NASA/CR-1999-208994

Facesheet Structures

Wrinkling

in Sandwich

Robert Northrop

P.

Ley,

Weichuan

Lin,

and

Uy El

Mbanefo Segundo, California

Grumman

Corporation,

January

1999

The

NASA

STI Program

Office

... in Profile

Since its founding, NASA has been dedicated to the advancement of aeronautics and space science. The NASA Scientific and Technical Information (STI) Program Office plays part in helping NASA maintain this important role. a key

CONFERENCE

PUBLICATION.

Collected papers from scientific and technical conferences, symposia, seminars, or other meetings sponsored co-sponsored by NASA. SPECIAL technical,

or

The NASA STI Program Office is operated by Langley Research Center, the lead center for NASA's scientific and technical information. The NASA STI Program Office provides access to the NASA STI Database, the largest collection of aeronautical and space science STI in the world. The Program Office is also NASA's institutional mechanism for disseminating the results of its research and development activities. These results are published by NASA in the NASA STI Report Series, which includes the following report types: TECHNICAL PUBLICATION. Reports of completed research or a major significant phase of research that present the results of NASA programs and include extensive data or theoretical analysis. Includes compilations of significant scientific and technical data and information deemed to be of continuing reference value. NASA counterpart of peer-reviewed formal professional papers, but having less stringent limitations on manuscript length and extent of graphic presentations. TECHNICAL MEMORANDUM.

PUBLICATION. Scientific, or historical information from

NASA programs, projects, and missions, often concerned with subjects having substantial public interest. TECHNICAL TRANSLATION. language translations of foreign and technical material pertinent NASA's mission. Englishscientific to

Specialized services that complement the STI Program Office's diverse offerings include creating custom thesauri, building customized databases, organizing and publishing research results ... even providing videos. For more Program information Office, about the NASA STI

see the following: Home

Access the NASA STI Program Page at http'//www.sti.nasa.gov E-mail your question [email protected]

via the Internet

to

Fax your question to the NASA Desk at (301) 621-0134 Phone the 621-0390 Write to: NASA STI Help Desk

STI Help

at (301)

Scientific and technical findings that are preliminary or of specialized interest, e.g., quick release reports, working papers, and bibliographies that contain minimal annotation. Does not contain extensive analysis. REPORT. Scientific and

NASA STI Help Desk NASA Center for AeroSpace 7121 Standard Drive Hanover, MD 21076-1320

Information

CONTRACTOR

technical findings by NASA-sponsored contractors and grantees.

NASA/CR-1999-208994

Facesheet Structures

Wrinkling

in Sandwich

Robert Northrop

P.

Ley,

Weichuan

Lin,

and

Uy El

Mbanefo Segundo, California

Grumman

Corporation,

National

Aeronautics

and

Space Administration Langley Research Center Hampton, Virginia 23681-2199 Prepared for Langley Research under Contract NAS1-19347 Center

January

1999

Available

from: Information (CASI) National Technical Information Service (NTIS)

NASA Center for AeroSpace 7121 Standard Drive Hanover, MD 21076-1320 (301) 621-0390

5285 Port Royal Road Springfield, VA 22161-2171 (703) 605-6000

ABSTRACT

The analysis

purpose

of this paper wrinkling

is to provide mode

a concise

summary

of the state-of-the-art This document Instead, with

for the is not

of the facesheet review

of failure

in sandwich

structures. wrinkling. and

an exhaustive number

of the published are reviewed

research in order Designers

related to provide

to facesheet designers should

a smaller a working their design

of key papers

analysts

understanding judgment formulas when

of the state-of-the-art. deciding which one

and analysts variety

use this survey facesheet

to guide

of a wide problem.

of available

wrinkling

is applicable

to a specific

design

CONTENTS
Section 1 2 INTRODUCTION ASSESSMENT FACESHEET 2.1 FACESHEET 2.1.1 ........................................ OF THE STATE-OF-THE-ART WRINKLING WRINKLING Structures FAILURE FOR PREDICTING ......................... .................... Facesheets and 7 Facesheets and 10 Composite Facesheets... 12 13 15 Correlation with 15 with 19 28 33 34 36 3 3 Page 1

ANALYSES with Isotropic

Sandwich Solid

Cores

................................... with Isotropic ................................. with Laminated

2.1.2

Sandwich Cellular

Structures Cores Structures

2.1.3 2.1.4 2.2

Sandwich Summary

..................................... RESULTS Exhibiting Predicitions Exhibiting Predictions ............................ Reasonably Good

EXPERIMENTAL 2.2.1 Test Results Theoretical 2.2.2 Test Results Theoretical

........................... Generally Poor Correlation

............................ ...................

2.3 2.4 3 4

EFFECTS EFFECTS

OF INITIAL OF COMBINED REMARKS

IMPERFECTIONS LOADS

........................

CONCLUDING REFERENCES

................................

..........................................

ii

ILLUSTRATIONS
Figure Page

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Global

and Local Buckling Strut Under of a Sandwich Models Models

Modes

in Sandwich

Structures

................

4 4 5 7 9 11 14

Sandwich Buckling Wrinkling Wrinkling

Uniaxial Strut

Load

........................................

................................................. Elam, and de Bruyne 4 .......................

of Gough,

of Hoff and Mautner 7 ..................................... Mode in Sandwich Mathematical (1) Failure (1) Failure (2) Failure (2) Failure Panels Models ...................... .................... Core Model Core Model Core Model Core Model ........ ...... ........ ......

Non-Harmonic Summary Reference Reference Reference Reference Reference Reference Reference Reference Reference Reference

Wrinkling

of Facsheet

Wrinkling

11 Test Data Mode 11 Test Data Mode 11 Test Data Mode 11 Test Data Mode

- Isotropic - Anti-Plane - Isotropic - Anti-Plane

18 18 20 20 21 21 23 23

12 Test Data - Isotropic 12 Test Data - Anti-Plane 20 Test Data - Isotropic 20 Test Data - Anti-Plane 21 Test Data - Loading 21 Test Data - Loading

Core Model Core Model Core Model Core Model Parallel Normal

............................. ........................... ............................. ........................... Direction Direction Half.... ....

to Core Ribbon to Core Ribbon

25 25

Reference 12 Test Data Plot of k z Versus Ratio of Wrinkling Wavelength to Core Cell Size ..................................................

26 in 31

19

Typical Facesheet Imperfection Manufactured Into Test Specimens Reference 26 ....................................................................

iii

TABLES
Table Page

Summary of Theoretical Wrinkling Stresses for Sandwich Cores ............................................................................. 2 Important Assumptions Underlying Facesheet Wrinkling Predictions ....................................................................... Published Wrinkling Test Data Correlation to Theoretical Expressions Stress .................................................................

Struts With

Thick 14

Theoretical 15 for 29

iv

NOMENCLATURE

strut or panel length notch depth

(see Figure 19)

2)

Ao b

(see Figure

strut or panel width facesheet D12, facesheet laminate laminate bending bending stiffness stiffnesses in the direction of the applied load

DI
Dll,

D22, D66 through-the-thickness Young's modulus of the core

EI

Young's flatwise

modulus strength shear

of the facesheet of the core modulus

(isotropic)

G_
kl, k2

transverse coefficients empirically notch number in-plane critical width

of the core 16 and 17, respectively

in Equations determined (see Figure

K0 L m P
Pcr

constant 19) (see Equation 2) 15)

of wrinkling

half-waves

load on sandwich buckling load load buckling

strut (see Figure

RE
Ps
S

Euler buckling shear crimping

load (see Equation material

2)

cell size of honeycomb sandwich thickness

tc tl
W o

core thickness facesheet

(see Figure

2) 2) 21)

thickness

(see Figure

initial imperfection

shape (see Equation (see Figure 5)

limit of core deformation

NOMENCLATURE

(CONT'D)

x,y,z 60

sandwich

strut or panel

coordinate

system

(see Figure 21)

2)

initial imperfection buckling mode wrinkling laminate

amplitude

(see Equation

half-wavelength half-wavelength major Poisson's ratio in the direction of the applied load

_'cr

critical facesheet

'V

Vc

core Poisson's facesheet dimpling wrinkling applied major stress stress

ratio (isotropic) Poisson's ratio (isotropic)

vl
(_Tdimp

Crwr crx, Cry Crwrx,Crwry

compressive stress

stresses allowables

in the x- and y- directions, in the x- and y- directions,

respectively respectively

compressive smallest

of Crwr _ and Crwry stress compressive compressive compressive compressive shear stress stress allowable stress stress stress allowable

core flatwise or1 _rwrl or2

major major minor minor

principal principal principal principal

"_core

through-the-thickness

vi

SECTION

INTRODUCTION
The the cost development of aerospace for primary of fiber structures structures. facesheets from composite has materials resulted and the drive in renewed in this paper, to reduce both the weight use and

interest

in the

of sandwich consists of core its foam

construction

As considered bonded on either

a sandwich

structure thick,

two thin load-bearing that prevents rigidity wood, the

side of a moderately The sandwich

lightweight attains from

facesheets mainly

buckling

individually.

structure made

bending or balsa Sandwich weight). particularly shapes facesheet need

by separating cores very such

the facesheets. as those made

Solid cores from

such as those honeycomb, of strength composite into

or cellular

aluminum (ratio fiber

may

be used. to is

structures Furthermore, attractive easily

exhibit

high

structural construction facesheets

efficiencies in laminated can

or stiffness applications complex

use of sandwich since laminated

be manufactured In addition,

curved the the can

more

than equivalent

metallic

facesheets. bond structures and skins

with composite operation stiffening

materials without elements

laminates

and the facesheet-to-core tooling. With discrete sandwich stiffeners

can be cured

in a single

for complex

the use of discrete is expensive

be minimized. tooling; damaging

Cocuring

due to the need stresses

for complex

furthermore, to laminated

these

stiffeners

give rise to high localized

that can be particularly

composite

structures. and lightweight wrinkling. Since failure cores are prone to a type of local may exhibit by facesheet wrinkling Throughout is this

Sandwich failure little known or no

structures asfacesheet

with thin facesheets wrinkling load catastrophic. or simply

sandwich of these

structures structures

post-wrinkling is typically to the

carrying

capability, accurate efficient

wrinkling important report,

Hence, and

prediction sandwich

of facesheet structures.

development wrinkling is used refers

of reliable to modes

the term buckling

having sense,

wavelengths referring

up to the thickness to instability modes,

of the core. regardless of

The term the mode's The analysis

in a more

general

wavelength. purpose of this paper wrinkling summary the purpose is to provide mode a concise summary of the state-of-the-art The objective wrinkling a smaller of the here is

for the facesheet an exhaustive Instead,

of failure

in sandwich related

structures. to facesheet from

not to provide in the literature. key papers Designers

of all the research is to present with more

published number of

information understanding

in order and analysts

to provide should

designers

a working to guide

of the when

state-of-the-art. which one

use this survey

their judgment

deciding

of a wide variety of available facesheetwrinkling designformulas is applicableto a specific designproblem. This work was performedunder Task 13of NASA ContractNAS1-19347,in supportof NASA's Environmental ResearchAircraft and SensorTechnology(ERAST) Program. The technicalmonitorwasMr. JuanR. Cruz.

SECTION ASSESSMENT

2 FOR PREDICTING FAILURE

OF THE STATE-OF-THE-ART FACESHEET WRINKLING

This assessment of treating describes to calculate wrinkling facesheet the results necessary as a strength

is divided wrinkling of various

into four failure

as a short

experiments factors." rather

performed Section

to evaluate 2.3 describes problem,

the analytical the treatment

"knockdown problem,

of facesheet the presence of

than a stability Section

by considering a brief

manufacturing effects

irregularities loads.

(or imperfections).

2.4 contains

discussion

of the

of combined

2.1

FACESHEET Facesheet

WRINKLING has

ANALYSES been treated as a local, modes short wavelength by buckling sandwich mode in of

wrinkling that is one 1).

traditionally

phenomenon structures results all but wrinkling Results procedure loads stress starts is then (see

of a number The extremely

of possible small

buckling

exhibited in the

Figure

buckling

wavelength boundary analyses uniaxially

wrinkling

in the buckling a few are from special based the strut

load being cases.

insensitive useful

to structural theoretical of the

conditions used

and curvature the sandwich The onset

Many

to predict flat

on a mathematical analyses are then

model

loaded complex

strut. usual

extrapolated

to more

structures. subjected compressive loaded

used with

to predict

the onset

of wrinkling

in sandwich principal using

structures facesheet

to combined stress. This

the calculation

of the maximum stress

compared structures

to an allowable with isotropic

derived

the uniaxially and experimental compressive results

strut model. (Ref. 1,

For sandwich pp. 44-48) and the

facesheets,

analytical principal strut

evidence

indicates wrinkling

that use of only the maximum expressions derived from the

stress

in the facesheet accurate

model

in reasonably

wrinkling

load estimates. a sandwich a length strut loaded a, and in uniaxial compression as shown in Figure 2. The strut The

Consider is assumed sandwich

to have

simply t:and shear Euler

supported

at boundary

conditions

at its ends.

has facesheets have

of thickness transverse as a classical

a core of thickness rigidities column. a. as well

tc. If it is assumed as infinite

that the core

and facesheets stiffness,

infinite

through-the-thickness into a mode load of the

the strut behaves equal

In this case the strut buckles how the buckling in Figure 3.

with half-wavelength strut,

to the strut length mode

A plot showing )_, appears

varies

with the buckling

half-wavelength,

A.

GLOBAL

BUCKLING

MODE

B.

SHEAR

CRIMPING

MODE

C.

FACESHEET

DIMPLING

SYMMETRIC D.
B.Ley-97.07/BTI -03

A NTY SYMMETRIC FACESHEET WRINKLING

Figure

1. Global

and Local

Buckling

Modes

in Sandwich

Structures

tF

3_
z Y

.2_P

/4
B.Ley-97-07/BTI -04

_/

Figure

2. Sandwich

Strut Under

Uniaxial

Load

PCR

Curve 1 Curve 2 _

........... Euler Beam, GC "-) o% EC --) oo Timoshenko Beam, GC Finite, EC --) oo

Curve 3 --.-------------------Sandwich Model, Symmetric Wrinkling Curve 4 Curve 5 PS _ _ Sandwich Model, GC, EC Finite, Facesheets Extensible Sandwich Model, GC, EC Finite, Facesheets Inextensible

B. Ley-97-07/BTI -05

_,

Figure Curve structures, must 1 in Figure the transverse 3 shows shear

3. Buckling the buckling

of a Sandwich of the classical

Strut Euler relative strut Pcr, can column. In most sandwich and

flexibility analysis. 132-135),

of the core is large Treating the

to that of the facesheets shown in Figure

be considered beam equation

in the (Ref.

sandwich load,

2 as a from the

Timoshenko following

2, pp.

the buckling

be determined

1
Per

1
PE

+
tbGc

(1)

where

PE is the Euler

buckling shear mode

load, modulus

t is the thickness of the core. appears

of the sandwich, A plot showing as curve

b is the width

of the strut, in Equation

and G c is the transverse 1 varies approaches with buckling

how Pc, as given 3.

wavelength

2 in Figure 1 vanishes

As the wavelength leaving (2)

zero,

the first term on the right hand side of Equation P_ = tbG c

This shear

load

is usually mode

referred

to as the lb.

shear Since

crimping shear

load

in sandwich is actually

structures. a short

The

crimping

is shown

in Figure

crimping

wavelength

form of antisymmetric wrinkling, it can be calculated either using Equation 1 with short wavelengthsor Equation 2. In general,Equation 1 is generally usedfor the calculation of buckling loadsassociated with longerwavelengthmodesonly, while Equation2 is usedfor the calculation of the crimping load. Up to this point, somewhatsimple expressionshave been derived for buckling of the sandwich strut owing to the fact that the through-the-thickness Young's modulus of the core has been assumedto be infinite. The problem becomes substantiallymorecomplicatedwhenthis assumption is eliminated. If the through-the-thickness stiffnessof the core is consideredin the buckling analysis, short wavelengthbuckling modes,usually known asfacesheet wrinkling modes, arise. Two
facesheet symmetric simple wrinkling case, surface wrinkling with cases must be considered. to the resting varies middle surface In the first case, of the sandwich foundation. appears that the wrinkling and of the facesheets using is the respect can be predicted how 3.

model load

of a plate typically may

on an elastic with wavelength

A plot showing as curve 3 in Figure with

the symmetric In the second to the wrinkling symmetric middle in an and

the facesheets of the

wrinkle If the

in a mode core flatwise at any

is antisymmetric

respect large, The

sandwich. mode does

stiffness

is sufficiently (Ref. 3, p. 188).

antisymmetric antisymmetric

not occur

wavelength ld.

wrinkling

modes

are shown is a short

in Figure wavelength

Antisymmetric accounting

wrinkling

buckling

mode shear

of the strut calculated flexibilities. at short

while

for the core through-the-thickness wrinkling wavelength load appears appears buckling

and transverse

A plot showing in a plot of the to be

how the antisymmetric of buckling facesheets inextensible), for long respect load during versus

as a local minimum as curve 4 in Figure (the 3.

wavelengths

If the end shortening are assumed

antisymmetric

is neglected cannot

facesheets overall

the resulting

approximation

be used to predict

buckling wrinkling

of the strut load with

wavelengths. to wavelength

A plot of this approximation appears as curve 5 in Figure 3.

to the antisymmetric

All of the analyses provides may continuous

described

up to this point The

are based assumption

upon

the assumption

that the core support cores. In may

support for sandwich

to the facesheets. structures if the This with

of continuous types thin,

facesheet of cellular the

not be valid core locally lc).

honeycomb are

or other sufficiently

cellular buckle

sandwich

structures,

facesheets

facesheets

into the core cells.

type of local

instability

is known

asfacesheet

dimpling

(see Figure

2.1.1

Sandwich The earliest only by

Structures theoretical isotropic Elam,

With studies

Isotropic of the and

Facesheets wrinkling isotropic They

and Solid

Cores were such performed study was were (3) the of the

of sandwich cores. that The

struts first

considering performed inextensible, effect

facesheets and

solid, 4

Gough,

de Bruyne. directly

assumed surface

(1) the

facesheets and

(2) the core

attached

to the middle

of the facesheets, load

of the core compressive could the be neglected. core having solving was the the

stresses Gough,

in the direction Elam,

of the applied

on the stability wrinkling

facesheets struts problem (assuming The analysis 156 through with

and de Bruyne 4 considered shown in Figure

of sandwich solution stress to the function conditions. on pages

boundary biharmonic of plane

conditions equation stress)

4.

The core

involved the core

of elasticity and enforcing as their

for the the core

in a state Elam,

boundary

of Gough,

and de Bruyne 3.

4 as well

results

are summarized

164 of Reference

(1)
B .Le y.-97.0 7/BTI

(2)
ANTIS YMMEYRIC WRINKLING MODE

-06

IMMOVABLE MID DLE SURFACE

Figure

4. Wrinkling

Models

of Gough,

Elam,

and de Bruyne 4

When

the core is sufficiently

thick,

that is when

(3)

where

tlis the thickness and

of the facesheet,

tc is the thickness

of the core, modulus thickness.

Eiis the Young's of the core, In this case, ratio,

modulus

of the facesheet, can be treated which given wrinkling in Reference

E c is the through-the-thickness on an elastic foundation occur,

Young's of infinite

the facesheet the stress at

as if it rests would 4 as

theoretically

Crwr,assuming

the core

Poisson's

v c, is zero is

O'wr _ 0.794(E:EcG

c)1/3

(4)

where thickness struts

G c is the

core

transverse

shear

modulus.

Note

the

lack

of dependence thickness.

of (Jwr on the For sandwich

of the core

due to the assumption when

that the core has

infinite

with thinner

cores,

(5)

there

is some

interaction in this

between case

the faces

on opposite of the core

sides and

of the core. facesheet

While

the theoretical as shown in

wrinkling Equations where

load

is a function lower bound 3, page

properties of t: and

3 and 5, a theoretical

for O'wrthat is independent 166 as

tc for the case

v c is zero is given

in Reference

O'wr

--0.630(E:EcGc)

1/3

(6)

The analysis when assumed plotted,

of Gough, would

Elam, appear

and de Bruyne 4 results as curve 5 in Figure

in an expression 3 due to the fact

for the wrinkling that the

load that, were

facesheets

to be inextensible.

Williams, antisymmetrical analysis accounts

Legget, buckling

and

Hopkins

5 were

the

first

to solve of a strut

the

more

general core

problems

of Their as

and symmetrical shear

wrinkling

of finite

thickness.

for the transverse of the facesheets. 3 and 4 in Figure surface stress

and through-the-thickness loads calculated

flexibilities using

of the core,

well as the stretching be similar attached core axial general short to curves to the middle compressive allowed

Plots of critical 3. They

this analysis

would

also dispensed but retained

with the assumption the assumption could

that the core of the

of the facesheets on the wrinkling Legget, mode with

that the effect

of the facesheets 5 to account wavelength

be neglected.

This more of the strut. at a and small study more stress

model

Williams,

and Hopkins the long

for possible buckling wrinkling with solid,

interaction mode of the occurred cores

wavelength Legget, than

wrinkling and Hopkins

Williams, lower that load the

5 concluded

that antisymmetric constructions

always

symmetric of Gough,

wrinkling Elam, load. similar

in sandwich and

isotropic

analysis

de Bruyne 4 produced

accurate the results and

estimates

of the

wavelength performed suitable derived

facesheet using

wrinkling

Cox and Riddell 6 present to Williams, struts with Legget, thick

of a theoretical 5 in a format wrinkling

an approach

Hopkins

for use in design. by Cox and Riddell

For sandwich 6 is given by

cores,

the facesheet

O'wr _ 0.760(EIE

cG c)1/3

(7)

Hoff

and Mautner struts

7 proposed with isotropic decay one half

the simpler facesheets linearly

models and

of symmetric cores a small

and antisymmetric in Figure of width 5.

wrinkling They

for

sandwich

solid

depicted zone

assumed to be

that core deformations the smaller wrinkling energy of either stress

to zero within

w that is chosen

the thickness a total

of the core potential

or a value

that minimizes The core, are 3.

the facesheet strain in the

calculated

using as well

energy

formulation. of the

extensional neglected

of the facesheets, hence,

as the axial

strain

energy

formulation;

the theory

is related

to curves 7 generally results with

3 and 5 shown depend presented a thick

in Figure

The expressions ratio of the expression provides for a

for O'er developed strut. However,

by Hoff based

and Mautner specific

on the width-to-thickness in Reference core (one by where 7, the

on the

symmetric reasonable

wrinkling estimate

of a sandwich of O'er in all cases.

strut

w < tc/2)

This value

of O'eris given

O'wr =0.910(EIE_Gc)

v3

(8)

S
LIMIT OF CORE SHEAR AND EXTENSION

i
'%
LIMIT OF CORE EXTENSION

CORESHEAR THROUGHOUT ENTIRE THICKNESS

7
(2)

(1)
SYMMETRIC WRINKLING ANTISYMMETRIC

WRINKLING

B.I_ey-97-O7/BTI -07

Figure

5. Wrinkling

Models

of Hoff and Mautner

2.1.2

Sandwich Based

Structures

with

Isotropic Williams

Facesheets 8 reasoned

and Cellular

Cores wrinkling to

on his own observations, failure would mode have of the

that in order in Figure the assumption that the axial were based

for facesheet 2, the

be a critical thickness thickness the applied stress load was ratio

sandwich enough

strut

depicted

core-to-facesheet core of

to be large is based was zero.

to render

of a semi-infinite stress (in the direction

valid. load) small

His analysis in the core but nonzero. as the

on the assumption Previous analyses

on the assumption

that this

The assumption stress of elasticity. cores

of zero core

stress

in the direction to the classical

of the applied plane stress

is known

anti-plane

assumption

as opposed

assumption of sandwich infinitely

used in the theory structures thick

This assumption such as honeycomb. deformations yielded

is directly Williams decayed the

applicable s assumed exponentially

to the analysis the core to be away from

with cellular

and reasoned into the core.

that the wrinkling The resulting would

the facesheet facesheet stress

analysis occur:

following

expression

for the

at which

wrinkling

0.825

](

1/3

(9)

where

vlis the Poisson's It can be shown

ratio of the facesheet. that the form of the wrinkling deformations consistent assumed with by Williams the He anti-plane considered 8 result in

core

stresses

that

violate

the

equilibrium was

equations first

stress both

assumption. symmetric Direct

A consistent and antisymmetric of the core leads

formulation wrinkling equilibrium

presented

by Hemp. 9 a core after

of a sandwich equations

strut having that remain

of finite

thickness. stress

solution

the

anti-plane

assumption that are

is imposed simple stress

to expressions

for the deformations

of the core during z (see by Figure

wrinkling 2). The

polynomials

in the through-the-thickness mode presented

coordinate,

wrinkling

in the symmetric

by Hemp 9 is given

_wr

(10)

Note shown

that

a simple

beam stress

on an elastic

foundation

with

foundation

modulus

2Ec/tcCan

be

to have a buckling

O'wr=O.82Ef

IEctz E/tc

(11)

10

This purely structures showed 1.732 written

theoretical

value cores.

is suggested After

in Reference

10 for use in the design wrinkling with anti-plane

of sandwich mode, Hemp 9 was as

with honeycomb that the associated higher than 10. the

considering stress

the antisymmetric structures

wrinkling stress

in sandwich with

cores given

times

associated

wrinkling

in the symmetric

mode

in Equation Norris

et al. 11 extended solid

the analysis

of Gough,

Elam, stress how, modulus

and de Bruyne in the cross by assuming is infinitely equation on

4 to include section

allowance or a

for an orthotropic, state of plane flatwise wrinkling Goodier facesheets to results neglected. anticipated consideration Goodier could and occur strain.

core in either Boller,

a state of plane

of the core

Norris,

and Voss 12 show Young's

that the ratio of the core large, the equation for

Young's stress and

modulus developed

to core in-plane in Reference the effect

11 simplifies of the core

to the axial

derived the more axial

by Hemp. 9 of the analysis could be those a

Neou _3 evaluated

stress

wrinkling general stress

and verified published Addressing from

that it was small based on the

by comparing that

the results the effect panel

of their of core failure

assumption

the results theory, modes

of some Goodier that

tests indicating

at loads

well below

wrinkling

and Hsu 14 extended not necessarily

the work harmonic strut were

of Hemp 9 to include in the free axial direction. wrinkling

of wrinkling
Hsu
TM

were

showed

that if the ends

of the sandwich

to rotate, at one-half shape.

in a mode

localized based

at the ends of the strut (see Figure on the assumption of a purely

6 below)

the load

predicted

using

an analysis

harmonic

mode

(1)
ENDS RESTRAINED AGAINST ROTATION-HARMONIC WRINKLING
B .Le y.-97.0 Z,BTI -08

(2)
ENDS FREE TO ROTAqENON HARMONIC WRINKLING MOD E LOCALIZED ATEND S

MODE

Figure Yusuff wrinkling stresses. width, zero. _5 combined model used

6. Non-Harmonic the anti-plane by Hoff and

Wrinkling

Mode

in Sandwich

Panels s and Hemp 9 with the facesheet wrinkling

core assumptions Mautner stress 7 to estimate derived

used by Williams symmetric mode

The expressions W (see Figure The decay

for wrinkling which

by Yusuff

15 are functions

of the calculated to decay linearly extensional to

5), over

the deformation by equating 11

of the core is assumed the sum of the core

width

W is calculated

shear

and

strainenergies storedwithin thewidth W to the energystoredin anequivalentextensional spring of modulusK, thenminimizing K with respectto W. For thick cores,W < 0.5tc, and
Cr_r=0.961(EIEcGc) 1/3 (12)

for thin cores,

W > 0.5t c, and

(13)

and for cores

where

W = 0.5t c (14)

2.1.3

Sandwich In the early

Structures 1970s, attention The

With

Laminated

Composite

Facesheets fabricated of facesheet extended loaded using laminated with

focused

on sandwich

structures

composite composite to calculate with

facesheets. facesheets

earliest

theoretical by Pearce

investigation and Webber. wrinkling

wrinkling Hemp's

was performed and

16 They of uniaxially core

9 analysis panels for as

both symmetric facesheets. facesheets

antisymmetric applied long the

sandwich and

orthotropic of the

They so that

anti-plane

assumptions buckling using

accounted as well

stretching short The

wavelength loads could panel

antisymmetric be calculated with specially

loads the same

wavelength symmetric to be

antisymmetric wrinkling stress

wrinkling

analysis. was

of a sandwich

orthotropic

facesheets

shown

Dllm2+ 2(/)12 +

2066

]{a_

Jib]

2 .t_ O22(l!{a_4"

\m

]_b)

2Ec a2 mZ YgZ t ft c

(15)

where

the D_,

D12 , D22 ,

and

D66

are

the facesheet load, and

laminate

bending

stiffnesses, dimension 10--Hemp's

a is the panel transverse 9 Equation to the for

dimension applied symmetric solved with present

in the direction load.

of the applied that Equation

b is the panel to Equation

It can be shown

15 reduces facesheets

facesheet

wrinkling--when of Pearce that

isotropic

are considered. general coupling; case

Webber of sandwich however,

and Stuart 17 structures they did not

the equations laminated

and Webber exhibit

_6 for the more

facesheets results. Webber

bending-extension

any numerical Gutierrez and

TM

extended beams

the subject

analysis

of Pearce They

and

Webber

_6 to study

the stress

facesheet

wrinkling

of sandwich

to bending.

compared

the facesheet

12

on the calculated example model loaded

compressive based presented was

side

of the

beam loaded

necessary strut model

to cause (both

wrinkling facesheets

to the

wrinkling

stress For the beam

on a uniaxially in Reference

in compression). the more using enough adhesive accurate

18, the wrinkling than

stress

calculated stress

using calculated

approximately

16% higher

the wrinkling and Webber

the uniaxially to allow layer. for They

strut model.

The analysis laminated

of Gutierrez

_8 was also general

unsymmetrically showed stress 50%. stress

facesheets

and the effect

of a facesheet-to-core adhesive layer

that including of a 0.010-in-thick Hence, ignoring

the effect facesheet the effect

of a 0.005-in-thick on a 1.0-in-thick of an adhesive thin facesheets wrinkling

on the theoretical this wrinkling exists)

wrinkling stress by

core was to increase layer (if such a layer

on the wrinkling

of sandwich Shield, Kim,

panels

with very

may be very conservative. of an isotropic plane with was strain those the first facesheet elasticity calculated one resting model using on a semi-

and Shield 19 considered core using calculated

infinite,

solid,

isotropic stresses

a two dimensional using this model Their study

and compared a model to include of an the

the wrinkling Euler effect The beam

on an elastic deformation

foundation.

of its kind of axial adequate

of shear results

of the facesheet that

in addition beam

to the effect model provides

stress

in the core. of the

generally stress

indicated

the simple

estimates

wrinkling

for thin isotropic

facesheets

resting

on solid cores.

2.1.4

Summary Theoretical studies performed since wrinkling in Figure 7. 1940 have mode These were yielded equations and used to design on one sandwich of the three

structures mathematical

against

the facesheet indicated

of failure equations, derived

are based

models describing

the model are listed is given

they are based

on, and the

the Reference

how these

equations

in Table by

1. Generally,

axial compressive

stress

in the facesheet

of a strut (or beam)

(16)

for sandwich

struts with solid,

isotropic

cores,

or by

(17)

13

THIN THIN PLATE REACTING CORE CORE TRACTIONS FROM CORE ELASTICITY EQUATIONS REACTIN

PLATE G CORE

uc=uo
Z Assumed Deformation Core -_X

_X

TRACTIONS

SOLUTION

._ .................._I w

(1)
GENERAL THIN REACTING PLAFE COR E LIMITED CORE (2) ZONE OF

DEFORMATION

i
x, Oy, Txy = NEGLECT IN PLANE OF CORE STRESSES SOLUTION ELASTICITY B.Ley-9707/BTI -09

(3)
ANTI PLANE CORE

SIMPLIFIES OF CORE EQUATIONS

Figure

7. Summary

of Facesheet

Wrinkling

Mathematical

Models

Table

1. Summary
Swr

of Theoretical

Wrinkling

Stresses

For Sandwich

Struts

With

Thick

Cores

MATHEMATICAL MODEL (see Figure 7) Gough,

REFERENCE Elam, and de Bruyne 4

0.79(EIEcGc) v3
0.7 6(Ef 0.91(ERE 0.96(El E, G_ y/3 G_ y/3 E c G c )V3

(1)

(1) (2) (2) (2)

Cox and Riddell 6 Hoff and Mantner 7 Yusuff, is W < t]2

0.82(EI<G_y/3 0.82E I I_
Eftc

W = t]2

(2)

Yusuff,

15

>

t J2

00.82EI

(2) + (3)

Williams s

(3)

Hemp 9

14

for sandwichstruts with coresfor which the anti-planeassumptionis valid (e.g., honeycomb cores). Equation 17was derived assumingthe core to be in an anti-planestateof stresswhile Equation16wasderivedwithout the anti-planestressassumption.For sandwichstructures with orthotropic laminatedcompositefacesheets, an equivalentmembraneYoung's modulusshould not beused. Rather,E I should
be replaced by

where design

D I is the facesheet purposes, the

laminate

bending

stiffness

in the direction k] and in Section

of the applied generally been

load.

For

constants

of proportionality results are discussed

have 2.2.

determined

experimentally.

Experimental

2.2

EXPERIMENTAL Developing a test

RESULTS to evaluate a highly localized instability stresses 17. failure such as facesheet by tests will empirically to with with cores). that

wrinkling

is a difficult

task.

In this section, values k] and suggested by Gough, given

facesheet

wrinkling

determined In particular,

be compared derived values

to the theoretical of the parameters values k] = 0.76

in Equations

16 and

indicated

in Equations Riddell

16 and 6 (thick

17 will be compared structures structures

the theoretical solid cores), solid cores)

by Cox and Elam,

sandwich

k] = 0.63 suggested and k 2 = 0.82

and de Bruyne 4 (thin sandwich sandwich predictions, are valid in Table design for structures

suggested

by Yusuff

_5 (thick

with cellular

In order

to correlate

test measurements during the

with theoretical

it is necessary during 2. the

to ensure test.

the assumptions some violated violation assumption. terms of the

made more

theoretical

development appears

A list of is the

important

of these of the test, the test

assumptions it is necessary

If an assumption test in which relaxation

in the performance is removed Whenever

to either

a new the should

or refine possible,

theoretical results listed

analysis presented in Table

to allow in this 2.

of the in

section

be evaluated

of the relevance

of the assumptions

2.2.1

Test Results Predictions Hoff and Mautner panels

Exhibiting

Reasonably

Good

CorrelationWith

Theoretical

7 tested were

51 flat, rectangular of 0.006-

sandwich

panels

4- to I 1-in wide paper

and

10.5-in (papreg)

long.

These

made

to 0.02-in-thick acetate edge

laminated core. failures

plastic

facesheets performed support,

with a 0.066were premature declared core

to 0.741-in-thick invalid failure due due

cellulose to premature to the

The results caused air bubbles

of 39 of the 51 tests by insufficient in the core, edge and

presence

of large

15

Table2. ImportantAssumptionsUnderlyingFacesheet Wrinkling TheoreticalPredictions ASSUMPTIONNO. 1 2 ASSUMPTION Thewrinkling load is independent of boundaryconditionsandhasan associated harmonicmodeshape. Symmetricwrinkling failure in sandwichstructures alwaysoccursat loadslower thanthosenecessary to causeantisymmetricwrinkling failure. Thecoreprovidescontinuoussupportto the facesheets. Neitherthe sandwichnor the individual facesheets exhibit shearextension, bending-twisting,or membrane-bending materialcoupling behavior. Any effect of the facesheet-to-core adhesivelayermaybe neglected. Thecoremay betreatedasif it wasattached to the middle surfaceof the facesheets.

3 4

5 6

difficulties encountered in creatinga uniform stateof stressin the specimens.Furthermore,test datausedto determinethe propertiesof the materialsusedto fabricatethe specimens exhibited extremelyhigh scatter. Using the resultsof the 12 valid testsandacknowledgingthe variability in the materialpropertiesof their specimens, Hoff andMautner 7suggested using a valueof k 1 =
0.50 in Equation 16. This value of kl is 34% lower than the theoretical value, k_ = 0.76, suggested bound by Cox and Riddell k_ = 0.63, derived 6 for thick from core sandwich structures, Elam, and 20% lower and than the lower

value,

the work stress assumed stress

of Gough,

de Bruyne 4 for thin cores. of the core and a conservative The facesheet

Note that the formula facesheet approximation wrinkling of sandwich Norris and core stress thicknesses; of the given

for the wrinkling it is tacitly true wrinkling

given

in Equation Equation sandwich

16 is independent 16 represents

that for all

configurations. used today

in Equation

16, first proposed

in 1945, is widely

in the design

structures

with solid cores. hundreds facesheets of sandwich were made struts made of various steel, combinations and glass cloth of facesheet laminates. hard failure

et al. H tested The

materials. were The made authors

of aluminum, cellulose

The cores rubber. modes

of granulated observed four

cork, cellular distinct

acetate,

balsa wood, the tests.

and cellular These

modes

of failure

during

were:

Elastic facesheet

wrinkling material

of the facesheets

at stresses

below

the

proportional

limit

of the

Core failure

due to "initial

irregularities

in the facesheets"

16

3. 4.

Core failure Compressive facesheet

at stresses strength

above failure

the proportional of the

limit of the facesheets at stresses insufficient to cause

facesheets

wrinkling. focus is on wrinkling here. failure, the results in Reference (3) was of tests on specimens 11, the authors poor, of these observed that failed that the

Since in mode preparation scatter Euler

the present

(4) are not considered of the specimens

As described in mode

that failed bond results

generally several

yielding specimens

unacceptable failed in an

in the facesheet-to-core buckling mode. either. Hence,

strength. from

In addition,

the tests of the specimens

that failed

in mode

(3) will

not be considered

All failed

specimens

made

of aluminum failure values mode

and

steel

facesheets the correlation

with

solid

granulated

cork

cores and

in mode

(1)--the determined

for which

between

the theoretically

experimentally wrinkling

of wrinkling

stress extracted

is expected from

to be the best. 2 through

A plot of the 4 of Reference

stresses

of the mode

(1) specimens, Young's through yielding

Tables

11, normalized slope applied stress. for thick bound bound

by the facesheet line passing

modulus these data

versus

(EIE c Gc)I/3/EI appear the appropriate for the derived

in Figure value

8. The

of a straight to Equation

indicates

of kl to be wrinkling

16, thus

a semi-empirical value very

expression of k_ = 0.76 well.

facesheet

As can be seen in Figure solid cores

8, the theoretical data

by Cox and Riddell 6 the theoretical cores provides lower a lower

fits the experimental by Gough, data. that failed was Elam,

Furthermore,

of k_ = 0.63 derived to the experimental While the specimens cork

and de Bruyne 4 for thin solid

in mode three that

(1) were orders

made

of solid

cork lower

core, than

the Young's that of the an

modulus facesheets. expression wrinkling Reference Figure 9.

of the

material

to four the

of magnitude core assumptions

Hence,

it is anticipated stress

anti-plane

are valid

so that

for the facesheet stresses

of the form shown that failed

in Equation

17 is appropriate. from Tables (Ect I/Eitc)

A plot of the 2 through _/2 appears 4 of in value for the

of the specimens

in mode Young's through

(1), extracted modulus these data versus

11, normalized The slope

by the facesheet line passing 17, thus

of a straight to Equation stress.

indicates

the appropriate expression

of k 2 to be applied facesheet wrinkling

yielding

a second

semi-empirical value

As can be seen in Figure a generally

9, the theoretical estimate

ofk 2 = 0.82 derived stresses

by Hemp 9 (assuming in these specimens.

vI = 0) provides

conservative

of the wrinkling

17

0.0025

0.0020

OWR

K I(E

FE CGC )1/3

0.0015

aWR EF 0.0010

0.00050

0 0 0.00050 0.0010 0.0015 (E F E C GC)I/3

I 0.0020 0.0025 0.0030

EF

Figure
0.0025

8. Reference

11 Test Data Mode

(1) Failure

- Isotropic

Core Model

Owr

K 2 E F A [ ._.LF_ ' EFt C

O0 0

0.0020

_ o
-

o o
_q_ "

0.0015 aWR EF

0.0010

0 0

0.00050

I
0 0.00050 0.0010 A_ctF 0.0015 0.0020

I
0.0025

Figure

9. Reference

11 Test Data Mode

(1) Failure

-Anti-Plane

Core Model

18

The Norris stresses facesheet stresses shown

specimens

that failed

in mode

(2) were characterized irregularities from Tables

by failure

of the core due to what A plot of the wrinkling 11, normalized by the

et al. 11 perceived of these Young's of these in Figure

to be "initial extracted versus

in the facesheets." 6 and 7 of Reference appears in Figure Young's wrinkling except

specimens, modulus specimens 11. The

(EIE c GJ/3/E_

10. A plot of the wrinkling versus (E tI/E I tJ 2 is close to the of 16,

normalized experimentally wrinkling

by the facesheet determined stresses (0.25

modulus stresses

are reasonably point that

the corresponding wrinkling this single as suggested stress. provides

theoretical

for a single

represents the exception to Equation

of the specimen point,

with the smallest from Figure 7 provided from bound

in) core thickness

used.

With

it can be seen

10 that applying a reasonable 11, applying

the factor bound

k_ = 0.50 estimate
k2 = 0.60

by Hoff

and Mautner,

lower

of the wrinkling
to

Similarly, another

as can be seen reasonable and lower

Figure estimate

the factor

Equation

17

of the wrinkling work were

stress. performed performed by Norris et al. H to struts cores from (EIE _ by from to core

Norris, sandwich having made Tables GJ3/EI

Boller,

VOSS

12

extended cores. steel

the experimental A total of 63 tests with 0.375-

struts

with honeycomb tempered paper.

on sandwich honeycomb extracted versus normalized are omitted failed due

0.010-in-thick of resin-treated 2 and

facesheets

to 2.00-in-thick of the specimens, Young's

A plot of the wrinkling 12, normalized 12. by the

stresses facesheet stresses

3 of Reference in Figure

modulus

appears

A plot of the wrinkling versus (Ec tI/Eity

of the specimens 13. Data

the facesheet Figures shear

Young's

modulus

2 appears indicated of a straight

in Figure that

12 and instead

13 in cases

where

the authors The slope value plotted

the specimens

of facesheet

wrinkling.

line passing

through 16.

the data plotted The slope of a

in Figure straight

12 indicates line passing

the appropriate through 17. the data

of k_ to be applied in Figure from 13 indicates

to Equation

the appropriate value

value ofk_

ofk 2 to

be applied results lower

to Equation

As can be seen estimates fits the data slope

Figure

12, the theoretical stress;

= 0.76

in generally bound value

unconservative of k_ = 0.63

of the wrinkling very well. In Figure

however,

the theoretical that the data

13, it can be seen

are fit extremely

well by a line having

k 2 = 0.82.

2.2.2

Test Results Predictions Further empirical by Jenkinson panels cores. having

Exhibiting

Generally

Poor

Correlation

With

Theoretical

studies

of the facesheet 2 and Harris

wrinkling

of honeycomb 21 Jenkinson and steel

sandwich

panels

were

performed sandwich honeycomb

and Kuenzi 0.012tested

and Crisman. aluminum

and Kuenzi

2 tested

to 0.031-in-thick six replicates

facesheets

on aluminum of 168

They

of 28 different

configurations

for a total

19

0.0040

0.0030

0 0
_'_K 0

I = .63 --

(JWR EF 0.0020

8
0

KI

= .50

0.0010 OWR = K I(E FE CGC)I/3

0 0.0042

I
0.0044

I
0.0046 (E v EC GC ) 1/3

I
0.0048

!
0.0050

B. Ley-97-08/BTI

-o3

EF

Figure
0.0040

10. Reference
m

I 1 Test Data Mode

(2) Failure

- Isotropic

Core Model

E C tF

// "_

OWR 0.0030

K2EF

_E--_

09/

o'-'/

aWR EF 0.0020

0.0010

0 0

I 0.0010

I 0.0020

I 0.0030

I 0.0040

I 0.0050

B. Ley-97-08/BTI -04

Figure

11. Reference

11 Test Data Mode

(2) Failure

- Anti-Plane

Core Model

20

0.0060

o
aWR = K I(E FE CGC)I/3

0.0050

0.0040

aWR EF 0.0030

0.0020

0.0010

0 0 0.0020 (Ev B.Ley-97-O8/BTI -05 0.0040 Ec Go) 1/3 0.0060 0.0080

EF

Figure
0.0060 --

12. Reference

12 Test Data - Isotropic

Core Model

O'WR = K2 EF 0.0050 --

_/E

c tv EF tc

O .,,_q,

._

//

0.0040

--

O'WR EF 0.0030

0.00 20

O
0.0010

0 0 0.0010 0.0020 0.0030 d Ec tF %/ EF tc 0.0040 0.0050 0.0060

I 0.0070

-06aBLey'97"OS/BTI

Figure

13. Reference

12 Test Data - Anti-Plane

Core Model

21

tests.

Of the 28 different failed

configurations wrinkling

tested, away

Jenkinson

and

Kuenzi

2 reported

that

only

10

configurations of the wrinkling

by facesheet

from the loaded extracted

ends of the specimens. Table 2 of Reference in Figure

A plot 20,

stresses

of these Young's

10 specimens, modulus versus

from

normalized

by the facesheet stresses

(EIE c Gc)V3/Ei appears by the facesheet

14. A plot versus data to

of the wrinkling
(E c tf El

of the specimens

normalized slopes

Young's

modulus through

tJ/2 appears the appropriate 17 (Figure

in Figure value 15).

15. As before, of kl to be applied and Kuenzi

of best fit lines passing 16 (Figure k_ = 0.044 in Figure 0.82.

these

indicate Equation

to Equation 2 suggested

14) or k 2 to be applied as shown 15. in Figure values and

Jenkinson fits the data

14. A differ 2

line of slope substantially attributed

k 2 = 0.125 from the

approximately

as shown and

These

theoretical

values

of k_ = 0.76

k 2 =

Jenkinson

Kuenzi

these

discrepancies and Kuenzi

to initial 2 theorized load

waviness (as did

of the facesheets. Norris et al. H and Norris, Boller, flatwise strength, and

Jenkinson before They them) presented

Voss

12

that

the wrinkling showing stress.

was a function

of the facesheet-to-core flatwise

strength. the higher

data

that the higher Unfortunately, 1 of Reference

the facesheet-to-core close 20 reveals inspection wide

the facesheet reported to-core specimen strengths crude when

wrinkling

of the flatwise scatter

strength

data

in column flatwise

12 of Table

in the measured

facesheetfor

strengths. one

For example, standard were

an average deviation

flatwise

strength data was

of 90 psi was reported equal to 37 psi.

20; however, of some

of the test

Flatwise

specimens

as low as 20 psi. and

This is no doubt technology

attributable available

to the relatively in the late 1950s

facesheet-to-honeycomb the work was performed. in higher and more

adhesive

bonding

The availability repeatable 2_ sandwich

of modern

film adhesives This

and bonding modern

processes technology

has resulted was used

bondline

strengths.

more

by Harris and

and Crisman. Crisman 2_ tested

Harris facesheets develop investigate normal ribbon Only

panels

having

0.020cores.

to 0.040-in-thick Their objectives

fiberglass were to

on 0.40a more

to 1.00-in-thick semi-empirical

aluminum analysis

honeycomb accounting stress versus

reliable

for initial when when with

facesheet

waviness

and to

possible

differences direction

in facesheet

wrinkling core

a compressive it is applied 18 different

load is applied parallel to the

to the ribbon direction. average

of the honeycomb were conducted stress 21. values

A total of 96 tests wrinkling

on panels

configurations. in the test

measured

are presented;

the amount wrinkling facesheet direction

of scatter

data is not presented presented (Ectf/E f in Table

in Reference 1 of Reference

Plots

of the measured by the ribbon

stresses Young's of the

of the specimens, modulus core, versus in

21, normalized and normal

loading

parallel

to the

appear

22

0.0040

0 0

0.0030 0 (IWR EF 0.0020 0

0 0.0010

OWR

K 1 (E FE CGC )1/3

I
0 0.020 0.040 (Er, E c Gc)
B.Ley-97-O8/BTI -07

I
0.060 1/3

!
0.080

EF

Figure
0.0040

14. Reference

20 Test Data - Isotropic

Core Model

0.0030 O OVeR EF 0.0020

0.0010

J
J

7
......

K2EF
_ EFtf C

I
0
B. Ley-97-O8/BTI -08

I
0.020 Ec tF EF tc

!
0.030

0.010

Figure

15. Reference

20 Test Data - Anti-Plane

Core Model

23

Figures16and 17,respectively.Straightlines boundingthesedataalsoappearin Figures16and 17. The slopeof a straightline passingthroughthe dataplotted in Figures16 and 17 indicates the appropriate valueof k 2 to be used in Equation 17.
Two bound 0.82. conclusions can be drawn in these stresses from figures Figures 16 and 17. First, both the upper the theoretical and lower ofk 2 = are,

values Second,

of k 2 indicated the wrinkling 30% lower

are less than one-half loaded normal

value

of the panels stresses ribbon

to the core ribbon loaded parallel properties

direction

on average, direction. modulus, derived their

than the wrinkling

of the panels versus

to the core ribbon is the core shear to those of

Since

the only difference and Boller, Crisman

in core

transverse

G c, Harris by Norris,

21 chose

the equations

derived

by Yusuff 15 as opposed derived

and Voss _2 (similar This using choice

to the equations was made of Norris, since

by Hemp 9) as the basis symmetric

semi-empirical stress

formulation. equation derived

the theoretical

mode

wrinkling of G c while possibility symmetric Why panels reported

the theory

Boller,

and Voss _2 was independent 2_ do not mention mode rather the a

Yusuff's that mode. was the some

_5 equation

was a function failed

of G c. Harris

and Crisman (core shear)

of the panels

in an antisymmetric

than

theoretical-experimental cores reported

correlation by Norris, Boller,

of the and

wrinkling

stress

of sandwich than that

with

honeycomb

Voss 12 so much As stated earlier,

better

by Jenkinson

and Kuenzi 2 and Harris is generally can

and Crisman2_?

poor theoreticalhowever, of the

experimental another wrinkles Jenkinson specimens possible more

correlation explanation closely.

attributed

to manufacturing

imperfections; wavelength Boller,

be found

by investigating information

the theoretical by Norris,

Using

the specimen

provided

and Voss, _2 of the

and Kuenzi,

2 and Harris from

and Crisman,

2_ the critical

wrinkling

half-wavelengths

can be calculated

(19)

Equation using

19 is derived expression

in Hemp. 9 The dimpling suggested in Chapter

load

of the specimens 22

was also estimated

the following

4 of Reference

(20)

where

s is the

cell since

size

of the

honeycomb

core

material.

Equation in cell

20 is generally size caused

somewhat by the fillets

conservative formed along

it does

not take

into account

the reduction used to bond

the walls

of the cells by the adhesive

the facesheet

to the core.

24

0.020
/ (JWR ---- K2 EF i/EC tF

J
_f

0.015

Owr
EF 0.010

0.0050

I 0 0.020 0.040
Ec tF EF tc

0.060

0.080

Figure 16. Reference

21 Test Data-

Loading Parallel to Core Ribbon Direction

0.010

(JWR

EF

0.0050

,F
B.Ley-97-08/BTI Y "_F

tF
_"C

Figure 17. Reference

21 Test Data-

Loading Normal to Core Ribbon Direction 25

Using and Kuenzi size

Equation

19, it was

determined

that in all of the specimens wrinkling

reported was

by Jenkinson less than the

2 and Harris cell

and Crisman,

21 the critical

half-wavelength dimpling In cases stress where valid;

of a single

(_,/s < 1.0); than core

furthermore,

the theoretical wrinkling stress.

calculated

using

Equation smeared likelihood

20 was lower value of the

the theoretical flatwise occurring

_,/s < 1.0, use of a furthermore, is very the strong. 17 to with that for the the

Young's between

modulus the wrinkling

is no longer and dimpling stress

of an interaction

modes given

A plot of the value match the various core

of k 2 needed experimental types versus

for the expression results reported

for wrinkling Boller,

in Equation

by Norris,

and Voss 12 for specimens 18. This plot indicates

two different each specimen

the ratio the ratio

_,/s appears

in Figure conservative stress.

set, the larger

_,/s, the more of wrinkling

the correlation it is highly 21 violated

is between likely that

test data tests

and the theoretical by Jenkinson in Table 2.

predictions and

Hence, and

performed

Kuenzi

2 and

Harris

Crisman

assumption

(3) indicated

1.5

0 0
1.0

0
VALUE THEORETICAL O O []

K 2

[] []

SPECIMEN O 0.5 SPECIMEN []

SET

1 psi 2

EC = 68,600 SET

E C = 16,700psi

O'WR

= K2E F

I/

-_-_ EFt C

0.0 0.0 [3.Leg-97-08/BTI -11a

I 0.5

I 1.0

I 1.5

I 2.0

CRITICAL

WRINKLING

HALF-WAVELENGTH/CORE

CELL

SIZE

Figure

18.

Reference 12 Test Data Core Cell Size

Plot of

k2

Versus

Ratio

of Wrinkling

Half-Wavelength

to

26

Another wrinkling stress

possible reported

explanation

for the relatively and Kuenzi loads.

poor experimental-theoretical 2 and Harris Norris and Crisman

correlation 21 is their neglect Boller,

of of

by Jenkinson

the antisymmetric indicated failure that many of the core.

buckling specimens

and wrinkling predicted be the result

et al. 1_ and Norris, wrinkling

and Voss _2 due to shear mode mode or

to fail by facesheet of buckling

in fact failed

This could

of the panel Failure such compared

in an antisymmetric

the presence certainly specimens loaded core

of initial why

antisymmetric Harris and

imperfections. Crisman 2_ observed direction since

in the antisymmetric a marked drop

could load of

explain loaded

in failure load

normal

to the core ribbon direction

to the failure buckling performed

of specimens of and

parallel shear

to the core ribbon Hence,

the antisymmetric that the tests

load is a function by Jenkinson

modulus.

it is also possible z_ violated 10-in-square

Kuenzi z and Harris Pearce composite compression. Reference support wrinkling

and Crisman 23 tested on 0.25-in results struts of these tested

assumption panels made

(2) indicated of 0.01-

in Table

2. laminated in edgewise developed in

and Webber facesheets The 16. Since

in to 0.02-in-thick cores

to 0.500-in-thick tests were

aluminum

honeycomb

to be used

to validate

the theory

in previous were

studies observed

(e.g., References to fail

7, 11, 12, 20, and 21) with immediately of Pearce stable upon 3

only

on the

loaded

ends a major sides

catastrophically work exhibited at loads 20% panel

of the facesheets, with

objective supported

of the experimental to see if panels tested but below was never occurred

and Webbe_ post-wrinkling than The gauge

was to test panels behavior. Failure

all four

of all four

of the specimens wrinkling load

to 30% higher buckling load. strain

the theoretical authors readings (symmetric) areas at loads of the

(symmetric)

the theoretical observed

note that wrinkling seemed to indicate load. the was

of the facesheets some Hence, failure form

directly; occurred

however, close

of local

instability concluded panels

to the theoretical in isolated wrinkling ever by

wrinkling below panel

the authors load

that wrinkling indicating as to whether that

occurred the post

final

of the The

behavior occurred Camarda.

indeed

stable.

evidence were

or not wrinkling tests performed

was not conclusive. 24

Similar

difficulties

encountered

during

Camarda isotropic cores. tests,

24 tested

12.0-in-square facesheets were tested

simply on 0.50in three

supported

panels

with

0.024-in-thick

quasi-

graphite-polyimide A total of nine panels

to 1.00-in-thick different stress

glass-polyimide The

honeycomb results of these

configurations. calculated using

along

with theoretical 24.

predictions Note

of wrinkling

Equation

11, appear

in Table predictions facesheet. Reference

7 of Reference of wrinkling Using the

that the effective Msi) theory accounting does and not

facesheet reflect

modulus the true

used bending used

in the theoretical stiffness in Section of the 4.3 of rather

load same

(7.538 laminate

material

properties stiffness

24, the effective

modulus

for the bending

of the facesheet

27

than the membranestiffness is 11.405 Msi resulting in a 23% increasein the theoretical wrinkling loads listed in Table 7 of Reference24. Basedon the new theoreticalwrinkling load estimates, it canbeseenfrom Table7 thatthe measured wrinkling loadsareprecisely50%lower thanthe theoreticalpredictions. This is in striking contrastto the resultsreportedby Pearceand Webber. 23 Camarda 24also statesthat all of the specimenswrinkled very close to a supported edge.The authorwentto greatlengthsin designinga testfixture thatwould imposeno rotational restrainton the paneledges.As pointedout in Section2.1.2,Goodierand Hsu TM showed that a
nonsinusoidal formulas mention sinusoidal violated based is made mode wrinkling mode can occur near a supported mode edge under at one-half such support the load predicted conditions. Since by no on the assumption of Goodier was missed. of a sinusoidal
14 work

and Hsu's Hence,

in Reference likely

24, it is possible

that a critical,

non_4

it is highly in Table 2.

that the tests

performed

by Camarda

assumption

(1) indicated

Bansemir panels antennae pure cited. 0.42. test shear

and Pfeifer 25 conducted thin laminated communications a loading concluded about condition

a theoretical-experimental composite satellites. absent facesheets Their from value work every

study and cores included other

of honeycomb of those

sandwich used in the subjected previously 0.33 and the on the to

with extremely of modern loading,

typical

tests of panels study

experimental

The authors No information specimens

that an appropriate the core

for k 2 in Equation specimens

17 is between

cell size of their failed

is provided. wrinkling could

Furthermore, mode based

were

all assumed indicating of the failed unsymmetric

to have that

in a symmetric mode Finally, strong zs violated

authors' detailed constructed Hence, Table 2.

calculations descriptions using

the antisymmetric are given. that exhibit

be neglected.

No specific were

specimens laminates

the facesheets

themselves

bending-stretching assumption

coupling. in

the tests

performed

by Bansemir

and Pfeifer

(4) indicated

A summary expressions

of the correlation stress,

of the test results Equations

described

in this section in Table 3.

to the theoretical

for wrinkling

16 and 17, appears

2.3 As

EFFECTS was

OF INITIAL

IMPERFECTIONS 2.2.2, poor correlation loads These has between generally theoretical been are estimates and

mentioned measurement

in Section

experimental manufacturing however, since these

of facesheet in the

wrinkling facesheets.

attributed random

to initial in nature;

imperfections

imperfections of the facesheet

they can be expressed mode shapes

as a linear

combination The usual

wrinkling

mode

shapes shape

are orthogonal.

assumption

made

is that the mode

28

Table

3. Published

Test Data

Correlation

To Theoretical SUGGESTED k2** 0.82

Expressions

For Wrinkling COMMENTS --

Stress

SOURCE Theory

SUGGESTED kl* 0.76 (upper bound) 0.63 (lower bound) 0.50 0.63-0.76 0.50 0.63 0.044

Hoff and Mautner 7 Norris et al. 11 Norris et al. 11 Norris, Boller,

-0.82 0.60 0.82 0.125

Uncertain

material

properties.

Nearly perfectly solid cores. Imperfect

flat facesheets, solid

facesheets, cores.

and

Honeycomb

Jenkinson and Kuenzi 2 Harris and Crisman 21

Tests probably violated assumptions (2) and (3) in Table 2. Tests probably violated assumptions (2) and (3) in Table 2. Tests probably violated assumption (1) in Table 2. Tests violated assumption (4) in Table 2.

--

0.21-0.38 (ribbon direction) 0.14-0.28 (transverse direction) 0.41 0.33-0.42

Camarda 24 Bansemir and Pfeifer 25

---

wr

O'wr = k2Ef _ Eitc

corresponding other words,

to the lowest it is assumed

wrinkling

load

is the dominant

term

in this linear of the facesheet

combination. of a sandwich

In

that the undulations

of the true surface by

strut, for example,

can be reasonably

approximated

w = 6sin(

_x ]
\_,cr]

(21)

where critical trigger

6 is the amplitude wrinkling premature

of the undulations Experimental

(waviness),

x is the axial coordinate, indicate that initial

and _'cr is the imperfections flatwise failure

wavelength. failure

observations by causing

of the sandwich or a core shear containing an

either failure

a facesheet-to-core imperfections). shape

(symmetrical

imperfections) strut

(antisymmetrical in the
Yusuff 26)

If a sandwich considered, lateral

imperfection for example

of a wrinkling expression

mode

is

it has been

shown

that (see, is

the resulting

for the

displacement

of the facesheet

29

6 sin / W-Dcr 1
P where P is the applied _,. load If the and Pcr is the wrinkling facesheet contains load associated with a wrinkling in the _, is given form by mode

(22)

of

half-wavelength wrinkling mode,

an imperfection flatwise stress,

of a symmetric

the maximum

facesheet-to-core 2EcW
tc

oz -

2E c 6
k P (Pcr -1) tc

(23)

where

E c is the core flatwise If the facesheet contains

Young's

modulus

and tc is the core thickness. in the form 1, page of an antisymmetric wrinkling mode,

an imperfection is (see Reference

the maximum

core shear

stress

163)

"lJ ....

kT/)

---_--k

tc

)l_-pc-__i_

(24)

kkP
where G c is the core flatwise shear modulus and tI is the facesheet the allowable core thickness. shear

]
If either o-z exceeds panel the will

allowable fail.

stress

or "Vcore exceeds

stress,

the sandwich

This wrinkling sandwich notch L, was wrinkling good values

assumption mode struts was

of the tested

criticality

of imperfections fashion by Rogers.

in the
27 The

shape author

of the tested

symmetrical honeycomb 19, the width, critical

in a controlled intentionally and controlled

with notches

built into the facesheets. during the fabrication Given be calculated this approach these

As shown of the panel.

in Figure The notch and the

depth, then

A 0, was known measured

following _'cr' the value

fabrication. 6 o could

two

parameters ratio.

wavelength,

by a simple given

Rogers 27 obtained in the measured of such in the a study open

theoretical-experimental of facesheet-to-core on panels Various with authors

correlation flatwise

using

the scatter The not results

strength

of the test specimens. imperfections could

performed literature. measurements

antisymmetrical have made

be located

attempts

to estimate

6 in Equation

21 based

on direct

of surface

profiles

and by comparisons

of test data to theoretical

predictions.

30

A o

B.Ley-97-O8/BTI -12

.,9------XCR-----._

Figure

19. Typical

Facesheet an

Imperfection value with

Manufactured for

Into Test Specimens to assume

in Ref 26 that For larger

In determining imperfection constructions wrinkling amplitudes with wavelength,

appropriate are associated cores,

6 , it is reasonable wavelength that

longer

imperfections. 6 was

sandwich to the

honeycomb i.e.,

Williams

8 suggested

proportional

6 -

KZcr

(25)

where

K 0 is a constant to the thickness square where

to be determined of the wrinkling

experimentally. wavelength and

Wan 2s suggested inversely

that

6 was to the

proportional facesheet

proportional

(26)

Note

that if the wrinkling

loads

associated 25 and

with a wide 26 indicate

variety

of different

modes of failure

(wavelengths) in the mode with a higher

are not too far apart, associated wrinkling

Equations

that

the assumption since a mode a larger

with the smallest load and longer

wrinkling wrinkling

load may be invalid wavelength form may have for the initial theorized

associated

imperfection

amplitude. of sandwich by pressure likely of

Norris constructions applied

et al. 11 suggested with solid cores.

a different The

imperfection

amplitude was caused would

authors

that the waviness

during

the bonding from

of the facesheets loading

to the core so that stiffer compliant to define cores.

cores

be more

to "rebound" core "rebound,"

the pressure

than more

Consistent

with this idea

Norris

et al. H found

it convenient

31

(27)

whereF_ is

the flatwise

strength

of the core. on the determination that all three facesheet of the constant forms K 0 by a large number of

Equations destructive correlation

25 through Test

27 rely

tests. of theory

results

indicate By assuming

of 6 o result failure Boller,

in reasonably occurred when

good the

to test.

that

wrinkling Norris, of sandwich

facesheet-to-core the following cores

allowable formula

flatwise

stress

was exceeded, stress

and Voss 12 developed with honeycomb

for the facesheet of initial

wrinkling imperfection:

structures

that includes

the effect

_O'wr

0.82
I_ O/ [ IT-'c rT" \ I

/ Ect:
f_Eftc (28)

1+

kF_to)

Note mode

that

Equation than testing, waviness

28 is inaccurate a flatwise Norris, stress Boller,

if the imperfection failure mode.

precipitates

a facesheet the need

fracture

failure

rather

In order

to eliminate to directly

for extensive of of the

destructive facesheet wrinkling Camarda surface 2.2.2,

and Voss 12 attempted wavelength to a large

measure

the amplitude predictions

at the wrinkling

to be used number

in the theoretical The results

load and compared 24 simply and obtained used

the result

of tests. from

were inconclusive. of the panel in Section for the poor

the largest

amplitude correlation the neglect

available between

profile and test. was

measurements As discussed responsible

reasonable that

theory

however,

it is likely theory

of boundary by Camarda. load 24

effects

correlation Instead simply

between

and test obtained the critical facesheet developed

of calculating that the

wrinkling with

and

wavelength, equal for the

Harris to the waviness

and cell

Crisman size

2_

assumed core.

wrinkled

a wavelength expression number very

of the 6 o,

honeycomb appearing fitting" and test.

They

an empirical

parameter,

in Equation technique However,

28 using Harris

the results

of a large

of destructive good correlation that other

tests.

This "curve theory not poor

allowed

and Crisman in Section

2_ to obtain 2.2.2,

between phenomena,

as discussed

it is possible been

accounted correlation k2 = 0.82.

for in the derivation of test data

of Equation

28, may have stress

responsible made

for the initially using Equation

with theoretical

wrinkling

predictions

17 with

A note based on the

of caution assumption

is warranted of criticality

here.

Expressions

such

as that given symmetric

in Equation wrinkling

28 are mode. 24, it can

of imperfections mode 32 should

in the

However,

imperfections

in an antisymmetric

not be ignored.

In Equation

be seen inversely shear

that the

core

shear

stress

generated

in a panel wavelength, shear which failure

with

an antisymmetric values

imperfection of 2,, very

is

proportional

to the imperfection Consider wavelength, core shear

2,. For small In Figure not

high core load, shear

stresses

may be generated. with a zero a sudden magnitude the classical

crimping.

3, the shear possible.

crimping

Ps, is associated crimping Depending substantially

is of course triggered this

In reality,

is actually on the

by an antisymmetric core failure 2. may

imperfection. occur at loads

of this shear

imperfection,

below

crimping

load given

in Equation

2.4

EFFECTS A rigorous

OF COMBINED treatment

LOADS of sandwich As discussed the largest panels in Section principal stresses When subjected to combined 1 showed stress loads is

of the wrinkling

conspicuously panels with

absent isotropic Others and

from the literature. facesheets, taking only

2.1, Plantema compressive measured the two from

that for need be

considered. loaded struts

suggest using

allowable

wrinkling equations.

tests of uniaxially stresses be used are

them

in interaction

principal equation

compressive,

it is suggested

in Reference

29 that the following

interaction

(29)

where

o._and o._r_ are the major stress and

principal

compressive principal always ribbon

stress

and the

corresponding

allowable

wrinkling allowable coordinate

o-2 and o._r2 are the minor stress.

compressive working

stress with

and the corresponding stresses rotated into a

wrinkling system

Bruhn 3 suggests parallel to the core ignoring

with axes

and transverse plasticity effects

directions,

then applying

the interaction

equation

of the form,

facesheet

( 2
0

q- Oyy

3)
+ _ = 1 (30)

Kow_

where direction

o-xxand of the stress

o-yy are the in-plane greatest for panels compressive loaded

normal stress, in uniaxial

stresses

such

that the x direction shear stress,

is aligned

with the

Txy is the applied compression stress direction.

o-wr is an allowable ribbon direction, ribbon and

wrinkling K=I.0

along is parallel

the core

if the axis of the largest if it is not parallel

applied

compressive

to the core

direction,

else K=0.95

to the core ribbon

33

SECTION CONCLUDING
There predicting can occur have been has been extensive analytical in sandwich and structures

3 REMARKS
experimental subjected mode shape treatment to compressive as shown of the loads. problem Wrinkling 1. Analyses The first of

facesheet either

wrinkling

in a symmetric based a rigorous upon

or antisymmetric two different

in Figure

developed involves

treatments elasticity isotropic loaded and

of the response equations. cores

of the core. This treatment

treatment applicable

solution made

of the core with solid,

is most

to sandwich

structures stress

and results shown and

in an expression in Equation moduli. away 16

for the theoretical where Other from

wrinkling

of a uniaxially of the facesheet

strut of the form core thicknesses

k 1 is in general derivations the facesheets based

a function on assumed

Young's

core displacement linearly to zero over

functions a small

that decay region

exponentially

or that decay expressions analytical

adjacent

to the facesheets in Equation involves core 16. use of of a

also lead to theoretical The the second main core

for the wrinkling treatment

stress

of the form shown wrinkling stresses zero.

of the facesheet that the

problem

anti-plane panel

assumptions.

It is assumed loads

in an anti-plane shown

sandwich that

in the plane constructions than

of the applied with

are exactly cores always The

It has been

theoretically stress in the

sandwich mode

anti-plane

wrinkle

at a lower

symmetric stress

in the antisymmetric of a uniaxially 17 where

mode. loaded

expression

for the theoretical

wrinkling

(symmetric

mode)

sandwich

strut with an anti-plane

core is of the

form shown The facesheets loads have

in Equation results and been

k 2 is a constant. of tests of uniaxially cores loaded sandwich to determine of far fewer and supported to control possible failure and struts with isotropic wrinkling

of hundreds both solid and

honeycomb

conducted

facesheet

reported facesheets These

in the open literature. subjected tests one are

The results loads

wrinkling

tests on panels edges have sudden can

with anisotropic been published.

to combined extremely

on all four perform

difficult failure modes

since structures

catastrophic impair

failure

in any

of several of the

in sandwich initiates. in some While

accurate

determination correlation has

mode

in which

theoreticalit has been Poor

experimental shown

been

shown cases,

to be surprisingly particularly initial there the facesheet is strong when

good

cases, cores

to be surprisingly is generally

poor in other attributed failure. correlation

honeycomb

are used.

correlation

to small However, of

imperfections evidence stress in

that trigger in the literature

a flatwise that poor with

core or core-to-facesheet theoretical-experimental

wrinkling

sandwich

structures

34

honeycombcoresis dueto lack of validity during test of one or more of the basicassumptions, listed in Table 2, that facesheet wrinkling theoryis generallybasedupon. Test resultsreported for sandwichstrutshaving(1) predictedwrinkling wavelengths greaterthanthecorecell size,(2) predictedfacesheetdimpling loads much higher than the predicted wrinkling loads, and (3) predictedantisymmetricbuckling loads much higher than the wrinkling loads generally show correlationwith theoreticalpredictionsto within 20%or better. Initial facesheetimperfections can seriously lower facesheetfailure loads. Analyses incorporating the effects of initial facesheetwaviness generally rely on a parameter that characterizesthe amplitude of this waviness. Attempts have been made to measurethis parameterdirectly; however, the best results have been obtained when the parameter is determinedempirically from a largenumberof destructivetests. Thesetestsshouldincludethe effects of combined loads, especially in the case of sandwich structures with anisotropic facesheets.Rigoroustheoreticalandexperimental treatmentof the wrinkling of sandwichpanels subjectto combinedloadsis conspicuously absentfrom the literature. There aremany waysin which the state-of-the-art of predictingfacesheet wrinkling loads in sandwichstructures may beadvanced.Amongthe effectsthat shouldbe investigatedfurther are:
.

The effects The effects The effects The effects

of initial

imperfections loads.

in the symmetric

and the antisymmetric

mode.

2. 3. 4.

of combined of facesheet of facesheet

bending-twisting transverse shear

coupling. flexibility.

35

SECTION REFERENCES
.

Plantema, Timoshenko, 1961.

F. J., Sandwich

Construction,

John Wiley of Elastic

and Sons, Stability,

New York,

1966. New York,

2.

S. P. and Gere,

J. M., Theory

McGraw-Hill,

Allen, 1969.

H. G., Analysis

and Design

of Structural

Sandwich

Panels,

Pergamon

Press,

Oxford,

Gough,

G. S., Elam, Supporting

C. F., and de Bruyne, Medium," Journal

N. D., "The

Stabilization Aeronautical

of a Thin Society,

Sheet Vol.

by a 44,

Continuous

of the Royal

1940, pp. 12-43.

Williams, Compressive

D.,

Leggett, End Loads,"

D. M. A., ARC

and

Hopkins, Report

H. G., R&M

"Flat

Sandwich

Panels

Under

Technical

1987,

1941. III: Instability of

Cox, H. L. and Riddell, Sandwich Struts and

J. R., "Sandwich ARC S. E.,

Construction Report

and Core Materials R&M 2125, 1945. Panels,"

and Beams," Mautner, Vol.

Technical "Buckling

Hoff,

N.J.

of Sandwich-Type

Journal

of the

Aeronautical
.

Sciences, D., "Sandwich

12, 1945, pp. 285-297. Construction: 2, 1947. of Sandwich Construction," ARC Technical Report R&M A Practical Approach for the Use of Designers,"

Williams, RAE Report

No. Structures a Theory

Hemp, 2672, Hexcell Norris, of the Report

W. S., "On 1948. Corporation,

"The

Basics

of Bonded

Sandwich

Construction," C. B., and Boller, to Edgewise

TSB

124, 1986.

C. B., Erickson, Facings No. 1810,

W. S., March,

H. W., Smith, Subjected

K. H., "Wrinkling Compression," FPL

of Sandwich 1949. K. H.,

Constructions

12.

Norris,

C. B., Boller, Subjected

and

Voss,

A. W.,

"Wrinkling

of the

Facings

of Sandwich 1953. in

Constructions 13. Goodier, Sandwich

to Edgewise I. M., "The

Compression," Evaluation

FPL Report of Theoretical Vol.

No. 1810-A, Critical

J. N. and Plates,"

Neou, Journal

Compression 664.

of the Aeronautical

Sciences,

18, 1951, pp. 649-656,

36

14.

Goodier, Journal

J. N. and

Hsu,

C. S., "Nonsinusoidal Sciences,

Buckling

Modes

of Sandwich

Plates,"

of the Aeronautical S., "Theory Society,

Vol. 21, 1954, pp. 525-532. in Sandwich Construction," Journal of the Royal

15.

Yusuff,

of Wrinkling

Aeronautical 16. Pearce, Plates," 17. Webber, Panels

Vol. 59, 1955, pp. 30-36. J. P. H., "Buckling of Sandwich Panels with Laminated Face

T. R. A. and Webber, Aeronautical Quarterly,

Vol. 23, 1972, pp. 148-160. I. B., "A Theoretical Plates Using Solution for the Buckling System," of Sandwich No. 407,

J. P. H. and Stewart, With Laminated of Bristol, Face 1989.

a Computer

Algebra

Report

University 18. Gutierrez,

A. J. and Webber, Faces,"

J. P. H., "Flexural Journal

Wrinkling of Solids

of Honeycomb and Structures,

Sandwich Vol.

Beams

with Laminated 645-651. 19. Shield,

International

16, 1980, pp.

T. W., Kim, Substrate

K. S., and Shield, in Plane Strain,"

R. T., "The Journal

Buckling of Applied

of an Elastic Mechanics,

Layer Vol.

Bonded 61,

to

an Elastic pp. 231-235. 20. Jenkinson, Steel 21. Harris, ASCE 22.

1994,

P. M. and Kuenzi, Subjected

E. W., "Wrinkling

of the Facings

of Aluminum No. 2171,

and Stainless 1959. Panels,"

Sandwich

to Edgewise W. C., "Face

Compression," Wrinkling

FPL Report Mode of Buckling Vol.

B. J. and Crisman, Journal

of Sandwich

of the Engineering of DC, Defense, 1968.

Mechanics

Division, Sandwich

91, 1965, pp. 93-111. MIL-HNDBK-23A,

Department Washington,

"Structural

Composites,"

23.

Pearce, With

T. R. A. and Webber, Carbon Fibre Face Plates,"

J. P. H., "Experimental Aeronautical Investigation TM-81895, Quarterly,

Buckling

Loads

of Sandwich

Panels

Vol. 24, 1974, pp. 295-312. Polyimide Sandwich Panels in

24.

Camarda, Edgewise

C. J., "Experimental Compression," H. and Face Pfeifer, Skins for NASA

of Graphite 1980. of Sandwich

25.

Bansemir, Reinforced Chapter

K., "Local Space Conference and Core

Stability

Structures of the

With SAMPE

Thin

Fiber

Application," London, Strength

Proceedings England,

European

Third Technology S., "Face Aeronautical Wrinkling Society,

1983, pp. 421-426. Construction," Journal of the

26.

Yusuff, Royal

in Sandwich

Vol. 64, 1960, pp. 164-167.

37

27. Rogers, C.W., "Face Wrinkling As a Function of Surface Waviness," Report ERR-FW-196, GeneralDynamicsCorp., 1964. 28. Wan, C. C., "FaceBuckling andCoreStrengthRequirements in SandwichConstruction,"
Journal of the Aeronautical R. T., Smith, Plates Sciences, Vol. 14, 1947, pp. 531-539. for Structural Stability Analysis of

29.

Sullins, Sandwich

G. W., and

Spier,

E. E., "Manual 1969. Vehicle

and Shells,"

NASA

CR-1457, of Flight

30.

Bruhn,

E. F., Analysis

and Design

Structures,

Tri-State

Offset

Company,

1973, pp. C12.1-C12.52.

38

REPORT

DOCUMENTATION

PAGE

OMB No. 0704-0188

Form

Approved

Public reporting burden for t_is collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing t_e collection of information. Send comments regardingtNs burden estimate or any other aspect of t_is collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503. 1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE

Janua_
4. TITLE AND SUBTITLE

1999

3. REPORT

TYPE AND Contractor Report

DATES

COVERED NUMBERS

5, FUNDING

Facesheet

Wrinkling

in Sandwich

Structures 529-10-11-01 NAS1-19347, Task 13

6. AUTHOR(S)

Robert

P. Ley, Weichuan

Lin, and Uy Mbanefo

7. PERFORMING

ORGANIZATION

NAME(S)

AND

ADDRESS(ES)

Northrop Grumman Corporation Military Aircraft Systems Division One Hornet Way El Segundo, CA 90245-2804
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

8. PERFORMING ORGANIZATION REPORT NUMBER

10. SPONSORING/MONITORING AGENCY REPORT NUMBER

National Aeronautics and Space Administration Langley Research Center Hampton, VA 23681-2199

NASA/CR-

1999-208994

11. SUPPLEMENTARY NOTES Langley Technical Monitor: Juan R. Cruz

12a.

DISTRIBUTION/AVAILABILITY

STATEMENT

12b.

DISTRIBUTION

CODE

Unclassified-Unlimited Subject Category 39 Distribution: Availability: NASA CASI (301) 621-0390

Nonstandard

The purpose of this paper is to provide a concise summary facesheet wrinkling mode of failure in sandwich structures. This published research related to facesheet wrinkling. Instead, a smaller provide designers and analysts with a working understanding of the use this survey to guide their judgment when deciding which one of design formulas is applicable to a specific design problem.

of the state-of-the-art for the analysis of the document is not an exhaustive review of the number of key papers are reviewed in order to state-of-the-art. Designers and analysts should a wide variety of available facesheet wrinkling

14. SUBJECT TERMS Sandwich Structures, Facesheet Wrinkling, Honeycomb Core, Foam Core

15. NUMBER

OF

PAGES

49
18. PRICE CODE

A03
17. SECURITY CLASSIFICATICN OF REPORT 18. SECURITY OF THIS CLASSIFICATICN PAGE 19. SECURITY CLASSIFICATION OF ABSTRACT 20. LIMITATION OF ABSTRACT

Unclassified
NSN 7540-01-280-5500

Unclassified

Unclassified
Standard Prescribed 298-102

UL
Form 298 (Rev. 2-89) by ANSI Std. Z-39-18