NASA-CR-t94603 /M _ (/'-CM

EFFECT OF IMPACT DAMAGE AND OPEN HOLE ON COMPRESSIVE STRENGTH OF HYBRID COMPOSITE LAMINATES

by

Dr Clem Hiel, Research Scientist Division of Engineering The University of Texas at San Antonio San Antonio, TX 78249

(NASA-CR-194603) EFFECT OF IMPACT DAMAGE AND OPEN HOLE ON COMPRESSIVE STRENGTH OF HYBRID COMPOSITE LAMINATES 31Mdy FinB| 1993 Report, (Tex_s 1Jun. Univ.) 1992 77 p G3/24

N94-16836 --THRU-N94-Io837 Unc|as

0191357

Prepared

for NASA Cooperative Agreement NCC2-724 Period: June 1, 1992 - May 31, 1993

CONTENTS
I Effect of Impact Damage and Open Hole on Compressive Strength of Hybrid Composite Laminates

II Appendix:

Papers published during grant period June 1, 1992- May 31, 1993

- Low and High Velocity Impact Response of Thick Hybrid Composites Proceedings of the American Society for Composites, Seventh Technical Conference, pp 1149-1159, Oct 1992 - Damage Tolerance of a Composite Interleaved Foam Core, Journal of Composites Fall 1992, pp 155-168 - Composite Core, Composites, Sandwich Technology Sandwich with

& Research,

Construction

with Syntactic

foam

Volume 24, Number 5, 1993, pp 447-450

6 October,

1993

Effect of Impact

Damage

and Open

Hole on Compressive Laminates

Strength

of Hybrid

Composite

Clement NASA Ames Research

Hiel 1 Moffett Field, California, 94035

Center,

Abstract
Impact hardware. strength hybrid damage tolerance of is a frequently impact damage beam core. listed and open specimens Three design hole which requirement size for composites compressive

The effect was skins studied

on laminate

on sandwich foam

combine

CFRP(*)-GFRP(**) have loaded been in pure

and a syntactic

test specimen

configurations were

investigated bending series strength uniaxially comparison

for this study. The first two were sandwich flexure).

beams which

(by four point

One series had a skin damaged

by impact,

and the second of compressive a third baseline series of

had a circular hole machined with increasing open damage

through one of the skins. The reduction (hole) size was coupons beams. compared. were tested Additionally to generate

loaded

hole compression

data for

with both series of sandwich

(*)

CFRP

: Carbon : Glass

Fiber Fiber

Reinforced Reinforced

Plastic Plastic

(**) GFRP

1 Research Scientists, NASA cooperative

Agreement

NCC 2-724 with Division of Engineering

at the

University of Texas at San Antonio. (Prof. H.F. Brinson Principal Investigator)

It wasconcluded thatpost-impact trengthof sandwichskinscanbe predictedby using an s openhole analyticalmodelin which theobserved (measured) damage sizeis usedasinput. The resultsrevealthe samedependencyf strengthon-damageize (or hole size)for both o s sandwichbeamseries.This can be attributedto the local natureof the impact damage within the skin. Suchdamage typically observed sandwichbeams is in with syntacticfoam core.The baseline datafor sandwichskincouponsindicateslowerstrengthascomparedto sandwichdatafor thesameholesize. The useof an empirical-analytical procedure predict the to
sandwich loading, sandwich support buckling beams, based on open hole skin laminate analysis leads to conservative beam provided skins, by strength predictions. to skin syntactic The higher residual and strength test data of impacted for uniaxial of skin micro

post-impact

performance to the lateral and

as compared the structural modes.

coupons, foam

is attributable which prevents

global-

induced-failure

INTRODUCTION Impact hardware. combined aimed damage tolerance is a frequently materials listed design requirement in secondary intensive for composites

The success

of composite

applications

loaded structures, research programs

with its potential materials

for primary

structures damage

has spurred

at integrating

into hybrid-,

tolerant-structural

configurations.

It can be generally low, matrix damage and in comparison are sensitive and other

stated

that the damage

tolerance metals. due

of structural

composite with

materials thermoset

is

to (homogeneous) to stress material Unlike notch concentrations or geometric to homogeneous sensitive

Especially to surface

composites cuts, notches,

holes,

impact

discontinuities materials, than

which fiber

promotes reinforced This

crack-initiation composites is due edge, are to the acting of

-propagation. more

significantly development perpendicular

in compression

in tension. of a notch major reason

of tensile

stresses

which

are, in the vicinity Another

or free is the

to the fibers

and to the lamina.

formation

delaminationdue to impact, which enhances sublaminate buckling mechanismsunder compressiveloading.Most of the research work aimed at ranking different composite materials,basedon their damagetoleranceperformancehas been motivated by
("Compression After Impact" type of testing CAI

The effect

of lateral

impact

on composite

laminates

can as a first approximation on the effects based plates on of holes

be looked and other classical studies under

at as a hole (only geometric analytical have

when

damage

is localized).

Early studies have been

discontinuities solution

in composites distribution

Lekhnitski's [1,2]. Follow-up composites

for the stress on the effect

in anisotropic

concentrated tension

of hole size on the strength [3,4].

of laminated

uniaxial

and compression

Numerous stacking

analytical sequence

and

empirical

studies

have

dealt

with

parametric

effects,

such

as

or material [5-12]. These

composition investigations

on compressive have been

strength

and stability

of open

hole specimens The

motivated

by two major

objectives: of the behavior [13-17].

first was to provide performance composite

a database of bolted joints was

and analytical (or pinned) investigated

models

for design joints. during

and prediction

mechanical of fastened The

composite extensively stresses

The mechanical the last decade

information

available

to date on bearing of the hole provides joint design impact)

at the hole contour required

as well as stress assessment the effect of structural damage [19-21 multiple effects ]. This is failure material dispersed of

distribution and optimized damage composite on composite mainly mode due

in the vicinity composite

the data

for strength was to assess strength impact

[18]. The second on the residual

objective

(mainly

due to lateral Attempts strength,

compressive to predict

elements. residual

to use the open hole model on the other hand,

were not so successful pattern, damage and the mixed in most and

to the highly which

complicated are typical

geometric for

characteristics [22-26

impact

composite sizes are

laminates

]. Multiple

delaminations

of different

shapes

randomlythroughoutthe laminatewidth andthickness.This is combinedwith extensive matrix- and inter-fiber cracks and with fiber fractures.Furthermore,the compressive strengthof impact damagedlaminatesis mainly controlledby a sublaminatebuckling mechanism [27-30]. Suchfailure characteristicsouldhardly berepresented the simple c by andwell definedopenhole geometry.Similarcommentsmay be relevantfor the caseof impactdamaged composite sandwich panelswith honeycomb cores[31-33].

This reportprovidesadditionalinformationon thedamage toleranceandresidualstrength predictionof a newstructuralconfigurationwhich wasdeveloped the authors[34,35]. by This compositesandwichsystemutilizes a syntacticfoam core which has considerably morestrengthandstiffnessascompared the commonpolymericfoams.This systemis to able to sustain significant flexural loading as comparedto thin laminateswhich are designed solely for in-planeloading.Additionallythe performance composite of sandwich panelswith syntactic foam core has beenproposedas the basic building block for a compositecompressor blade[35,36]. It was shownthat in suchsandwichconstruction, damageis locally confinedwithin a well definedboundaryandmay thereforebe treated like anopenhole.This is attributedto the localenergyabsorption capacityprovidedby the syntacticfoam core [37]. This appliedcomposites technologyprogramhashadso much bite thatit hasdrawninterestfrom across thegeneralconsumer roductsector. p

Threedifferent testconfigurations havebeencompared this report.The first in

sandwich the skin beams was which were by loaded impact, in pure and for bending the (four point flexure). hole For was

two were one series carefully

damaged

second,

a circular

machined damage

through (hole) size

one of the skins. The reduction was compared. were Additionally to generate

of compressive a third series baseline

strength of uniaxially

with increasing loaded with open both

hole compression series of sandwich

coupons beams.

tested

data for comparison

Thethreemainobjectivesof thepresent esearch r were: 1. Experimentallyverify the applicabilityof an open hole model for the predictionof residualstrengthafterimpact. 2. Comparethe compressivestrengthof laminatesandwichskins with open holes by loading a sandwichbeamin pure flexure with the compressive strength of uniaxially loadedskin coupons. 3. Developan empirical-analytical procedure which canbe utilizedto predictthe residual strengthof impactdamaged sandwichbeamsunderflexure,based a simpleopen hole on skin laminateanalysisandtestdatafor uniaxialloading.

The information in
basic section associated obtained model section materials details test used

this report to build

has been organized the test samples,

in four sections. and their were the material compared test

The first describes properties. The

the

second and the the

the three procedures.

test configurations, The third for

which

in this study, and

section

gives

results an

organizes

experimental is discussed lists three

results which can

interpretation. for

Additionally, predictive purposes.

empirical-analytical The fourth and last

be utilized

conclusions

which

are supported

by this research.

MATERIALS The structural beams

AND

PROPERTIES proposed by the authors, is shown in Figure 1. These

configuration,

sandwich

were cut from larger panels overall consists thickness of five

to a length of 355 mm (14") As indicated for which

and a width of in Figure 1, the and

76 mm (3"). The sandwich properties Materials beam

is about 34 mm (1.34"). different basic materials

more

details

are given below.

The Core supplied having

(1)

for the sandwich Syntactic density

specimens

was made

of precast

syntactic

foam

(SYNTAC micro

350)

by Grace the average

Company.

It consists

of epoxy

resin filled

with glass

balloons

of 0.6 gr./cm 3.

The

skin laminate weave to the fiber

consists on both

of a central

Carbon and

Fiber film

Reinforced adhesive was

(CFRP)

laminate bond

(2), layers the glass

of glass weave carbon

sides of the skin, The central prepreg

(3) to intimately fabricated from

center reinforced

laminate.

CFRP

laminate,

unidirectional by BASF. laywere layers of the

Bismaleimide

tapes (rigidite average epoxy

G40-600/5245C)

supplied

It consists up. Two placed FM300 CFRP

of 18 plies, layers

with .14 mm (.005") fabric reinforced laminate

ply thickness, (GFRP)

and with a (0/+30/-30)3s (7781/5245C) Two

of glass and below adhesive

(4) prepregs protection corp.)

above prepreg laminate

the CFRP film (made

for external

of the skin. were placed

by American

Cyanamid

between

and the GFRP. with a total of 22 plies, the so called average "standard thickness was cured 350 at 177C cycle" (350F) in a press by with the heated prepreg

The skin platens,

laminate, following

F cure

as supplied

manufacturer.

The measured

of the cured

skin laminate

was 2.52 mm (.099").

The adhesive curable referred

used to bond

the skins to the core was a Hysol beam obtained,

EA9394

room

temperature 1 has also been

adhesive

(5). The sandwich as a "Thick

as shown (THC).

in Figure

to by the authors

Hybrid Composite"

Material The basic

Properties material foam, properties of the cured, layers unidirectional in Table CFRP 1. They lamina, the GFRP fabric, the

syntactic state the

and the adhesive temperature literature (RT.) and

are given

are designated data the was

for the cured obtained foam C365 from were for

at room available

dry condition.

Most of the constituents' The properties ( D638 of

supplier

information.

syntactic properties,

obtained

independently

following

ASTM

test standards

for tensile

compressive strengthandC273 for shearproperties). Most of the CFRP skin propertieswere computedbasedon the respective laminainputs,using compositelaminateanalysis,exceptfor the compressive strength(Ftc) andthe coefficientsof thermalexpansion (at,
obtained experimentally. a 2) which were

TEST Figure

CONFIGURATIONS 2 gives an overview

AND

PROCEDURES which were used for this study.

of the three test configurations, and the test procedure

This section

contains

a description

for all three configurations.

Series beam

1: Impact flexural

damaged

sandwich which (drop

skin

laminates;

The upper

skin

of a sandwich above, After was the

test configuration, a low velocity a visible test machine as the

was shown weight)

in Figure instrumented

1 and discussed impact system.

impacted impact, energy.

by using there The remains impact

indentation

with a diameter

D which

depends

on the impact (9.8 feet) and is

has a maximum Impac 66 test

drop height machine diameter

of 3 meters by

commercially Laboratories. attached required circular

known

made

Monterey (hardened

Research steel) tip

The impactor base

is a 16 mm (.625") the assembly which

hemispherical 86N. The

to a rigid height columns

with

weighing is driven

impactor motor.

is raised Two

to the

by a chain guide and by

winch

by an electric

lubricated is released is arrested

the impactor the

during

its fall. Subsequently The rebound of the

the weight impactor The

pneumatically automatically impact variables

impacts a braking

test-sample. to insure

system

a single to account

impact

event.

values during two photo calculated The

of the falling. cells. from

were

defined

experimentally by measurement impact energy drop impact

for the friction time between was 0.88g.

The velocity The actual

was determined maximum The kinetic average

of the elapsed just before acceleration process was

the collision was about

this velocity. response Signal

calculated the

dynamic Dynamic to the

of the Analyzer

system (type

during 3562A)

monitored accelerometer

by an H.P. was

An Endevco

type 2252

attached

top of the impactor.The impacted sandwichbeamspecimens weresimply supportedon two rollershavinga spanof about200mm(8") asindicatedin Figure3.

After impactthe visible damagesizealong the beamwidth andthe damagedepthwere measured. ubsequently damaged S the sanwichbeamswere loadedto failure in a 4-point flexure set-upas illustratedin Figure4. The loadingconfigurationputs the sidewith the damaged skin in compression. constantcross-head A speedof 1.84 mm/min (.07"/min) wasmaintained duringthis test.

Series2:
skins

Open

hole sandwich

skin laminates; (Figure

were obtained 1) before

by pre drilling

one

of the Hole

in the sandwich

beam configuration between

bonding

it onto the core.

diameters

in this case varied strength

3.05 and 22.2 mm (.12" by using the four-point The cross-head

and .87"). which now put 1

The residual

was obtained

bend procedure, was the same

the side with the hole in compression. (1.84 mm/min or .07"/min).

speed

as for series

Series only

3: Open the skin

hole skin coupon laminate. method 3-88) The which

laminates; test procedure

These

coupons

were the

obtained 3-88

by utilizing (SACMA SRM RP

followed

SRM special is based

recommended 3-88 1092

is detailed in Figure.

in reference 5. The

[38].The

SACMA on NASA

compression [39]. Specimen from

fixture

is shown

test method

dimensions

were 38.1 x 305 mm (1.5" to .44") as shown strength is shown were

x 12") and the hole diameters 5. Specimens prepared 6. The mm without and tested is as

(D) varied hole

1.5 to 11.1 mm (.06" of reference 1-88. The special D695. 14c. [40]. Respective

in Figure data were

for determination to SRM ASTM

compressive test fixure dimensions

according based shown on

in Figure x 80.8

test method x 3.18")

12.7

(.5"

in Figure

In both tests,

specimens

were loaded

to failure

in compression

at constant

cross

head speed

of 1.27 mm/min

(.05"/min)

TEST

RESULTS

Most Due

of the test results to the complex strength laminates

deal with the effect and data different obtained

of damage

(or hole)

size on laminate involved (series in the

strength. present for

hybrid for sandwich

compositions skin

investigation, skin coupon

laminates

1 & 2) and to the

(series

3) could

only be compared

by referring (the CFRP

maximum in this in

stresses case). Figure which

at failure Stress

(crcf_m,x) acting through

on the same material the different phases

phase

laminate are

distributions

of the skin coupon

shown

7. The stress is obtained shown the carbon is ct times analysis.

tr x shown by dividing on the right phase higher The

on the left hand side is the average the load at failure hand

axial compressive area. stiffness the stress was coupon CFRP The

stress stress

by the cross-sectional 7 accounts for the

distribution between carbon laminate considered Figure factor

side in Figure phase.

contrast in the by is

and the glass than the average level the

Therefore, A value carbon

as indicated, for a =1.45 at skin of the

stress.

obtained failure

stress for

O-xc in the f in-situ

phase strength

representative the stress

compressive

laminate.

8 shows y converts analysis phase by found

distribution

in the sandwich stress

skin laminate.

Here

the correction from simplified stress on

the maximum to a stress at skin failure. substituting in Table-1

compressive ?r_f, which

O'ef which

is obtained

sandwich the carbon configuration properties

is the accurate

maximum found

compressive

A value for y .882 has been the dimensions specified program. are reported

for the basic sandwich 8 and the material

in Figure

into a laminate

analysis which

The test results carbon phase

for all three configurations, at skin failure. An interpretation

below, is given

give the stress subsequently.

in the

of these results

Series 1:
Impact

Impact

damaged

sandwich

skin laminates; before, within impact damage a zone which of a sandwich

damage

characteristics: foam zone core

As was mentioned

with syntactic circular. white imprint which view This imprint

is local and well confined easily coating due

is approximately indentation boundary zone and a

can be inspected fabric

to a damage in Figure visible

induced 9. The

at the GFRP representation

as shown non-

of this beyond sectional of

is a good

of the internal as non-damaged in Figures

CFRP

damage

the skin may of the local

be considered shown as damage

as can be seen

by the cross

damage

10 and 11. Hence,

the transverse

measure

this imprint configuration energy,

was defined

size to be compared to be directly results

to the open

hole diameter

in test

2. The damage in Figures

size seems

related

to the increase

of the impact study was [35] also

as shown

11 and 12. Similar syntactic beams

were obtained Additionally

in another this trend

for sandwich demonstrated tests [36]. Effect from defined

beams

with interleaved of sandwich

foam core. subject

for the case

to high velocity

(ballistic)

impact

of damage the flexural here as

on residual test results the

strength of impact skin

: Nominal damaged stress

Compressive sandwich beams. nominal

Strength Nominal

was

obtained is area size size less at

Strength

maximum

per

unit failure.

cross-sectional of the damage damage less and

(disregarding causes range

the presence

of damage) effect larger

at sandwich on nominal damage

Increase

a pronounced shown

deterioration 13. For

strength

for the small becomes following in Figure beams

in Figure Shear failure,

sizes

this effect boundary, (Shown

significant. the 30-deg results velocity were

originating

from the damage failure

the fibers

angle was seen as the controlling also found impact in earlier studies

mode.

14a). Similar subject to high

for the case of sandwich

(ballistic)

tests [41].

Series

2: Open hole sandwich

skin laminates;

i0

The relationshipbetweenNominal Compressive Strengthand hole diameter,shown in Figure 15, was obtainedexperimentallyfrom four point flexure tests on open hole sandwichbeams. hedatawasseento follows a trendsimilarto thatobservedin Figure 13 T for the
post- impact sandwich specimens.

The

observed

failure at this

mode, stage,

shown that

in Figure size damage

14b, was definition on residual

also similar. for the

It may

therefore sandwich

be is

concluded, justified

damage

post-impact

and that the effect open

of impact

strength

can be evaluated

based

on respective

hole sandwich

data and analysis.

Series The

3: Open relationship

hole

skin coupon Nominal

laminates; Compressive Strength and hole diameter, shown in

between

Figure coupon except range. observed along

16 was obtained specimens. that The

experimentally data follows degradation mode,

from uniaxial a trend similar

compressive to that obtained especially

tests on open for series hole

hole skin #1 and #2 diameter to the one failure to

the strength

rate is higher, shown in Figure beam

in the small

The observed for series fiber

failure

14c was found

to be similar

#1 and #2 sandwich orientation which

specimens

i.e, predominantly mode

shear

30-deg

at the CFRP was shown

laminate. earlier

This failure 6.

is also similar

that for virgin

specimens,

in Figure

INTERPRETATION

A. Analytical The analytical

background formulations hole which were developed in references[2, to examine 3 & 41] for the

prediction results beams

of open

compression analytical

strength

will be used

the experimental of sandwich defined

and to provide based on damage

tool for prediction

of post-impact

strength

size measurements.

This is justified

in light of the well

11

localized foam failure, Figure core

impact

damage

which

was found

to be typical

for the sandwich

having

syntactic stress at in

in the present

investigation.

According

to this model

the compressive width

crr_ is a function 17. in Equation by dividing

of the hole diameter

(D) and the specimen

(W) as shown

As indicated be obtained Y

1, the notched the strength

strength

(which wide

is experimentally laminate

measurable)

can

of an infinitely

by a correction

function

oO

(1)
trN -Y(D/W) The correction function can be calculated as follows

YD/W

2 -I-(1 -D/W) D 3(1 -_)

(2)

Strictly called

speaking the "isotropic

this equation width

is only correct factor".

for isotropic Gillespie

laminates

and therefore shown

Y is

correction is applicable

et al [42] have laminates

nevertheless values smaller

that the above than .25, which

expression

to orthotropic

for D/W

was the case in this investigation.

According orthotropic

to Whitney plate is related

and

Nuismer

[3]

the

notched

strength

of

an

infinitely

wide

to the unnotched

strength

by the following

equation;

2o- 0 1-_

-3o'-]
with

(3)

D 2j -D+2a

(4) i

12

with i=s for the sandwich skin i=c for the skincoupon and K_ =1 +_/2 _
Equation 1. was originally / E, -Vxy +Ex/2Gxy used to predict in a multi-ply the variation laminate. zone of tensile The (5) strength due to a through introduced adjacent by criterion fitting has to

the thickness represent to the

hole (or notch)

parameter

a i was stressed

a distance, hole. The data

chracterizing distance assuming

the damage as a free stress loaded

in the highly to be

region

is used

parameter over

determined zone. This

experimental been extended

an average compression

the damage

to include

laminates

by Nuismer investigation

and Labor

[4]. Two

damage

size-parameters

have

been

used for the present configuration.

a_ for the sandwich

configurations

and a c for the coupon

B. Comparison

of the effect

of impact

damage

with that of an open

(drilled)

hole. 18 The The

The residual strength as a function data for both

data which was shown (open

in Figures hole) size

13 and 15 is replotted (D) and specimen dependence curve, using

in Figure (W).

of the ratio of damage series results of flexural were

width

test beams

show

a similar

on D/W.

experimental

therefore

represented

by a single

the analytical

open hole model

of Equation

1, with a_ = 9.3 ram. as the curve fitting

parameter.

This result compression

indicates strength

that the use of an open hole model is justified required for the present case.

for the prediction It also means history that

of post-impact the damage velocity, size

(D) is the only load, etc.) does

parameter not need

and that the impact to make

(impact

energy,

to be known

quantitative findings

residual

strength

predictions.

This latter point

is also substantiated

by earlier

[33,35].

13

C. Comparison
laminates. The stress

of the

effect

of open

holes

in sandwich

skins

and

skin

coupons

at failure

of

unnotched specimen

samples

was

used

to normalize

the strength

data 19

obtained compares which

on the two different this normalized

geometries

used in test series #2 and #3. Figure of the normalized trends

stress

at failure

as a function

hole size (D/W), loss with skins

was also used to represent D/W can be observed. more gradually

the data in Figure The general trend

18. Similar appears

of strength

increasing

to be that the sandwich

lose their strength

than the skin laminate

coupons.

The

respective

analytical

plots

based

on

Equation

1 reveal

two

different

empirical

parameters; coupons the stress

a higher (ao=2.7mm) distribution Hence, skins

one for the sandwich The shape based parameter and

skin ( as=9.3mm a physical level and

) and a lower significance may

one for the skin by quantifying of notch

a_ may have

its singularity

be a measure

sensitivity. that sandwich

on comparing

the a values

in the two cases and impact

it may be concluded damage as compared

are much

less sensitive

to open-hole

with skin laminates. D. Net Strength The Net Strength in the cross drilling. Comparisons (NS) is defined as the load carrying capacity of the material that remains or as a

section

of the skin material

after part of it has been effects

taken out by impact

It's advantage

is that local stress concentration size or hole diameter (NNS). Changes

can now be compared

direct function "Normalized for values coupons

of the damage Net Strength"

(D). This was done in Figure for sandwich in NNS skins were is noticeable

20 for the only minor for skin

in NNS

of D above with values

4 mm (.16").

A continuous

decrease

of D up to 10 mm (.39").

Additionally (NSL=

a Normalized It is plotted

Net Strength as function

Loss (NLS)

variable

may be derived (Figure

from NNS 21 ).

1-NNS).

of hole diameter

for both cases

14

The sandwichskins,with NSLmax=26%, clearly havea substantiallybetter performance thanthe skincouponswith NSLmax--42%.

The betterresidualcompressive strengthof damaged

as compared possible 1) The with the respective performance

(open-hole) coupons

sandwich

skins

(in-situ) to two

of skin

is attributable

reasons: presence of a supportive buckling induces structural mechanism in-plane [44,45]. core stabilizes the skin resistance against a

compressive 2) The which

sub-laminate

[43]. bi-axial compressive stress state in the skin

presence

of the core

may improve

its axial strength.

CONCLUSIONS
An experimental sandwich loaded beams open-hole investigation (post-impact laminated and analytical was conducted and open-hole) coupons on three loaded test configurations; and a series two series of

in flexure composition

of uniaxially skins.

with the same

as the sandwich conclusions:

The test results

consideration

lead to the following

The

assumption with

that

the

localized foam core,

impact can

damage, be modeled strength

which

is typical hole

for

sandwich is justified. based on a

construction Engineers simple input.

syntactic

as an open of sandwich

can therefore

calculate model,

the post impact in which

skins

open hole analysis

the observed

(measured)

damage

size is used as

- The similar

residual

compressive

strength (hole)

of post-impact

and open-hole

sandwich

skins

show

dependency

on damage

size. This can be attributed

to the local nature

of the

15

impactdamage within
foam core.

the skin, as is typically

observed

in sandwich

beams

with syntatic

- The

baseline data skins

data

for sandwich hole

skin coupons size.

indicates

lower

strength Strength trend

as compared for open-hole

to

sandwich sandwich diameter.

for the same

Normalized follows

Compressive a similar

and for laminate

skin coupons

with increasing

hole

- The Net

Strength

Loss (NSL)

(derived

from

net stress

at failure)

is significantly

higher

for open-hole

skin laminate

coupons

than for its sandwich

skin counterparts.

-The higher to the better state

performance resistance in the

of open hole (or impact of the skin to compressive skin, both effects which

damaged) sub-laminate are due

sandwich buckling

skins

is attributable

and to the biaxial of the syntactic

of stress

to the

presence

structural

foam core.

ACKNOWLEDGMENT

The author the test

would

like to thank and

Dr. Howard Branch

Nelson,

Roy Hampton, Ames

and Dave Center

Chappell for Branch

of their at the are

Engineering

Analysis

at NASA

Research

support the

and encouragement for the

and to Paul Scharmen manufacturing

of the Model of the

Development

Center

high-level

specimens.

Additionally, plots

contributions greatly

of Mike

Luft in keeping

track of the experimental

data and making

appreciated.

16

References
[1] Lekhnitskii, by Tsai, S. G., " Anisotropic Plates " translated and Breach in russian, from the second Science 1957. Translated in russian, criteria Materials from russian, 1968. for laminated Vol. 8 July composites 1974, pp. NASA 'rT russian New edition York.

S.W.,

and Cheron, Original

T., Gordon publication

Publishers,

1968, pp. 171-190. [2] Savin, G.N.,

" Stress

distribution Original R.J.,

around holes" publication "Stress

F-607,1970, [3] Whitney, containing 253-265. [4] Nuismer, J.M.,

pp. 227-324. and Nuismer,

fracture

stress concentrations".

J. of Composite

R.J. and Labor,

J. D.,

" Application

of the average Vol.

stress

failure

criterion:

Part II - Compression," [5] Sun, C.T. and Lao,

J. of Composite J., " Failureloads Science

Materials, for notched

13, Jan.

1979, pp. 49-60. laminates with a

graphite/epoxy

softening [6] Lubowinski,

strip, " Composite E. G., Guynn,

and Technology,"

Vol. 27, 1985, pp. 121-133. rate Sensitivity 1988, 30p. of

W. E., and Whitcomb, NASA

J. D., "Loading August

open hole composites [7] Guynn, of

in compression,"

TM 100634, Investigation

E. G. and Bradley, failure 23, May

W. L., "A detailed in open hole

of the micromechanisms J. of Composite

compressive Vol.

composite

laminates,"

Materials," [8] Guynn, zone

1989, pp. 479-504. W. L., Measurments laminates 8, March loaded of the stress supported by the crush

E. G. and Bradley, in open hole

composite Vol. " On Structure, Kao, Vol. C.S.,

in compression,"

J. of Reinforced

Plastics [9] Larson, holes," [10] Lin, Composite

and Composites, Per-Lennart, Composite C.C. and

1989, pp. 133-149. of orthotropic stretched plates with circular

buckling

Vol. 11, 1989, pp. 121-134. "Buckling of laminated plates with holes," J. of

Materials,

23, June

1989, pp. 536-553.

[11]

Chang,F. K. andLessard,L. B., "Damagetoleranceof laminated composites containingan openhole andsubjected compressive to loadings:PartI--Analysis," J.
of Composite Materials, L. B.and an, open Vol. Chang, hole 25, Jan. F. K., and 1991, pp. 2-43. "Damage to tolerance of laminated loadings: composites Part II:

[12]

Lessard, containing Experimental,"

subjected Materials,

compressive

J. of Composite R.A.

Vol. 25, Jan. 1991, pp. 44-64. G. S., "Strength Vol. of Mechanically 1982, Fastened

[13]

Chang, Composite

F. K., Scott, Joints,"

and Springer,

J. of Composite Scott, R.A.,

Materials, " Failure

16, Nov.

pp.470-494. comntaining 18, May 1984, pin

[14]

Chang, loaded

F. K. and

of composite

laminates Vol.

hole-method

of solution,"

J. of Composite

Materials,

pp. 255-289. [15] Mahajerin, an orthotropic [16] Tsujimoto, joints," [17] E. and plate," Sikarskie, D. L., "Boundary Materials, element study of a loaded hole in

J. of Composite

Vol. 20, July failure analysis

1986, pp. 375-389. of composite bolted

Y. and Wilson,

D., "Elasto-plastic Vol. 20, May of bolted

J. of Composite I., " On Materials, S. W.,

Materials, bearing Vol.

1986, pp. 236-252. graphite/epoxy laminates," J. of

Erikson, Composite

strength

24, Dec. Design" 18.

1990, pp. 1246-1269. Fourth Edition, Think Composites: Dayton,

[18]

Tsai, Paris,

"Composite

and Tokyo.

1988, Section

[19]

Williams, holes

J. G. "Tough

Composite strength

Materials;

Effect

of impact strain fiber

damage laminates."

and open NASA

on the compression Publication C. C. and

of tough

resin/high

Conference [20] Chamis, damage 355.

2334, Ginty,

1984, pp 61-79 C. A., predictive " Fiber composite ASTM structural STP IOI2, durability and

tolerance:

Simplified

methods",

1989, pp. 338-

[21]

Jegley, D. C., "Compression behavior of graphite-epoxy and graphitethermoplastic panels with circular holes or impact damage" NASA Conference Publication3087Part2, 1990,pp. 537-558.

[22]

Hsi-Yung T.W. and Springer, G. S., " Measurmentsof matrix cracking and delaminationcaused impacton composite by plates"J.
22, June 1988, pp. 518-532. T.W. and Springer, plates" G. S., " Impact induced Materials, stresses, Vol. strains, 22, June and 1988, of Composite Materials, Vol.

[23]

Hsi-Yung delaminations pp. 533-560.

in composite

J. of Composite

[24]

Lesser, composites"

A. J. and Filippov, 36th International J.H. and Williams, loaded

A. G., " Kinetics SAMPE J. G.,

of damage April

mechanism

in laminated

Symposium,

1991, pp. 886-899. of Technical graphite-epoxy Memorandum

[25]

Starnes, structural 84552.

Failurecharacteristics NASA

components 1982, 24 p.

in compression"

[26]

Chai, H. and Babcock in delaminated laminates", R. D.,

C. D., "Two-dimensional J. of Composite Materials,

modelling Vol. K. N., J.

of compressive

failure

19, Jan. 1985, pp. 67-98. "Buckling of a damaged Technology &

[27]

Marshall, sublaminate Research Vol.

Sandorff,

P. E. and Lauraitis, laminate" ASTM

in an impacted 10, No. 3, Fail Shragai, A.

of Composite

1988, pp. 107-113. " Effect of impact beams", loading on damage and residual Vol. 14,

[28]

Ishai, compressive

O. and

strength

of CFRP

laminated

Composite

structures,

1990, pp. 319-337 [29] Shalev, composite [30] Soutis, composite pp. 536-558. D. and Reifsnider, laminates", C. and Fleck, plate K.L., "Study of the onset of deIamination at holes in

J. of Composite A.F.,

Materials,

Vol. 24 Jan. failure

1990, pp. 42-71. fiber T800/924C 24, may 1990,

" Static compression hole", J. of Composite

of carbon Vol.

with single

Materials,

[31] Gottesman,T., Bass,M. and Samuel,A., "Critically of Impact Damagein Composite SandwichStructure," 6th InternationalConferenceof CompositeMaterials,Vol. 3, 1988, pp. 327-335. [32] Sommers, M., Weller, T., and Abramovich, H., "Influence of predetermined dalaminations bucklingbehaviorof compositesandwichbeams," Composite Structures, on
Vol. [33] Kim, 17. 1991, pp. 292-329. C. G., AND Jun, E. J., " Impact Materials, resistance of composite laminated sandwich plates",

J. of Composite [34] Abrate, April [35] Ishai, core,"

Vol. 26, No. 15, composite

1990, pp. 2247-2261. materials", Appl. Mech. Rev. Vol. 44, No. 4

S. "Impact 1991. O., and Hiel, ASTM

on laminated

C. "Damage

tolerance Technology

of a composite & Research

sandwich Vol.

with interleaved 14, No. 3, Fall

foam 1992,

J. of Composite

JCTRER,

pp. 155-168. [36] Hiel, C. and Ishai, O., "Design of highly March damage-tolerant sandwich panels," 37th

International [37] SACMA Fiber-Resin [38] Standard Publication [39] SACMA Composites. [40] Whitney, materials," [41] Daniel,

SAMPE Recommended Composites. Tests for

Symposium, Test

1992, pp. 1228-1242. Compression 3-88. Revised Edition, NASA Reference Properties of Oriented

Method

for Open-Hole method, Composites SRM

Recommended Resin 35 pages.

Toughened 1983,

1092, July Recommended

Test Method Method,

for Compressive SRM 1-88. modes

Properties

of Oriented

Fiber-Resin

Recommended J.M. and Guihard

S. K. " Failure SAMPE

in compression April

testing

of composite

36th International I. M., " Behavior Mechanics, "Biaxial

Symposium,

1991, pp. 1069-1078. holes under biaxial loading,"

of graphite/epoxy

plates

with

Experimental [42] Daniel, I. M.,

Vol. 20, No. 1, Jan. 1980, 8 pages. testing of [02/+45]s graphite/epoxy plate with hole," Experimental

Mechanics, [43] Gillespie,J.W., predictions,"

Vol. 22, No. 5, may. and Carlson, Composites L.A.,

1982, 8 pages. "Influence of finite width on notched laminate strength

Science

and Technology,

Vol 32, 1988.

List Detailed description

of Captions (Virgin) sandwich beam configuration.

of non-damaged configurations

Illustration Flexural

of specimen test set-up. Impact Loading Loading

for the different

test series

Fi__ig_. Low velocity 4 Compression Compression Stress Stress External Fig. _: 10 :SEM Micro

test set-up: device device for standard for standard testing testing of virgin skin specimens. skin specimens. under under impact uniaxial flexure. loading. energy: 90J ) compression.

of open-hole

Distribution Distribution damage of Typical graphs

and formulation and formulation imprint on sandwich

for hybrid for hybrid

skin laminate sandwich beam

skin due to low velocity impact damage damaged

Cross-section

with low velocity of sandwich

(Impact under

of cross-section

specimens

different

levels

of impact Fig. 12; The effect

energies of low foam velocity core. size on Nominal residual strength of post-impact sandwich impact energy on visible damage size of sandwich beams

with syntactic Fig. 13: The effect (series Fig. 14: Typical 1). failure

of damage

modes

for:

a) Impact-damaged b) Open-hole c) Open-hole

sandwich

beam beam

under under

flexure flexure

skin of sandwich skin coupon of open-hole of open-hole under

uniaxial

compression. specimens (series (series 3) 2).

Effect Fig. Fig. Fig. 16: Effect

of hole diameter of hole diameter model

on nominal on nominal

strength strength

sandwich

skin coupons strength. strength

17: Open-hole 18: Comparison beams.

and formulation damage

for prediction

of compressive on residual

of impact

and open

hole effects

of sandwich

Fig.

19; Comparison strength

of open-hole beams

size per unit width vs. skin coupons.

effects

on normalized

nominal

compressive

for sandwich

Comparison the effectof open-holesizeon normalizednetcompressivestrengthfor of sandwichbeamsvs. skincoupons.

Fig. 21" Comparison beams of the effect of open-hole size on net compressive strength loss for

sandwich

vs. skin coupons.

<_,-_

. ,._

C 0 0
Wi(

om

8
= v
1 i 0

<

el,i_

"lee
o0

0
_m _m

0 L

g_dW

.lie

E
i

em

0
el

E
ee o_

el

rj_

c_
_ 0

E
om

.
_

_o H

_,_

olml

!-.

'

2_

It

Figure 5. Compression Loading device for standard testing of virgin specimen ORIGINAL P-A_
BLACK AND WHITE PHOTOGRAPH

"i"ii;ii

;_ _

Figure 6. Compression loading device for standard OmGrNAL testing of open-hole skin specimens PAGE
BLACK AND WHITE PHOTOGRAPH

II

II

II

Figure 9. External

damage imprint on sandwich impact loading

BLA,.,K AND

skin due OmG_NAL _:AGr_

WHITE PHOTOGR.AP_-_

to low velocity

I ORtG,NA,.

PA_ PHOTOGRAPH

BLACK

AND

WHIt"E

Figure 10. SEM of typical cross section with low velocity impact damage (impact energy=90 J)

47 J (34 ft-lb)

69 J (51 ft-lb)

90 J (67 ft-lb)

136 J (100 ft-lb)

180 J (132 ft-lb)

Figure 11, Micrographs of cross-section of sandwich specimens damaged under different levels of impact energy

7_

O O0 T.-

O v--

.=
i..

O .D

_J Ol

I
,.. CO x-k0 T.O,I ,r05 D t_ O

zO *mm

(ram)

_zis

oSt_mEfl

_lqISIA

Q0

a_

o_

,m

_._.

e_

(gdIA[) _Jn!!_l

q_!A_puus _u ss_lS

_A!SS_Jdmo_) leu!moN[

U!_lS "X_l_

I
!

i
|

i
i

i
|

Figure 14. Typical failure modes for:a)Impact-damaged sandwich beam/b)open hole skin/c)open hole coupon

om_

T.-

C,d

emm

o
Xl
s.Im I

0 r_2 0

oo

k_

_
_

_
l_u!moN u!_IS "x_IA/

(_dlA[) oJnl!_I q_!_pu_s

ssoJ_S OA!SSOadmoD

o/
/
o

w-

om

8
(VdIAI)

8
osnl.t_l

8
upls _

8
ssoJ1S

8
OA!SsoJdmo_)

_
IBu!moN

8
"X_IA[

Figure

17 : Open Hole Model and Formulation for Prediction of Compressive Strength.

o
r..)*.=
l

/
0 o oo 0 0 0 0 0 0 0 0 0 0 0 0

1
0 0 0

o
o

_m

(gdIAL) oanp._I

m._iS lv d_I_ID u! ssoalS

aA!ssoadmoD

mnm!xvIAI

C_

.L
_0 rle_

_.-r_

=
e_

_"cl
............................. i_/_ .............. _ ........

_r.f)

ee

f_ 0 ! II
oU

I
o _Jnp.e_l u_s _ d

V
o o

_v d-H_ID u! ssoJ_ S _A!SS_Jdmo D lvm.mol_ P_..igm.mNI

oJ

N
oi

em

E E

!
......... i ........ C;

=
0 i
o_

l
o d

,-

oJnl.tUd

upis

l_ d_l_I_3 u.t ssoJ_S OA!SsoJdmo_)

1_H POZt.lumJoH

_
0

emm

CD 0
i

0
T-

OO

0 o

(%) ssol q_uo_S oA!ssoJdmoD _oN

_o

II Appendix
Papers published during grant period June 1, 1992- May 31, 1993

PROCEEDINGS OF THE AMERICAN SOCIETY FOR COMPOSITES

SEVENTH TECHNICAL CONFERENCE

COMPOSITE

MATERIALS,

MECHANICS

AND PROCESSING

Co-Sponsored PENNSYLVANIA THE COMPOSITES STATE and BOEING MANUFACTURING

by UNIVERSITY TECHNOLOGY CENTER

HELICOPTERS

October Keller University

13-15,

1992 Center

Conference Park,

Pennsylvania

IECHNOMIC
PUBLISHINGCO., INC [ ,ANCASTER BASEL

N94-16837
Low and High Velocity Impact Thick Hybrid Composites Response of

C. HIEL AND O. ISHAI NASA Ames Research Center MS 213-3 Moffett Field, CA 94035

ABSTRACT The effects of low and high velocity impact on thick hybrid composites

(THC's) were experimentally compared. which were bonded onto an interleaved (350 F). The impactor of the similar were order tip for both cases In spite weights, strength damage laminates damaged of magnitude between dependence

Test Beams consisted syntactic foam core and was a 16 mm (0.625") in velocity ranges

of CFRP skins cured at 177C hemisphere. and impactor and residual strength on composite for the

steel

difference

relationships found. The

impact energy, of the skin

damage size, compressive models for performance

size agree and may composite,

well with enable the based

analytical prediction

open hole of ultimate

on visual

inspection.

NOMENCLATURE aoc Ex : Free parameter modulus modulus Ratio. in Average Stress Criterion for compression.

: Young's : Young's

in x-direction. in y-direction.

v,y G, KT R W

: Poisson's : Shear : Stress : Hole

Modulus. Concentration Radius. Panel Width. Factor for Infinite Width.

: Sandwich

1i49 PRECEDING PAGE BLANK NCT F}LI_ED

1150

THICK COMPOSITES

: Finite

Width

Correction

Factor.

ON oN _n

: Unnotched : Notched : Notched

Strength. Strength Strength. for Infinite Width.

INTRODUCTION

Extensive these materials damage which

research can only is mainly of the

on

carbon-epoxy

laminates

has

clearly

shown

that

accommodate in the form

impact energy by of a delamination

developing internal failure mode. The impaired, and An additional by visual impact a 6 mm thicknesses low velocity

residual compressive may limit the use drawback examination. response of (0.236") varying from is that

strength performance these laminates to damage, in most

is therefore severly secondary structures. cases, is not detectable

Publications laminates are 4 to 64 plies.

which compare low and rare. Cantwell and Morton Grafil They XA-S/BSL914C found that for

high-velocity (1989) choose with of

hemisphere

to impact

laminates conditions

impact, the size and the shape of the target determines its energy absorbing capacity and therefore its impact response. High velocity impact loading induces a localized form of target response and the level of damage incurred does not, therefore, further generally velocity appear to be governed by the areal size of the component. They concluded that high velocity impact loading by more detrimental to the integrity of a composite drop-weight impact loading. Moon and Shively laminates made Their findings a small projectile is structure than low(1990) choose a 12.7 AS4-934, to those

mm (0.5") hemisphere to impact 48 ply and IM7-855I-7 prepregs respectively. reported by Cantwell and Morton. A composites (1992). more comprehensive was literature published by

of AS4-1806, were similar

review, Abrate

on (1991)

damage and by

tolerance Ishai and

of Hiel

in general

Traditionally, sandwich constructions consist of three main parts; two thin, stiff and strong skins separated by a thick, light, and weaker core. The skins are adhesively bonded onto the core to enable load transfer between the components. Composite efficient way to utilize and very successfully sandwich composite in industry. construction has been found to be a very laminates and is therefore used extensively Until recently, the main emphasis was on which require damage tolerance carbon-epoxy to the the low and best high studies layers strength have been and impact and high conducted or no of

secondary structural components stiffness-to-weight ratios. Several on work sandwich was constructions foam found core. that having Nevertheless, compares lightweight

skin of the

honeycomb knowledge, response

author's

high-velocity

Low and High Velocity Impact Response of Thick Hybrid Composites

1151

sandwich subsequently

panels with referred

a structural to as a thick

(syntactic) foam hybrid composite

core. This (THC). recently details

type

of material

is

(Ishai and size with both

Studies on the and Hiel 1992). conditions characterization and residual

impact response of THC's have This paper discusses the relevant for of the low impact and high-velocity damage. by The energy an analytical on

on and

been performed fabrication, the the The inspection damage closes for paper strength

experimental

impact, relationship model. the

between residual

strength

is represented

a comparison of the low and high-velocity AND

effect of impact impacts.

MATERIALS

FABRICATION

An consists 1. 2. 3. 4. 5.

illustration of the thick hybrid of the following components: laminate, composed with of 18 plies protection, C prepreg

composite

is shown

in Figure

l a. It

A skin

of prepreg layup. composed layers. of two

(G40-600/5245C)

a (0/+30/-30)3s

An external layer for skin glass fiber fabric 7781/5245 A layer of FM300 adhesive. A layer of 7781/5245 orientation. Three The layers

C prepreg foam

at +45/45 (Syntac First, 350). the layers la, mold The of syntactic are laid-up foam into an core are cut. and to a

of syntactic

fabrication different

is as follows: parts, shown

Then

the

in figure

aluminum

mold. After the layup is completed, the transferred to a press with heated plattens. 350F cure-cycle after which it is demolded.

is closed, vacuum bagged whole assembly is subjected

It should be noted that this fabrication process has great significance since it is also applicable to sandwich constructions geometries because the foam beams, with can be cast dimensions tipped into any desired in figure The shape. lb, were were

technological with complex

Sandwich sandwich panel with a diamond IMPACT Low

shown bandsaw.

cut then

from

the

using a diamond coated sander.

edges

polished

LOADING Impact velocity test rig. tip was impact An 86 allowed tests were N (19.3 lbs) to fall freely conducted impactor from heights using with a ranging a conventional 16 mm (0.625") from 0.30 m (1

velocity

Low dropweight hemispherical

1152

THICK COMPOSITES

ft) to 2.13

m (7 ft) thereby

creating

impact

velocities

ranging

from simply

2.4 m/sec supported

(7.9 with

ft/sec) to 6 m/sec (19.7 ft/sec). The sandwich the distance between the supports being 0.203 High Velocity Impact

beams were m (8").

High pressure restrained resistance and the up by

velocity

impact

tests

were

performed

using

an

airgun.

Air

with

to 1.03 Mpa (150 psi) a plastic diaphragm. value, a small located at the of the air.

was fed to a chamber. When the pressure electric current, the diaphragm expansion

At this point in the chamber through precipitated the air m ( 70" a tapered

the air was reached a a piece of a

pre-determined wire release end

passed of 1.79 by

center of The rapid

its rupture accelerated ) barrel. tube Upon (sabot-

sabot/projectile reaching the

combination along the length of the of the barrel, the sabot is stopped the 17 gram (0.04 supported sandwich Ibs) projectile beam. The

catcher) allowing strike the simply ranged measured locations geometry DAMAGE

to continue free terminal velocities

flight and obtained

from 40 m/sec (130 ft/sec) to 160 m/sec by digital clocks which were activated in the barrel. Both the impactor and the as in the low velocity impact tests. INSPECTION AND CHARACTERIZATION

(525 ft/sec). The velocity was by trip wires located at three sandwich beams had the same

The

design

of the

sandwich

panels

allowed

for

the

extent

of damage

to be

easily differentiated speed impact causes glass-epoxy. The therefore assessment. conducted between causes surfaces high-velocity impactor). appears impact the case skin of impact 1983). damage an sophisticated Cross to relate

by visual inspection. a localized damage circular delamination NDT equipment

It was observed that any low or highand delamination of the surface layer of is easily visible in both cases and is not needed for an zone actual initial damage damaged and the was routinely delamination velocity of the shape impact contact for of the a

sectional cutting through the the observed surfacedamage the core. Figure 2 indicates while the tangential elastic of a cone the same indentation that induced there impact. velocity, on crack. energy with (and

the skin and indentation cause the impact

that the low displacements for the same

formation

Figure

3. is representative

The permanent to be deeper than energy. of high a THC damage Aditionally, velocity at low inflicted

induced by the by the high speed more

low speed impactor impactor at the same present in the impacted similar to the and Williams 3.has It is in Figure laminates.

is substantially

delamination

In summary one can state that as shown in Figure 2. is very laminates (Starnes as shown thermoset

thermoplastic

The impacted skin of a THC at high velocity, which is very similar to that infliced on rate dependence of stiffness and in future mathematical models

therefore likely that needs to be introduced

strength in the for THC's.

z-direction

Low and High Velocity

Impact

Response

of Thick

Hybrid, Composites

1153

Further damage caused energy. caused size Final by both by high Following

evaluation to the low impact and

of the energy

damage

mechanism in Figure have at this impacts scatter the

is obtained 4. As a similar time, higher beams can

by

relating

the on the

as shown be formulated has more

be seen, the

damage damage

high-velocity cannot impacts

dependence because impact were

conclusions velocity damage

at the sandwich

energies. subject to

characterization,

four point bending. The distance between the supports ( 13" ) with a distance between the loads of 0.076 m (3"). with Skin the damaged skin on Stress at Failure (SSF). STRENGTH low and high-velocity impact damage was the compressive side. Strength

was chosen Each THC was

as 0.33 m was loaded as the

defined

RESIDUAL The expected as stress This

localized,

and

is therefore however, strength. difference The solid failure

to have only a limited raisers and can therefore from figure

effect on the beam have a significant the residual

stiffness. They act, effect on laminate is plotted

is evident

5, where

strength

as a function

of damage diameter. Again it can be seen that there is basically no between reduction in strength due to low and high-velocity impacts. curve criterion was obtained which leads by using the Whitney-Nuismer Equation: (1974) average to the following

stress

O'N Y(2R/ W) =
which states that the notched strength (which is experimentally an infinitely as follows; wide measurable) laminate by can a

be obtained by correction factor

dividing Y, which

the strength of is can be calculated

Y(2R/W)=

strictly speaking, this equation therefore Y is called the "isotropic (1988) have shown nevertheless orthotropic this laminates for d/W investigation. According infinitely following wide equation; to Whitney and plate

is only correct for isotropic laminates finite width correction factor". Gillespie that the above expression is applicable values smaller than .25, which was the case

and et al to in

Nuismer is related

(1974) to the

the

notched

strength strength

of by

an the

orthotropic

unnotched

1154

THICK

COMPOSITES

2o'o(1-)

=
with R R+a._ and

3)('')]

Kr=l+

Ex

The strength The highly due quantity

equations were originally to a through the thickness aoc was region determined damage introduced adjacent by zone. by to

used to predict hole (or notch) a characteristic hole. The the

the variation in a multi-ply damage is used assuming extended

of tensile laminate. zone as in the a free

to represent

stressed

distance data been (1979).

parameter to be stress over the compression Our is that the

fitting experimental This criterion has Nuismer the over and Labor

an average to 'include

loaded basic impact

laminates

assumption damages

in using material

described a radius

analytical R, and

approach that this

to THC's, material no Therefore and be data was by obtained

longer participates in the load transfer the damaged material can effectively considered-as found Nuismer to be and Figure both may factor the low therefore a hole 6.09 Labor 6 relates mm with (1979) the radius R. The (0.24"), which

process within be tought of parameter is very close laminate. to the impact to

the laminate. as nonexistent present the result

aoe for the

on a carbon'epoxy residual strength

energy,

and

shows

that It

and high-velocity be concluded residual (provided core the

data can be merged that impact energy reduction impactor same

onto a single is the single

master curve. most important panels with

to control

strength

of structural tip is used).

sandwich

interleaved

CONCLUSIONS
O

Damage having

size the

was same

found energy.

to be similar

for both

low

and

high

velocity

impacts

Damage low impact

microstructure velocity and

was

found

to resemble laminates

thermoplastic at high impact

materials velocity.

at

thermoset

Low and High Velocity Impact Response of Thick Hybrid Composites

1155

Reduction in residual while impact velocity The Whitney-Nuismer an appropriate size to residual

strength is directly plays a minor role. average presentation strength. stress

controlled

by

the

impact

energy,

criterion, of the

for

open

hole data

laminates, relates

provides damage

experimental

which

ACKNOWLEDGMENT The and Dave Research authors Chappell Center for express of the their their Test support appreciation Engineering and expertise and to Michael Analysis to Luft, Branch They this also Howard at NASA project by Nelson Ames Paul

encouragement. dedicated

acknowledge

the substantial Scharmen of the REFERENCES

manufacturing Ames Modelshop.

Abrate, Mechanics Cantwell, Impact 551.

S.,

I991, Reviews,

"Impact vol

on

Laminated 1991, pp

Composite 155-190. of the Vol.

Materials,"

Applied

44, #4, April J., 1989

W.J., Response

and

Morton

"Comparison Composites,

Low

and

High

Velocity 6, pp 545-

of Sandwich

Panels,"

20, Number

Freeguard, G.F., and Marshall, product and process technology-," Gillespie, Laminate Hiel, C., J.W., Strength and Ishai, and Carllson, Predictions," O., L.A.,

D., 1980, "Bullet-Resistant Composites, January, pp (1988) Composites and "Influence Sdence High-Velocity Core,"

glass25-32. Width

A review

of

of Finite and

on

Nothced 32.

Technology, Impact

1991"Low.Syntactic Ga Dec 1992,

Response ASME Winter

of

Sandwich Panels Annual Meeting, Ishai, O. and Hiel,

with Atlanta, C.C.,

Foam 1-6.

Proceedings

"Damage Journal

Tolerance of Composites

of Composite Technology

Sandwich and

Panels Research,

with Interleaved Vol 14, #3 ,Fall Masters, Interleaved Composite Moon, D., Damage International J.E.,

Foam 1992 1987,

Core,"

"Correlation Proceedings Eds.

of

Impact of the

and Sixth

Delamination International

Resistance Conference

in on

Laminates," Materials, and Shively, in Thick SAMPE

F.L. Matthews

et al., Vol 3, pp 3.96-3.107. of Causing of the Impact 35-th

J.H., 1990, Laminated symposium.

"Towards a Unified Method Composites," Proceedings

1156

THICK COMPOSITES

Nuismer, Criterion: Starnes,

R.J., and Part J.H.,

Labor,

J.D.,

1979,

"Applications

of the

Average

Stress

Failure

U-Compression," and Williams,

J. Composite J.G., "Failure

Materials, Characteristics

13: 49-60,. of Graphite-Epoxy of Composite Pergamon Press,

Structural Components Loaded in Materials-Recent AdvancesHashin, 1983, pp 283-306 J.M. and R.J. Nuismer, Stress

Compression," Mechanics Z., and Herakovich, C.T.,

Whitney, Composites p.253.

1974,

"Stress

Fracture

Criteria

for

Laminated vol 8,

Containing

Concentrations,"

J. Composite

Materials,

.2

3
4,

5
4,

core

core

4,

5
4,

co re
_

Fig. l(a)

Identification

of materials

In interleaved

sandwich

panels

Fig. l(b)

Principal

dimensions

of Interleaved

beam

Fig. 2

Low.velocity

impact

damage:

cross-sectional

view

1157

........

--5

......

\_-

.......

-_

-_--

_-_

...........

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

=--2_Z_

..........

_'21_'2_'Y__ ............

_" ._52 _-_

"-'_

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

Fig. 3

High-velocity

Impact damage: cross-sectional

view

20

el el B el el el 4e

g
10. G

el

.L

PJ

0 300

Energy-Uf-(J)

Fig. 4 Dependence of damage size on impact energy el low velocity impact high velocity impact

1158

t
U

I
II611al__OacI _

1
N_Om.

1
17ti'_

oa

01

02 o*moge

43 _e

o4 7 epeeJm4_

o| wJdqb ( O|Wl 711ram 07i OI O|

Fig. 5 Dependence of residual strength on damage size (normalized by specimen width) = low velocity Impact o high velocity impact

|,.
L: . --_

J
CNll ul_m ILIII,I_ _

!
I_t4tl I I_t/Jlfo uIIIIN | eepiclu ueqnl.

I
" IIIN 117111 I

i
m

!| '"
20 40 _0 8(l Impeel 06 ee, oql I t20 J|

Cento

to lee Oep

Nr4_l

dSll

t40

110

too

200

220

Fig.

6 Dependence of residual low velocity Impact o high velocity impact

strength

on Impact

energy

1159

Author zed Reprint 1992 from Journal of Composites Technology & Research, FALL 1992 Copyright American Society for Testing & Materials, 1916 Race Street, Philadelphia, PA 19103

Ori

lshai t attd

Clenwnt

Hiel 2

Damage Tolerance of a Composite Sandwich with Interleaved Foam Core

REFERENCE: lshai, O. and lticl, C., "Damage Tolerance of a Composite Sandwich with Interleaved Foam Core," JournalofComposites Technology & Research, JCTRER, Vol. 14, No. 3, Fall 1992. pp. 155- 168. ABSTRACT: A composite sandwich panel consisting of carbon fiberreinforced plastic (CFRP) skins and a syntactic foam core was selected as an appropriate structural concept for the design of wind tunnel compressor blades. Interleaving of the core with tough intcrlaycrs was done to prevent core cracking and improve damage tolerance of the sandwich. Simply supported sandwich beam specimens were subjected to low-velocity, drop-weight impacts as well as high-velocity. ballistic impacts. The performance of the interleaved core sandwich panels was characterized by localized skin damage and minor cracking of the core. Residual compressive strength (RCS) of the skin, which was derived from flcxural test, shov,'s the expected trend of decreasing with increasing size of the damage, impact energy, and velocity. In the case of skin damage, RCS values of around 5(1U of the virgin interleaved reference were obtained at the upper impact energy range. Based on the similarity between low velocity and ballistic impact effects, it was concluded that impact energy is the main variable controlling damage and residual strength, where as velocily plays a minor role. The superiority (in damage tolerance) of the composite sandwich with interleaved foam core. as compared with its plain version, is well established This is attributable to the toughening effect of the intcrlayers which serve the dual role of crack arrestor and energ.,, absorber of the impact hmding. KEYWORDS: damage, damage tolerance, impact, ballistic impact. impact velocity, impact energy, sandwich beam, interleaving, syntactic foam, residual strength, carbon fiber-reinforced foam

RT RCS SSSF b d, EII,E,_,_

Fif,F2t

Room Residual

temperature compressive

conditions strength stress at sandwich failure

Skin maximum compressive Sandwich width Impactor diameter Lamina longitudinal spectively Lamina respectively Lamina strength, longitudinal respectively in-plane in-plane shear shear and longitudinal and

and transverse transverse

elastic tensile

moduli,

re-

strength,

transverse

compressive

F_
Gt.: g H /t L,, I go

Lamina Lamina Constant

strength modulus

of gravity

Drop-weight height Sandwich thickness Sandwich span Skin thickness Ply thickness lmpactor Lamina l.amina weight longitudinal hmgitudinal and transverse Poisson's ,atio CTE, respectivcly

w,
O[ 1_O{ 2 1)12

Introduction Composite replacing lions. This as: high materials metals is duc strength are considered and per superior to be good compressor mechanical unit weight, candidvtcs blades properties long fatigue for such life,

in helicopter to their and stiffness

applica-

Nomenclature BVD CFRP CTE DTC DTE FRP GFRP ttC Barely Carbon Damage Damage Glass visible damage fiber-reinfl_rced of thermal tolerance tolerance fiber-reinforced core

durability, The last plastic expansion by past promoted input data ternatives sandwich plastic chosen subject

and better damage tolerance characteristics (DTC). advantage has been shown to be of major importance failures of aluminum wind composites compressor of CFRP The strength concept. residual tunnel and rotor skins effect were stage blades. aluminum blades. and foam of impact selected NASA Ames ala research for wind structure and development ttmne] (R&D) project It) provide dcsign A composite core was on damas a major it was a full loadand

Coefficient

characteristics evaluation plastics

for comparing composed

Fiber-reinforced Honeycomb

as an appropriate for investigation.

age and consequential found depth poor ened

At an early

of the research

TPrcsently, visiting scientist, NASA Ames Research Center, Mail Stop 213-3, Moffctt Field, CA 94035: permanently, professor, Tcchnion-lsracl Institute of Technology, l laifa, Israel. :Principle investigator, Composite Material Research Program, NASA Ames Research Center, Mail Stop 213-3, Moffett Field, CA 94035. g: 1992 by the American Society for Testing and Materials 155

that an elevated-temperature-cured plain syntactic foam, was highly was manifested strength. with by extensive To reduce adhesive and residual

sandwich, with sensitive to impact cracking of the core

ing. This

this effect,

the core was toughintcrlayers.

by interleaving

glass/epoxy

PRE_NG

PAGE BLANK

NOT

FILMED

156

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

The

main

objective data

of the prcsenl for damage sandwich

investigation system. for Structural

was

to provide (DTE) of

may

absorb

the

impact

energy.

The

weak

link

in sandwiches

in

experimental this complex

tolerance

evaluation

composite

many cases is the core material, stresses induced under flexural

which may fail by shear or tensile impact. Most of the publications

Damage Tolerance Laminates Most objectives: The assessment investigations

Methodology

Composite

on this topic deal with sandwiches composed of honeycomb core and CFRP skins. Similar to the DTE of laminates, the evaluation of sandwiches is treated flaws, cracks artificial types in the at three flaw's, of flaws core levels, and namely: the effect may of be fabrication The detected: impact by thermal damage. curing stresses,

dealing

with

DTE

are aimed

at three

main

following

as a result interfaces cracking,

_ff fabrication

caused

of structural

performance

under

static

or

cyclic loads or both elements, which were pact. To provide guidelines assurance of composite impact Ranking, damage and

as well as survivability of structural previously damaged by accidental inaand allowables structure_which where DTE for design and quality are likely to sustain

partial separations at bonded core and skins, skin transverse flaws the were reduction found to affect in its shear

in the core and between and delaminations. Core performance as a result [14,I51. of modulus Interabove [161. of flaw

sandwich strength

and

facial separation critical debonding To enable

also has a significant length and depends and

effect on strength on skin configuration of the effects

has to be considered. purpose, of different to impact comand their

the evaluation

prediction

for material

selection

posite systems based on their residual structural performance. The first certification assessments and damage visible inates impact

response

size and location on the composite sandwich performance, artificial flaws are inserted into the sandwich structure. Information from and these the studies may lead to the definition is essential of flaw criticality for sandwich design related strength which

issue is of major concern for aircraft industries and authorities. For this purpose, some specifications based Another on DYE related DTE have been to critical classification proposed levels is defined [I 2]. These energy lamafter refare mainly size. of impact

and quality ded within flexure models

assurance. one of the

In most cases, artificial flaws are embedskins in a sandwich which is subjected to loading up to failure on [19,20]. [17,181. Anal}ileal buckling claimed cases, composites the sublaminate It has been

and requirements

or compressive are based, of delaminated

in many

as "barely

mechanism

damage" indicate

(BVD) threshold. Data on carbon-epoxy that at BVD level, compression strength to as low as 4()5 of the undamaged respective level of residual to be the accepted allowable carbon-epoxy composites

that damage caused effect on laminate ducted that, fiber on CFRP at BVD breakage level

by low velocity impact has the most severe and sandwich performance [1]. Tests con(IIC) core have skin indicated by local [21]. Reand above, damage of the is characterized impacted

may decrease

skin and honeycomb and delamination

erence strength. The strain seems, nowadays, for high performance craft applications. material selection DTE such testing the high boundary sensitivity conditions

compressive design limit air-

in structural

Most investigations are based on several [3,4]. diameter, [5-71 . This effort of the composite

that are concerned with attempts to standardize is essential because test variables and its of

sidual strength in most cases is below reference. Analytical model predictions results than experimental data. It was

50fi of the nondamaged gave more conservative concluded, in other in-

methods

to the impact

vestigations, that impact energy to failure increases with skin thickness and its rigidity [22]. Increasing honeycomb density tends to improve damage minor effect. Several when investigations a small diameter tolerance, dealt impactor but with cell the dinaension effect has only impact a

as: the impactor

the specimen

geometry,

The effect of impact velocity has also been considered. There is a clear distinction between the low velocity drop-weight test and the high velocity (ballistic) test as a result of their probable different effects on damage characteristics [8-10]. The effect of material composition on DTC a uniform test method. Several clamped indicated plate [8,11,12] effect a strong can only be evaluated by keeping investigations that have used the beam configurations parameters [13] have on DTC, material

of ballistic

was used

[23-251.

In most cases,

the damage was characterized local internal delaminations. as a hole through on this model, [261. Investigation [25,27] terfacial has indicated as a result separations be reduced the skin. are in good into that which

by combined fiber fiaeturcs and This failure mode may be modeled of residual strength, based findings loading life and may inwith experimental of cyclic at BVD formed compressive level, during fatigue impact. on sandmaterials in and of

Predictions agreement cvcn were

or narrow of different

the effect

namely: variation in layer stacking sequence, using thermoplastic rather than thermoset resin as a matrix, interleaving the laminate with tougher plies, and so forth. During the last decade, most

of propagation deal with combinations

of dclaminations the effect of skin of impact and core

of the publications on DTE were limited to composite laminates. Studies on the effect of impact on damage and residual performanee of substructural elements such as sandwich panels have been less and frequent, possibly as a result of the numerous parameters the complexity involved.

Several investigations withes with different

such as: aluminum, glass-phenolic and Nomex" honeycombs, three-dimensional (3-D) fabric, and Rohacell _ foam. Skins, most shown cases, that are composed by proper with tougher of graphite-epoxy of core fibers, widely [28-31]. material, Tests properties improveselection adhesive,

have

hybridization the sandwich Damage Tolerance skins entirely Evah_ation in sandwich different of Composite panels than Sandwich Panels imfor ment in impact toward impact used a more guidelines tion

the mechanical with corresponding Recently, and predict elements Such attention

may be varied energy

absorption. composite [32,33]. approach

has shifted under are mainly and provide selec-

Composite pact behave

subjected plate

to flexural mainly

attempting of basic in aircraft

to understand structural applications DTE

the behavior which

laminates

the following two reasons. First, loads when the sandwich is under shear Second, stresses are confined mainly a relatively the core provides

the skin flexure, to the

is under phme axial hence, interlaminar local impacted which zone. locally

studies

try to establish by proper

standardized

for structures tolerance variable.

for improving

the damage layup

soft substrate

of materials

and composite

ISHAI AND HIEL ON TOLERANCE

OF COMPOSITE

SANDWIC H

157

The Effect

of huerleaving decade, fracture with many efforts have been dedicated toughness and damage tolerance of brittle epoxy matrices techniques, One designated review test methods, promfor of A comprehensive

Materials Sandwich ponents, together. layers,

and Specimens structures skins, present All In the is added. are usually core, case, composed of three which the main bonds comthem above to

During the last toward improving advanced elevated this topic composites temperature

namely,

and an adhesive a fourth phase,

interleaved

applications. the different

of the constituent

materials

for the

[34] summarizes of toughened

and properties ising approaches inates by softer was found critical laminar damage, tended moplastic [38-40]. that fracture locations

composites.

of the most

components were cured at 177C (350F) and are designated be used under service conditions of up to 122C (250F).

was the interleaving and tougher materials interleaving [35], hereby toughness, may reduce

of the carbon-epoxy lamsuch as adhesive films. It interlaminar increasing and controlling This approach materials such tougher FRP stresses the at interimpact was exits ther-

Constituent The

Materials skins supplied ply thickness Two layers were fabricated (CFRP) Each from unidirectional tapes consisted skin of car(Rigidite 18 layup: epoxy protecsolid It is having FM300 interleaved oriented at of adhesive i+

significantly decreasing

structural

bon fiber-reinforced G40-600/5245C) plies (average (0/+ 30/-30)3_.

bismaleimide

prepreg

and improving RCS [13,36,37]. to include different interleaving films It was and hybridization also used using

by BASF. of BASF

of 0.14 ram) with the following glass fabric-reinforced were added for external was made of prefabricated supplied with The by Grace adhesive Corp. Syntactic. used The was glass microballoons

interlayers element

successfully

at the structural

level [41,42]. To the best knowledge of the authors, the interleaving method hits not been used in conjunction with syntactic foams. While this is probably a result of the limited application to date of these foams in high performance it is, however, reasonable to assume that nique may significantly strength improve impact sequent residual materials. Conchtding Based chanics of sandwiches sandwich structures, the interleaving techeffect and subcore of these

(GFRP) prepregs (7781/5245C) tion of each skin. The core syntactic composed the density prepreg foam (Syntac resin 350) 0.6g/cmL of epoxy of about film made filled

by American

Cyanamid

damaging composed

phases consisted of one ply of glass fabric prepreg + 45 to the beam axis embedded between two plies film.

Remarks on the above literature structures, in relevance skin is the review with and information gcneral on mecomments

Sandwich the following A typical core bonded of plain the skins to core cated gether process were cut of local the GFRP

Specimens sandwich fabric specimen and skins (FM300). skin of about panels 350 plies pieces configuration of two core three foam with CFRP layers interleaved skins which with are

of sandwich

may be concluded The ture The awfid impact. for the loading. composite

the present

investigation: struc-

is illustrated together core with adhesive the

in Fig. coating with the

1. It consists by four

backbone

of the slmdwich

interleaves. were panels

In the case bonded were prepregs to the fabrito-

and provides its strength and stiffness. main function of the core is to support local buckling and to absorb energy It must also possess enough transfer of shear and tensile

reference film

specimens,

the skins Sandwich and by means 6 atm). Two as follows: by 76 b}

as a result

by cocuring with (under from

of the solid core

interleave types

strength and stiffness stresses under flexural

of a press-molding of specimens

pressure the cured of

Syntactic foams, which are composed of epoxy resin reinforced with glass microballoons, have higher density than other foams and I tC cores. They possess, however, better strength performance and foams stiffness structural characteristics sandwich temperature as required applications. applications (350F for high

Long

beams

about

30

mm

for

residual

strength tests. Short beams of about damage assessment.

210 by 76 by 30 mm for cross-sectional

Syntactic

for elevated

[176.6C1) may be cracked their relative high brittleness due to their high laminates, characteristics coefficient which offer leaving composite tolerance tures. Objectives The objectives the of thc present effect of impact techniques,

undcr impact loading because of and induccd curing tensile stresses of thermal have been promise expansion. successful sandwich Interfor strucCFRP Skin

proven fimm

for improving

damage
Syntactic Foam core _"_'1

_/I/I///I///////I't'//.4

of syntactic

_'_ ,,.t

Tough interlayers

_'__/1/11"//I////I/1."//;

research loading

are as follows: on damage sandwiches core sandwich and subsesyn-

Protective

glass-fabric

coating

Study quent tactic Develop taking DTC

residual strength foam cores. a database into account the effect design

of composite

with

for interleaved damage tolerance of core

structure
Plain core sandwich reference

considerations. parameters on sandwich

Investigate postimpact

composition

to provide

guidelines performance.

for optimizing

structural

FIG. l -- Typical configuration of composite core and plain core sandwich reference.

sandwich

with interleaved

158

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

After edges diamond

rough

cutting

with

a carbide-coated under smooth

saw, water

the specimens" by means of a surfaces.

Fabrication

of

I-I_ /

I Effect of core I I composition I

sandwich specimens

,,,,'ere machined powder-coated

and polished disk to attain

and parallel

Low vemoelty__1 tl

E.ect ofim_ct

Characteristics The CFRP basic lamina,

of Sandwich mechanical the GFRP

Constituents properties fabric, of the cured, the syntactic unidirectional foam, and the
I ,.__] _

I I . sua,.ns.actionofdam , ctdam ge I I coreandinte"a ts 1

o.,statictes, I IE.ectotlmpact on damag_

adhesive layers are given in Table cured state at room temperature the constituents' and were supplier obtained data were obtained The information. independently

1. They are designated (RT) dry condition. from the available ASTM test of the syntactic [D 638], Sandwich

for the Most of literature foam ASTM Cores standards

I $
_ J '[__[ j 1

I InterelatiOnsh'ps between damage [ tCer_ n_e _rial:les_ Concluslon.s and [ [ tnpu s lor aes0gn l

Effect of impact on residual strength [ strengt_ Effect of damage on nes for optimal DTC

properties following

Data base and des gn guide sandwich configuration

aiming at improved

(ASTM Test for Tensile Properties of Plastics Test for Flatwise Compressive Strength of

FIG. 2--Scheme

of research program

and test procedure.

[C 365], and ASTM Test for Shear Properties in Flatwise Plane of Flat Sandwich Constructions or Sandwich Cores [C 273]). Most of the CFRP skin properties were computed based on the rehnpact Two range during are high Test Procedure A flow shown were chart of the research program two identical and ballistic and test series impact procedure of specimens loading. After loaded in flexDamage tolcore on the reference relationand is Testing types of impact events defined (ballistic tests which as low test), were may designated occur to represent the blades they and and test) spective lamina inputs, using composite laminate analysis, except for the compressive strength (F,.) and the coefficients of thermal expansion (cq,c_:) which were obtained experimentally.

of impact installation, velocity commonly

to compressor tunnel operation, (drop-weight An illustration

maintenance,

and wind velocity respectively.

basic specifications of these tests for the present investigation are shown in Fig. 3. There is a large difference in impact velocity and impactor weight between the two tests; however, to get a reliable comparison between low and high velocity tests the impactor head geometry was kept identical in the two cases.

in Fig. 2. Accordingly, subjected to low velocity assessment, for residual

visual damage ure to failure erance sandwich ships between

these specimens were strength determination. and plain based strength. were evaluated

characteristics configurations impact these parameters between

of interleaved variables

Drop-Weight The machine

Impact

Test impact system comprises of a Impact The 66 test maxi-

and damage

characteristics

instrumented made

and residual

by Monterey

Research

Laboratories.

TABLE
ELASTIC PROPERTIES GPa

t--Sandwich

con._tituettts properrie._.
PROPERTIES MPa C.T.E C "Ix 10 _ Thick. (ram)

STRENGTH

uNri_ MATERIAL CFRP G40-600 5245C GFRP Fabric 7781 5245C Syntactic Foam 350(2 Adhesive FM300 .08psf CFRP Skin
97.2 14.8 El I

F_.22

Gt2

V12

Fit

Flc

F2t

F2c

1=6

ill

(3.2

t 0

170

11.8

5.2

.33

2070

1380

75

251

102

-.3

28

.14

30.3

30.3

5.4

.17

374

560

374

560

99

9.9

9.9

.24

2.26

2.26

.84

.31

27

54.6

27

54.6

25

48*

48*

2.45

2.45

.88

.38

53

98

53

98

35

77

77

.26

24.5

1.21

936

660

70

289

153

-3.3*

15.1"

2.52

(0/'30/-30)3= (**)
*) **) Coficient Most of Thermal Expansion Values were were determined based experimentally at temperature lamina inputs, range except _t, of 20-120C (D. and Fie which were derived experimentally.

of CFRP

skin properties

computed

on the respective

ISHAI AND HIEL ON TOLERANCE

OF COMPOSITE

SANDWICH

159

Low

velocity, (Drop weight)

set-up

High velocity

(balistic),

Air- gun

set-up

/
ImpactorHeight- H< 2.5m

Impactor Weight- W = I
w

W = .177 N v-p _ IB, d= 16ram

f
200ram

Impactortip diameter-d = 16ram

_111111111111111111111111111111HHH_

l 200mm "_

JL

Velocity Energy

range : up to 6 m/sec range : up to 160 J

Velocity Energy

range : up to 160 m/sec range : up to 220J

FIG. 3--Illustration

of .,etup._ for two types of impact tests.

mum a rigid raised system pactor was call)'

tower base

height with

is about hemispherical

3.0 m. The weighing

impactor

comprises steel) attached impactor

of a to is

lower

values to

of the the

measured ones weight. 0.88g. process made Most

velocity The

and energy mainly response average

variables to frictional

as drop

16-mm-diameter

tip (hardened by the hydraulic

compared resistance acceleration tem during analyzer ometer imens about versus plain The

to the predicted falling was about the impact Type Type 3562A 2252 simply responses sandwiches

are attributed The dynamic

the assembly height

86 N. The system

calculated by a dynamic using

to the required to insure is guided derived by

and released by a braking im-

of the syssignal accclcrto the specof and for the

pneumatically.

Its rebound

is arrested event. the exact The elapsed time

automatically During values was between columns.

was monitored by Hewlett which two rollers the impact 5, which

a single-impact by two lubricated falling, the

the fall, the of impact two

Packard sandwich

circular _,elocity

To account variables opticells.

made

by Endevco on

was attached beam having integrated

for the friction

during

top of the impactor. were time core 200 mm.

of the impacted and in Figs.

experimentally.

determined photo

supported recorded

a span velocity

measuring

Typical

acceleration during in Fig. are shown

The actual maximum before the collision in Fig. 4 in comparison

velocity and the derived arc plotted as functions with the respective

kinetic energy just of the drop height curves. The

of interleaved is typical

5 and 6, respectively.

predicted

acceleration

response

7.5

220

200"--

//

180"

t60
-"n-140

/
_

L0.
80 ......... -_

I
60 40 Analytical ---_ ___O' Exlm'tmental Analytical

0,0 0,0 0,5 1.0 Height

T
1.5 [ml 2,0 2.5

20 0

......

0.0

0.5

1.0 1.5 Height [ml height.

2.0

2.5

FIG. 4--Calibration

curves of impact variables as function

of drop-weight

160

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

0.4

0.2

--_

+-

h iL
_

,.O2 _-0.4

i+
+10

.:---,
-2.0 -1,2 -0.4 e 04 1.2 210 Tim4 2 .0 (nloc) 36 4,4 $2 8.0

4).1t

\
\ _l L__J -I.2 -2.0 -t.2 -0.4 0.4 1.2 20 Tkne 2.g 1nl4c) 3.8 4.4 5+2 6,0

FIG. 5--Typical acceleration and velocity response impact test (input energy." 156 J) (skin damage only)+

curves fi)r interleaved

core sandwich.

Recorded

during

low veloci O'

0.rm

"
__

-I
_____._ --

0.00 -0.06

-0,16 -O24

\ \
-I.0 -O.2" 0.6 1,4 2.2 3.0 3,8 4.6 5.4 6.2 7.0

-0.32 -0.40

Ii
-I.0 -0.2 0.8 1.4 2.2

_ ____
3.0 3.8 4+6 5.4 6.2 7.0

-0+48

-0.56

Time (m_) FIG. 6--Typical acceleration and velocity response energy: 68.7J) (core damage mainly). curves for plain core sandwich. Recorded

Time (reset) during low velocity impact test (input

noncracked minimum

interleaved (or maximum

core

sandwich,

shows which

a tend

towards

the

air

was

released.

The

abrupt

air expansion

accelerated

deceleration)

is not easy to define

quantitatively as a result of the graph curve, which is obtained by integration is smooth and continuous and allows the minimum between for computation acceleration impactor of energy from loss the input velocity

fluctuations. The velocity of the acceleration graph, precise determination of slope. The difference velocity is used absorbed,

sabot/impactor combination along the 1.79-m tapered barrel which caught the sabot at its end. After a short free flight, the 17-g impactor collides with a simply supported imen. The terminal velocities obtained, by the air pressure, measured positions sandwich used by digital close beam ranged clocks from sandwich beam specwhich were controlled The velocity was

its extreme as a result

40 to 160 m/s. by trip wires

and response

activated

located

at three and the as those

of energy

mainly by the skin local damage during the impact process. On the other hand, the response of the plain core sandwich to the impact tered velocity is different acceleration curve impact. with (Fig. graph attributable Here, the 6). It is characterized with no trend to the cracking the lower control loss, upward mainly specimens by a highly scatat all and a discontinuous of the noninterleaved velocity with the after impact prointerleaved

to the barrel had the same

edge. Both composition

the impactor and geometry test.

for the drop-weight,

low velocity

impact

Impact

Damage

Characterization loading, each specimen dimensions of the damage was inspected were measured, visually namely:

core during (as compared

After impact and the external

core) indicates higher energy cess in the sandwich core.

due to the failure

the damage size and its depth. was close to circular and the to be a measure sured by a special of its size. indicator

In most cases, the damage shape average diameter was considered Maximum damage depth was meaof 0.01 ram. Different

to an accuracy

High

Velocity tests

(Ballistic) were

Impact conducted

TesO by using an air-gun device. Air it was presand

Ballistic pressure restrained sure level,

specimens, representative of the overall impact range, were sectioned through the damage center for internal inspection of the damage sandwich. Typical photographs of external and internal in Figs. damage surfaces for the interleaved specimens 7 and 8 and will be discussed later. are shown

(up to 1.03 MPa) by a thin plastic the diaphragm

was fed to a chamber in which diaphragm. At a predetermined was ruptured by electrical heating

ISHAI AND HIEL ON TOLERANCE

OF COMPOSITE

SANDWICH

161

,,, ............

Residual

Strength

Testing damage failure inspection, was the the flexure placed test. An specimens using were an MTS speed illustrative of

Following loaded
155._ 115.11 t_.tRI 19.7

external

to ultimate skin was

in four-point the sandwich during Constant

test 1.84

system. mm/min

In all cases,

so as to load

the damaged

in compression. maintained

cross-head

|35.6

qg?- I

558

IS..t

description of the flexural system is shown in Fig. 9. The relevant values of the skin compressive stress at sandwich failure (SSSF) and the core shear stress at sandwich failure (CSSF) were derived from the ultimate beam formulations. the derivation due to the high core load P, value A classical stiffness (above ratio 30). failure value (SSSF) represents skin skin The is the the laminate relationship Hence, maximum cross residual and is treated under was linear maximum effective section side of the skin laminate based on the simplified sandwich sandwich analysis was used for given in Fig. 9. It is justified the CFRP skin and the between

of stress

formulation

113.1

83A

5.1It+

Ih,7

interleaved Skin stress strength stress acting

at skin SSSF

on the upper The

at failure.
'RL3 6_.1, ,+57 15.t+

compressive the residual here mauniaxial to failstress

(RCS)

of the damaged

laminate material

strength stress

of the sandwich. loading. The

croscopically

as a quasi-homogeneous load-deflection

ure which was catastrophic and brittle. criterion was found to be adequate.
46.5 34.3 3.2fi Ill. _

Test

Results results and groups, and namely: strength 10. their evaluation which variables, A detailed are may involved be classified damage list of these with into several three

Test variables
Z-LO 17.7

characteristics impact variables.

main

characteristics, variables

and residual is given FIG. 7--External specimens subjected (top) view of damage for interleaved core sandwich to different low velocity impact energy levels.

in Fig.

Impact The

Damage protective were left clear

Assessment glass found imprints fabric-epoxy whose layers sensitive on the varied external skin

D.IPI_CT DAMAGE IN V_TI'II 1NTI(RI.I';AVED

COMPOSITE FOAM CORE

SANDWI(7tl PANEI,S t <'uoss.sl.:<l'mX vlt:w)

=

surfaces which

to be highly

to the impact

loading

dimensions

with the impact

magnitude (see Fig. 7). The boundaries to be dictated by the contact surface and the specimen. area measured to match provide failure found may The dimensions from approximately levels. an excellent bonded 8) were imprint of coating

of these imprints seems between the impactor tip of internal the tool interfacial specimens respective that impact skin for de(Fig.

the cross-sectioned It was concluded

external this type damage impact

sizes at all impact and assessment

113.1

83.4

5.09

16.7

inspection

in a real structure, mode. In all cases,

where tested

skin damage

is the predominant

68.7

50.7

3.96

13.0

damage was confined to a well-defined almost circular. The predominant failure cracking externally specimens and delaminations defined from cracking which findings for either damage with interleaved the skin process were damage seems slightly which zone core, did not (see Fig. could initiation

local zone which was modes were transverse propagate of core beyond cracking (Fig. internal the orig11). in11). In the case of the be detected by the at high

46.5

34.3

3.26

10.7

inating This terleaves

zone

to be arrested damaged or delaying core the effectiveness specimens,

impact

levels.

24.0

17.7
m'

2.35

7.7

These process

demonstrate preventing (noninterleaved)

of the interleaving In the case developed cracks

core failure.

of the plain

FIG. 8--Internal sandwich specimens

(cross-sectional) view of damage for interleaved core subjected to different low velocity impact energy levels.

through the core depth which action of tensile curing stresses flexural impact. A typical

were activated by the combined and shear stresses induced by the of such cracking is shown in

pattern

162

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

Front view of loaded sandwich beam

Cross sectional view

P
CFRP Skin I Load Fixture

Syntactic Foam

2 I d= h+t
f/////////I////ll//tt/////2

= 330 mm Approximate sizes (mm) t=2.5 h = 30

(_'max skin
I

----

Peh / 2btd 2
=

- Maximumnormaleffectivestress acting on skinlaminate - Averagecore stressshear stress for derivation

d = 27 b = 76 of residual strength.

r_average core

P / 2bd

FIG. 9--Flexural

test setup and formulation

iii

IrH

E_772_:5.,

Impact Variable - Impact Energy (input) -Impact Velocity (input) -Energy Loss (response) Damage Characteristics - Damage Size (diameter) - Damage Depth - Damage Area - Failure modes

W_
Ui Vi A U - Derived - Derived - Computed from from experimental experimental input and energy velocity output plot ( fig. 4 ) plol ( fig. 4 ) difference

10 mm

_I

from

velocity

D,
d

- Average - Maximum

diameter depth

of visual of skin

external

damage ( fig.

( fig. 11 )

7 )

damage

crater

Ad - _E)_'4
(Skin,Core or Interracial )

Residual Strength Variables - Skin max. compressivestress at skin failure -Skin max. compressive stress at core failure Core max. shear stress at skin failure - Core max. shear stressat core failure FIG. lO--List

SSSF SSCF CSSF CSCF

of test variables. FIG. 11 -- Typical cross-sectional view of tow velocity impact damage for composite sandwich with interleaved syntactic foam core (impact velocity. 6m/s; impact energy. 156 ]).

Figs. 12 and 13 for the interleaved counterpart.

plain

core version

as compared

with

its

The

Effect

of Impact three

Variables parameters

on Damage may be used

Characteristics to define skin dam-

Ultimate Loaded

Failure and Residual in Flexure of interleaved Such modes

Strength

of Damaged

Specbnens

In general,

age geometry, namely: size (average diameter), area, and depth (see Fig. 10). In the case of interleaved core sandwiches, all of these and were energy found loss. to increase Figures area the continuously 15 show which with the impact effect energy of these from these its and proeffect 14 and and depth energy

In most cases combination fracture fabric always mature sandwich of skin computed along interleaf from shear interlaminar

core failure

sandwiches, was found

ultimate in-plane

failure shear and

was due to skin damage. of three separations and core the impact failure Residual stress also the

to be a complex and buckling,

(see Fig. 16), namely, delamination the CFRP Failure With predominant mode failure was determined site. the failure layers. was

30 , sublaminate between damage prevalent external

variables relationships energy on damage deceleration nounced exceed

on damage damage that

may be representative and especially proportional,

laminate seems

and GFRP to originate precore was (see Fig.

for the overall

geometry. impact

It may be concluded variable, almost

few exceptions, mode for the plain (SSSF) given which

loss component,

have a direct,

area and depth. The effect of impact velocity were less at low levels but become much more at the higher range. Damage tip. size did not of impactor

16). This was

version. compressive

strength at sandwich

by the value in Fig. 9. The

generally

the diameter

by the approximate

formulation

ISHAI

AND

HIEL

ON TOLERANCE

OF COMPOSITE

SANDWICH

163

plain core
FIG. 12--Comparison of impact cross-sectional view (impact energy." damage 68.7 J). for interleaved versus

interleaved
plain foam core

core
composite sandwiches,

II

II [1/I_l iI/_ -- ..........

plain core
FIG. 13--Comparison side view (impact energy: of impact 68.7 J). damage for interleaved versus

interleaved
plain foam core

core
composite sandwiches,

240

240

22O 2OO ...... m -_--

220

200

180 .... _/ 7_

/
'

E
_ 140

_ /

i E

180

140

<_I_' g eo
8o

/
/

a_
4O

?- ......

t'-

Y
,/
J
0 20 40 60

4o

L/
/
o 20

0 80 100 Energy 120 (J) 140 180 180

40 Energy

BO Less (J)

80

100

120

Impsct

FIG.

14--The

effect

of low

velocity

impact

energy

and

energy

loss on

damage

area.

164

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

3.0

./

3.0

23

2.7

2.4 .....

/
/

2.4

2.1

2,1

g
_. 1.8

,/ /
.f

/,

1,8

./
t 1.5'

E
_t &
1.5

1.2'

/ 4
=/

1.2

0.9-

0.9

0.6

0.6

0.3

0.3-

0.0

/20 40 60 80 ImpeCl 100 Energy 120 (J) 140 180 180

0.0

/
20

/!

40 Energy

60 Loll

80 (J)

1 O0

120

FIG.

15--The

effect of low velocity impact energy and energy loss on damage depth.

FT:'

Upper view

Skin failure
--.(impact energy : 90.3J )

Side view

I
_e
(impact

failure
energy

_
: 156J ) .....

Side view
__.. ....

FIG. 16-- Typical failure

modes

in residual strength

test

effects of damage characteristics interleaved sandwich specimens pact these impact specimens for cases are shown in Figs. values Residual failure and 17 and relationships range.

and impact energy on SSSF for damaged under low velocity im18. The trend common strength for all reduction sandwich obtained of residual levels

Evaluation Three the effect

of Experimental main topics are dealt

Findings with in the present between study, namely:

is the high rate strength (Fig. 18).

of interleaving,

the comparison

low and high

at low impact

the tendency

to level

off at the upper

of interleaved are close to those

impact velocity, and mainly, the dependence of residual strength on damage and impact variables. The significant beneficial effect of interleaving onstrated in Fig. on improving limited residual strength is clearly data dem19. The and scattered for the plain

that failed of skin

by core cracking

ISHAI AND HIEL ON TOLERANCE

OF COMPOSITE

SANDWICH

165

8OO

800

_700

\
\

_'700

\
600.

\ -%
; 40O r .... { 500-_

\
i

|
400

;_Nx)

1;300-

! ez_

!
l
200"

!
100

100,

i
0 2 ' 4 I 6 Dam|0o , I 8 I 10 SIze I ]2 (mm) 14 t6 18 0 O0 0.3 06 09 Damage 1.2 1.5 Deplh l.B (mm) 2.1 24 2.7 3.0

FIG. 17--The effects of damage size and depth on residual compressive interleaved core specimens subjected to low velocity impact.

strength

of sandwich

skin for

core core =..

700 = 600---_ ....... '.............. -_--

sandwich which was

specimens not only stresses curing

was

due

to premature by low impact such In most

cracking energy cases,

in the but also the dis-

caused

by residual

as well.

integrated core could not support the skins and was unable to transfer stresses. Consequently, the sandwich had a very low stiffness =I
=1

5OO

and

residual

strength

that

did not

reflect

the structural

potential the other than 50% of impact

of the CFRP skin. The interleaved core specimens, on hand, retained the expected residual strength and stiffdamaged energy from skin even in cases of core impact tests of failure. range More level of the original applied drop-weight in spite rather between than damage the effect strength of impact of the large velocity was retained velocity order seems at the higher on residual (Fig. magnitude

E 300 ---

ness of the

i
g,

at low velocity and ballistic the two tests. and residual

(155 J). strength (_25) indicates premise

200

. i m

.......:........... Skin failure Coce failure -._

--

Comparing as derived similar trends

100

?-_

"--

___

20) revealed

| 20

1
40

, 60 Impal

, 80 Enaroy

i
100

. 120

I I 140

, 160

difference that impact variable

in velocity energy affects that

This finding strength. This

(J)

to be the prevailing

FIG. 18--The effect of impact energy on residual strength of sandwich panels with interleaved core subjected to low velocity impact,

is supported pact energy on the same

by plotting the data of residual strength versus imderived from both low and high impact velocity tests coordinates derived One as shown within a single main at widely of the in Fig. curve 21. Both fit in spite range sets of data and a of the fact

are well intermingled

i

Inlerleavedcore-skin failure Interleavedcore-corefailure Plaincore-corefailure

that

they were

different

of velocities in maintaining

impactor

weights.

concerns

damage-sensitive structural occurrence, location, and mation replace residual is needed the strength. hole model" of the impact the use in Ref normalized tests Refs is plotted 43 and formulation for the damaged

element is the ability to detect the size of an impact event. This infordecision whether based tool to ignore, on repair, or of is the circular as forstrength imprediction of the 4.5. in Ref mainly evaluation

element,

400[]

An appropriate damage found

for this prediction in Ref 26. The of this model of residual

i
> 2OO --

\_

"open shape justifies mulated versus pact

as was demonstrated analytical solution

in the present data

investigation

of the 43. The

3
_e

experimental

damage 44 (see for

size for low velocity Fig. the 22). present A full case

and ballistic description

d
0 .... 20 40 60 Impoct I 80 Enorgy , t00 (J) 120 i40 160

and compared

with the analytical is given

from

analytical

FIG. 19--The effect of low velocity impact energy on residual strength of sandwich panels (with plain versus interleaved foam core).

Good agreement to the diameter

between experiment and analysis for sizes up of the impaetor is evident. Note that the ana-

166

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

]mpactor 800

weighl

= 86N _. 800 I

Impactor

weight

= .177N

=v 7oo600" 1= .................... t ............. i

!
ii

700

L__

.....

BOO

--

-.,
O

,I

,5oo.

".a.2o

400

J
300

4011

...........

!
i 200 .... ,.... t ............... i 300 200 .............................. ..........

100

100 .......

g, 00,0 .!0
8

! ...... 1.6
2.4 3,'2 Velocity 4.0

i 4.B i.
(m/see)

5.6

6.4

o 0 20

. 40 Ballistic

* 60 Impact 80 Velocity 100 120 (mlsec} 140 160

LOW

Impact

FIG, 20-- The effect of impact velocity on residual velocity versus ballistic impact tests).

compressive

strength

of sandwich

skin (low

= O

Drop

weight

impact

( impactor

weight

= 86N )

Ballistic

impact

[ impaclor

weighl

= .177N

[yticaI model is for a plate of infinite element is a finite-strip skin supported reservations, damage sizes it appears smaller that than one fourth

width, whereas the present by a core. In spite of these prediction is valid width. for of the sandwich

-_-m--

the analytical

Conclusions Based
Cuwe fit for drop we_hl Ir data

on

the

experimental

results

and relating

their

evaluation, system

the with

following

conclusions syntactic

may be drawn foam core

mainly sandwich

to the damtemper-

age tolerance an interleaved

....

performance

of a composite suitable

fl_r elevated

_1
40 60 eo Impll 1oo Inlr_y 12o IJ) 140 160 18o 2O0 220

ature

applications. tolerance failure performance is significantly improved which by can

Damage core Impact

interleaving. is controlled visually. by local skin damage,

FIG. 21--The effect of impact energy on residual strength of sandwich panels with interleaved cores.

be inspected

BOO

--

700

Dropweight impact (impactor weight= 86N) Ballisticimpact (impactor weight = .177N) Analysis (open hole model)

|=
|
; Z
m 300

400

!
I
200-_ E 100

k
Impaclor eaddiameler h

\
0 0.0 0.1 0.2 Olmmge 0.3 alze 0.4 / Ipeclrnen 0=.5 width ( 0.6 b = 78ram 0.7 } 0.8 0.9 t .0

FIG. 22-- The effect of normalized damage size on residual compressive (low velocity and ballistic impact test results versus analytical solution).

strength of sandwich

skin

ISHAt

AND

HIEL

ON TOLERANCE

OF COMPOSITE

SANDWICH

167

Residual about 50%

strength of and energy size both the

decreases original (at

with an

impact energy

energy level of

down 160 J).

to [18]

Damage impact

residual rather and low

strength than impact strength and be

are

directly

dependent

on weight. same

velocity are

or impactor affected impact in the energy.

Damage way by

residual velocity may

ballistic by

Residual of damage

strength size.

predicted

visual

measurements

[19]

[20] Acknowledgment The and at authors wish to thank of the Research Dr. Test Howard Engineering for of the of the their Model Nelson, and Roy Analysis and at the specimens. [23] References [24] [1] Demurs, E., "Damage Tolerance of Composites," in Proceedings of the American Society for Compos#es, 4th Technical Conference, 1990, pp. 425-433. Papoff, A. J., Dill It. D., Sanger, K. B., and Kautz, E. F., "Certification of Damage Tolerance Composite Structures," in Proceedings of 8th DODLYASA/FAA Conference on Fibrous Composites in Structural Design, 1989, pp. 499-514. "Standard Tests for Toughened Resin Composites," NASA Reference Publication 1092, revised edition, 1983, 32 pp. "'Boeing Specification Support Standard: Advanced Composite Compression Test," BSS 7260, revised edition, 1986, 21 pp. Falabella, R., Olsen, K. A., and Boyle, M. A., "Variations in Impact Test Methods for Tough Composites," in Proceedings of 35th International SAMPE Symposium, 1990, pp. 1454-1465. Moon D. and Shivel, J. It, "Toward a Unified Method of Causing Impact Damage in Thick Laminated Composites," in Proceedings [25] Hampton, Branch encourCenter [22] [21]

28th Israel Annual Conference of Aviation and Astronautics, 1986, pp. 186-190. Ishai, O. and Rosenzweig, A., "The Effect of Interlaminar Flaw Characteristics on Residual Compressive Performance of GFRP Sandwich Specimens," Annual Report on Damage Tolerance of Composite Materials, Technion R&D Foundation Contract 893/9748/ 3-101, May 1990, Chap. 2, pp. 21-37. Kassapoglou, C., "Buckling, Post-Buckling and Failure of Elliptical Delaminations in Laminate under Compression," Composite Structures, Vol. 9, 1988, pp. 135-159. Sommers, M., Weller, T., and Abramovieh, H., "Influence of Predetermined Delaminations on Buckling Behavior of Composite Sandwich Beams," Composite Structures, Vol. 17, 1991, pp. 292329. Gottesman, T., Bass, M., and Samuel, A., "Critically of Impact Damage in Composite Sandwich Structure," in 6th International Conference of Composite Materials, Vol. 3, 1988, pp. 327-335. Steinman, A. E., "'Damage Tolerance of Thin Skin Sandwich Panels,'" Report DFML-TR-76, Airforee Materials Laboratory, WrightPatterson AFB, OH, Feb. 1977. Farely, G. L., "'Effect of Low Velocity and Ballistic Impact Damage on the Strength of Thin Composite and Aluminum Shear Panels," NASA Technical Paper 2441, May 1985. Rhodes, M. D., "Impact Fracture of Composite Sandwich Structure," NASA Technical Paper 75-718, 1975. Thart, W. G. and Wanhill, R. J., "Impact Damage Effects on Fatigue of Composite Materials," NASA Report NLR-MP-8201 lU, 1982. Williams, J. G., "Effect of Impact Damage and Open ttoles on the Compression Strength of Tough Resin/High Strain Fiber Laminates," in NASA Conference Publication 2334, Proceedings of a Workshop Sponsored by NASA Langley Research Center, May 1983, pp. 61-79. Nettles, A. T., Lance, D. G., and Hodge, A. J., "'A Damage Tolerance Comparison of 7075-T6 Aluminum Alloy and IM7/9772 Carbon/Epoxy," in Proceedings of the 36th International SAMPE Symposium, April 1991, pp. 924-931. Akay, M. and Hanna, R., "A Comparison of Honey-Comb Core and Foam Core Carbon Fiber-Epoxy Sandwich Panels," Composites, Vol. 21, No. 4, 1990, pp. 325-333. Verpoest, I., Wevers, M., lyons, J., and De Meester, Fabric for Compression and Impact Resistant Composite Structures," in Proceedings of 35th International SAMPE sium, April 1990, pp. 296-307. Nettles, A. T. and Hodge, A. T., "Impact tloney-Comb panels with Graphite-Epoxy ceedings of 35th 1430- 1440. International SAMPE Testing Face Symposium, P., "3DSandwich Sympo-

Dave NASA

Chappell Ames and

Center

support Shop

agement for the

to Paul

Scharmen

high-level

manufacturing

sandwich

[2]

[26]

[3] [4] [5]

[27]

[28]

[6]

of the 35th h_ternational SA MPE Symposium, 1990, pp. 1466-1478. Chen, C. H., Chert, M. Y., and Chen, J. P., "The Residual Shear Strength and Compressive Strength of C/E Composite Sandwich Structure After Low" Velocity Impact," in Proceedings of the 36th h_ternational SAMPE Symposium, April 1991, pp. 932-943. [8] Starnes, J. tl. and Williams, J. G., "'Failure Characteristics of Graphite-Epoxy Structural Components Loaded in Compression," NASA Technical Memorandum 84552, 1982, 24 pp. [9] Cantwell, W. J. and Morton, J., "'Comparison of the Low and High Velocity Impact Response of CFRP," Composites, Vol. 20, No. 6, 1989, pp. 545-551. [10] Zee, R. H., Wang, C. J., Mount, W. A., Jang, B. Z., and Hsich, C. Y., "Ballistic Response of Polymer Composites," Polymer Composites, Vol. 12, No. 3, 1991, pp. 196-202. [11] Williams, J. G., O'Brien, T. K., and Chapman A. J., Ill, "Comparison of Toughened Composite Laminates Using NASA Standard Damage Tolerance Tests," NASA CP-2321, presented at ACEE Composite Structures Technology Conference, Aug. 1984, 72 pp. [12] Dempsy, R. L. and Horton, R. E., "Damage Tolerance Evaluation of Several Elevated Temperature Graphite Composite Materials," in Proceedings of 35th brternational SA MPE Symposium, 1990, pp. 1292- 1305. [7] [13] Ishai, O. and Shragai, A., and Residual Compressive Composite Structures, Vol. "Effect of Impact Loading on Damage Strength of CFRP Laminated Beams," 14, 1990, pp. 319-337.

r29]

[30]

of Glass-Phenolic Sheets," in ProApril 1990, pp.

[31J

[32]

[33]

[34]

[35]

[14]

[15]

Minguet, P., Dugudji, J., and Legace, A. P., "Buckling and Failure of Sandwich Plates with Graphite-Epoxy Faces and Various Cores," in Proceedings" of 28th Structures, Structural Dynamics and Materials Conference, 1987, pp. 394-401. Zenkert, D., "Strength of Sandwich Beams with Mid-Plane Debonding in the Core," Composite Structures, Vol. 15, 1990, pp. 279299. Zenkert, D., "Strength of Sandwich Beam with Interfacing Debonding," Composite Structures, Vol. 15, 199l, pp. 311-350. Bass, M., Gottesman, T., and Fingerhut, U., "Criticality of Delamination in Composite Materials Structures," in Proceedings of

[36]

Lee, L. J., Huang, K. Y., and Fan, Y. J., "'Dynamic Response of Composite Sandwich Plate Subjected to Low Velocity Impact," in Proceedings of the Eight International Conference of Composite Materials, July 1991, Paper 32-D. Kan, H. P., Whitehead, S., and Kautz, E., "'Damage Tolerance Certification Methodology for Composite Structures," in NASA Conference Publication 3087, 8th DOD/NASA/FAA Conference on Fibrous Composites in Structural Design, Nov. 1989. pp. 479-498. Wong, R. and Abbott, R., "Durability and Damage Tolerance of Graphite-Epoxy Honey-Comb Structures," in Proceedings of 35th International SAMPE Symposium, April 1990, pp. 366-380. Sela, N. and Ishai, O., "Interlaminar Fracture Toughness and Toughening of Laminated Composite Materials: a Review," Composites, Vol. 20, No. 5, 1989, pp. 423-435. Altus, E. and Ishai, O., "The Effect of Soft Interleaved Layers on Combined Transverse Cracking-Delamination Mechanism in Composite Laminates," Composite Science and Technology, Vol. 39, 1990, pp. 13-27. Sun, T. C. and Rchak, S., "Effect of Adhesive Layers on Impact Damage in Composite Laminates," in Composite Materials: Testing and Design (Eighth Conference), ASTM STP 972, J. D. Whitcomb, Ed., American Society for Testing and Materials, Philadelphia, 1988, pp. 97- 123. Sun, T. C. and Norman, T. L., "Design of a Laminated Composite with Controlled-Damage Concept," Composite Science and Technology, Vol. 39, 1990, pp. 327-340. Frazier, J. L. and Clemons, A., "Evaluation of Thermoplastic Film tnterleaf Concept for Improved Damage Tolerance," in Proceedings

[37]

[16] [17]

[38]

168

JOURNAL

OF COMPOSITES

TECHNOLOGY

& RESEARCH

of 35th 1627. [39]

International

SAMPE

Symposium,

April

1990,

pp.

1620-

[42]

Avery, J., Allen, M. R., Sawdy, D., and Characteristics of Composite Compression

Avery, S., "Survivability Structure,'" NASA ConDOD/NASA/FAA Design, 1989, pp.

[40]

[41]

Jang, B. Z., Chen, L. C., Wang, C. Z., Lin, H. T., and Zee, R. It., "Impact Resistance and Energy Absorption Mechanism in Hybrid Composites," Composite Science and Technology, Vol. 34, 1989, pp. 305-335. Wang, C. J. and Jang, B. Z., "Impact Performance of Polymer Composites: Deformation Process and Fracture Mechanisms," in Proceedings of the Eighth International Conference of Composite Materials, Paper 32-B, 1991. Busgen, A. W., Erring, M., and Scholle, M., "Improved Damage Tolerance of Carbon Fibre Composite by Hybridization with Polyethylene Fibre, Dyneema SK 60," Conference of the American Society 424. in Proceedings for Composites, of 4th Technical 1989, pp. 418-

ference Pubfication 3087, Proceedings of 8th Conference on Fibrous Composites in Structural 455- 478. [43]

[44]

[45]

Whitney, J. M. and Nuimer, R. J., "Stress Fracture Criteria for Laminated Composites Containing Stress Concentrations," Journal of Composite Materials, Vol. 8, July 1974, pp. 253-265. Nuismer, R. J. and Labor, J. D., "Application of the Average Stress Failure Criterion: Part II-Compression,'" Journal of Composite Materials, Vol. 13, Jan. 1979, pp. 49-60. Hiel, C. and Ishai, O., "Low and High Velocity Impact Response of Sandwich Panels with Syntactic Foam Core," in Proceedings of the ASME Symposium on Recent Advances in Structural Mechanics, PVP Vol. 225/NE-Vol. 7, Dec. 1991, pp. 137-141.

Designer's

Corner
selection approaches for the architectures and processes; choice of failure grity of composite effective concurrent criteria used products; engineering various available fibre for establishing inte-

Short contributions of less than 1000 words plus key illustrations are being invited, covering topical issues associated with the design and application of composites. Notable designers from a broad range of industries including aerospace, automotive, civil, bioengineering and recreational are encouraged to submit a contribution to this section. Communications may cover, but not necessarily be restricted to, the following subjects: novel and innovative and fabrication; concepts in composites design

approaches.

economics issues and other exploitation of composites;

impediments

to the wider

Contributions will be subject to a rapid review and publication process. Prospective contributions, marked for the 'Designer's Corner', should be submitted to: Dr Keith T. Kedward, Department of Mechanical & Environmental Engineering, University of California, Santa Barbara, CA 93106, USA. Fax: I (805) 893 8651

Composite sandwich with syntactic foam

construction core

A practical assessment of post-impact damage and residual strength

C. H/EL, D. DITTMAN (NASA Ames Research and O. ISHAI Center, USA)

Most composite sandwich constructions with a lightweight core are difficult to reliably inspect for postimpact damage. Additionally the residual strength cannot easily be estimated, and therefore aeronautical

designers tend to prefer a skin stringer for primary load-bearing structures. The purpose of this note is to report

type arrangement

on

a successful

34 mm

1
Skins

Foam

RT Adhesive ;and glass-fabric interphases

composition

: CFRP + GFRP

- Rigidite 5245C/G40-600 Lay-up Fabric - 7781-5245C - 2 external layers foam - Syntac 350 ( glass micro

(0/+30/-30)3s (for surface

protection) resin )

Core

composition

Syntactic

balloons

in epoxy

Interphases Fig. 1
Sandwich

composition
configuration

with

Hysol
syntactic

EA9394
foam core

Adhesive

+ GFRP

Fabric

0010-4361/93/050447-04

O 1993

Butterworth-Heinemann

Ltd NUMBER5.1993 447

COMPOSITES.VOLUME24

inspection method for sandwich panels foam core and to summarize a procedure cal assessment of post-impact damage strength.

with syntactic for the practiand residual

A syntactic foam core is a composite itself, since it often contains 50% (by weight) of hollow glass or ceramic microspheres in a thermoset matrix. A disadvantage is that its weight is typically four to eight times higher than that of the traditional foams used in aerospace applications. One main advantage is that the mechanical properties of syntactic foams are several orders of magnitude higher than those of the lighter (traditional) foams _. Sandwich construction with syntactic foam core therefore provides a sensible approach for land- or marinebased applications, where damage tolerance and residual strength, rather than weight savings, dominate the design requirements. After a feasibility study conducted at NASA Ames Research Center, the concept shown in Fig. I was selected as the basis for the design of highly damagetolerant composite wind tunnel compressor blades. Hybrid glass fibre-reinforced plastic/carbon fibre-reinforced plastic (GFRP/CFRP) composite skins were bonded onto a syntactic foam core. Details of the materials together with manufacturing and test procedures are given elsewhere _.2. Extensive low- and high-velocity impact tests revealed that the damage was always localized and confined. This confinement, as shown in Fig. 2, is due to the energyabsorbing capacity of the glass microspheres which are part of the syntactic foam core. Additionally, as shown in Fig. 3, the imprint formed at the GFRP external surface is localized and clearly visible to the unaided eye. This visibility is due to local delamination, over an area which is slightly elliptical (with major axis D), at the hybrid GFRP/CFRP skin interface and has a practical significance, as is demonstrated below. This technical note will address two specific issues: First, what makes this sandwich system damage tolerant? Second, how can the residual compressive strength after impact be determined? Analytical models to predict the residual strength of open-hole composite samples as a function of hole size are available in several publications 3 5. Fig. 4(a) shows an impact-damaged skin and Fig. 4(b) shows a skin in which a hole of diameter D was drilled. The residual strengths of both specimens were found to be equivalent for D ranging between 10 and 20 mm. This in turn suggests that the imprints on the GFRP skin coating are a replica of the damage; hence, a measure of the imprint size will allow the prediction of the residual strength of an impact-damaged sandwich. The localized and confined nature of the impact damage is attributed to the high energy-absorption capacity of the syntactic foam. Scanning electron microscopy (SEM) reveals that most of the impact energy is consumed through crushing of the glass microspheres. This failure mechanism reigns within a hemispherical zone, which is centred at the point of impact and spreads downwards into the syntactic foam core material. This zone is defined by the discolouration of the core, as shown in Fig. 2, which is evidently due to the failed microspheres.

Fig. 2 Confined damage after low-velocity levels of: (a) 47 J; (b) 69 J; (c) 90 J; (d)

impact at impact 136 J; (e) 180 J

energy

448

COMPOSITES.

NUMBER

5. 1993 PRIECEq34NG PAGE BLANK NOT FILMED

Fig. 3

Damageimprint at the externaIGFRPsurface

Fig. 5 SEM micrographs taken (a) inside discoloured zone and (b) outside discoloured zone evident whereas the core damage size (defined extent of discolouration) is consistently larger. by the

--b

Thus highly damage-tolerant sandwich constructions can be obtained by using hybrid composite skins and a syntactic foam core. This is achieved by localization of the damage due to the high absorption of impact energy via crushing of the glass microballoons. The local region of skin failure may be represented by an external imprint that is clearly visible to the unaided eye. Post-impact strength can be predicted by direct measurement of the imprint size using available open-hole theories. The concept which was suggested for the design of highly damage-tolerant wind tunnel compressor blades combines three material phases with specific purposes:
1) CFRP

Fig, 4 Comparison of sandwich skins with impact damage and open hole This is seen from the enlarged micrograph of Fig. 5(a), which was taken inside the discoloured zone, in contrast to Fig. 5(b) which was taken outside this zone. SEM was also used to observe the microstructural pattern of the impact damage. Micrographs of cross-sections in Fig. 6 show the damage for five (low-velocity) impact energy levels. The CFRP skin damage zone can be clearly observed and compared with the GFRP imprint size and the core damage size. Results of these measurements are shown in Fig. 7. A good correlation between external (GFRP) imprint size and internal (CFRP) damage size is

2)

3)

skins, which are the structural backbone, to provide high specific strength and stiffness; syntactic foam core which has high mechanical properties and therefore provides an excellent shear tie between the skins. Additionally it supports the skins against buckling, localizes the impact damage and absorbs energy through a microballoon crushing mechanism; and GFRP fabric which acts as a sacrificial protective coating for the CFRP and serves as a visual enhancement of impact damage for residual strength assessment.

COMPOSITES

. NUMBER

5. 1993

449

,oo i ...... !
25 .. 20,

.[

.......

Io

/
o _5 5O Impacl 100 Energy (J) 125 15o 175 2o0

_v

.....<

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

71

Fig. 7 Effect of low-velocity GFRP, CFRP and foam core

impact

energy on damage size in

aerospace-type transferring

constrtictions technology

need

to

be modified application.

when

to a land-based

REFERENCES 1 2 lliel, C. and lshai, O. "Design of highly damage-tolerant sand'xich panels" 37th lnt SAMPE Symp. March 1992 pp 1228 1242 lshai, O. and lliel, C. "Damage tolerance of a composite sandwich with interleaved foam core" J Composite Techmd Rc._ 14 No 3 (Fall 1992) pp I55 168 Whitney, J.M. and Nuismer, R.J. 'Stress fracture criteria for laminated composites containing stress concentrations' J Composite
Alater 4 Nuismer, faiIure 8 (July R.J. criterion: 1979) C. and pp lshai, strength on Compression 1992 1974) and pp Labor, Part 49 O. 60 'Effect of hybrid Re._7_onse of impact composite of Composite damage skin and open hole ASTM 16 17 on II 253 265 J.D. 'Application J of the average Mater shess 13

Compression'

Composite

d
5

(January lliel,

compressive S.vmp November

laminates" Structurex.

AUTHORS

Dr

Clement

Hiel, R&D Center,

to

whom

correspondence with the EEM Field,

should Composite

be Ames Dan

addressed, Fig. 6 SEM micrographs of localized damage for five (low-velocity) impact energy levels: (a) 47 J; (b) 69 J; (c) 90 J; (d) 136 J; (e) 180J Materials Research Dittman failure The design with syntactic where than by concepts foam may be appropriate for many applications tolerance rather here indicate that the design is driven weight. The findings and design notions by damage presented valid for EEM nently, specializing

is a Principal Program, Moffett and NASA Haifa, Dr

Investigator

Branch, NASA CA 94035, USA. specializing is a Visiting Center. Israel composite

is a Research analysis, in Ishai Branch, Dr

Engineer, Ori Ames Israel. Ishai of

in materials Scientist, at PermaInstitute materials,

mechanics a Professor

Research at Technion,

of Technology,

450

COMPOSITES.

NUMBER

5.

1993