Beruflich Dokumente
Kultur Dokumente
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
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
6 October,
1993
Effect of Impact
Damage
and Open
Strength
of Hybrid
Composite
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
on laminate
on sandwich foam
combine
and a syntactic
test specimen
configurations were
beams which
by impact,
through one of the skins. The reduction (hole) size was coupons beams. compared. were tested Additionally to generate
loaded
hole compression
data for
(*)
CFRP
: Carbon : Glass
Fiber Fiber
Reinforced Reinforced
Plastic Plastic
(**) GFRP
Agreement
at the
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
coupons, foam
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
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
of structural
composite with
materials thermoset
is
Especially to surface
holes,
impact
which fiber
-propagation. more
in compression
of tensile
stresses
which
or free is the
to the fibers
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
when
damage
is localized).
discontinuities solution
in composites distribution
in anisotropic
concentrated tension
of laminated
uniaxial
and compression
Numerous stacking
analytical sequence
and
empirical
studies
have
dealt
with
parametric
effects,
such
as
composition investigations
strength
and stability
of open
motivated
by two major
models
and prediction
information
available
as well as stress assessment the effect of structural damage [19-21 multiple effects ]. This is failure material dispersed of
the data
objective
(mainly
compressive to predict
elements. residual
were not so successful pattern, damage and the mixed in most and
geometric for
characteristics [22-26
impact
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
damaged
second,
a circular
machined damage
one of the skins. The reduction was compared. were Additionally to generate
strength of uniaxially
coupons beams.
tested
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
the
which
section
gives
results an
organizes
interpretation. for
be utilized
conclusions
which
are supported
by this research.
AND
configuration,
sandwich
more
details
(1)
specimens
was made
of precast
syntactic
foam
(SYNTAC micro
350)
Company.
It consists
of epoxy
resin filled
with glass
balloons
of 0.6 gr./cm 3.
The
consists on both
of a central
Carbon and
Fiber film
(CFRP)
laminate bond
center reinforced
laminate.
CFRP
laminate,
Bismaleimide
G40-600/5245C)
supplied
of 18 plies, layers
for external
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
laminate, following
F cure
as supplied
manufacturer.
The measured
of the cured
skin laminate
used to bond
EA9394
room
adhesive
as shown (THC).
in Figure
to by the authors
Hybrid Composite"
Properties material foam, properties of the cured, layers unidirectional in Table CFRP 1. They lamina, the GFRP fabric, the
are given
at room available
dry condition.
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
AND
This section
contains
a description
Series beam
1: Impact flexural
damaged
skin
laminates;
The upper
skin
in Figure instrumented
indentation
with a diameter
D which
depends
of 3 meters by
known
made
Monterey (hardened
with
weighing is driven
impactor motor.
is raised Two
to the
winch
by an electric
during
impacts a braking
test-sample. to insure
system
a single to account
impact
event.
were
defined
calculated 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
by pre drilling
one
of the Hole
in the sandwich
bonding
diameters
The residual
was obtained
speed
as for series
Series only
These
coupons
were the
obtained 3-88
followed
is detailed in Figure.
in reference 5. The
[38].The
SACMA on NASA
fixture
is shown
test method
dimensions
x 12") and the hole diameters 5. Specimens prepared 6. The mm without and tested is as
1.5 to 11.1 mm (.06" of reference 1-88. The special D695. 14c. [40]. Respective
in Figure x 80.8
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 damage
(or hole)
compositions skin
laminates
(series
3) could
only be compared
maximum in this in
at failure Stress
phase
laminate are
distributions
shown
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
contrast in the by is
stress.
obtained failure
stress for
phase strength
compressive
laminate.
distribution
skin laminate.
Here
O'ef which
is obtained
is the accurate
maximum found
compressive
A value for y .882 has been the dimensions specified program. are reported
in Figure
into a laminate
analysis which
below, is given
in the
of these results
Series 1:
Impact
Impact
damaged
sandwich
damage
As was mentioned
induced 9. The
as shown non-
is a good
CFRP
damage
as can be seen
by the cross
damage
the transverse
measure
was defined
to the open
hole diameter
in test
size seems
related
to the increase
as shown
beams
to high velocity
(ballistic)
impact
Strength Nominal
was
Strength
maximum
per
unit failure.
the presence
Increase
a pronounced shown
strength
sizes
originating
the fibers
angle was seen as the controlling also found impact in earlier studies
mode.
(ballistic)
tests [41].
Series
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
It may
therefore sandwich
be is
concluded, justified
damage
post-impact
of impact
strength
can be evaluated
based
on respective
hole sandwich
Series The
3: Open relationship
hole
between
the strength
in the small
failure
to be similar
specimens
shear
30-deg
laminate. earlier
This failure 6.
is also similar
specimens,
in Figure
INTERPRETATION
background formulations hole which were developed in references[2, to examine 3 & 41] for the
of open
compression analytical
strength
will be used
of post-impact
strength
size measurements.
This is justified
11
impact
damage
which
was found
to be typical
having
syntactic stress at in
in the present
investigation.
According
to this model
(W) as shown
As indicated be obtained Y
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)
Strictly called
laminates
Y is
correction is applicable
expression
to orthotropic
for D/W
According orthotropic
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
parameter
a i was stressed
in the highly to be
region
is used
parameter over
an average compression
the damage
to include
laminates
by Nuismer investigation
and Labor
[4]. Two
damage
size-parameters
have
been
configurations
B. Comparison
of the effect
of impact
damage
(drilled)
in Figure (W).
width
test beams
show
a similar
on D/W.
experimental
therefore
represented
by a single
the analytical
of Equation
parameter.
indicates strength
that the use of an open hole model is justified required for the present case.
(impact
energy,
to be known
quantitative findings
residual
strength
predictions.
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
geometries
stress
at failure
as a function
of strength
increasing
coupons.
The
respective
analytical
plots
based
on
Equation
1 reveal
two
different
empirical
a_ may have
its singularity
be a measure
on comparing
the a values
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
It's advantage
(D). This was done in Figure for sandwich in NNS skins were is noticeable
in NNS
4 mm (.16").
A continuous
decrease
of D up to 10 mm (.39").
Additionally (NSL=
a Normalized It is plotted
Loss (NLS)
variable
from NNS 21 ).
1-NNS).
of hole diameter
14
The sandwichskins,with NSLmax=26%, clearly havea substantiallybetter performance thanthe skincouponswith NSLmax--42%.
(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
sub-laminate
presence
of the core
may improve
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.
consideration
The
assumption with
that
the
impact can
which
is typical hole
for
syntactic
as an open of sandwich
can therefore
calculate model,
skins
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
of the
15
impactdamage within
foam core.
observed
in sandwich
beams
with syntatic
- The
data
indicates
lower
to
Normalized follows
Compressive a similar
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
skin counterparts.
of open hole (or impact of the skin to compressive skin, both effects which
sandwich buckling
skins
is attributable
of stress
to the
presence
structural
foam core.
ACKNOWLEDGMENT
would
Nelson,
Engineering
Analysis
at NASA
Research
support the
Development
Center
high-level
specimens.
Additionally, plots
contributions greatly
of Mike
Luft in keeping
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.,
Publishers,
" Stress
fracture
stress concentrations".
J. of Composite
J. D.,
" Application
stress
failure
criterion:
13, Jan.
graphite/epoxy
and Technology,"
in compression,"
TM 100634, Investigation
compressive Vol.
composite
laminates,"
1989, pp. 479-504. W. L., Measurments laminates 8, March loaded of the stress supported by the crush
in compression,"
J. of Reinforced
buckling
Vol. 11, 1989, pp. 121-134. "Buckling of laminated plates with holes," J. of
Materials,
23, June
[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]
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
and Springer,
16, Nov.
[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
Y. and Wilson,
Erikson, Composite
strength
[18]
Tsai, Paris,
"Composite
and Tokyo.
1988, Section
[19]
Williams, holes
J. G. "Tough
Composite strength
Materials;
Effect
damage laminates."
of tough
resin/high
2334, Ginty,
1984, pp 61-79 C. A., predictive " Fiber composite ASTM structural STP IOI2, durability and
tolerance:
Simplified
methods",
[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]
in composite
J. of Composite
[24]
Lesser, composites"
of damage April
mechanism
in laminated
Symposium,
[25]
Failurecharacteristics NASA
components 1982, 24 p.
in compression"
[26]
of compressive
failure
[27]
Sandorff,
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,
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",
1990, pp. 2247-2261. materials", Appl. Mech. Rev. Vol. 44, No. 4
on laminated
C. "Damage
tolerance Technology
sandwich Vol.
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,
Symposium, Test
1992, pp. 1228-1242. Compression 3-88. Revised Edition, NASA Reference Properties of Oriented
Method
Toughened 1983,
Properties
of Oriented
Fiber-Resin
in compression April
testing
of composite
Symposium,
of graphite/epoxy
plates
with
Vol. 20, No. 1, Jan. 1980, 8 pages. testing of [02/+45]s graphite/epoxy plate with hole," Experimental
Science
and Technology,
of non-damaged configurations
Illustration Flexural
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
Cross-section
(Impact under
of cross-section
specimens
different
levels
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:
sandwich
beam beam
under under
flexure flexure
uniaxial
on nominal on nominal
strength strength
sandwich
for prediction
of compressive on residual
of impact
and open
hole effects
of sandwich
Fig.
of open-hole beams
effects
on normalized
nominal
compressive
for sandwich
sandwich
<_,-_
. ,._
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
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)
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/
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
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
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
oJ
N
oi
em
E E
!
......... i ........ C;
=
0 i
o_
l
o d
,-
oJnl.tUd
upis
1_H POZt.lumJoH
_
0
emm
CD 0
i
0
T-
OO
0 o
_o
II Appendix
Papers published during grant period June 1, 1992- May 31, 1993
COMPOSITE
MATERIALS,
MECHANICS
AND PROCESSING
HELICOPTERS
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
analytical prediction
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
: Sandwich
1150
THICK COMPOSITES
: Finite
Width
Correction
Factor.
ON oN _n
INTRODUCTION
on
carbon-epoxy
laminates
has
clearly
shown
that
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
which compare low and rare. Cantwell and Morton Grafil They XA-S/BSL914C found that for
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
review, Abrate
on (1991)
damage and by
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
author's
high-velocity
1151
sandwich subsequently
a structural to as a thick
type
of material
is
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
between residual
strength
is represented
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
(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
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.
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
shown bandsaw.
cut then
from
the
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
1152
THICK COMPOSITES
ft) to 2.13
m (7 ft) thereby
creating
impact
velocities
ranging
from simply
(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
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
passed of 1.79 by
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
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
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
formation
Figure
3. is representative
The permanent to be deeper than energy. of high a THC damage Aditionally, velocity at low inflicted
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
z-direction
Impact
Response
of Thick
Hybrid, Composites
1153
Further damage caused energy. caused size Final by both by high Following
of the energy
damage
by
relating
the on the
be seen, the
damage 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
defined
localized,
and
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
dividing Y, which
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
equations were originally to a through the thickness aoc was region determined damage introduced adjacent by zone. by to
to represent
stressed
fitting experimental This criterion has Nuismer the over and Labor
an average to 'include
laminates
assumption damages
in using material
described a radius
analytical R, and
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
energy,
and
shows
that It
to control
strength
sandwich
interleaved
CONCLUSIONS
O
Damage having
size the
was same
found energy.
to be similar
for both
low
and
high
velocity
impacts
was
found
to resemble laminates
materials velocity.
at
thermoset
1155
Reduction in residual while impact velocity The Whitney-Nuismer an appropriate size to residual
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
S.,
I991, Reviews,
"Impact vol
on
Laminated 1991, pp
Materials,"
Applied
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
of
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.,
Core,"
of
Impact of the
and Sixth
Delamination International
Resistance Conference
in on
F.L. Matthews
1156
THICK COMPOSITES
Labor,
J.D.,
1979,
"Applications
of the
Average
Stress
Failure
Materials, Characteristics
Structural Components Loaded in Materials-Recent AdvancesHashin, 1983, pp 283-306 J.M. and R.J. Nuismer, Stress
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__ ............
"-'_
...................
Fig. 3
High-velocity
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_
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.
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
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
Room Residual
temperature compressive
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
elastic tensile
moduli,
re-
strength,
transverse
compressive
F_
Gt.: g H /t L,, I go
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,
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
Fiber-reinforced Honeycomb
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
ing. This
this effect,
by interleaving
glass/epoxy
PRE_NG
PAGE BLANK
NOT
FILMED
156
JOURNAL
OF COMPOSITES
TECHNOLOGY
& RESEARCH
The
main
objective data
was
to provide (DTE) of
may
absorb
the
impact
energy.
The
weak
link
in sandwiches
in
tolerance
evaluation
composite
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
_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
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
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
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
erence strength. The strain seems, nowadays, for high performance craft applications. material selection DTE such testing the high boundary sensitivity conditions
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
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
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
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 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.
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
subjected plate
to flexural mainly
laminates
the following two reasons. First, loads when the sandwich is under shear Second, stresses are confined mainly a relatively the core provides
is under phme axial hence, interlaminar local impacted which zone. locally
studies
standardized
for improving
soft substrate
of materials
and composite
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
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,
interleaved
of the constituent
materials
for the
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
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-
1. It consists by four
backbone
of the slmdwich
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,
as a result
interleave types
of a press-molding of specimens
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
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
Protective
glass-fabric
coating
residual strength foam cores. a database into account the effect design
of composite
with
structure
Plain core sandwich reference
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
rough
cutting
with
saw, water
Fabrication
of
I-I_ /
sandwich specimens
and parallel
Low vemoelty__1 tl
E.ect ofim_ct
Constituents properties fabric, of the cured, the syntactic unidirectional foam, and the
I ,.__] _
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
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
aiming at improved
(ASTM Test for Tensile Properties of Plastics Test for Flatwise Compressive Strength of
FIG. 2--Scheme
of research program
[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.
maintenance,
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.
these specimens were strength determination. and plain based strength. were evaluated
of interleaved variables
Impact
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
OF COMPOSITE
SANDWICH
159
Low
set-up
High velocity
(balistic),
Air- gun
set-up
/
ImpactorHeight- H< 2.5m
Impactor Weight- W = I
w
f
200ram
l 200mm "_
JL
Velocity Energy
Velocity Energy
FIG. 3--Illustration
tower base
height with
is about hemispherical
impactor
of a to is
lower
values to
of the the
velocity The
variables to frictional
as drop
16-mm-diameter
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
86 N. The system
pneumatically.
Its rebound
Packard sandwich
circular _,elocity
made
by Endevco on
during
experimentally.
determined photo
supported recorded
a span velocity
measuring
Typical
velocity and the derived arc plotted as functions with the respective
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
T
1.5 [ml 2,0 2.5
20 0
......
0.0
0.5
2.0
2.5
FIG. 4--Calibration
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)+
core sandwich.
Recorded
during
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
noncracked minimum
core
sandwich,
shows which
a tend
towards
the
air
was
released.
The
abrupt
air expansion
accelerated
deceleration)
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
and response
activated
located
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
low velocity
impact
Impact
Damage
Characterization loading, each specimen dimensions of the damage was inspected were measured, visually namely:
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
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
was fed to a chamber in which diaphragm. At a predetermined was ruptured by electrical heating
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
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
at skin SSSF
at failure.
'RL3 6_.1, ,+57 15.t+
(RCS)
of the damaged
laminate material
strength stress
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
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
Assessment glass found imprints fabric-epoxy whose layers sensitive on the varied external skin
surfaces which
to be highly
to the impact
loading
dimensions
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.
113.1
83.4
5.09
16.7
inspection
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
zone
impact
levels.
24.0
17.7
m'
2.35
7.7
These process
core failure.
of the plain
(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
P
CFRP Skin I Load Fixture
Syntactic Foam
2 I d= h+t
f/////////I////ll//tt/////2
(_'max skin
I
----
Peh / 2btd 2
=
d = 27 b = 76 of residual strength.
r_average core
P / 2bd
FIG. 9--Flexural
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
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 ]).
plain
core version
as compared
with
its
The
Effect
of Impact three
Variables parameters
Ultimate Loaded
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
ultimate in-plane
was due to skin damage. of three separations and core the impact failure Residual stress also the
(see Fig. 16), namely, delamination the CFRP Failure With predominant mode failure was determined site. the failure layers. was
laminate seems
geometry. impact
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
version. compressive
strength at sandwich
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
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
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
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
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
__.. ....
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
of interleaving,
the comparison
at low impact
the tendency
to level
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
by core cracking
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
was
due
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
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,
200
. i m
--
100
?-_
"--
___
20) revealed
| 20
1
40
, 60 Impal
, 80 Enaroy
i
100
. 120
I I 140
, 160
(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,
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
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[]
i
> 2OO --
\_
investigation
3
_e
experimental
size for low velocity Fig. the 22). present A full case
d
0 .... 20 40 60 Impoct I 80 Enorgy , t00 (J) 120 i40 160
and compared
from
analytical
FIG. 19--The effect of low velocity impact energy on residual strength of sandwich panels (with plain versus interleaved foam core).
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
!
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
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
mainly sandwich
to the damtemper-
performance
of a composite suitable
fl_r elevated
_1
40 60 eo Impll 1oo Inlr_y 12o IJ) 140 160 18o 2O0 220
ature
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
with an
impact energy
energy level of
to [18]
Damage impact
are
directly
dependent
on weight. same
velocity are
Damage way by
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
Center
support Shop
to Paul
Scharmen
high-level
manufacturing
sandwich
[2]
[26]
[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]
[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
International
SAMPE
Symposium,
April
1990,
pp.
1620-
[42]
Avery, J., Allen, M. R., Sawdy, D., and Characteristics of Composite Compression
[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.
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
construction core
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
composition
: CFRP + GFRP
- Rigidite 5245C/G40-600 Lay-up Fabric - 7781-5245C - 2 external layers foam - Syntac 350 ( glass micro
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
COMPOSITES.VOLUME24
inspection method for sandwich panels foam core and to summarize a procedure cal assessment of post-impact damage strength.
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)
energy
448
COMPOSITES.
NUMBER
Fig. 3
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
impact
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,
laminates" Structurex.
AUTHORS
Dr
Clement
to
whom
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
Investigator
mechanics a Professor
Research at Technion,
of Technology,
450
COMPOSITES.
NUMBER
5.
1993