706

ACKNOWLEDGEMENT AUTHORS
© Copyright Controller, HMSO, London, UK, 1986. The authors are with the Division of Materials
Applications, National Physical Laboratory,
REFERENCES Teddington, Middlesex T W l 1 0LW, UK. Inquiries
l US Department of Defense "Structural Sandwich Composites' should be addressed to Dr Johnson in the first
(MIL-HDBK 23A, Washington, DC, USA, 1968) instance.

EXHIBITION

SECOND EUROPEAN CONFERENCE

ON COMPOSITE MATERIALS
AND
SIXTH INTERNATIONAL CONFERENCE
ON COMPOSITE MATERIALS

July 20-24, 1987

IMPERIAL COLLEGE, LONDON - ENGLAND

COMJ
PRELIMINARY ANNOUNCEMENT
ASSOCIATION EUROPEENNE DES M A T E R I A U X COMPOSITES
2, Place de la Bourse, 33076 B O R D E A U X - F R A N C E
T61. : [Link] - Telex : 560047 Chareci

328 C O M P O S I T E S . OCTOBER 1 9 8 6