See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/24304015

Damage Characteristics and Residual Strength of Composite Sandwich Panels

Impacted With and Without a Compression Loading

Article  in  Collection of Technical Papers - AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference · February 1998
DOI: 10.2514/6.1998-1783 · Source: NTRS

CITATIONS READS

17 134

2 authors, including:

David Mcgowan
NASA, Langley Research Center
20 PUBLICATIONS   180 CITATIONS   

SEE PROFILE

All content following this page was uploaded by David Mcgowan on 13 March 2015.

The user has requested enhancement of the downloaded file.

 DAMAGE CHARACTERISTICS AND RESIDUAL STRENGTH OF
COMPOSITE SANDWICH PANELS IMPACTED WITH AND WITHOUT
COMPRESSION LOADING

David M. McGowan and Damodar R. Ambur

NASA Langley Research Center
Hampton, VA 23681-0001

Presented at the 39th AIAA/ASME/ASCE/AHS/ASC Structures,

Structural Dynamics and Materials Conference
Session No. 15 - Damage Tolerance

AIAA Paper No. 98-1783

Long Beach, California

April 20-23, 1998
 DAMAGE CHARACTERISTICS AND RESIDUAL STRENGTH OF COMPOSITE
SANDWICH PANELS IMPACTED WITH AND WITHOUT A COMPRESSION LOADING

David M. McGowan* and Damodar R. Ambur 

NASA Langley Research Center
Hampton, VA 23681-0001

Abstract
The results of an experimental study of the impact damage characteristics and residual strength of composite
sandwich panels impacted with and without a compression loading are presented. Results of impact damage
screening tests conducted to identify the impact-energy levels at which damage initiates and at which barely visible
impact damage occurs in the impacted facesheet are discussed. Parametric effects studied in these tests include the
impactor diameter, dropped-weight versus airgun-launched impactors, and the effect of the location of the impact
site with respect to the panel boundaries. Residual strength results of panels tested in compression after impact are
presented and compared with results of panels that are subjected to a compressive preload prior to being impacted.

Introduction structure that need to be addressed are the criterion for

Sandwich structures have been used extensively in visible impact damage and the effect of impacting a
stiffness-critical aircraft secondary structures. The preloaded structure on its residual strength. The
application of sandwich structures in aircraft primary existing visible impact-damage criterion for thick
structures has been limited mainly because of monolithic laminates is a 0.1-in. dent depth or 100 ft-
undesirable moisture absorption and moisture retention lbs of impact energy. This criterion has been used
characteristics, and due to an insufficient understanding successfully for thick laminated structures such as wing
of the effects of various types of damage on the cover panels. Sandwich structures which are proposed
structural response of sandwich structures. One such to be used in commercial transport aircraft fuselage or
source of damage is caused by low-speed impact events wing structures use facesheets of 8 to 16 plies (0.04 in.
that result from dropped tools during manufacturing to 0.08 in.) thick for which a 0.1-in. dent depth
and from runway debris and hail stones during service. corresponds to a penetration. Reference 4 indicates that
Sandwich structures have the potential for improved a dent depth of 0.1 in. is not an appropriate criterion for
structural efficiency and reduced manufacturing cost. visible impact damage for 8- to 16-ply-thick facesheet
This potential can be further increased by using sandwich plates. A 0.05-in. dent depth has been
advanced composite material systems to construct the identified as a potential visible damage criterion in Ref.
facesheets of the sandwich structure. These advantages 8 for relatively thick facesheet (up to 36 plies thick)
will make composite facesheet sandwich concepts very sandwich structures. A similar criterion which can be
attractive for application to aircraft primary structures1,2 used for field service inspections of transport aircraft
if the low-speed impact damage tolerance issues for this made from thin-facesheet (8 to 12 plies thick)
type of structure can be understood and resolved. composite structures is needed. This criterion has to be
Most of the reported experimental research on the applicable for different impactor diameters, facesheet
impact of sandwich plates was conducted to determine ply stacking sequences, impactor masses, sandwich
the effectiveness of different core materials, to study core densities, and impact sites away from the center of
the influence of different facesheet materials and the panel. The results from this study may be used to
stacking sequences on damage resistance, to simulate establish the damage tolerance criteria for thin-
impact tests using static indentation tests, to identify the facesheet sandwich structures.
damage modes due to impact, and to determine the The current practice for determining damage
compressive or tensile residual strength and failure tolerance of structures with skin-stringer construction is
modes associated with final failure.3-7 The two main to impact the unloaded skin laminate and then load it to
damage tolerance issues associated with sandwich failure in compression to determine its residual
strength. This approach is acceptable for thick
*
Aerospace Engineer, Structural Mechanics Branch, monolithic laminates such as those studied in Ref. 9
Structures Division. Senior Member, AIAA.
 
Assistant Head, Structural Mechanics Branch, where the residual strength results for specimens
Structures Division. Associate Fellow, AIAA. impacted while preloaded and then loaded to failure do
Copyright Ó 1998 by the American Institute of Aeronautics and
Astronautics, INC. No copyright is asserted in the United States under
Title 17, U.S. Code. The U.S. Government has a royalty-free license to
exercise all rights under the copyright claimed herein for government
purposes. All other rights are reserved by the copyright owner.
1
American Institute of Aeronautics and Astronautics
not differ significantly from the results obtained from honeycomb core with 1/8-in. hexagonal cells and a
compression-after-impact tests for impact-energy levels density of 4.5 lb/ft3. The two facesheets are co-bonded
of practical interest. Similar data is needed for to the core using a film adhesive. The test matrix for
sandwich structures. Only Ref. 7 addresses the the study is summarized in Table 2. A diagram of a
influence of compression preload on the residual typical test panel that identifies the impact site in the
strength of impact-damaged sandwich plates. This panel is shown in Fig. 1. The impact-energy levels
work, which was limited to (0/90)s graphite-epoxy selected for the CAI and the impact-under-load tests
facesheet sandwich beam specimens, suggested that were based on the results from the impact-damage
these 3.0-in.-wide sandwich beams fail on impact while screening tests to be presented subsequently.
preloaded in compression at more than 40 percent of As part of the test panel preparation, the core at the
the undamaged specimen compression failure load. At 5-in.-wide loaded ends of the panels was removed to a
the visible damage threshold, which was identified to depth of 0.5 in., and epoxy resin material was potted
occur for this sandwich panel at 0.88 ft-lb of impact between the facesheets to prevent an end-brooming
energy7, the ratio of the preload to the undamaged failure. These potted ends were machined flat and
specimen failure load was determined to be 0.52. parallel to each other to assure that a uniform load is
Finite-width effects influence the impact damage applied to the panel. For the CAI tests, the impacted
characteristics, the failure strength, and the failure facesheet of the panel was painted white so that a
mode of structures. Thus, there is a need for studying shadow moirŽ interferometry technique could be used
the residual strength of compression-loaded sandwich to monitor out-of-plane displacements and to observe
plates with more practical facesheet stacking sequences any local response such as delamination buckling and
and with a well quantified visible damage criterion to growth at the impact sites. The moirŽ fringe patterns
fully understand the behavior of preloaded sandwich were recorded using still and video photography.
structures subjected to low-speed impact in service. For all tests, the panels were loaded in uniaxial
The present paper presents and discusses results of compression in a 120-kip-capacity hydraulic test
a series of impact-damage screening tests and an machine. A typical test setup is shown in Fig. 2.
experimental study of the compression-after-impact Clamped boundary conditions were provided at the
(CAI) and impact-under-load responses of composite loaded edges of the panels by a steel support fixture.
facesheet sandwich panels subjected to low-speed The unloaded edges of the panels were supported with
impact damage. The impact-damage screening tests knife-edge supports. Surface strains were measured
establish the impact-energy levels for damage initiation using electrical resistance strain gages. Direct-current
and barely visible impact damage (BVID) for the test displacement transducers (DCDTÕs) were used to
panels used in this study. Compression tests on panels measure out-of-plane displacements of the facesheets as
impacted using dropped-weight and airgun-propelled well as the stroke of the test machine loading platen.
impactors with predetermined impact energy levels The axial load applied to the panels was measured
values between damage initiation and BVID levels are using the test machine load cell. The strain gage,
conducted to obtain the CAI results. Additional impact DCDT, and load data were recorded using a digital data
tests with the same predetermined impact-energy levels acquisition system. Data were taken at one second
are conducted on compression-preloaded panels. intervals while loading the panels to failure.
Preloaded panels which survive the impact events are
loaded to failure to determine their residual strength.
Impact Damage Screening Tests
The impact-damage screening tests were performed
Test Panels, Apparatus, and Conditions on samples machined from the same panel as the test
The eleven sandwich test panels of the present panels to establish the impact-energy levels necessary
study were machined from a larger panel and are 5-in. to initiate damage and to inflict barely visible impact
wide and 10-in. long with thin composite facesheets damage (BVID) in the impacted facesheet. These
and honeycomb core. The facesheets are 8-ply-thick screening tests were performed using a dropped-weight
laminates manufactured using Hercules, Inc. impact apparatus10 and an airgun-launched impactor.
preimpregnated AS4/8552 graphite-epoxy tape and Prior to being impacted, each sample was placed into a
woven fabric materials. The facesheet laminate steel support fixture with a 4.0-in. by 5.0-in. opening
stacking sequence is [O f/45/-45/0/90/-45/45/Of], where that was secured to a rigid table. The screening tests
an ÒfÓ indicates a fabric ply. Nominal elastic properties were also conducted study the parametric effects of the
of the two materials used in the facesheet laminates are impactor diameter and the location of the impact with
given in Table 1. The core material is Korex¨ respect to the boundaries of the samples.

2
American Institute of Aeronautics and Astronautics
 The dropped-weight impact apparatus uses a 2.5-lb dropped-weight impacts with a 0.5-in. tup and for the
impactor with either a 0.5-in.- (which is a common airgun impacts.
practice) or a 1.0-in.-dia. spherical instrumented steel Contact force profiles obtained during dropped-
tup. The 1.0-in.-dia. tup is used to understand the weight impacts with the 0.5-in.- and 1.0-in.-dia. tups at
sandwich panel impact response and damage impact energy levels of 1.0 and 1.5 ft-lbs are presented
characteristics for a larger tup diameter. The airgun in Fig. 7. Although the maximum contact force for
uses a 0.5-in.-dia. aluminum sphere as the impactor. impacts with the 1.0-in.-dia. tup are comparable to
The 2.5-lb dropped-weight impactor was determined to impacts with a 0.5-in.-dia. tup, the associated damage
be acceptable for the impact tests based upon the area is much smaller. This result is due to a larger
analytical results presented in Fig. 3 that were contact area of the panel experiencing a smaller contact
generated using the analysis approach presented in Ref. pressure. These results suggest that the damage
11. Maximum contact force values as a function of resulting from dropped-weight impacts with a 0.5-in.-
impactor speed are presented in Fig. 3 for a sandwich tup and from the airgun are very similar at these low
panel subjected to an impact energy of 7.0 ft-lbs. The impact-energy levels and that a 0.5-in.-dia. tup initiates
results in this figure represent different combinations of damage in the facesheet at much lower values of impact
impactor mass and speed that must be used to result in a energy than the 1.0-in.-dia. tup.
constant impact energy level. These results suggest that
for impactor speeds less than 289 in./sec, (i.e., impactor Barely Visible Impact Damage in a Facesheet
weight values less than 1.0 lb), variations in the A series of impacts was also performed to
impactor speed (or mass) do not significantly influence determine the impact-energy level necessary to inflict
the contact force magnitude, which is primarily barely visible impact damage (BVID) in the facesheet
responsible for the resulting damage. of the sandwich panels. Impact-energy levels greater
than or equal to 5.0 ft-lbs were used. The criterion used
Initiation of Damage in a Facesheet for BVID in the present study is that the dent depth at
Several impacts were performed to establish the the impact site should be greater than or equal to 0.05
minimum impact-energy level necessary to initiate in. As shown in Fig. 5, BVID occurred for 7.0 ft-lbs of
damage in the impacted facesheet of the samples. impact energy for the dropped-weight impact using the
Impact-energy levels from 0.5 ft-lb to 3.0 ft-lbs with 0.5-in.-dia. tup and the airgun impact. However, as
energy increments of 0.5 ft-lb were used. The initiation shown in Fig. 6, the damage area resulting from the
of damage is determined from ultrasonic C-scan airgun impact is slightly larger than that from the
inspection of the impacted facesheet. The results of a dropped-weight impact. This result is also supported
typical C-scan inspection are shown in Fig. 4 for a by the photomicrographs shown in Fig. 8. These
dropped-weight impact energy of 4.0 ft-lbs using a 0.5- photomicrographs are from machined and polished
in.-dia. tup. As shown in the figure, the damage area is sections through impact sites in two samples impacted
taken to be the light-colored area in the C-scan. From at an impact-energy level of 7.0 ft-lbs. One sample was
similar C-scans of the impact-damage screening dropped-weight impacted using a 0.5-in.-dia. tup and
samples, damage is considered to have initiated at an the other sample was airgun impacted. As seen in Fig.
impact-energy level of 1.5 ft-lbs for the dropped-weight 8, the damage resulting from the airgun impact is more
impacts using the 0.5-in.-dia. tup and at 1.0 ft-lb for the severe than that resulting from the dropped-weight
airgun impacts. The extent of damage that resulted impact even though the dent depths are equal and the
from the dropped-weight impacts using the 1.0-in.-dia. damage areas are similar. This observation is also
tup was considered to be negligible for impact-energy supported by the results in Fig. 3 which suggest that for
levels less than or equal to 3.0 ft-lbs. a given impact energy level, the contact force
Measured values of the residual dent depth and the magnitude increases marginally as the impactor speed
damaged area at each impact site are plotted as a increases (i.e., for decreasing impactor mass). These
function of the impact energy in Fig. 5 and Fig. 6, results indicate as the impact-energy level increases,
respectively, for dropped-weight impacts using a 0.5- low-mass, high-speed airgun impacts can cause more
in.- and a 1.0-in.-dia. tup and the airgun impacts. The extensive internal damage that may be difficult to detect
damaged area measurements are approximated from C- visually. Thus, using dent depth only as a criterion for
scan inspections of the impacted samples. As seen in measuring the severity of impact damage in sandwich
Fig. 5, the values of dent depth for all three impact structure may not always be reliable.
methods are very similar to each other for impact-
energy levels of up to approximately 3.0 ft-lbs. The Effect of impactor diameter on BVID. Due to
values of the damaged area are also very similar for the restrictions in the impact test machine, the largest

3
American Institute of Aeronautics and Astronautics
impact-energy level that can be attained with the Fig. 12. As shown in the figure, the two contact force
dropped-weight apparatus using the 2.5-lb impactor profiles have very similar duration and magnitude.
used in this study is 12.7 ft-lbs. The residual dent depth Photomicrographs taken of the two impact sites also
at this impact-energy level for an impact using the 1.0- indicate that the damage states of the two impact sites
in.-dia. tup is 0.0425 in., which is slightly less than the are very similar. These results indicate that the location
0.05-in. criterion used in this study. An impact at the of the impact in sandwich structure with respect to the
same level of 12.7 ft-lbs with the 0.5-in.-dia. tup boundaries of the structure does not significantly
resulted in impactor penetration through the facesheet. influence the damage state produced by the impacts and
These results, as shown in Fig. 5, indicate that, as the that the impact event is very local.11
impact-energy level is increased to a value greater than
3 ft-lbs, the tup diameter does influence the dent depth Impact-Energy Levels for Residual Strength Tests
resulting in a significant difference in the impact- Ten 5.0-in.-wide by 10.0-in.-long sandwich panels
energy values at which visible damage occurs. were impacted for residual strength testing. To allow
Contact-force profiles for dropped-weight impacts for a direct comparison of results from the airgun
with the 0.5-in.- and 1.0-in.-dia. tups at impact-energy impacts which use a 0.5-in.-dia. sphere, only the 0.5-
levels of 6.0 and 7.0 ft-lbs are presented in Fig. 9. in.-dia. tup was used for the dropped-weight impacts for
Although the maximum contact force value for the 0.5- these tests. Based upon the results of the impact-
in.-dia. tup remains relatively constant for these energy damage screening tests, the CAI panels were impacted
levels, the dent depth increases from approximately at energy levels of 1.5 ft-lbs, 4.0 ft-lbs, and 7.0 ft-lbs
0.04 in. to 0.05 in. (see Fig. 5). Similarly, the using the dropped-weight impactor and at 1.1 ft-lbs, 3.8
maximum contact force value for the 1.0-in.-dia. tup ft-lbs, 4.2 ft-lbs, and 6.9 ft-lbs using the airgun
remains nearly constant while the dent depth increases impactor. The minimum and maximum impact-energy
from 0.0055 in. to 0.007 in. Furthermore, although the levels represent the thresholds for damage initiation and
maximum values of contact force for the impacts using BVID in the impacted facesheet. The variations in the
the 1.0-in.-dia. tup are higher than those for the impacts impact-energy levels for the airgun impacts occur due
using the 0.5-in.-dia. tup, the values of dent depth and to slight fluctuations in the speed of the airgun-
damage area are lower for the larger diameter tup. launched impactor. The impact-energy level used for
Photomicrographs of machined and polished sections the preloaded panels was selected to be 7.0 ft-lbs,
through the impact sites of two samples impacted at an which corresponds to a BVID. These impacts were
impact-energy level of 6.0 ft-lbs using the 0.5-in.- and performed using the airgun only. The procedure for
1.0-in.-dia. tups are shown in Fig. 10. As shown in this establishing the preload value for the impact-under-load
figure, the damage states resulting from the two impacts panels is discussed in a later section.
are very different. Significant internal damage resulted
from the impact using the 0.5-in.-dia. tup, while only
minor damage in the form of a delamination between Residual Strength Characterization
the central 0° and 90° plies resulted from the impact Results from residual strength tests of eleven
using the 1.0-in.-dia. tup. This result further illustrates sandwich panels are presented in this section. Panel 1
that tup diameter does affect the dent depth and it is, in was an undamaged control panel. Results for CAI
addition to the contact force magnitude, an important panels impacted using a dropped-weight impactor with
parameter in determining the threshold impact-energy a 0.5-in.-dia. tup and an airgun impactor are compared.
level for BVID in this sandwich structure. Due to a technical problem, the airgun impactor did not
strike the center of Panel 6. This impact was repeated
Effect of Impact Location for Panel 7, and Panel 6 was tested without knife-edge
Two dropped-weight impacts were performed to supports along the unloaded edges to investigate the
gain insight into the effects of the location of the impact effect of boundary conditions on the structural response
with respect to the boundaries of the sample. The tests of the panel. The effect of applying a compressive
were performed using the 0.5-in. tup and an impact- preload prior to impacting the panel is also discussed in
energy level of 7.0 ft-lbs. Impact sites shown in Fig. 11 this section. The panel numbering and the type of test
that were 0.75 in. and 2.5 in. inward from the boundary performed on each panel is given in Table 2.
of the fixture were selected. The values of the residual
dent depth for the two impacts sites are 0.0415 in. and Compression After Impact (CAI) Tests
0.0455 in., respectively, and the values of the damage Typical load-end-shortening curves for Panels 1
area are 0.2155 in 2 and 0.2355 in2, respectively. The and 5 through 8 are shown in Fig. 13. The failure load
contact force profiles for the two impacts are shown in for each panel is shown as a filled circle. A comparison

4
American Institute of Aeronautics and Astronautics
of the data of Panels 5 - 8 with that of Panel 1 indicates level that produces BVID, this difference increases to
that the presence of impact damage has no effect on the 13.7 percent, with the airgun-impacted panels failing at
global axial stiffness of the panels. The load-end- lower loads. This result is consistent with the results of
shortening data for Panels 1 - 4 also support this finding. the impact-damage screening tests. The damage
Typical axial surface strain results that show the associated with airgun impacts was shown to be more
redistribution of load around the impact damage severe than that associated with dropped-weight
location are shown in Fig. 14 for Panels 1 through 4 and impacts as the impact energy level is increased.
in Fig. 15 for Panels 1 and 5 through 8. In these
figures, far-field strains at the location marked SG A Impact Under Load Tests
are plotted using solid lines, and strains at a location The effect of applying a compressive preload to a
adjacent to the impact site (SG B) are plotted using panel before impacting it was also investigated in the
dashed lines. As shown in Fig. 14, Panels 3 and 4 present study. Panels 9 through 11 were first loaded to
exhibited local strain concentrations (ratio of SG B to a predetermined load level and then impacted with the
SG A data) near the impact site. Each panel exhibited a airgun-launched projectile at impact energy levels of
maximum measured strain concentration of 2.54. Panel 7.2, 7.5, and 7.1 ft-lbs. Again, the variation in the
2 did not exhibit any local strain concentration near the impact energy levels occurs due to variations in the
impact site. Panels 5, 7, and 8 exhibited maximum speed of the airgun-launched projectile.
measured strain concentrations of 1.61, 2.83, and 2.62, The value of the preload was selected based upon
respectively. The strain concentrations exhibited by the the design stress for the original panel from which the
airgun-impacted panels were slightly higher than those test panels were machined. The original panel was
for the dropped-weight-impacted panels for a given designed to study the effect of discrete-source damage
impact energy level. These results support the observed on this sandwich structure. The design stress for this
severity of the damage conditions for airgun impacts as panel, which would correspond to a design limit stress,
compared to that for dropped-weight impacts. was 14.4 ksi. This stress corresponds to a load of 8,500
Panel 6 was tested without knife-edge supports, lbs for the present test panels. Therefore, Panel 9 was
and it exhibited a strain concentration of 3.55 just prior loaded to 8,500 lbs before being impacted. Since Panel
to failure. However, DCDT measurements normal to 9 did not fail at impact, the preload value was increased
the undamaged facesheet indicate that bending occurred to 12,800 lbs for Panel 10. This 150 percent increase
just prior to failure in a direction that increased the represents the difference between design limit and
compressive strain in the impacted facesheet. Thus, the design ultimate stress. Panel 10 did not fail at impact.
absence of knife-edge supports appears to only have In an attempt to establish a narrow load and strain range
affected the panel response just prior to failure. The at which failure would occur at impact, Panel 11 was
global axial stiffness of the panel was not affected and preloaded to 18,000 lbs. This preload is 200 lbs less
the failure load was very close to that of Panel 7. than that at which Panel 8 (CAI panel airgun impacted
The failure mode for all impacted panels was a at 6.9 ft-lbs impact energy) failed. Panel 11 failed at
compression failure across the width of the panel impact. This result indicates that there is a threshold
passing through the impact site. The failure mode for value of strain in the loaded panels above which impact
Panel 6 is shown in Fig. 16. This figure also shows the at an energy level high enough to cause BVID will
deformation of Panel 6 that contributed to the increased result in failure. A summary of the preload value,
strain concentrations at the impact site. The failure impact energy level and final failure load for Panels 9
loads for all panels tested are plotted in Fig. 17 as a through 11 is presented in Table 2. These results also
function of the impact energy level at which each panel suggest that for the BVID criterion chosen in this study,
was impacted. In Fig. 17, the experimental results for the sandwich panel can support approximately 47
the undamaged and CAI panels (1 through 8) are percent of its undamaged failure load before failing
plotted as open symbols, and the results for the impact- when subjected to airgun impact.
under-load panels (9 through 11) are plotted as the Load-end-shortening curves for Panels 8 through
filled circles. Polynomial curves fitted through the data 11 are presented in Fig. 18. The failure load for each
for panels 1 through 8 are also shown to assist in panel is indicated by a filled circle. Panel 8 is included
identifying trends in the data. There is only a 5.1 to determine if differences in response occur between a
percent difference in the failure loads for the panels CAI panel and an impact-under-load panel when both
impacted with the dropped-weight apparatus as panels are impacted with the same impact-energy level.
compared to those impacted with the airgun at the These data indicate that the effect of impacting
impact energy level at which damage initiates. compressive preloaded panels on the global axial
However, as the impact energy level is increased to a stiffness of the panels was negligible.

5
American Institute of Aeronautics and Astronautics
 Axial surface strain results for Panels 8 through 11 tup. Results also indicated that the damage states
measured at locations away from the impact site and resulting from dropped-weight impacts with a 0.5-in.-
adjacent to the impact site are shown in Fig. 19 and Fig. dia. tup and the airgun were very similar to each other
20, respectively. The results in Fig. 19 indicate that the for low values of impact energy. As the impact-energy
far-field response of the impacted facesheet is relatively level was increased to that which causes BVID, the
insensitive to the order in which the panel is impacted airgun impacts caused more extensive internal damage
and loaded. However, the results in Fig. 20 indicate than the dropped-weight impacts even though the dent
that the local response of the impacted facesheet is depths for the two methods of impact were very similar.
sensitive to this order. Upon impact, Panel 9 exhibited This result suggests that using dent depth only as a
a 54 percent increase in the strain adjacent to the impact measure of the extent of damage in an impacted
site (approximately 1,200 m -in/in), and Panel 10 structure may not always be reliable. Finally, dropped-
exhibited a 43 percent increase (approximately 1,500 m- weight impact test results with a 0.5-in.-dia. tup in the
in/in). Panel 11 failed upon impact at a strain level of center and near the edge of a panel suggest that the dent
approximately 4,900 m-in/in. The strain at this location depths, damage area, and contact force profiles for each
in Panel 8 just prior to failure was 11,600 m-in/in, impact are very similar to each other, indicating the
which is more than twice as large as that in Panel 11 local nature of the impact response.
when it was impacted. This result also supports the Tests of CAI panels impacted with 0.5-in.-dia.
conclusion that there is a threshold value of strain in the dropped-weight and airgun impactors were conducted
loaded panels above which impact at an energy level using impact-energy levels determined from the impact-
high enough to cause BVID will result in failure. damage screening tests. Results of the CAI tests
Although this strain threshold is larger than the indicate that the global response of the panels was not
relatively low design strain for the present sandwich affected by the presence of impact damage. The
panels, a more aggressive structural design could easily residual strength of the panels impacted using both
require design strain values that approach this methods was similar to each other for low-energy-level
threshold. The existence and value of this threshold impacts. However, as the impact energy level was
must therefore be taken into account when using increased, the residual strength of the airgun-impacted
residual strength results determined from CAI tests to panels was less than that of the dropped-weight-
design aircraft structures. impacted panels. This result is consistent with the
results from the impact-damage screening tests.
Compression tests were also conducted on panels
Concluding Remarks that were preloaded prior to being impacted with the
The present paper describes the results of impact- airgun. The impact energy level used was that required
damage screening tests and an experimental study of to inflict BVID in the impacted facesheet. Preloaded
the compression-after-impact (CAI) and impact-under- panels that survived the impact events were loaded to
load responses of composite facesheet sandwich panels failure to determine their residual strength. Results of
subjected to low-speed impact damage. These panels these tests indicate that, for the BVID criterion used in
were 5-in.-wide and 10-in. long with thin composite this study, the global response of the panels was
facesheets co-bonded to honeycomb core. Dropped- insensitive to the order in which the panel is impacted
weight impactors with 0.5-in. and 1.0-in.-dia. steel tups and loaded. However, the local response near the
and an airgun-launched impactor (0.5-in.-dia. aluminum impact site is very sensitive to this order. Furthermore,
sphere) were used in the present study. these tests indicated that there is a threshold value of
A series of impact-damage screening tests was strain in the preloaded panels above which impact at an
conducted to determine the impact-energy levels at energy level high enough to cause BVID will result in
which damage initiates and at which barely visible failure. Therefore, to allow for more aggressive designs
impact damage (BVID) occurs in the impacted that utilize composite facesheet sandwich structure, the
facesheet. The criterion used for BVID in the present existence and value of this strain threshold must be
study is a residual dent depth at the impact site greater taken into account when using residual strength results
than or equal to 0.05 in. This criterion seems to be determined from CAI tests to design aircraft structures.
practical for both thick-facesheet sandwich panels used
in Ref. 8 and the thin-facesheet sandwich panels used in
the present study. Results of these tests indicate that References
damage initiates and BVID occurs at lower impact 1
Smith, P. J.; Ilcewicz, L. B.; and Olson, J. T.:
energy levels for dropped-weight impacts with a 0.5- Advanced Technology Composite Fuselage.
in.-dia. tup than it does for impacts with a 1.0-in.-dia. Proceedings of the Fifth NASA/DoD/ARPA Advanced

6
American Institute of Aeronautics and Astronautics
Composites Technology Conference, NASA CP-3294, 7
Rhodes, M. D.: Impact Fracture of Composite
1995, pp. 1-31. Sandwich Structures. AIAA Paper No. 75-748, May
2
Avery, W. B.; Flynn, B. W.; Hanson, C. T.; Scholz, D. 1975.
B.; Murphy, D. P.; Carberry, D. J.; Dopker, B.; 8
McGowan, D. M. and Ambur, D. R.: Damage
Williams, A. M.; and Ilcewicz, L. B.: Design and Tolerance Characteristics of Composite Fuselage
Structural Development of a Composite Fuselage Keel Sandwich Structures With Thick Facesheets, NASA
Panel. Proceedings of the Fifth NASA/DoD/ARPA TM 110303, Feb. 1997.
Advanced Composites Technology Conference, NASA 9
Williams, J. G.; Anderson; M. S.; Rhodes, M. D.;
CP-3294, 1995, pp. 463-496. Starnes, J. H., Jr.; and Stroud, W. J.: Recent
3
Gottesman, T.; Bass, M.; and Samuel, A.: Criticality Developments in the Design, Testing and Impact-
of Impact Damage in Composite Structures. Damage Tolerance of Stiffened Composite Panels.
Proceedings of the Sixth International Conference on Fibrous Composites in Structural Design (Proceedings
Composite Materials, London, 1987, pp. 3.27-3.35. of Fourth Conference on Fibrous Composites in
4
Rix, C.; and Saczalski, T.: Damage Tolerance of Structural Design), Plenum Press, New York, NY,
Composite Sandwich Panels. Proceedings of the Eighth 1980, pp. 259-291.
International Conference on Composite Materials, 10
Ambur, D. R.; Prasad, C. B.; and Waters, W. A.: An
Honolulu, 1991, pp. 3-I-1 to 3-I-10. Internally Damped, Self-Arresting Dropped Weight
5
Tsang, P.H.W.; and Dugundji, J.: Damage Resistance Apparatus for Studying the Low-Speed Impact
of Graphite-Epoxy Sandwich Panels Under Low Speed Response of Composite Structures. Journal of
Impacts. Journal of the American Helicopter Society, Experimental Mechanics, Vol. 33, No. 1, March 1995,
Vol. 37, No. 1, January 1992, pp. 75-81. pp. 64-69.
6
Ferri, R.; and Sankar, B.V.: A Comparative Study on 11
Ambur, D. R. and Cruz, J. R.: Low-Speed Impact
the Impact Resistance of Composite Laminates and Response Characteristics of Composite Sandwich
Sandwich Panels. Proceedings of the American Society Panels. AIAA Paper No. 95-1460, April 1995.
for Composites - Ninth Technical Conference, Newark,
DE, September 20-22, 1994, pp. 482-490.

Table 1. Nominal elastic properties of AS4/8552 graphite-epoxy pre-impregnated tape and cloth materials.
Material E11, msi E22, msi G12, msi G13, msi G23, msi n12 n13 n23 Ply thickness, in.
Tape 16.3 1.36 0.766 0.766 0.52 0.32 0.32 0.40 0.0073
Fabric 8.70 8.70 0.766 0.766 0.52 0.046 0.32 0.40 0.0078

Table 2. Test matrix and summary of failure loads.

Panel Impact energy, Type of impact Dent depth, Damage area, Load at impact, Failure load,
ft-lbs in. in2 lbs lbs
1 Undamaged - - - - 38,500
2 1.5 Dropped weight 0.0030 0.0285 - 26,200
3 4.0 Dropped weight 0.0135 0.0873 - 22,500
4 7.0 Dropped weight 0.0501 0.2565 - 20,700
5 1.1 Airgun 0.0022 0.0160 - 27,600
6* 3.8 Airgun 0.0210 0.1140 - 20,800
7 4.2 Airgun 0.0230 0.1781 - 20,400
8 6.9 Airgun 0.0450 0.3010 - 18,200
9 7.2 Airgun - - 8,500 18,300
10 7.5 Airgun - - 12,800 17,200
11 7.1 Airgun - - 18,000 18,000
*
Projectile did not strike center of panel. Impact was repeated on Panel 7. Panel 6 was tested without edge supports.

7
American Institute of Aeronautics and Astronautics
 Potted ends

Composite
facesheet
5.0
2.5 Honeycomb
10.0 x core

Impact
site

Damage area
5.0 Fig. 4. Typical C-scan results forPanel 3 dropped-
weight impacted at 4.0 ft-lbs using a 0.5-in.-
Fig. 1 Schematic diagram of a typical test specimen.
diameter tup.
(Dimensions are in inches)
Loading 0.06 (0.5-in. tup)
platen (1.0-in. tup)
0.05 Airgun
Clamping
fixture 0.04
Dent
Edge support depth, 0.03
in.
DCDT 0.02

0.01
Test
specimen 0 2 4 6 8 10 12 14
Impact energy, ft-lbs
Fig. 2 Typical test set-up. Fig. 5. Residual dent-depth values at each impact site.

3000 0.4
Impact energy level = 7.0 ft-lbs (0.5-in. tup)
2500 (1.0-in. tup)
Analysis of Ref. 11 0.3 Airgun
2000
Damage
Maximum area,
1500
contact in2 0.2
force, lbs
1000

500 0.1

0 1000 2000
Impactor speed, in/sec 0 2 4 6 8 10 12 14
Fig. 3 Maximum contact force as a function of impactor Impact energy, ft-lbs
speed for the present sandwich panels. Fig. 6. Damaged area values at each impact site.

8
American Institute of Aeronautics and Astronautics
 1200
500 1.5 ft-lbs
7.0 ft-lbs
(0.5-in. tup) 1000
400 (1.0-in. tup)
1.5 ft-lbs
800
300 (1.0-in. tup) 6.0 ft-lbs
Contact 1.0 ft-lbs Contact (1.0-in. tup)
force, 200 (0.5-in. tup) force, 600
lbs lbs
400 7.0 ft-lbs
100 1.0 ft-lbs (0.5-in. tup)
(1.0-in. tup) 6.0 ft-lbs
200
0 (0.5-in. tup)
0
-100
3000 4000 5000 6000
1000 2000 3000 4000 5000 6000
Time, m-sec.
Time, m-sec.
Fig. 7. Contact force profiles for dropped-weight
Fig. 9. Contact force profiles for dropped-weight
impacts at 1.0 and 1.5 ft-lbs of impact energy.
impacts at 6.0 and 7.0 ft-lbs of impact energy.

Impact Impact

(a) dropped-weight, 0.5-in.-diameter tup, 7.00 ft-lbs (a) 0.5-in.-diameter tup, 6.0 ft-lbs of impact energy
of impact energy (32x magnification) (32x magnification)

Impact
Impact

0° ply

90° ply

(b) airgun, 7.10 ft-lbs of impact energy (b) 1.0-in.-diameter tup, 6.0 ft-lbs of impact energy
(32x magnification) (25x magnification)

Fig. 8. Photomicrographs comparing damage state Fig. 10. Photomicrographs comparing damage state
resulting from dropped-weight and airgun resulting from dropped-weight impacts using
impacts. a 0.5-in.- and a 1.0-in.-diameter tup.

9
American Institute of Aeronautics and Astronautics
 5.5 40000 SG A
4.0 SG B Panel 1

30000
Panel 2
x Load, Panel 4
2.5
0.75 lbs.
20000
x
6.5 Load
5.0 2.0 SG A
10000
Panel 3
SG B
boundary of fixture sandwich panel
0 -0.004 -0.008 -0.012 -0.016
Strain, in/in
Fig.11. Locations of impacts to study boundary effects
on impact response (dimensions are in inches). Fig.14. Axial surface strain results for Panels 1 - 4.

800 40000 SG A
Panel 1
SG B
7 ft-lbs, middle
Panel 5
600 30000
Panel 7 Panel 6
Contact Load,
force, lbs.
lbs 400 20000
7 ft-lbs, edge
Load
200 10000 SG A

Panel 8
SG B

0 10002000 3000 4000 5000 0 -0.005 -0.010 -0.015 -0.020

Time, m-sec. Strain, in/in
Fig.12. Effect of impact location on contact force Fig.15. Axial surface strain results for Panels 1 and
profiles for dropped-weight impacts. 5-8.
40000 Panel 1
Load
failure
d
30000
Panel 5
Load, failure
lbs. Panel 8
20000 Failure across
failure Panel 6 width of panel
failure
10000 Panel 7
failure

00.02 0.04 0.06 0.08 0.10 0.12

End shortening (d), in.
Fig.13. Load-end shortening response for Panels 1
and 5-8. Fig.16. Failed Panel 6.

10
American Institute of Aeronautics and Astronautics
 40000 20000 Panel 11 Panel 8
Airgun
Dropped weight
30000 Preloaded, airgun Panel 9
15000 Panel 10
Failed at impact
7 ft-lbs
Failure 20000 impacts
10000
load, Load,
lbs. Impacted at 8,500 lbs lbs. Load
10000 5000 Strain gage
Impacted at 12,800 lbs location

0 2 4 6 8 0 -0.004 -0.008 -0.012

Impact energy, ft-lbs
Strain, in/in
Fig.17. Failure loads as a function of impact energy.
Fig. 20. Axial surface strain results for Panels 8-11.

20000 Panel 8

Panel 11 Panel 9
15000
Panel 10

Load, 10000
lbs.
Load
5000
d

0 0.01 0.02 0.03 0.04 0.05 0.06

End shortening (d), in.
Fig.18. Load-end shortening response for Panels 8-11.

20000
Panel 8 Panel 9

15000
Panel 10 Panel 11

10000
Load,
Load
lbs.
5000 Strain gage
location

0 -0.002 -0.004 -0.006

Strain, in/in
Fig.19. Axial surface strain results for Panels 8-11.

11
American Institute of Aeronautics and Astronautics

View publication stats