experimental characterization of phenolic-impregnated ... · experimental characterization of...

Composite Structures 93 (2011) 2910–2924

Contents lists available at ScienceDirect

Composite Structures

journal homepage: www.elsevier .com/locate /compstruct

Experimental characterization of phenolic-impregnated honeycomb sandwichstructures for transportation vehicles

A. Zinno a,⇑, A. Prota a, E. Di Maio b, C.E. Bakis c

a Department of Structural Engineering, University of Naples ‘‘Federico II’’, Italyb Department of Material and Production, University of Naples ‘‘Federico II’’, Italyc Department of Engineering Science and Mechanics, Pennsylvania State University, United States

a r t i c l e i n f o a b s t r a c t

Article history:Available online 27 May 2011

Keywords:Phenolic sandwich structuresEnvironmental degradationExperimental testingImpact behavior

0263-8223/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.compstruct.2011.05.012

⇑ Corresponding author. Tel.: +39 0817686336.E-mail address: [email protected] (A. Zinno).

In the present paper the main outcomes of an experimental characterization on phenolic impregnatedhoneycomb sandwich structures are presented. The experimental investigations addressed both the sta-tic and dynamic properties of novel sandwich material, manufactured expressly for transportation indus-try and both the structural and impact behavior of the sandwich configuration. Moreover in order tofulfill design requirements, the prediction of the material properties and structural behaviors of sandwichstructures due to environmental degradation have been assessed using accelerated aging tests. The out-comes herein presented provide information for the modification of design parameters to minimize theinfluences of the environmental factors and the adverse effect of in-service impact events.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction smoke and combustion products during a fire, compared to other

As designers in the transportation industry strive to reduce fuelconsumption and improve safety, composite sandwich structuresthat provide improved stiffness-to-weight ratio, are becoming anattractive alternative to metals for mass transport applications. Areduction in structural weight of one large component usually trig-gers positive synergistic effects for other parts of the vehicle. Forexample, a reduction of the mass of a railway car body could leadto weight savings in the traction system, suspension, brakes andother subsystems. Therefore, using composite sandwich structuresnot only reduces weight, thereby improving fuel economy andincreasing payload capacity, but also enables the design of aerody-namic, stable vehicles with a low center of gravity. Examples ofsandwich composites employed in transportation vehicles can befound in [1,2] as structural floor and roof panels, in [3] as frontstructure, in [4,5] as body panels, and in [6] as nonstructural inte-rior panels. However, problems with the fire resistance of organicmatrix composites are seen by many as the most significant factorhindering their rapid expansion into a wide range of engineeringapplications in transport and infrastructure [7].

Among the currently available sandwich composite materials,phenolic resin-impregnated glass fiber reinforced plastics and ara-mid paper honeycomb (Nomex) are considered for the structuraldesign of load carrying components of civil transportation vehiclesin the present work. Phenolic resin is mainly considered for itsinherently fire-retardant properties that evolve low levels of

ll rights reserved.

types of resins—i.e., epoxy, polyester and vinylester—that burnand release large amounts of heat, smoke and toxic fumes thatpose a risk to people, especially in the confined space of vehicles,and make it difficult for fire fighters to extinguish the fire. Despitethe great interest on composite structures, only few publicationscontain information about mechanics and material characteristicsof phenolic impregnated sandwich structures [8,9]. In particular,impact and environmental durability data for sandwich panelsinvolving glass/phenolic skins and Nomex honeycomb core havenot been reported before, to the authors’ best knowledge.

Since the mechanical properties of composite structuresexposed to environmental influences such as humid air, tempera-ture and ultraviolet radiation may be degraded with time, adecrease in performance over time can be expected. The environ-mental effects are, in general, peculiar to the polymeric matrix[10–12] as well as the fiber–matrix interaction and are linked toa wide variety of phenomena that can ultimately lead to swellingor even to the dissolution of the polymeric matrix of the compos-ites. Moreover, attention should be focused on environmental deg-radation of the interface between core and facesheets [13,14]. Asresult, environmental effects must be considered during the designprocess. Since the degradation process of a composite sandwichstructures depends on the environmental conditions, type of skinsand core material, and production process, it is therefore necessaryto evaluate the mechanical properties through accelerated agingtests in order to predict long-term performances of sandwichcomposites.

In addition to the aforementioned issues, in order to reliablypredict the structural safety of composite sandwich structures,

http://dx.doi.org/10.1016/j.compstruct.2011.05.012

mailto:[email protected]

http://dx.doi.org/10.1016/j.compstruct.2011.05.012

http://www.sciencedirect.com/science/journal/02638223

http://www.elsevier.com/locate/compstruct

A. Zinno et al. / Composite Structures 93 (2011) 2910–2924 2911

understanding the adverse effect of in-service impact events (e.g.impact and penetration damage) has become important in thetransportation industry. Generally, structural sandwich compo-nents have low resistance to out-of-plane impact due to the thinouter composite skins and the highly deformable cores. The com-mon damage mechanisms in sandwich composites (e.g. matrixcracking, debonding, fiber failure) may appear individually orinteract, resulting in complex skin failure modes under impact.After the fracture of the skin, the impacting object may damageand penetrate into the core. If impact speed is low, sandwich pan-els may respond by bending and little damage occurs if the kineticenergy of the impacting object is absorbed elastically by the panel.At higher impact velocities a critical condition is reached when lo-cal contact stress exceeds local strength, leading to laminate bend-ing failure, core/skin interface delamination and core compressionfailure [15]. Core deformation and failure are therefore decisivefactors for the energy absorption capability and impact behaviorof sandwich panels [16]. Experimental and analytical studies havebeen conducted to understand the mechanical response of honey-comb sandwich structures under various loadings [17,18]. To de-sign sandwich panels for short-term dynamic loads, it isnecessary to have information about the influence of loading rateon the material properties. It is well known that in case of highloading rate an increase in material stiffness and strength com-pared to the static behavior may occur, which is referred to asthe strain rate effect. When this effect is neglected, dynamic finiteelement (FE) simulations and theoretical derivations based on sta-tic material data often do not agree with experimental behavior.Consequently, design approaches using static data can be too con-servative, which inhibits potential weight savings. The strain rateeffect on axial behavior of both aluminum [19–22] and Nomex[23,24] honeycomb structures have been experimentally investi-gated through dynamic compressive tests performed with differenttechniques (i.e. drop weight, gas gun, and split Hopkinson bar). Theresults have shown that dynamic loading leads to: (i) a marginalincrease of the initial stiffness and the peak compressive strength;(ii) a significant increase of the crush strength (plateau stress afterthe peak compressive strength).

Both experimental and numerical studies [25–28] have beenpresented on low-velocity impact on Nomex honeycomb sandwichstructures, but only a small amount of literature concerns impacttests where complete penetration occurs [29].

The present paper describes the results of a material character-ization program run on phenolic-impregnated honeycomb sand-wich structure to be used in the construction of light vehicles intransportation fields. The testing program is therefore devised toinvestigate the role of the core and skin materials and their inter-action in order to satisfy the structural safety in terms of stiffnessand strength requirements, environmental influences, and in-ser-vice impact events. In particular: (i) static tests of the skin and corematerials and sandwich elements are presented in order to estab-lish the mechanical properties of the selected materials and to pro-vide the structural response and failure modes of the impregnatedsandwich structures; (ii) accelerated aging tests, involving bothtensile and flexural behaviors, are presented in order to analyzelong-term performances of both laminate material and sandwichelements exposed to outdoor environments; (iii) dynamic test re-sults for the compression behavior and the impact response ofsandwich panels are presented to explain the strain rate effectsand damage mechanisms due to low velocity impact events.

Fig. 1. Shape geometry of the hexagonal honeycomb core.

2. Materials

The materials selected to manufacture the phenolic-impreg-nated sandwich structures are:

� 48 kg/m3 hexagonal honeycomb core with a nominal cell size of3.18 mm and made of phenolic resin-impregnated aramidpaper;� Pre-impregnated satin-weave E-glass fiber reinforced phenolic

resin skins with a cured ply thickness of 0.25 mm.

Since the skins are pre-impregnated, there is no additionaladhesive used to bond the skins to the core. The materials werelaminated in an autoclave at a temperature of 135 �C, a vacuumpressure of 2.5 bar and a curing time of 90 min. All the sandwichspecimens considered in the following experimental activitieshave been assembled with the ‘‘L’’ direction of the honeycomb core(Fig. 1) along the primary direction.

The skin behavior can be modeled in a relatively simple mannerwhen their properties are known, whereas the mechanical model-ing of the honeycomb core material is less straightforward. The re-sponse of the core to shear loading mainly depends on the materialused in the core and on the ratio of the core density to that of thesolid material constituting the core. Zhang and Ashby [30] give athorough overview of the literature on model the elastic and col-lapse behavior of honeycomb materials, while Aktay et al. [31]present numerical modeling of transverse crush behavior of honey-comb cored materials using both detailed micromechanics andhomogenized models.

3. Static tests

3.1. Laminate tests

The experiments involved both tensile and three-point bendingtests to evaluate the in-plane tensile and shear properties and theinterlaminar strength of the selected face sheet laminates.

Quasi-static tensile tests (Figs. 2a and b) were run on1 � 15 � 250 mm coupons, tested in one series with the warp fi-bers parallel to the load and in a second series with warp fibersperpendicular to the load. These tests were performed in accor-dance with the ASTM D3039M standard [32]. Ultimate tensilestress and strain, elastic modulus, and Poisson’s ratio have beenmeasured for the warp and fill directions. Due to the nearly bal-anced nature of the fabrics, laminates with the warp fibers perpen-dicular to the load are characterized by values close to those withwarp fibers perpendicular to the load. Quasi-static shear tests(Fig. 2c) were run on 2 � 25 � 250 mm tensile coupons with a[+45/�45]2s stacking sequence in accordance with the ASTMD3518M standard [33]. Shear modulus, ultimate shear stress andstrain have been derived by these tests. Three strain gauges wereapplied to each coupon in order to monitor the longitudinal andtransversal strains and check the system alignment of coupons.Short-beam tests (Fig. 2d) in accordance with the ASTM D2344Mstandard [34] were run on 6 � 12 � 36 mm specimens, made byparallel laminating of prepreg.

0

100

200

300

400

0 0.005 0.01 0.015 0.02

Strain [-]

Stre

ss [M

Pa]

0

100

200

300

400

0 0.005 0.01 0.015 0.02

Strain [-]

Stre

ss [M

Pa]

0

20

40

60

0 0.01 0.02 0.03 0.04 0.05

Strain [-]

Stre

ss [M

Pa]

0

10

20

30

0 0.2 0.4 0.6 0.8 1 1.2 1.4

Crosshead displacement [mm]

Stre

ss [M

Pa]

(a) (b)

(c) (d)

Fig. 2. Experimental results of laminate tests: (a) tensile tests in wrap direction; (b) tensile tests in fill direction; (c) shear tests; (d) short beam tests.

Table 1Results for face sheet laminate tests.

Glass/phenolic

Av. C.v. (%)

Elastic modulus (GPa) 25.5 warp dir. 2.323.0 fill dir. 3.3

Tensile strength (MPa) 325 warp dir. 2.9288 fill dir. 3.0

Ultimate strain (%) 1.53 warp dir. 7.01.56 fill dir. 6.0

Shear modulus (GPa) 3.41 7.5

Shear strength (MPa) 43.3 6.2

Shear strain (MPa) 2.47 3.1

Interlaminar strength (MPa) 21.3 7.7

Poisson ratio (–) 0.15 6.0

2912 A. Zinno et al. / Composite Structures 93 (2011) 2910–2924

All tests were run on a 10 kN universal test frame controlled byan electronic control unit which allows monitoring the appliedload and the stroke of the top cross head. Strain signals were ac-quired by a digital data acquisition system. Tests were conductedat a constant cross head velocity of 2 mm/min for tensile and sheartests and 1 mm/min for the short-beam tests. Table 1 reports elas-tic properties and strength values of face sheet material calculatedfrom the quasi-static tests.

3.2. Sandwich tests

The sandwich experiments involved both compressive and flex-ural tests to evaluate the compressive and shear core propertiesand to provide the bending response in terms of stiffness and fail-ure modes of the selected phenolic sandwich laminates. Thereforeindentation tests were performed in order to investigate themechanical behavior of honeycomb sandwich structures exposedto localized point loads, for example due to interaction with at-tached structures.


3.2.1. Uniaxial compressive testsOut-of-plane crushing behavior of Nomex honeycomb has been

investigated by flat-wise stabilized compressive tests according toASTM C365M [35] standard. The tests were run on five60 � 60 � 32.2 mm coupons bonded between two 1-mm-thickphenolic skins with a constant cross head speed of 0.5 mm/min.The specimens were laminated with external skins in order to pre-vent local crushing at the edges of the honeycomb cores. Specimenand test set-up details are shown in Fig. 3.

Compressive modulus, stabilized compressive strength andstrain, crush strength and strain values at which densification oc-curs have been determined by these tests. The stress–strain rela-tionship (Fig. 4a) consists of three stages: the elastic regime up tothe stabilized compressive strength, the crushing regime at nearlyconstant plateau stress (crush strength), and finally the densifica-tion regime, where the cellular structure is fully compactedresulting in a steep stress increase. The mean behavior of fivespecimens is shown in Fig. 4a. Fig. 4b shows the three deforma-tion stages of the honeycomb core during a flat-wise compressiontest.

3.2.2. Flexural testsNomex honeycomb core shear properties and flexural behavior

of the sandwich configuration were investigated by flexural testsaccording to ASTM C393M [36], and ASTM D7250M [37] respec-tively. Both three-point and four-point bending loading configura-tions were used to test sandwich specimens involving differentskin and core geometries as reported in Table 2. Deformation ofthe core in the vicinity of the loading and support points could lim-it the use of fundamental analyses [38,39] to interpret the experi-mental results. However, in the present case, local indentation ofthe core was avoided by using:

(i) Loading bars consisting of 25-mm-wide flat steel blocks. Thebars were designed to allow free rotation of the specimen atthe loading and support points, such as to have sufficientstiffness to avoid significant deflection of the bars underload.

(ii) A size of specimen that satisfies the following relation [36]:

Fc P2ðc þ tÞrtðS� LÞlpad

ð1Þ

Fig. 3. Static flat-wise stabilized compression tests: (a)

where Fc is the core compression strength; S is the supportspan length; L is the loading span length, (L = 0 for 3-pointloading); r is the expected facing ultimate strength; t is thefacing thickness; c is the core thickness; lpad is the dimensionof loading pad in the lengthwise direction of the specimen.

In the three-point bending tests (Fig. 5), skin deformations weremonitored by three longitudinally-oriented electrical resistancestrain gauges—one at the mid-span of the bottom skin and one ata 25-mm distance from the mid-span on the top and bottom skins.In the four-point bending tests, skin deformations were monitoredby four longitudinally-oriented electrical resistance strain gauges—two at the mid-span of the bottom skin and two at the mid-span ofthe top skin. The displacement of the mid-span was monitoredusing a LVDT for both loading configurations. All the flexural testswere conducted in stroke control with a cross-head speed of 6 mm/min.

In order to compare the results of the different specimen geom-etries and loading configurations, the flexural response of the se-lected sandwich panels has been reported in terms of bendingmoment–curvature curves, as shown in Fig. 6a where each curveis the mean data of the three replicate specimens. The sandwichcurvatures have been calculated by the strain signals of the topand bottom skins at the mid-span and at a 25-mm distance fromthe mid-span for the four-point and three-point bending tests,respectively. Fig. 6a allows comparing the flexural stiffness (slopeof the curves) of the sandwich elements obtained by varying thecore and/or skin thicknesses. As expected, a higher increase inthe sandwich stiffness can be obtained by increasing the corethickness (from GN_3/GN_4 to GN_1/GN_2) as opposed to the skinthickness (from GN_1/GN_3 to GN_2/GN_4).

Based on the measured sandwich skin properties, the sandwichtransverse shear rigidity and the Nomex honeycomb shear modu-lus were calculated as functions of midspan deflections and appliedforces in flexural tests of sandwich beams. First, the flexural stiff-ness, D, of the sandwich configuration is found as:

D ¼ Eðd3 � c3Þb12

ð2Þ

where E is the elastic modulus of the skin found in the laminatetests and b and d are the width and total thickness of the sandwich,respectively. Then, for each specimen, the shear rigidity, U, and the

test set-up; (b) specimen geometry.

0

1

2

3

0 0.2 0.4 0.6 0.8

Strain [-]

Stre

ss [M

Pa]

(a) (b)

Fig. 4. Experimental results of static flat-wise stabilized compression tests: (a) mean stress–strain curve; (b) deformation stages — initial (top), crushing regime (middle),densification (bottom).

Table 2Set of flexural tests on sandwich specimens.

Specimen code No. of specimens Skin lay-up Width (mm) Skin thickness (mm) Core thickness (mm) Support span (mm) Load span (mm)

GN_1 3 [0/90]s 100 1 21.5 420 140GN_2 3 [0/90]2s 100 2 21.5 420 140GN_3a 3 [0/90]s 100 1 10.5 325 0GN_3b 3 [0/90]s 100 1 10.5 265 0GN_4a 3 [0/90]2s 100 2 10.5 150 0GN_4b 3 [0/90]2s 100 2 10.5 170 0


core shear modulus, G, for at least ten sets of applied forces anddeflections were found as:

U ¼ PðS� LÞb4 D� Pð2S3�3SL2þL3Þ

96D

h i ð3Þ

Fig. 5. Three-point bending test set-up.

G ¼ Uðd� 2tÞðd� tÞ2b

ð4Þ

where D is the beam mid-span deflection corresponding to the forcelevel P. The shear rigidity and core shear modulus are finally evalu-ated as the average of the values calculated at each force level foreach specimen.

Since all specimens failed by transverse shear in the core, theshear strength of the selected Nomex honeycomb core has alsobeen calculated from the results of the flexural tests. Table 3reports the mechanical properties of the core material derivedfrom the quasi-static tests.

Closed-form analytical and numerical tools are able to capturethe experimental behavior both in terms of stiffness and in regardto failure modes. Fig. 6b shows load displacement curves derivedfor the GN_3a and GN_3b specimens and the comparison withresults of simulations based on both higher-order sandwich beamtheory (HOSBT) [41,42] and on the finite element approach. The FEanalysis has been performed on 3D-models (Fig. 7) using Nastran�

finite element codes [43]. The skins were meshed using 4-nodeshell elements, while the core was meshed using 8-node solidelements. A 2D-orthotropic material was used to define the com-posite fabric prepreg, and the composite function was used tocreate the stacking sequence of the face sheets. On the other hand,3D-orthotropic material elements were used to define the honey-comb core. Skins and core material properties are defined only inthe linear elastic range. Tied constraints were applied to obtain afull bond between the skins and the core.

0

0.05

0.1

0.15

0.2

0.25

0.00E+00 2.00E-04 4.00E-04 6.00E-04 8.00E-04

Curvature [mm-1]

Ben

ding

Mom

ent [

KN

m]

GN_1

GN_2

GN_3GN_4

0

0.5

1

1.5

2

2.5

0 5 10 15Displacement [mm]

Forc

e [K

N]

Exp.

Analy.

FEM

GN_3aGN_3b

(a) (b)

Fig. 6. Experimental results of flexural tests: (a) bending moment–curvature curves; (b) force–displacement curves of GN_3 sandwich specimens.

Table 3Nomex honeycomb mechanical properties.

Experimental Manufacturer

Av. C.v. (%) Data [40]

CompressiveModulus (MPa) 136 8.0 138Strength (MPa) 2.18 5.1 2.24Crush strength (MPa) 1.34 0.8 –Densification strain 0.76 1.2 –

ShearShear strength (MPa) – – 1.21 L DirectionShear Modulus (MPa) 45.11 0.7 44.8Shear strength (MPa) 0.73 1.6 0.69 W DirectionShear Modulus (MPa) – – 24.1


Fig. 8 shows the ‘‘failure mode map’’ concept [44–46] applied tothe phenolic-impregnated honeycomb sandwich elements underthree-point bending loading configurations. The failure modemap was generated analytically using ordinary sandwich beamtheory. The map shows the dependence of failure mode upon thesandwich geometry and the relative-core density and is able tocapture the actual experimental failure mode. The relative coredensity is defined as the ratio of honeycomb core’s density to thehoneycomb constituent material’s density.

3.2.3. Indentation testsIn order to investigate the crushing behavior of sandwich struc-

tures subjected to localized point loading, indentation tests wereperformed on 50 � 250 mm sandwich specimens manufactured

Fig. 7. FE model of sandwich specimens: (a) three-point

by laminating 32.2-mm-thick Nomex core between two 1-mm-thick glass/phenolic skins. The sandwich beams were supportedby a steel substrate. Thus, overall bending of the specimen wasavoided. The tests were carried out under displacement controlat a cross-head speed of 2 mm/min. The indentation load wasapplied through a steel cylinder (20 mm in diameter) across thewhole width of the beam cross-section (Fig. 9). The test was con-ducted on three replicate specimens for each of four different dis-placement levels. Afterwards the load was released at a cross-headspeed of 20 mm/min. During unloading, the face sheet flexed backbut did not recover completely its undeformed shape: thus, a resid-ual facesheet dent remained.

During the indentation tests, the load-indentation curve was re-corded for both loading and unloading steps. A typical load-inden-tation curve for a Nomex honeycomb sandwich specimen is shownin Fig. 10a. The curve shows a linear behavior up till the peak load.The emission of a noise (cracking sound) at the end of the lineardomain is believed to be associated with the onset of core crushing.

From an analytical point of view, if the back face of a sandwichbeam is supported by a rigid plate, the upper facing can be consid-ered as a beam of rigidity EI supported by a foundation which pro-vide a reaction r(x) per unit length. The equilibrium of the beam isgoverned by the equation:

EId4w

dx4 þ rðxÞ ¼ 0 ð5Þ

where w is the transverse displacement. When the transverse nor-mal stress remain low, the core behaves elastically and the reactionof the foundation is proportional to the transverse displacement

bending loading configuration; (b) deformed shape.

1.E-02

1.E-01

1.E+00

1.E-04 1.E-03 1.E-02 1.E-0

Skin thickness/support span

Rel

ativ

e co

re d

ensi

ty

Core Shear

Indentation

Skin failure

3a 4a3b 4b

Wrinkling

(a) (b)

Fig. 8. Failure modes map of sandwich specimens under three-point bending loading configuration. The N symbols identify experimental configurations.

x

y

Stiff steel substrate

Steel cylindrical indentor

Sandwich specimen

Fig. 9. Sandwich beam static indentation test set-up.


(r(x) = kw). In that case, the equilibrium of the top facing is gov-erned by the equation for a beam on linear elastic foundation:

d4w

dx4 þkEI

w ¼ 0 ð6Þ

The stiffness of the foundation k is related to modulus of the core inthe transverse direction Ec, the width of the beam b, and the thick-ness of the core c by:

k ¼ Ecbc

ð7Þ

The foundation behaves elastically as long as the compressive stressin the core does not exceed the value Fc (core compressionstrength). That is, when the contact force reaches the value:

P ¼ 2bFc

kð8Þ

where

k ¼ffiffiffiffiffiffiffik

4EI4

rð9Þ

In the present case, the peak indentation force derived by Eq. (8) isequal to 1.51 kN, those value is very close to the experimental one1.54 kN (C.v. = 8.16%).

After the peak indentation force, the force–displacement curvebecame nonlinear with a decrease in the stiffness. The nonlinearbehavior was due to the progressive honeycomb crushing in thearea under the indentor. Fig. 10b shows the residual dent magni-

tude and the length of damaged area as function of the imposedindentation. Each data point reported in Fig. 10b is the average ofthree tests. The results of the residual dent measurements are verysensitive to the time passed after indentation test. In this study, themeasurement of the residual dent, equal to the displacement whenthe load dropped to zero, was performed directly in the test ma-chine and immediately after unloading to the zero load level, thus,characterizing the instantaneous residual dent magnitude. Theresidual dent magnitude and length can be theoretically foundapplying the principle of minimum total energy. Energy methodsbased on assumed displacement functions, such as that used inthe analysis of unloading, provide powerful tools for solving com-plex problems but do not in general lead to exact solutions as theconditions of equilibrium are only approximately satisfied [47]. Forthis reason, in Fig. 10b the residual dent magnitude and lengthhave been interpolated by third order polynomials as functionsof maximum indentation displacement.

4. Accelerated aging tests

In the present section, accelerated aging tests have been pre-sented to analyze long-term performances of both laminate materialand sandwich elements exposed to outdoor environmental condi-tions. The expected service life of the composite materials used intransportation fields is approximately between 20 and 30 years—along period of exposure for polymer materials. Generally, damageby aging or weathering—environmental degradation of structuralcomponents of vehicle—can be due to a combination of processes.

0

0.5

1

1.5

2

2.5

3

0 2 4 6 8 10 12 14

Displacement [mm]

Forc

e [k

N]

0

2

4

6

0 2 4 6 8 10 12 14

Maximum displacement [mm]

Mag

nitu

de o

f res

idua

l den

t [m

m]

0

10

20

30

40

50

Leng

th o

f res

idua

l den

t [m

m]

Magnitude of residual dentLength of residual dent

(a) (b)

Fig. 10. Experimental results of indentation test: (a) force–displacement curves; (b) residual dent depth at different imposed displacement values.


The main factors that can affect long term properties in the railwayfield [48], considered in accelerated aging tests, are:

� Hygrothermal conditioning.� Thermal cycles.� Ultraviolet radiation.� Chemical attacks.

The natural process of absorption of moisture in many compos-ite materials is usually very slow. To accelerate the effects ofmoisture, specimens have been conditioned at 80 �C and 85% RH(hot-wet conditioning), and 80 �C and 0% RH (hot-dry condition-ing) for 60 days. To assess the influence of freeze–thaw cycles,specimens have been subjected to an alternating decrease of tem-perature from 4 to �18 �C and increase from �18 to 4 �C in not lessthan 2 h and not more than 5 h. All the specimens have been sub-jected to 250 cycles. The UV environment was reproduced by sub-jecting the specimens to Cycle C, continuous UV with uninsulatedblack panel temperature at 50 �C, as described in the standardASTM D 5208 [49]. For the analysis of the degradation due to alkaliand acid attack, specimens were submerged in alkaline (NaOH 1 Msodium hydroxide solution) and acid (HCl 1 M hydrochloric acid)solutions, respectively, for 20 and 40 days at a temperature of60 �C.

After environmental conditioning, the specimens were tested inorder to evaluate in-plane tensile properties of the skin, core shearproperties and flexural behavior of the entire sandwich configura-tion. The test matrix of accelerated aging tests is summarized inTable 4.

4.1. Laminate tests

Static tensile tests were run on 1 � 15 � 250 mm conditionedcoupons with the warp fibers parallel to the load. These tests wereperformed as described in the static section. Elastic modulus, ulti-mate strain and stress have been determined by tensile tests foreach specimen. It is underlined that coupons dipped into the alka-line solution were completely deteriorated (Fig. 11a) and couldtherefore not be mechanically tested. Experimental results are pre-sented in Fig. 11b in terms of degree of degradation of tensile prop-erties for each environmental conditioning. The obtained valuesshow considerable reductions in ultimate strain and stress forthe hot-wet and acid conditionings.

The degradation due to hygrothermal conditioning is related tothe diffusion of water molecules in a polymer structure that causesan expansion and swelling of the structure: the increasing in dis-tance between the macromolecular chains involves a weakeningof the secondary intermolecular forces so that material becomessoft and more ductile. The chemical environment influences boththe glass fibers and the polymeric resin. In particular, the acid con-ditioning causes the cleavage of the macromolecular chains, whilethe effects of the alkaline conditioning include both the disintegra-tion of the matrix and the corrosion of the E-glass fibers.

4.2. Sandwich tests

Three point bending tests were run on 16.5 � 100 � 235 sand-wich panels (GN_4 specimen code) with equal laminated E-glass/phenolic composite face sheets, each consisting of eight 0/90 wo-ven plies (2 mm total thickness) stacked in the [0/90]2s arrange-ments, bonded to 10.5 mm honeycomb Nomex core. Each type ofsandwich structure was tested with two different support spans:S = 170 mm and S = 150 mm. The test procedure is the same as thatdescribed previously for unconditioned specimens. Two replicatespecimens are tested for each configuration.

Figs. 12 and 13 shows the average load–displacement curvesmeasured for each conditioned specimen families loaded in flex-ure. No premature failures (i.e., debonding) occurred during thesetests. All specimens failed due to shear in the core. Therefore, thecore shear ultimate strength has been obtained as a function ofthe maximum load. The experimental results show a reduction inthe core shear ultimate strength of about 20% for hygrothermal,freeze–thaw and acid conditionings, while no reduction is ob-served for UV and hot-dry conditionings. The experimental resultsshow a reduction in the sandwich flexural stiffness of less than15%. As expected, the flexural stiffness reduction is due to the mod-ulus reduction observed in the facesheets in each environmentalconditioning. There are no reductions in terms of core shear mod-ulus and sandwich transverse rigidity.

5. Dynamic tests

Dynamic tests at different loading were conducted using a dropweight tower apparatus that is controlled by an electronic controlunit which allows monitoring displacement measurements with alaser measurement system (Fig. 14a). The experimental investiga-

Table 4Specimen characteristics for accelerated aging tests.

Environmental conditioning Specimen code No. of specimens

Tensile test Flexural test

Ultraviolet: Cycle C at 50 �C UV 3 4Freeze/thaw: 250 cycles from �18 to +4 �C FT 3 4Acid solution: 1 M of HCl for 20 days at 60 �C HC_20d 3 4Acid solution: 1 M of HCl for 40 days at 60 �C HC_40d 3 4Alkaline solution: 1 M of NaOH for 20 days at 60 �C HC_20d 3 4Alkaline solution: 1 M of NaOH for 40 days at 60 �C HC_40d 3 4Hot-wet 80 �C and 85% RH for 60 days HW 3 4Hot-dry 80 �C and 0% RH for 60 days HD 3 4

0

25

50

75

100

UV Freez/Thaw Hot Wet Hot Dry HCl Solution40d

HCl Solution20d

Red

uctio

n [%

]

Ultimate StressUltimate StrainElastic Modulus

(a) (b)

Fig. 11. Experimental results of conditioned phenolic/composite specimens: (a) specimens after alkaline conditioning; (b) degradation of tensile properties for eachenvironmental conditioning.


tions addressed the dynamic compression behavior and the impactresponse of the sandwich panels. The compression tests were de-signed to provide the strain rate effect of the core behavior,whereas the impact tests assess the damage mechanisms occurringin the phenolic skins.

0

1

2

3

4

5

6

0 2 4 6 8 10Displacement [mm]

Not-Conditioned

Ultraviolet

Hot-Dry

Hot-Wet

Load

[KN

]

(a)

Fig. 12. Load–displacement curves derived by bending tests for conditioned

Since a drop weight tower cannot guarantee a constant strainrate during the all the tests, various techniques have been usedin the literature to determine the strain rate. In particular Hsiaoand Daniel [50] used falling weight impact system for dynamiccharacterization of composite materials in compression at strain

0

1

2

3

4

5

6

0 2 4 6 8 10Displacement [mm]

Not-Conditioned

Ultraviolet

Hot-Dry

Hot-Wet

Load

[KN

]

(b)

specimens: (a) support span of 150 mm; (b) support span of 170 mm.

0

1

2

3

4

5

6

0 2 4 6 8 10

Load

[KN

]

Displacement [mm]

Not-Conditioned

Freeze/Thaw

HCl 40d

HCl 20d

0

1

2

3

4

5

6

0 2 4 6 8 10

Load

[KN

]

Displacement [mm]

Not-Conditioned

Freeze/Thaw

HCl 40d

HCl 20d

(a) (b)

Fig. 13. Load–displacement curves derived by bending tests for conditioned specimens: (a) support span of 150 mm; (b) support span of 170 mm.


rates up to several hundred per second. The strain rate seemed toreach a nearly constant value at a strain level of approximately10%–20% of the ultimate strain in all cases. The authors measuredstrain rate by differentiating the strain–time curve at strain read-ings above 20% of the ultimate strain. Brown et al. [51] used a mod-ified instrumented falling weight drop tower for dynamiccharacterization to analyze the effect of strain rate on the tensile,shear and compression behavior of a E-glass/polypropylene wovenfabric composite over a strain rate range of 10�3–102 s�1. Thestrain rate during the tests was derived from the gradient of thestrain–time history data. Baker et al. [21] used gas gun dynamictests to determine the energy absorption properties of aluminumhoneycomb core. Since the impactor velocity was a nearly constantvalue until the core reached its compressive strength, the strainrate was computed based on the initial velocity of the impactor.

Fig. 14. Dynamic tests: (a) drop weight tower apparatus; (b) s

In the present experiments, it was observed that the impactorvelocity holds a nearly constant value until the core reaches itscompressive strength value. After that, the velocity decreased.However the gradient of the velocity curve is very low in the crush-ing regime, whereas it steepens in the densification regime. Basedon the mentioned observation, the strain rate value was calculatedbased on the initial impactor velocity.

5.1. Uniaxial Compressive tests

Out-of-plane compressive tests were conducted on the droptower facility at three different strain rates (60 s�1, 120 s�1 and200 s�1) on cylindrical specimens of 45 mm diameter and either32.2 mm or 10.5 mm thickness (Fig. 14b). The smaller thicknesswas considered to achieve higher strain rates. In order to prevent

pecimen details for flat-wise stabilized compression tests.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

0 5 10 15 20

Original data

Filter data

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

0 0.2 0.4 0.6

Stre

ss [M

Pa]

Strain [-]

Strain rate 60

Strain rate 120

Strain rate 200

-

Stre

ss [M

Pa]

Time [ms]

(a) (b)

Fig. 15. Dynamic compression tests on honeycomb structures: (a) filtering data; (b) stress–strain curves at different strain rates.

0

10

20

30

40

50

60

70

0 10 20 30 40

v=1 m/sv= 4m/sv= 8 m/s

End of v= 1 m/s impact test

0

1

2

3

4

5

6

7

0 10 20 30 40


Skin thickness 1 mmIndentor diameter 12.7 mm

Top skin failure Bottom skin failure

Ener

gy [J

]

Load

[kN

]

Displacement [mm] Displacement [mm]


(a) (b)

Fig. 16. Experimental results of 1 mm skin specimens impacted by 12.7 mm indentor: (a) energy–displacement curves; (b) load–displacement curves.


local crushing at the edge of the honeycomb structure, the corewas bonded to 1-mm-thick glass/phenolic skins. A fourth orderButterworth low-pass filter was used in order to filter out super-posed high frequency oscillations associated with dynamic loads(Fig. 15a). In this way, comparisons of dynamic and static test datacan be made. Fig. 15b shows the stress–strain-diagrams at theinvestigated strain rates. Each shown curve represents the meandata of five replicate specimens for each specimen family.

It can be seen that dynamic loading leads to a significant in-crease of both compressive and crush strength; in particular thecompressive strength presents a DIF (Dynamic Increase Factor, ra-tio of the dynamic value over the static one) of 1.20 (at 200 s�1),whereas the crush strength presents a DIF of 1.10 (at 200 s�1).The influence of dynamic loading on the densification point hasbeen also observed: the dynamic strain value is about 10% lowerthan the quasi-static one. On the contrary, no influence on theinitial stiffness has been observed. For aluminum honeycombstructures the DIF for crush strength was observed to be about

1.33 (gas gun test, 100 s�1) [21], 1.40 (split Hopkinson pressurebar, 800 s�1) [22] and 1.50 (gas gun test, 2000 s�1) [23] abovethe quasi-static value. For Nomex honeycomb structures the DIFfor the crush strength was observed to be about 1.00–1.30 in thestrain rate domain from 50 s�1 to 300 s�1 [24].

5.2. Impact tests

Impact tests were performed on specimens prepared with 11-mm-thick Nomex honeycomb sandwiched between either 1-mmor 2-mm glass/phenolic skins consisting four or eight fabric plies,respectively. The specimens were clamped using cylindrical ringsand impacted with a 16.8-kg mass at three different energy levels,achieved with three different velocities (v = 1 m/s, v = 4 m/s andv = 8 m/s). Tests at different impact velocities were performed toprovide top skin damage (v = 1 m/s) and complete penetration(v = 4 m/s and v = 8 m/s). Two different hemispherical tipdiameters (12.7 mm and 20 mm) were adopted to provide their

0

10

20

30

40

50

60

70

0 5 10 15 20 25 30 35 40Displacement [mm]

Ener

gy [J

] v=1 m/sv= 4m/sv= 8 m/s



0

1

2

3

4

5

6

7

0 5 10 15 20 25 30 35 40Displacement [mm]

Load

[kN

] v=1 m/sv= 4m/sv= 8 m/s


Top skin failure

Bottom skin failure(a) (b)

Fig. 17. Experimental results of 2-mm skin specimens impacted by 12.7-mm indentor: (a) energy–displacement curves; (b) load–displacement curves.

0

10

20

30

40

50

60

70

0 10 20 30 40


Skin thickness 1 mmIndentor diameter 20 mm


0

1

2

3

4

5

6

7

0 10 20 30 40


Skin thickness 1 mmIndentor diameter 20 mm

Top skin failure Bottom skin failure

Displacement [mm]

Ener

gy [J

]

Displacement [mm]

Load

[kN

](a) (b)

Fig. 18. Experimental results of 1-mm skin specimens impacted by 20-mm indentor: (a) energy–displacement curves; (b) load–displacement curves.


influences on the impact response. The main objective was toassess the influence of skin thickness, impactor diameter, impactenergy and impact velocity on the main outcomes of the impacttests—i.e. the impact damage (damaged area and through-thickness damage), the force history, and the energy absorption.

Load–displacement and energy–displacement curves areplotted in Figs. 16–18 for impact events which produce respec-

Table 5Sandwich panel impact test results: mean and coefficient of variation (�).

1 mm skin thickness 2 mm12.7 mm impactor diameter 12.7

v = 1 m/s v = 4 m/s v = 8 m/s v = 1

Absorbed energy (J) 8.51 21.7 21.7 7.94(1.2%) (3.7%) (7.1) (0.0%

Top skin peak force (kN) 2.16 2.13 2.19 3.79(5.0%) (6.7%) (9.5%) (0.7%

Bottom skin peak force (kN) – 2.16 2.42 –(6.3%) (2.1%)

tively visible damage of the top skin (v = 1 m/s) and completepenetration (v = 4 m/s and v = 8 m/s) of the sandwich specimens.Each curve is the mean of the three replicate specimens. All theresults in terms of energy absorption and peak force have beensummarized in Table 5.

At lower level of energy (v = 1 m/s), the load curves grow up to apeak level which is due to the penetration of the impactor through

skin thickness 1 mm skin thicknessmm impactor diameter 20 mm impactor diameter

m/s v = 4 m/s v = 8 m/s v = 1 m/s v = 4 m/s v = 8 m/s

59. 6 63.2 9.23 43.1 40.9) (4.8%) (1.2%) (0.5%) (1.1%) (6. 3%)

3.92 4.47 2.70 3.32 3.11) (7.3%) (10.4%) (0.9%) (7.0%) (7.1%)

5.44 6.09 – 3.36 3.15(12.5%) (10.4%) (4.2%) (2. 5%)

Fig. 19. Pictures of the impact face and cross-sectional view of sandwich specimens: (a) 1 m/s velocity, 1 mm skin thickness and 12.7 mm impactor; (b) 1 m/s velocity, 2 mmskin thickness and 12.7 mm impactor.


the top skin. Furthermore, it can be observed that, before the pen-etration occurs, the curves present a change of slope due to thecrushing of the Nomex honeycomb structures that happens whenthe Nomex attains the compressive strength.

At higher level of energy (v = 4 m/s and v = 8 m/s), after the firstpeak the load curves present a low-loading plateau, characterizingthe penetration of the honeycomb core, and a second sharp peaklevel due to the failure of the bottom skin. The absorbed impact en-ergy displays a constant initial slope up till a penetration of the topskin is achieved. The gradient of the curve reduces when the core ispenetrated and increases again when the weight impacts on thebottom skin.

The experimental trends show that, in higher-energy impacts,penetration resistance and absorbed energy are governed by theoverall rigidity of the target and the resistance of the facing to per-

Fig. 20. Pictures of the impact face and cross-sectional view of sandwich specimens: (a) 4skin thickness and 12.7 mm impactor.

foration [52,53]. However, the energy absorbed by each skin is al-most constant and increases proportionally to the skin thickness.The absorbed energy of the honeycomb core, related to the plasticstrain of the walls, is very low in relation the energy absorbed bythe skins. Therefore, the strain rate effects observed in the coreare not evident in the sandwich composite under ballistic impacts.

In case of 1-mm skin thickness, no significant strain rate effectsare observed for the peak forces and absorbed energies, whereas incase of 2-mm skin thickness a slight increase of the peak forces isobserved (18% for top skin from 1 to 8 m/s, 12% for bottom skin from4 to 8 m/s). In any case the energy absorbed by the facesheets andhoneycomb increases with impactor diameter. In order to furtherunderstand the damage mechanism involved in impact events, theinfluence of skin thickness and impactor size on the failure modesare presented in Figs. 19–21. The main failure mode in the composite

m/s velocity, 1 mm skin thickness and 12.7 mm impactor; (b) 4 m/s velocity, 2 mm

Fig. 21. Pictures of the impact face and cross-sectional view of sandwich specimens: (a) 1 m/s velocity, 1 mm skin thickness and 20 mm impactor; (b) 4 m/s velocity, 1 mmskin thickness and 20 mm impactor.


skins is fiber breakage, which represents the basic energy absorptionmechanism. However, when increasing the skin thickness, delami-nation mechanisms become relevant on the bottom skin.

6. Conclusions

The experimental activities presented in this paper providemechanical data of novel phenolic sandwich materials and validatethe manufacturing process and the mechanical behavior. The qua-si-static study, in particular, underlines the capability of analyticaland numerical models to capture the elastic flexural stiffness ofsandwich structures.

In order to evaluate the structural performances of phenolicsandwich structures subjected to environmental effects, the se-lected phenolic sandwich materials were exposed to acceleratedaging conditions including temperature, moisture, chemical at-tacks and ultraviolet radiation. In the evaluation of the degradationof the sandwich materials, the following main outcomes have beenobtained: (i) the E-glass/phenolic composite shows a considerablereductions in ultimate strain and stress for the hot-wet and acidconditionings; (ii) the E-glass/phenolic composite shows a reduc-tion in elastic modulus of less than 15% for each environmentalconditioning; (iii) the Nomex honeycomb core shows a reductionin shear ultimate strength of about 20% for hygrothermal, freeze–thaw and acid conditionings.

Moreover, the influence of the skin thickness, the materialstrain rate, the diameter of the impacting projectile and the impactvelocity on the dynamic behavior of the sandwich composite mate-rial have been experimentally derived in order to understand theresponse of the sandwich material and the damage mechanismsinvolved in an impact event on a sandwich structure. The dynamicimpact tests reveal that the global behavior and energy absorptionare mainly due to the contribution of the skins. As it was expected,in impact tests, in case of thicker skins, delamination occurs at thebottom skin.

References

[1] Vaidya UK, Samalot F, Pillay S, Janowoski GM, Husman G, Gleich K. Design andmanufacture of woven reinforced glass/polyprolene composites for masstransit floor structure. J Compos Mater 2004;38:1949–72.

[2] Zinno A, Fusco E, Prota A, Manfredi G. Multiscale approach for the design ofcomposite sandwich structures for the train application. Compos Struct2010;92:2208–19.

[3] Belingardi G, Cavarota MP, Duella R. Material characterization of composite-foam sandwich for the front structure of high speed train. Compos Struct2003;61:13–25.

[4] Ning H, Janowsky GM, Vaidya UK, Husman G. Thermoplastic sandwichstructure design and manufacturing for the body panel of mass transitvehicle. Compos Struct 2007;80:82–91.

[5] Kim JS, Lee SJ, Shin KB. Manufacturing and structural safety evaluation of acomposite train carbody. Compos Struct 2007;78:468–76.

[6] Potluri P, Kusak E, Reddy TY. Novel stitch-bonded sandwich compositestructures. Compos Struct 2003;59:251–9.

[7] George M. Fire-safe composites for mass transit vehicles. Reinf Plast2002:26–30.

[8] Torre L, Kenny JM. Impact testing and simulation of composite sandwichstructures for civil transportation. Compos Struct 2000;50:257–67.

[9] Manalo AC, Aravinthan T, Karunasena W, Islam MM. Flexural behavior ofstructural fiber composite sandwich beams in Flatwise and edgewise positions.Compos Struct 2010;92:984–95.

[10] Ito M, Nagai K. Degradation issues of polymer materials used in railway field.Polym Degrad Stab 2008;93:1723–35.

[11] Chung K, Yoshioka K, Seferis JC. Hygrothermal cycling effects on the durabilityof phenolic based composites. Polym Compos 2002;23:141–1524.

[12] Ishai O, Hiel C, Luft M. Long term hygrothermal effects on damage tolerance ofhybrid composite sandwich panels. Composites 1995;26:47–55.

[13] Siriruk A, Penumadu D, Weitsman YJ. Effect of sea environment on interfacialdelamination behavior of polymeric sandwich structures. Compos Sci Technol2009;69:821–8.

[14] Veazie DR, Robinson KR, Shivakumar K. Effects of the marine environment onthe interfacial fracture toughness of PVC core sandwich composites. ComposPart B Eng 2004;35:461–6.

[15] Abrate S. Impact on composite structures. Cambridge University Press; 1998.[16] Aktay L, Johnson AF, Holzapfel M. Prediction of impact damage on sandwich

composites. Comput Mater Sci 2005;32:252–60.[17] Gibson LJ, Ashby MF. Cellular solids: structure and properties. Cambridge

University Press; 1997.[18] Foo CC, Chai GB, Seah LK. Mechanical properties of Nomex material and

Nomex honeycomb structures. Compos Struct 2007;80:588–94.[19] Wu E, Jiang WS. Axial crush of metallic honeycombs. Int J Impact Eng

1997;19(5–6):439–56.[20] Zhao H, Gary G. Crushing behavior of aluminum honeycombs under impact

loading. Int J Impact Eng 1998;21:827–36.[21] Baker WE, Togami TC, Weydert JC. Static and dynamic properties of high-

density metal honeycombs. Int J Impact Eng 1998;21:146–63.[22] Yamashita M, Gotoh M. Impact behavior of honeycomb structures with

various cell specifications–numerical simulations and experiment. Int J ImpactEng 2005;32:618–30.

[23] Goldsmith W, Sackman JL. An experimental study of energy absorption inimpact on sandwich plates. Int J Impact Eng 1992;12:241–62.

[24] Heimbs S, Schmeer S, Middendorf P, Maier M. Strain rate effects in phenoliccomposites and phenolic-impregnated honeycomb structures. Compos SciTechnol 2007;67:2827–37.


[25] Meo M, Vignjevic R, Marengo G. The response of honeycomb sandwich panelsunder low-velocity impact loading. Int J Mech Sci 2007;67:2827–37.

[26] Herup EJ, Palazzotto A. Low-velocity impact damage initiation in graphite/epoxy/Nomex honeycomb-sandwich plates. Compos Sci Technol 1997;57:1581–98.

[27] Park JH, Ha SK, Kang KW, Kim CW, Kim HS. Impact damage resistance ofsandwich structure subjected to low velocity impact. J Mater Process Technol2008;201:425–30.

[28] Anderson T, Madenci E. Experimental investigation of low-velocityimpact characteristics of sandwich composites. Compos Struct 2000;50:239–47.

[29] Mines RAW, Worral CM, Gibson AG. Low velocity perforation behavior ofpolymer composite sandwich panels. Int J Impact Eng 1998;21:855–79.

[30] Zhang J, Ashby MF. The out-of-plane properties of honeycombs. Int J Mech Sci1992;34(5):475–89.

[31] Aktay L, Johnson A, Kroplin BH. Numerical modeling of honeycomb core crushbehavior. Eng Fract Mech 2008;75:2616–30.

[32] ASTM D3039M. Standard test method for tensile properties of polymer matrixcomposite materials. West Conshohocken (PA): American Soc for Testing andMaterials.

[33] ASTM D3518M. In-plane shear response of polymer matrix compositematerials by tensile test of a ±45� laminate. West Conshohocken (PA):American Soc for Testing and Materials.

[34] ASTM D2344M. Standard test method for short-beam strength of polymermatrix composite materials and their laminates. West Conshohocken (PA):American Soc for Testing and Materials.

[35] ASTM C365M. Flatwise compressive properties of sandwich cores. WestConshohocken (PA): American Soc for Testing and Materials.

[36] ASTM C393M. Core shear properties of sandwich constructions by beamflexure. West Conshohocken (PA): American Soc for Testing and Materials.

[37] ASTM D7250M. Determining sandwich beam flexural and shear stiffness. WestConshohocken (PA): American Soc for Testing and Materials.

[38] Allen H. Analysis and design of structural sandwich panels. Oxford: PergamonPress; 1969.

[39] Plantema FJ. Sandwich construction. New York: Wiley; 1966.[40] HexWeb HRH-10 Product Data. Dublin (California): Hexcel Corporation.[41] Frosting Y, Baruch M. High-order theory for sandwich-beam behavior with

transversely flexible core. J Eng Mech 1992;118:1026–43.[42] Frosting Y, Baruch M. Localized load effects in high-order bending of sandwich

panels with flexible core. J Eng Mech 1996;122:1069–76.[43] MSC Nastran 2001. Reference manual. Santa Ana (CA): MSC Software

Corporation; 2001.[44] Petras A, Sutcliffe MPF. Failure mode maps for honeycomb sandwich panels.

Compos Struct 1999;44:237–52.[45] Triantafillou TC, Gibson LJ. Failure mode maps for foam core sandwich beams.

Mater Sci Eng 1987;95:37–53.[46] Shenhar Y, Frosting Y, Altus E. Stresses and failure patterns in the bending of

sandwich beams with transversely flexible cores and laminates compositeskins. Compos Struct 1996;35:143–52.

[47] Zenkert D, Shipsha A, Persson K. Static indentation and unloading response ofsandwich beams. Compos Part B Eng 2004;35:511–22.

[48] CEI EN 50125-1. Environmental conditioning for equipment – part 1 –equipment on board rolling stock. The Italian Electrotechnical Committee.

[49] ASTM D5208. Standard test method for fluorescent ultraviolet (UV) exposureof photodegradable plastics. West Conshohocken (PA): American Soc forTesting and Materials.

[50] Hsiao HM, Daniel IM. Strain rate behavior of composite materials. Compos PartB Eng 1998;29:521–33.

[51] Brown KA, Brooks R, Warrior NA. The static and high strain rate behavior of acommingled E-glass/polypropylene woven fabric composite. Compos SciTechnol 2010;70:272–83.

[52] Shih WK, Jang BZ. Instrumented impact testing of composite sandwich panels.J Reinf Plast Compos 1989;8:298–370.

[53] Akay M, Hanna R. A comparison of honeycomb-core and foam-core carbon–fiber epoxy sandwich panels. Composites 1990;21:325–31.

experimental characterization of phenolic-impregnated ... · experimental characterization of...

Documents