Wrinkling of Sandwich Panels

for Marine Applications

Linus Fagerberg
KTH Aeronautical and Vehicle Engineering
SE-100 44, Stockholm, Sweden
E-mail: [email protected]

ISSN 1651-7660
ISRN KTH/AVE/DA-03/18-SE

Wrinkling of Sandwich Panels for Marine Applications

Preface
The work presented in this doctoral thesis has been carried out at the Department of
Aeronautical and Vehicle Engineering at the Royal Institute of Technology, between February
1999 and August 2003. The Swedish Defence Material Administration, FMV, has provided
the financial support through the contract 63823-LB126424. The research was initiated
through the design of the Visby class corvette and a close cooperation with Kockums AB
Karlskronavaret has been held through the more than four years of research.
I would like to express my gratitude to the department and all the staff that made me feel
welcome and appreciated. During my research I have always been able to exchange ideas with
fellow colleagues, both in their offices and also of course in the coffee room. Special thanks
to my supervisor Professor Dan Zenkert who has always been there for me when I have
needed support and new fresh ideas. My contact with FMV, and especially Mr. Anders
Lnn, has also provided valuable insights that has helped me throughout the work.
The diversity of the tasks I have been doing during the time spent at the department has been
very valuable to me personally. I have chosen to investigate wrinkling of sandwich plates, a
problem some people would argue is narrow but the more time I spent with it, the more vast
and intriguing it seems.
My deepest gratitude to my family and friends for being there and making my life fun to live.
I would not be the person I am if not my parents Mats and Malin had raised me and my
brothers Magnus and Jonas like they did. Letting us learn by our mistakes and encouraging us
to go through life with a curious mind.
Finally, all my love to Kristina for choosing to share her life with me. Not everybody would
appreciate my sometimes all too present absent-mindedness towards life and responsibilities,
when my mind was occupied with new interesting formulas coming to life.

Stockholm, August 2003

Linus Fagerberg

Wrinkling of Sandwich Panels for Marine Applications

iii

Abstract
The recent development in the marine industry with larger ships built in sandwich
construction and also the use of more advanced materials has enforced improvements of
design criteria regarding wrinkling. The commonly used Hoffs formula is not suited for the
highly anisotropic fibre reinforced sandwich face sheets of today.
The work presented herein investigates the wrinkling phenomenon. A solution to wrinkling of
anisotropic sandwich plates subjected to multi-axial loading is presented. The solution
includes the possibility of skew wrinkling where the wrinkling waves are not perpendicular to
the principal load direction. The wrinkling angle is obtained from the solution together with
the maximum wrinkling load. This method has been supported with tests of anisotropic plates
subjected to uni-axial and bi-axial loading.
The effect of the face sheet local bending stiffness shows the importance of including the face
sheet stacking sequence in the wrinkling analysis. The work points out the influence of the
face sheet local bending stiffness on wrinkling. Three different means of improving the
wrinkling load except changing core material is evaluated. The effect of the different
approaches is evaluated theoretically and also through comparative testing.
The transition between wrinkling and pure face sheet compression failure is investigated.
Theoretical discussions are compared with compressive test results of two different face sheet
types on seven different core densities. The failure modes are investigated using fractography.
The results clearly show how the actual sandwich compression failure mode is influenced by
the choice of core material, changing from wrinkling failure to face sheet micro buckling
failure as the modulus density increases.
Finally, a new approach is presented where the wrinkling problem is transferred from a pure
stability problem to a material strength criterion. The developed theory provides means on
how to decide which sandwich constituent will fail first and at which load it will fail. The
method give insight to and develop the overall understanding of the wrinkling phenomenon.
A very good correlation is found when the developed theory is compared with both finite
element calculations and to experimental tests.
Keywords: wrinkling, local buckling, imperfection, stability, anisotropy, sandwich

Wrinkling of Sandwich Panels for Marine Applications

Aims and scope

Throughout the more than four years of research the goal has been to investigate and further
increase the understanding of the wrinkling phenomenon. This has been done with the simple
theories and without complicating things more then necessary. The research should be
applicable and made known to the sandwich community. All findings should be presented so
that the largest possible audience can understand them.
In this thesis sandwich wrinkling is investigated. This is done in the context of linear-elastic
material constituents and isotropic core materials. Issues like dimpling, where the face sheet
buckles into the honeycomb cells, are not included, neither is the local buckling or local
indentation that can occur close to load introductions. The thesis specifically deals with the
effects of anisotropy in the face sheets typically derived from the use of aligned fibres and
different stacking sequences. It is investigated and exemplified how the wrinkling load can be
increased through altering the composition of the face sheet as well as changing core material.
The transition from wrinkling to the pure compression failure is also examined. Finally,
wrinkling is transferred from a traditional stability criterion into a material failure criterion in
order to study the effect of initial imperfections. All theories developed within the thesis are
thoroughly compared to experimental tests and numerical finite element models. Thus, the
contents of this thesis provide an extension of the classical work on wrinkling.

Wrinkling of Sandwich Panels for Marine Applications

vii

Dissertation
The thesis includes a short introduction to the common wrinkling theories where the most
well known and most often referred to formulae are derived and discussed. Some special
cases of wrinkling and their proposed solutions are presented. The introduction is followed by
four appended papers:
Paper A
Fagerberg L. and Zenkert D., Effects of Anisotropy and Multi-Axial Loading on the
Wrinkling of Sandwich Panels, Submitted for publication.
Paper B
Fagerberg L., The Effect of Local Bending Stiffness on Wrinkling of Sandwich Panels,
Journal of Engineering for the Maritime Environment, Vol. 217, pp. 111-119, 2003.
Paper C
Fagerberg L., Wrinkling and Compression Failure Transition in Sandwich Panels, Accepted
for publication in Journal of Sandwich Structures and Materials.
Paper D
Fagerberg L. and Zenkert D., Imperfection Induced Wrinkling Material Failure in Sandwich
Panels, Submitted for publication.

Wrinkling of Sandwich Panels for Marine Applications

ix

Division of work between authors

Paper A
Fagerberg performed the analytical calculations, tests and finite element calculations and
wrote most of the paper. Zenkert initiated and outlined the work and assisted in all analytical
calculations and helped writing the paper.
Paper D
Fagerberg initiated the work, developed the analytical model, performed the tests and
calculations. Zenkert supported throughout the work and the paper was written jointly by the
authors.

Wrinkling of Sandwich Panels for Marine Applications

xi

Contents
Preface

Abstract

iii

Aims and scope

Dissertation

vii

Division of work between authors

ix

Table of contents

xi

Introduction

Background

Introduction to wrinkling theory

No decay: Winkler foundation

. . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Linear decay: Hoffs method
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Exponential decay: Plantemas approach
. . . . . . . . . . . . . . . . . . . . . 9
Differential equation method: Allens approach
. . . . . . . . . . . . . . . . . . 10
Special situations

11

Wrinkling of sandwich panels . . . . . . . . . . . . . . . . . .

Wrinkling under multi-axial load
. . . . . . . . . . . . . . .
Face stress exceeding elastic limit
. . . . . . . . . . . . .
Wrinkling and initial imperfections
. . . . . . . . . . . . .
Wrinkling and finite element modelling
. . . . . . . . . . . .
Nonlinear response of compressively loaded sandwich structures

.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
.
.
.
.
.

11
11
12
13
13
15

Summary of thesis

17

Future work

18

References

18

Appended Papers
Paper A

A1-A16

Paper B

B1-B16

Paper C

C1-C14

Paper D

D1-D22

Introduction
A structural sandwich comprises of two stiff face sheets on a thick lightweight core. One of
the benefits is that the sandwich has a very high bending stiffness compared to its weight and
it is therefore often used within the aeronautical, vehicle and marine industry where low
weight and high load carrying capabilities immediately allow for higher payload. The high
bending stiffness comes from that the stiff face sheets are separated (in space) and the core
bonds them together, making them work in concert much more effectively then if they would
have been alone. However, the sandwich is not that good at carrying in-plane compressive
loads and it has a unique stability failure mode of its own wrinkling. In this thesis this
failure mode is investigated. The thesis starts with some notes on sandwich within the marine
industry and why wrinkling is nowadays getting more attention. The background to the
problem is discussed and thereafter follows an introduction to the classical wrinkling theories.
A discussion on wrinkling related problems is given, including finite element analysis
methods. The summary of the thesis providing information on how and where this thesis
makes a contribution to the subject is then given, followed by a short discussion on possible
future work.

Figure 1. The first Visby class corvette at the Karlskrona shipyard the 5 of March
2002.

Background
The sandwich concept, using two thin stiff face sheets in combination with a thick lightweight
core, has been utilised in structures for some time. During the last decade the development
within the Swedish marine industry has been expanding the boundaries of what previously
was believed possible. Larger vessels then ever before are now being built in sandwich
construction and this brings some new considerations to the attention of the engineer.
Kockums Karlskronavarvet, who is part of the HDW group, builds the Visby class corvettes
for the Swedish Navy, see Figure 1. Visby is a surface combatant, designed for a wide range
1

Linus Fagerberg

of roles: anti-surface warfare, anti-submarine warfare, mine countermeasures, patrol and

much more.
A small boat is usually designed to be tough enough to handle, i.e. it should feel comfortably
solid and it should not break when drawn up on a beach. A high-speed powerboat has to be
able to sustain high slamming loads in addition. All these loads are localised and mainly
perpendicular to the boat hull. Therefore sandwich boats of small to medium dimensions often
have a rather high-density core to prevent local face indentations and proportionally thick face
sheets to avoid face failure.
A larger boat hull is often more slender, since the drag through water is less for a slender hull
than for a bulky one. For longer hulls other types of forces have to be included in the design
of the ship.
A boat travelling through heavy seas is subjected to wave induced loads. Figure 2 shows what
the shipwright refers to as hogging and sagging motion. When the ship is subjected to
hogging, the midsection is lifted on a wave while the bow and aft is not supported. This
causes tensile stresses in the deck and compressive stresses in the bottom. In sagging it is the
other way around, the bow and aft is lifted, while the midsection is not supported. Sagging
therefore creates compressive stresses in the deck and tensile in the bottom. If the ship also
travels through the waves misaligned with the wave front the ship will be subjected to torsion
as well.

Figure 2. Hogging and sagging of ship hull.

The two types of loading are usually not of the same importance for different ship types. The
weight distribution compared to the volume distribution of the ship decides which of the cases
is worst. For the type of ship that Visby is (see Figure 1), sagging is more severe.
This wave induced hogging and sagging leads to high in-plane stresses in the deck and bottom
of the ship. These stresses in long and slender boats are much higher than almost any
localised load can give rise to. Therefore this kind of ship has to have deck and bottom panels
designed to sustain high in-plane loads, both compressive and tensile.
It is often more difficult to design a panel for compressive forces than for tensile. In
compression there are several different possible failure modes that have to be considered. One
of these is the compressive failure of the face sheet, which often for fibre composites takes the
form of microbuckling. In addition to compressive failure, a structure loaded in compression
can sometimes collapse through loss of stability. There are several classic examples of this

Wrinkling of Sandwich Panels for Marine Applications

and the most well known is possibly the buckling of a column (Figure 3). Columns are
however rather rare in ships but plates can of course buckle as well. The critical number of
waves is not necessarily the same for a sandwich plate as for a single skin composite or
isotropic metallic plate since the sandwich have much lower shear stiffness. A sandwich can
in addition to the previously described global form of buckling also loose stability through
local buckling, where the instability is within the sandwich, i.e. the sandwich face sheet
buckles but the symmetry plane of the sandwich remains more or less undeformed.

Figure 3. The first three buckling modes of a simply supported a thin column and the
first buckling mode of a simply supported long plate.
The local buckling can in turn be divided into several subgroups; Localised buckling, which
occur close to load introductions or loads applied perpendicular to the sandwich face sheets,
dimpling, where the face sheet buckles into the cavities of honeycomb cores, and finally
wrinkling, which in theory can occur simultaneously on the entire sandwich plate. Wrinkling
is sometimes referred to as natural wavelength buckling since it is associated with periodic
waves of a wavelength depending on the material properties, core- and face thickness, and not
on the in-plane dimensions of the sandwich. Wrinkling will be more thoroughly described
below.
The local types of buckling have not been much of a concern for shipwrights building in
sandwich using glass fibre composites and high density core materials. But the development
in the material field has led to a decrease in cost for materials which formerly where used only
in aerospace structures. The high performing materials allow for a more full-blown utilisation
of the sandwich concept with thinner face sheets and more lightweight cores. This
development has led to that wrinkling can be the active design constraint. The theories now in
use when designing against wrinkling is often rather old and they were derived when most
face sheets used were metallic and they do therefore badly suit the fibre composite face sheets
of today.

Introduction to wrinkling theory

In this section a short introduction to some of the most commonly referred theories on
wrinkling are presented together with the derivation of a few frequently used and refereed to
formulae.
The first reported work on sandwich face wrinkling was published in early 1940's by Gough
et al. [1] and Williams et al. [2]. The results from these investigations are in essence covered

Linus Fagerberg

by the following. Local instability of the sandwich face is of course possible in pure in-plane
compression of the sandwich, but also when it is subjected to ordinary bending, since one of
the faces will be in membrane compression.
The wrinkling wave is a periodic wave with the (half) wavelength l, see Figure 4.

Figure 4. Definition of the wrinkling wave.

For a beam type sandwich panel, or a plate in cylindrical bending, the governing differential
equation reads

Df

d 4w
d 2w
+
z = 0
P
dx 4
dx 2

(1)

where Df is the bending stiffness of the face sheet alone, P is the compressive load acting on
the face sheet and z the support pressure from the core, the elastic foundation. There are
several different cases of wrinkling, depicted in Figure 5, and ways that this problem has been
solved.

Figure 5. Wrinkling instability.

The occurrence of the different wrinkling modes depends on the geometry of the sandwich
(thickness of components) and the material properties. In the first case (Case I) rigid base or
single-sided wrinkling, only one face sheet of the sandwich wrinkles. This mode can appear
when the sandwich is under bending and only one face sheet under compression, or if the
sandwich has un-symmetric face sheets, one with lower buckling load than the other.
The second mode (Case II) is the anti-symmetrical, in-phase or snake mode where the two
face sheets wrinkle in phase. In this case the two face sheet deformations share the same
phase. This mode is common for anisotropic core materials where the shear stiffness is
considerably lower than the Youngs modulus (perpendicular to the face sheet), for example
honeycombs, and for sandwich configurations with a thin core. A sandwich configuration
prone to snake mode wrinkling must carefully be checked against all global buckling modes.
It is likely that a global buckling analysis will predict even lower buckling loads than Case II
wrinkling formulae.
The third case (Case III) is defined as symmetrical wrinkling, out-of-phase or hourglass
wrinkling. Hourglass wrinkling is a special case of rigid base wrinkling with a smaller

Wrinkling of Sandwich Panels for Marine Applications

distance to the rigid plane. A smaller distance to the rigid base provides a firmer support of
the face sheet and therefore makes the wavelength shorter and critical load higher. Hence
Case III wrinkling formulae always predict a higher wrinkling load then Case I formulae. This
mode is accordingly a bit theoretical and of less importance to the engineer.
In wrinkling the displacements of the faces are transmitted to the core in which they damp out
rapidly in the thickness direction. If the core thickness exceeds a certain value the mid-plane
of the core remains flat for all three wrinkling mode cases. In the real world foreign to
symmetry and perfect materials it is of course possible to have wrinkling where the waves are
not exactly in phase or anti-symmetrical and where the wrinkling amplitude differs between
the face sheets. Both face sheets of the sandwich will most probably have some sort of
imperfection or not being entirely symmetric and this will in most cases lower the wrinkling
load that we are about to derive.
The wrinkling load can be derived in several different ways and the following sections
describe some of them. The first and fourth methods described are based on solving the
governing differential equations, while the second and third are based on energy methods.
Each approach has a different assumption for the core stress. Although the approaches are
different they yield approximately the same result, apart from the first one, and the differences
are mainly due to different assumptions on the decay of the transverse stress.
No decay: Winkler foundation

When attempting to solve a buckling problem some assumptions must be made regarding the
displacement field. This is also done in ordinary beam theory where for example the beam
section is supposed to remain flat and perpendicular to the neutral axis of the beam. A simple
assumption regarding the core as a medium supporting the sandwich face is the Winkler
foundation model. This assumes that the support material consists of an array of
continuously distributed linear springs, as shown in Figure 6.

Anti-symmetrical

Symmetrical

Figure 6. Winkler foundation for wrinkling

In the snake mode case it is seen that the springs remain unloaded even after wrinkling. It is
further seen that the mode of deformation in the core is shear rather than tension/compression,
which the set of springs are unable to model, and hence no solution can be derived in this
case. In the symmetrical case, on the other hand, the model becomes more realistic since the
mode of deformation in the core is both tension/compression and shear. In this case it assumes
a core with some modulus perpendicular to the faces (Ecz) but with no out-of-plane shear
stiffness, i.e., Gcxz = 0. The Winkler model then suggests that the core stress due to the
wrinkling of the face is proportional to the face sheet deformation, and may be written as

cz ( x,0) = K z w( x,0) = K z w f ( x)

(2)

where Kz is the so called foundation modulus. Since the stress is constant through the core
thickness and core shear is neglected the wrinkling wave (w(x,z)) damps out linearly through
the core. It further has to be equal to zero at the symmetry plane (Case III wrinkling, see

Linus Fagerberg

Figure 4 and 6). Therefore the wrinkling wave can be written as w(x,z)=wf (1-2z/ tc).
Assuming further that one may treat the set of springs as a continuous media so that z =
dw/dz and cz = Eczcz, we may then for z = 0 write

cz ( x,0) = K z w f ( x) = Ecz

2E
2
w f ( x) so that K z = cz
tc
tc

(3)

To derive the governing equation following this approach, a small section of the sandwich
face sheet can be examined, see Figure 7.
Mx+

Tx

Mx

dMx
dx
Nx +

Nx

cz

Tx +

d Nx
dx

dTx
dx

Figure 7. Forces, bending moments and stresses acting on the face sheet.

By treating the face sheet as an ordinary beam or strut and neglecting transverse shear
deformations (a most reasonable assumption for a thin composite or metal face), the
governing equation for the beam is derived from the two equilibrium equations for the face
dTx
dM x
d 2w
+ zz + N x
=0
Tx = 0 and
dx
dx
dx 2

(4)

leading to the differential equation (also using that d2w/dx2 = Mx/Df and taking the in-plane
load as compressive, i.e., Nx = P)
d 4w P d 2w Kz
+
+
=0
dx 4 D f dx 2 D f

(5)

This is solved by assuming some appropriate function for the deflection field w that fulfils
Equation (5). One such is simply the usual wrinkling wave equation
w = W sin

(6)

Inserting this into Equation (5) yields an expression for P as

P = Df

+ Kz

(7)

To find the critical load one must further minimise P with respect to the unknown wavelength
l, which gives
2 Df 2 2 Kz l
Df 4
dP
4
0
=
+
=

l
=
dl
l3
2
Kz

(8)

with the critical load as

PWi = 2 D f K z = 2

E fx t 3f Ecz
6t c

or Wi = 0.8165

E fx t f Ecz
tc

(9)

Wrinkling of Sandwich Panels for Marine Applications

To derive a Winkler based formula for rigid base (Case I) wrinkling it simply to assume that
w must be zero for z = tc and to redo the derivation (the foundation modulus is in that case
equal to Ezc/tc providing only half the previous support).
The winkler approach suffers from the drawback that the shear in the core is neglected. This is
only reasonable either if the core truly has a very low shear modulus, or if the wavelength of
the wrinkling is sufficiently long. In all other cases, there will be some shear lag effect
smoothing out the stress cz so that it decreases with increasing z. The use of Equation (9)
may therefore be somewhat unreliable.
Linear decay: Hoff and Mautnerss method

The next method to assess the wrinkling problem is that of Hoff and Mautner [3]. In this case
the shear stress in the core is accounted for and the shear lag is modelled by a linear decay
function. The centre line of the core is in the first case assumed to be undeformed and this
model thus simulates Cases I and III.
The deformation of the face induces tensile and compressive stresses in the core perpendicular
to the face. If the wavelength, l, is short it would hardly effect the material in the middle of
the core and hence it is assumed that the core is only affected in a small zone with depth h.
Since the buckling is symmetrical the core deforms only in the z-direction. Assuming the face
to undergo a sinusoidal displacement (Equation (7)), and that the wave damps out linearly
(linear decay) with coordinate z, and we can then write
w=

W (h z )
n x W ( h z )
x
sin
=
sin
h
L
h
l

(10)

where L is the length of the beam or strut, l the wavelength of the wrinkles (see Figure 4) and
that n is allowed to be a large number. Assuming the core to have low (or zero) in-plane
modulus the tensile/compressive stress in the core can be written cz = Ec w/z and if the
deflections are in the z-direction only the core shear stress is czz = Gc w/x. It may be worth
noticing that the assumed stress field in the core not satisfies the 2-D equilibrium condition
ij,j = 0. Anyhow, using this assumption the strain energy stored in the core over the length l
will become
Uc =

1
2Ec

l h

cz2 dxdz +

0 0

1
2Gc

l h

2
dxdz =
cxz

0 0

E cW 2 l Gc 2W 2 h
+
4h
12l

(11)

The strain energy stored in the face due to bending is

Uf =

Df
2

d2w f

dx =

dx 2

W 2 4 t 3f E f
48l 3

(12)

The work done by the applied load is

l
dw f
1
Up =
P
dx
20

W 2 2 t f
W 2 2
dx =
f
P=
4l
4l

(13)

The energy equation U=Uf+Uc+Up can now be solved for f which after some rearrangement
results in

Linus Fagerberg
2
E l2
hG E f
f = 2c + c +
12
t f h 3t f

tf

(14)

The critical stress in this equation depends on the parameters l and h. The values of these are
those that minimises the critical stress. Therefore, they are found by
d f

2
2
d f
Gc
2 Ec l E f t f
= 2

=0
= 2
+
= 0 and
dl
dh
tf h
6l 3
t f h 2 3t f

Ec

(15)

Solution of the two equations are

E 2f
E f Ec
h
l
6
=
1
.
65
= 0. 913
and
tf
Gc2
tf
Ec Gc

(16)

Substitution into Equation (14) yields the critical face stress as

Ho = 0.913 E f Ec Gc

(17)

Now, this formula is only correct when the zone h, as given in Equation (16), is smaller than
half the core thickness tc/2 (or tc if Case I wrinkling is treated). If not, the assumed
displacement field in Equation (10) must be changed so that tc/2 is substituted for h and the
same derivation as outlined above is again performed. This will result in the following
relations [3].
E f tc
E f Ec t f
l
t
= 1. 42 4
and HoIII = 0.817
+ 0.166Gc c
tf
Ec t f
tc
tf

(18)

As seen, Equation (18) gives the same result as the Winkler model in the previous section but
with an extra term for the shear part. This is not surprising since even the Winkler model
assumes some linear decay of the deformation and that w = 0 for z = tc/2. Hence, Equation
(17) should be used for symmetrical (Case III) wrinkling with thick cores and Equation (18)
for Case III wrinkling on thin cores.
By making a similar analysis of the anti-symmetrical case (Case II) the following results are
obtained [3]. For h<tc/2 (thick cores)
E 2f
E f Ec
h
l
6
3
= 1. 5
and = 2.15
tf
Gc2
tf
Ec Gc

HoIIa = 0.513 E f Ec Gc + 0.33Gc

tc
tf

(19)

(20)

and for h = tc/2 (for thin cores)

E f tc
E f Ec t f
l
t
= 1. 67 4
and HoIIb = 0.59
+ 0.387Gc c
tf
Ec t f
tc
tf

(21)

Hoff and Mautner [3] concluded from their analysis that the very simple formula, Equation
(17), is a conservative estimate of the buckling load for all cases except for thin cores where
Equation (21) can predict a lower load. Hoff also found that the prediction of the buckling

Wrinkling of Sandwich Panels for Marine Applications

load trends agreed very well with tests but suggested that in practical cases the load should be
predicted using the conservative formula

Ho = 0.53 E f Ec Gc

(22)

This is the formula a large part of the industry uses today and is often referred to as the Hoffs
formula. According to this formula, the wrinkling stress is independent of the sandwich
geometry and a function only of face and core properties. It also follows that the core will
have the most influence on the wrinkling load with improving load bearing capacity for cores
with high elastic properties.
Exponential decay: Plantemas approach

Shear lag problems are usually described by some exponential functions in hyperbolic sines
or cosines. Therefore, another deflection function was used by Plantema [4] for the stress
decay assuming the waves to damp out exponentially (exponential decay) as
w = We kz sin

n x
x
= We kz sin
L
l

(23)

We further assume that there are no displacements in the x-direction, i.e. u = 0. The following
derivation based on this assumption also assumes basic wrinkling, or rather that the buckling
of one face has no effect on the face sheet on the other side. Using this function in the same
way as outlined above, one can write

cz = E c cz = E c

w
w
w u
, cx = 0, czz = Gc czz = Gc
+
= Gc
z
x
x z

(24)

Again, this assumption violates the 2-D equilibrium condition ij,j = 0 in the core. The
bending strain energy of the face sheet can from this assumption be written

Uf =

Df
2

d 2wf

dx =

dx 2

W 2Df 4

(25)

4l 3

and the strain energy in the core as

Uc =

1
2Ec

l h

cz2 dxdz +

0 0

1
2Gc

l h

2
cxz
dxdz =

0 0

E cW 2 kl Gc 2W 2
+
8
8kl

(26)

and finally
l
dw f
1
Up =
P
20
dx

dx =

PW 2 2
4l

(27)

which results in the total strain energy as

U =

4W 2 D f
4l 3

E cW 2 kl Gc 2W 2 PW 2 2
+

8
8kl
4l

(28)

Minimising with respect to W leads to an expression for the load as

U
= 0 W = 0 (trivial solution) or
W

(29)

10

Linus Fagerberg

P=

2Df
l2

E c kl 2 Gc
+
(non-trivial solution)
2k
2 2

(30)

and by then letting P/l = P/k = 0 one arrives at

PPl =

33
2 D f Ec Gc
2

(31)

Which is often referred to as the Plantema formula. If the face sheet is isotropic Equation (31)
can be rewritten to

Pl = 0.8253 E f Ec Gc

(32)

As seen, this is very close to the result obtained by Hoff and Mautner for the symmetrical case
and thick core (see Equation (17)). In fact, the only difference is that another decay function
has been used, otherwise the assumptions are the same. It must be understood that Equation
(31) is much more powerful then Equation (10) or (32). The Plantema formula incorporates
the effect of the local bending stiffness and not only the in-plane stiffness of the sandwich
face sheet. This is of the greatest importance when dealing with anisotropic and layered
composite face sheets commonly used today.
Differential equation method: Allens approach

The following derivation assumes a thin beam (the face) under compressive end loads,
continuously supported by an elastic medium, which extends infinitely on one side of the
beam. Hence, the face is unaffected by the opposite face. The differential equation for a
homogeneous beam or plate in cylindrical bending (the face) with Nx = P is
Df

d 4w
d 2w
+
=q
P
dx 4
dx 2

(33)

Now, the transverse load q equals z. Assume that the core is isotropic and the same shape of
the face as above, i.e., w = W sin(x/l). By assuming an Airy's stress function the stress, cz,
necessary to deform the core in this manner may as Allen [5] showed be written
a
l

z = wm sin

x
l

with a =

2 Ec
(3 vc )(1 + vc )

(34)

This leads to expressions for the natural wavelength and instability loads reading [5]
l=

2 4D f
a

and PAllen =

1
2

2
3

+ 23

D f a 2 0.88 3 D f a 2

(35)

If the face sheet is isotropic the bending stiffness can be substituted and the resulting formulae
for the wrinkling wavelength and critical stress are obtained as
l = tf

(3 vc )(1 + vc ) E f
12(1 v ) Ec
2
f

and Al =

3
3

12(3 vc ) (1 + vc )
2

E f Ec2

(36)

Wrinkling of Sandwich Panels for Marine Applications

11

This gives for c = 0.3

Al = 0.58 3 E f Ec2

(37)

and with Gc = Ec /2(1+c )

Al = 0.783 E f Ec Gc

(38)

which is very close to the result in Equation (17) and (32) and also the proposed formula in
Equation (22). Again the formula for anisotropic face sheets, Equation (35), is more powerful
then Equation (37) or (38), since it incorporates the effect of the local bending stiffness of the
face sheet.
Several papers have been written on this topic since these more classical works were
published. In general, the approaches have been to find a more unified approach to solve the
different wrinkling modes. One such approach was published recently by Niu and Talreja [6].
Others attempt to combine global buckling of sandwich columns and wrinkling, e.g. [11]. In
essence, there is little new to these approaches, they come to basically the same conclusion
that the "old" formulae in principal agrees well. Surprisingly, almost all the recent work on
sandwich panel wrinkling have in common that they treat uni-axial loading of sandwich
beams/columns, without initial imperfections and there is very, very seldom any experimental
verification.

Special situations
Wrinkling of sandwich panels

The above analyses have been derived for a state of plane stress, which is legitimate provided
that the column is narrow. When it is wide, the column must be treated as a panel and a state
of plane strain predominates. In an analysis similar to that of the column but for panels, Norris
et al. [7] derived the expression for the wrinkling stress as

cr = 3

E f Ec
(1 2f )(1 c2 )

Gc

(39)

using differential equations where the constant equals 0.72 for f = 0.3 and c = 0.2. This
indicates that the formula experimentally verified by Hoff and Mautner, Equation (22), can be
used even for panels, but with the E substituted by E/(12). In fact, the tests verifying this
formula were performed on panels under uniaxial compressive loads [3].
Wrinkling under multi-axial load

The results from previous sections are only valid for beams or plates under uni-axial loading.
Plantema [4] has presented results indicating that the bi-axial loading case for a plate can be
treated by simply comparing the lowest wrinkling load in any of the two main directions. For
multi-axial loading, i.e., including all edge forces Nx, Ny and Nxy, and generally anisotropic
sandwich plates, no solution or solution procedure is known except the suggested solution
presented in Paper A in this thesis.

12

Linus Fagerberg

In Paper A the problem is addressed using a different approach where the ratio between
applied load and sustainable wrinkling load is accessed for all directions within the plate. This
method also include the effect of skew wrinkling where the wrinkling wave is not
perpendicular to the highest compressive principal stress within the sandwich plate.
Sullins et al. [8] presented some recommendations regarding wrinkling under multi-axial
load. They may help the designer to pursue this task and the recommendations are stated as
follows;
1.

Calculate the principal stresses 1 and 2, either by means of analytical or numerical

methods.

2.

If only one of the principal stresses is in compression, then neglect the tension stress and
treat the problem as one-dimensional.

3.

If both principal stresses are in compression then use the interaction formula

1
2
+
1cr
2 cr

1 with 1 >2

(40)

to determine the wrinkling strength. The critical stresses 1cr and 2cr are the onedimensional wrinkling stresses calculated in the directions of the principal stresses.
The interaction formula of Equation (40) is taken from analyses and tests of global (Eulertype) buckling of sandwich shells and is just, without physical justification, translated to be
used for wrinkling as well. One can argue over the validity of this kind of approach but it does
anyway provide some means of determining the wrinkling strength for a more general case. A
scheme like the above is also very well suited for implementation in finite element codes or
used as a constraint in optimisation programs. The most serious criticism against this scheme
is that for highly anisotropic faces the directions for the principal stresses do not coincide with
directions for the principal strains. Furthermore, the anisotropy also implies that the wrinkling
strengths, here denoted 1cr and 2cr, may vary strongly with the in-plane direction. However,
for isotropic and moderately anisotropic faces, the approach should be more accurate.
Vonach and Rammerstorfer [9] has also addressed the problem of wrinkling of orthotropic
sandwich plates under general loading. They tackle the problem by assuming the core to be
infinitely thick and the wrinkling wave at the interface between the face sheet and core to be
sinusoidal. Thereafter they are able to solve the governing differential equation describing the
face sheets deformation.
Face stress exceeding elastic limit

The described methods all predict the wrinkling load and not the failure load. In practice, the
actual wrinkling failure will occur as one buckle becomes unstable. Whether it buckles into or
away from the core depends on the strength of the core and/or the adhesive joint. Buckling
into the core occurs if the compressive strength of the core is lower than the tensile strength of
the core or the adhesive joint and buckling outwards breaking the core or the adhesive joint if
the other way. If the wrinkling stress exceeds the yield strength of the face it is customary to
replace Ef by a reduced modulus just as in the overall buckling. One such reduced modulus
proposed by Norris et al. [7] is

Wrinkling of Sandwich Panels for Marine Applications

Er =

4 E f Et
( E f + Et )2

13

(41)

where Et is the tangent modulus. The application of Equation (22) then requires a trial-anderror operation to find the correct wrinkling stress. A more conservative approach would be to
use Er = Et.
Wrinkling and initial imperfections

The actual wrinkling load is very much affected by any initial imperfections or waviness of
the face. Some design guidelines to account for such effects are given in [7] and [8] where it
is found that the factor preceding the cube root expression in Equations (17), (22), (32) and
(39) decreases from between 0.5 - 0.9 down to such low numbers as 0.05 to 0.1 for very high
waviness. However, waviness in the faces is usually carefully avoided, not only because of a
reduction in the wrinkling stress, but mainly because a smooth surface is a sought after
characteristic when using a sandwich design.
The reason for using the conservative formula in Equation (22) is mainly due to the effect of
initial irregularities. Derivations of this effect are performed in [7] and [8] concluding that the
effect has a maximum when the irregularities have a wavelength equal to 2l. In practical
cases, initial irregularities are likely to reduce the wrinkling strength to about 80% of the
theoretical [4].
Within the higher order sandwich theory, first published by Benson and Mayers [11], the
effect of imperfections has been investigated by for example Frostig [12], Sokolinsky et. al.
[13] and Wadee [14]. Their work is interesting but sometimes difficult to understand and
implement.
The pure geometrical nonlinearity of wrinkling, which can be predicted both by traditional
analytical methods and with FEM, does not necessarily mean that the sandwich structure fails
at that point. It just predicts the load where the deformation starts to increase more rapidly.
The structure is not damaged until the material fails. If for example a sandwich panel in a boat
hull starts to wrinkle this panel deforms. If the surrounding panels still have not reach their
wrinkling load, the total load will be redistributed and the wrinkled panel may not break. The
wrinkling induced material failure is more difficult to predict than pure wrinkling. In nonlinear wrinkling analyses it is possible to monitor both strains and stresses in the sandwich
through the loading process. Therefore it is possible to predict when the material will fail and
also which constituent will fail first, the core or the face sheet. This information could be
utilised in the selection of material combinations. In Paper D herein a method is suggested
able to find the appropriate knockdown factor and corresponding failure load.
Wrinkling and finite element modelling

With the development of computer power during the last 20 years the use of numerical
methods provide new means for structural analysis. Finite element (FE) software is one such
tool nowadays available both to academic researchers and to the industry. Several articles has
been published in the wrinkling field where FE plays an important role. Vonach [15] has
published work on for example the supporting effects from the opposing face sheet where FE

14

Linus Fagerberg

was a key instrument. Several authors, e.g. Hadi [16], has also published work based on the
comparison between different wrinkling theories and FE calculations.
Finite elements is a useful tool when investigating the wrinkling phenomenon but a few issues
are good to bear in mind. The following ideas are based on the experiences from the FE
calculations performed in this thesis. If the sandwich face sheet is much thinner than the core,
different elements should be used to model the constituents. In a 2D model the face sheet
should be modelled with beam elements and the core with membrane elements. In a 3D model
the face should be modelled with shell elements and the core with solid (brick) elements. If
two different element types is not used a poor mesh is almost always obtained unless a very
fine mesh is used. Using mixed models, where elements with different number of nodal
degrees of freedoms are used, is sometimes questioned. However, the mathematical error in
using mixed models is small compared to modelling the thin face sheets and relatively thick
core using only one type of element, where a high number of elements is required to avoid too
high element aspect-ratio and distortion.
When modelling wrinkling, elements with quadratic shape (displacement) functions are
preferred over elements with linear shape functions. The quadratic displacement functions
model the buckling mode much better with fewer elements. To achieve a decent load estimate
the element length should be shorter than half the wrinkling wavelength. If also the shape of
the wrinkling wave is important and a better load estimate is wanted, at least four elements
per wave length should be used. The core should at least be modelled with four elements
trough the thickness. The sides of the core elements should be of equal lengths, forming a
square in 2D or cube in 3D. If fewer elements are used through the core thickness, or in the
length direction of the model, buckling modes might be missed and the predicted loads
erroneous.
The model needs to be long enough in the length direction to incorporate four wrinkling
waves or the length of the model must be equal to the natural wrinkling wavelength,
otherwise a to high buckling load will be obtained.
The boundary conditions of the model can be carefully modelled using periodic boundary
conditions as in [9] and [10]. But still the length of the model is important. The length must be
an even multiple of the natural wavelength or a too high buckling load is obtained. The model
boundaries can also be modelled with simpler non-periodic boundary conditions as in, Paper
A and D in this thesis. If a unit cell is used it should be starting and ending at the peaks and
bottoms of the wrinkling wave.
Figure 8 shows a few typical finite element models used to model wrinkling. The first model
one (Figure 8a) is a 2D model with beam and membrane elements. It has a fine mesh but
relatively few degrees of freedom making it fast to run. This model was used in nonlinear
calculations investigating the post-buckling behaviour of wrinkling (see Paper D for details).
The second model (Figure 8b) is similar to the first but it is a 3D model with solid elements in
the core and layered shell elements to model the face sheet. It has the same function as the
previously described model, is more computationally intense, but can model anisotropic face
sheets easily. Using coupled degrees of freedom can make the model less computationally
intense. The stiffness matrix will however still be larger then for the previous model.

Wrinkling of Sandwich Panels for Marine Applications

15

Figure 8. Different finite element models used in wrinkling analysis.

(a) 2D model of a unit cell. (b) 3D model of a unit cell with layered shell element
modelling the face sheet. (c) 3D model with periodic boundary conditions.
(d) 3D model of a small sandwich plate used for tests.

The third model (Figure 8c) is a plate model with periodic boundary conditions modelling a
sandwich face sheet on the sandwich core. This model can be used if the load is multi-axial
and it can be used to decide for example at which load and angle the face sheet wrinkles. The
last and fourth model (Figure 8d) is a FE model used to model an actual test situation. The
boundaries for this last model are modelled as rigid plates.
When using FE calculations always remember that the result is never better than the
underlying code and always compare the results obtained from a new FE model with some
known data from tests or analysis.
Nonlinear response of compressively loaded sandwich structures

The nonlinear response of compressively loaded sandwich structures can be studied either by
means of higher order theory, e.g. [12-14] or with nonlinear finite element calculations, e.g.
Paper D herein. Both of these methods can give insight into the post-buckling response of the
structure. If the structure is beam type a limit load can be expected. If the wrinkling load is
below the global buckling load the onset of wrinkling will trigger global buckling.
Figure 9 shows examples of nonlinear response of sandwich structures. The solid lines show
the equilibrium paths for a unit cell (basic wrinkling) in compression. The three lines arise
from three different amplitudes of initial imperfections in the wrinkling wave (0.25, 0.1 and
0.01 mm). This is a local form of small scale imperfection further investigated in Paper D in
this thesis. It can further be seen from the figure that even if the basic wrinkling case shows a
nonlinear behaviour and a stiffness change in the vicinity of the predicted buckling load it still
carries an increased load. This is explained with the fact that even if the face sheet buckles
and carries no increased load the core is continuously compressed and carries a larger portion

16

Linus Fagerberg

of the load. To avoid confusion it should be said that all equilibrium paths presented in Paper
D are for the face sheet alone without including the load carried by the core, as all paths in
Figure 9 does.

Figure 9. Nonlinear paths for two different FE models and a few different
degrees of initial imperfection. The solid lines show the nonlinear response for
the basic wrinkling case and The dashed lines show the equilibrium paths for
a pre-bent sandwich beam.

The dashed lines in Figure 9 show the equilibrium paths for a sandwich beam that, in addition
to wrinkling, can buckle in a global mode. The beam model has a global form of imperfection
and is not entirely straight prior to loading but has a pre-bent shape, or in other words, a
radius of curvature. The different degrees of global imperfection (and the three dashed lines)
is from three different beams with increasing radius of curvature. The lowest dashed line is
for a highly bent beam where the radius of curvature was five times the length of the beam.
The middle line corresponds to a radius of curvature twenty times the length of the beam and
the upper line is for a beam with a radius of curvature eighty times its length, being almost
perfectly straight. All beam models also has a very small local imperfection in the upper face
sheet that triggers wrinkling and mathematically makes it easier to find the onset of wrinkling
in the nonlinear analysis. The beams were modelled with an initial curvature because the
nonlinear response is easier to find with FE analysis when the buckling direction is known. It
can be seen from the equilibrium paths that the initial curvature of the beam models both
affect the initial stiffness of the beam and at which load wrinkling will occur. At the onset of
wrinkling a limit load is reached and the load drops for an increased strain. If this would
happen to a real structure it would collapse. If the structure is plate type the post buckling
behaviour (equilibrium path) is much more smooth and without a pronounced limit load. This
is also the case if wrinkling is the dominant stability failure mode and no large interaction
between global and local modes exists.

Wrinkling of Sandwich Panels for Marine Applications

17

Summary of thesis
In Paper A both the effect of through-the-thickness anisotropy and the effect of multi-axially
loaded plates are investigated. It is shown how the anisotropy of the face sheets should be
tackled and the wrinkling load decided. The developed method further provides means to
decide at which angle in the plane wrinkling occurs and that this angle does not have to be
perpendicular to the highest principal compression stress as been assumed in earlier work, e.g.
[7]. The proposed analytical method is verified with both comparison with detailed FE
calculations and experimental tests.
In the second paper, Paper B, three possible methods to increase the load bearing capacity of
wrinkling sensitive sandwich panels are developed and exemplified. This is done using simple
mathematics and several examples. All methods described are focused on the face sheet. It is
common to always tackle the wrinkling problem by choosing a core material with higher
elastic modulus. It is shown that this is often unnecessary and that by altering the face sheet
lay-up the load bearing capacity is increased with virtually no additional weight. It is also
shown with extreme clarity that the commonly used Hoffs formula (Equation (22)) is a blunt
instrument that should not be used unless dealing with isotropic (through-the-thickness) face
sheets.
In Paper C the transition from wrinkling failure to compression failure of the face sheet is
investigated. Theoretical discussions are compared with test results for sandwich panels with
two different face sheet types and seven different core densities. For the first time in the open
literature this transition from wrinkling to compression failure is shown through actual tests.
The failure modes are investigated using fractography. The results clearly show how the
actual sandwich compression failure mode is influenced by the choice of core material,
changing from wrinkling failure to face sheet micro buckling failure as the core density
increases.
Finally, in the last paper, Paper D, a new approach is presented where the wrinkling problem
is transferred from a pure stability problem to a material strength criterion. A theory is
presented providing means on how to decide which sandwich constituent will fail first and at
which load it will fail. This is accomplished using the effect of initial imperfections in the
shape of the wrinkling wave and the effect this imperfection has on the strains within the
compressively loaded sandwich panel. It makes it possible to estimate the required
knockdown factor and failure load for each initial imperfection amplitude and material
combinations. A very good correlation is found when the developed theory is compared with
both finite element calculations and to experimental tests. The additional information possible
to attain from this derived method give insight to and develop the overall understanding of the
wrinkling phenomenon.

18

Linus Fagerberg

Future work
The formulae now used for multi-axial loaded panels are only explicit fittings to the results
from tests that often were difficult to conduct. Some are even used with the only argument
that they predict the buckling load of single skin plates. One can thus seriously doubt the
validity of these approaches for practical purposes. Instead, most engineers found themselves
having to use one of the formulae derived for uni-axial loading, simply checking one or a few
critical directions. The recent findings regarding wrinkling behaviour in anisotropic multiaxially loaded sandwich plates, e.g. [10] and Paper A herein, are currently being implemented
as a failure criterion in FE codes and will be making the life easier for the engineer designing
large sandwich structures. The post-buckling behaviour and wrinkling-global buckling
interaction is still difficult to fully predict and has to be investigated further.
There is still some lack of experimental data available and the researchers and engineers
publishing work on wrinkling and stability should try to make their test results more available
to others. A lot more experimental work is still to be done in the field of sandwich stability
and wrinkling. High-speed photography and digital speckle photography are new tools
currently at disposal that should more often be used in combination with the traditional
compression testing.
Wrinkling is a sandwich specific and sometimes unknown failure mode to people starting out
designing sandwich structures. It is therefore important to the sandwich community, as a
whole, to understand the wrinkling phenomenon and to make simple-to-use guidelines
available to the engineers. If this fails, poor structures and catastrophic failures is the possible
result. This might give the sandwich concept a poor reputation and not the proper appreciation
it deserves.

References
[1] Gough C.S., Elam C.F. and deBruyne N.A., "The Stabilization of a Thin Sheet by a
Continuous Supporting Medium", Journal of the Royal Aeronautical Society,
January 1940.
[2] Williams D., Legget D.M.A. and Hopkins H.G., "Flat Sandwich Panels under
Compressive End Loads", Royal Aircraft Establishment Report No. A.D. 3174, June
1941.
[3] Hoff N.J. and Mautner S.E., Buckling of Sandwich Type Panels, Journal of the
Aeronautical Sciences, Vol. 12, No 3, pp 285-297, July 1945.
[4] Plantema F.J., Sandwich Construction, John Wiley & Sons, New York, 1966.
[5] Allen H.G., Analysis and Design of Structural Sandwich Panels, Pergamon Press,
Oxford, 1969.
[6] Niu K., Talreja R., Modelling of Wrinkling in Sandwich Panels under
Compression, Journal of Engineering Mechanics, Vol. 125, No. 8, pp. 875-883,
August 1999.

Wrinkling of Sandwich Panels for Marine Applications

19

[7] Norris C.B., Ericksen W.S., March H.W., Smith C.B., and Boller K.H., Wrinkling
of the Facings of Sandwich Construction Subjected to Edgewise Compression, U.S.
Forest Product Laboratory Report 1810, 1949.
[8] Sullins R.T., Smith G.W. and Spier E.E., "Manual for Structural Stability Analysis of
Sandwich Plates and Shells", NASA-Contractor Report No. 1467, 1969.
[9] Vonach W.K., Rammerstorfer F.G., Wrinkling of thick orthotropic sandwich plates
under general loading conditions, Archive of Applied Mechanics, Vol. 70, pp. 338348, 2000.
[10] Fagerberg L., Wrinkling of Anisotropic Sandwich Panels Subjected to Multi-Axial
Loading, Report 2000-27, Sandwich Construction 5 Proceedings of the 5th
International Conference on Sandwich Construction, Ed. H.-R. Meyer-Piening and
D. Zenkert, EMAS Publishers, London, pp. 211-220, 2000.
[11] Benson A.S., Mayers J., General Instability and Face Wrinkling of Sandwich Plates
Unified Theory and Applications, AIAA Journal, pp. 729-739, April 1967.
[12] Frostig Y., Buckling of sandwich panels with a flexible core High-order theory,
International Journal of Solids and Structures, Vol. 35, pp. 183-204, 1998.
[13] Sokolinsky V.S., Frostig Y. Nutt S.R. Special behaviour of unidirectional sandwich
panels with transversely flexible core under static loading, International Journal of
Non-Linear Mechanics, Vol. 37, pp. 869-895, 2002.
[14] Wadee M.A., Effects of periodic and localized imperfections on struts on nonlinear
foundations and compression sandwich panels, International Journal of Solids and
Structures, Vol. 37, pp. 1191-1209, 2000.
[15] Vonach W.K., A General Solution to the Wrinkling Problem of Sandwiches, VDI
Verlag, Dsseldorf, 2001.
[16] Hadi B.K., Wrinkling of sandwich column: comparison between finite element
analysis and analytical solutions, Composite Structures, Vol. 53, pp. 477-482, 2001.